From 63be96bdc7e9f388a5c49295bd7e150462fd003a Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 18 Sep 2007 19:29:26 -0600
Subject: temporarily set the FRAG_BIT_FOGC bit in InputsRead when fog is
 enabled

---
 src/mesa/main/texenvprogram.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 72b54b27d9..37f8fc8090 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1081,6 +1081,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
        * a reduced value and not what is expected in FogOption
        */
       p.program->FogOption = ctx->Fog.Mode;
+      p.program->Base.InputsRead |= FRAG_BIT_FOGC; /* XXX new */
    } else
       p.program->FogOption = GL_NONE;
 
-- 
cgit v1.2.3


From 3bf8d2ac7108a7f0f1722e411161e013bb8573f0 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 25 Sep 2007 15:18:51 -0600
Subject: Disable vertex shader fog, compute fog in fragment shader.

---
 src/mesa/main/texenvprogram.c       | 16 +++++++++++-----
 src/mesa/state_tracker/st_context.c |  5 +++++
 2 files changed, 16 insertions(+), 5 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 37f8fc8090..48c28d45ca 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1097,15 +1097,21 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    ASSERT(p.program->Base.NumInstructions <= MAX_INSTRUCTIONS);
 
    /* Allocate final instruction array */
-   program->Base.Instructions
-      = _mesa_alloc_instructions(program->Base.NumInstructions);
-   if (!program->Base.Instructions) {
+   p.program->Base.Instructions
+      = _mesa_alloc_instructions(p.program->Base.NumInstructions);
+   if (!p.program->Base.Instructions) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY,
                   "generating tex env program");
       return;
    }
-   _mesa_copy_instructions(program->Base.Instructions, instBuffer,
-                           program->Base.NumInstructions);
+   _mesa_copy_instructions(p.program->Base.Instructions, instBuffer,
+                           p.program->Base.NumInstructions);
+
+   if (p.program->FogOption) {
+      _mesa_append_fog_code(ctx, p.program);
+      p.program->FogOption = GL_NONE;
+   }
+
 
    /* Notify driver the fragment program has (actually) changed.
     */
diff --git a/src/mesa/state_tracker/st_context.c b/src/mesa/state_tracker/st_context.c
index e0304dd22d..7c20b036a4 100644
--- a/src/mesa/state_tracker/st_context.c
+++ b/src/mesa/state_tracker/st_context.c
@@ -26,6 +26,8 @@
  **************************************************************************/
 
 #include "main/imports.h"
+#include "main/extensions.h"
+#include "tnl/tnl.h"
 #include "vbo/vbo.h"
 #include "st_public.h"
 #include "st_context.h"
@@ -98,6 +100,9 @@ struct st_context *st_create_context( GLcontext *ctx,
    /* XXXX This is temporary! */
    _mesa_enable_sw_extensions(ctx);
 
+   /* we'll always do per-pixel fog in the fragment shader */
+   _tnl_allow_vertex_fog(ctx, GL_FALSE);
+
    return st;
 }
 
-- 
cgit v1.2.3


From 83fad68ec1989c719646a76f4cc5e0b3d23537ed Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 25 Sep 2007 15:20:04 -0600
Subject: include programopt.h

---
 src/mesa/main/texenvprogram.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 48c28d45ca..cdf35b4636 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -32,6 +32,7 @@
 #include "shader/prog_instruction.h"
 #include "shader/prog_print.h"
 #include "shader/prog_statevars.h"
+#include "shader/programopt.h"
 #include "texenvprogram.h"
 
 /**
-- 
cgit v1.2.3


From d781cdc8fadc802a1f2edbeb13ccb1ee768ce803 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 2 Oct 2007 16:55:21 -0600
Subject: Generate a texenv fragment program if there's a GLSL vertex shader
 but no GLSL fragment shader.

This allows Glean glsl1 test to get pretty far.
---
 src/mesa/main/texenvprogram.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index cdf35b4636..a554c033c4 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1247,7 +1247,8 @@ _mesa_UpdateTexEnvProgram( GLcontext *ctx )
 
    /* If a conventional fragment program/shader isn't in effect... */
    if (!ctx->FragmentProgram._Enabled &&
-       !ctx->Shader.CurrentProgram) {
+       (!ctx->Shader.CurrentProgram ||
+        !ctx->Shader.CurrentProgram->FragmentProgram) ) {
       make_state_key(ctx, &key);
       hash = hash_key(&key);
       
-- 
cgit v1.2.3


From e69943e6dda102df8418a8261b95155350181a2f Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 23 Oct 2007 10:23:01 -0600
Subject: bump up MAX_INSTRUCTIONS and add an assertion to catch emitting too
 many instructions

---
 src/mesa/main/texenvprogram.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index a554c033c4..2614440a74 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -36,10 +36,11 @@
 #include "texenvprogram.h"
 
 /**
- * According to Glean's texCombine test, no more than 21 instructions
- * are needed.  Allow a few extra just in case.
+ * This MAX is probably a bit generous, but that's OK.  There can be
+ * up to four instructions per texture unit (TEX + 3 for combine),
+ * then there's fog and specular add.
  */
-#define MAX_INSTRUCTIONS 24
+#define MAX_INSTRUCTIONS ((MAX_TEXTURE_UNITS * 4) + 12)
 
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
@@ -478,6 +479,8 @@ emit_op(struct texenv_fragment_program *p,
    GLuint nr = p->program->Base.NumInstructions++;
    struct prog_instruction *inst = &p->program->Base.Instructions[nr];
       
+   assert(nr < MAX_INSTRUCTIONS);
+
    _mesa_init_instructions(inst, 1);
    inst->Opcode = op;
    
-- 
cgit v1.2.3


From 112a1580f658948e553fe04399a20958dca67c16 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Tue, 23 Oct 2007 10:54:50 -0600
Subject: properly init dst reg's CondMask/Swizzle fields

---
 src/mesa/main/texenvprogram.c | 4 ++--
 src/mesa/tnl/t_vp_build.c     | 6 +++---
 2 files changed, 5 insertions(+), 5 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 2614440a74..abc2567134 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -462,8 +462,8 @@ static void emit_dst( struct prog_dst_register *dst,
    dst->File = ureg.file;
    dst->Index = ureg.idx;
    dst->WriteMask = mask;
-   dst->CondMask = 0;
-   dst->CondSwizzle = 0;
+   dst->CondMask = COND_TR;  /* always pass cond test */
+   dst->CondSwizzle = SWIZZLE_NOOP;
 }
 
 static struct prog_instruction *
diff --git a/src/mesa/tnl/t_vp_build.c b/src/mesa/tnl/t_vp_build.c
index cbffed5155..dc2d5e0065 100644
--- a/src/mesa/tnl/t_vp_build.c
+++ b/src/mesa/tnl/t_vp_build.c
@@ -488,8 +488,8 @@ static void emit_dst( struct prog_dst_register *dst,
    dst->Index = reg.idx;
    /* allow zero as a shorthand for xyzw */
    dst->WriteMask = mask ? mask : WRITEMASK_XYZW; 
-   dst->CondMask = COND_TR;
-   dst->CondSwizzle = 0;
+   dst->CondMask = COND_TR;  /* always pass cond test */
+   dst->CondSwizzle = SWIZZLE_NOOP;
    dst->CondSrc = 0;
    dst->pad = 0;
 }
@@ -512,7 +512,7 @@ static void debug_insn( struct prog_instruction *inst, const char *fn,
 
 
 static void emit_op3fn(struct tnl_program *p,
-		       GLuint op,
+                       enum prog_opcode op,
 		       struct ureg dest,
 		       GLuint mask,
 		       struct ureg src0,
-- 
cgit v1.2.3


From 2a3f3679eba802dcb4b46f90c66326c9195cdbcb Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Mon, 29 Oct 2007 09:23:46 -0600
Subject: Disable the else clause which assigns the default fragment program to
 ctx->FragmentProgram._Current

The _Current field should either point to the fragment program which is to be
run (GLSL, ARB_f_p, fixed-func-generated, etc) or be NULL if conventional
fixed-function code is to be used.  Matches TNL program code.
---
 src/mesa/main/texenvprogram.c | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index abc2567134..21a290402f 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1279,9 +1279,11 @@ _mesa_UpdateTexEnvProgram( GLcontext *ctx )
             _mesa_printf("Found existing texenv program for key %x\n", hash);
       }
    } 
+#if 0
    else {
       ctx->FragmentProgram._Current = ctx->FragmentProgram.Current;
    }
+#endif
 
    /* Tell the driver about the change.  Could define a new target for
     * this?
-- 
cgit v1.2.3


From 83ce6c51d7189094cf2a13fdcc0882a051a3bd41 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Mon, 29 Oct 2007 11:23:02 -0600
Subject: Refactor _mesa_UpdateTexEnvProgram()

Will be replaced by _mesa_get_fixed_func_fragment_program().
---
 src/mesa/main/texenvprogram.c | 68 +++++++++++++++++++++++--------------------
 src/mesa/main/texenvprogram.h |  3 ++
 2 files changed, 39 insertions(+), 32 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 21a290402f..d8f3314c70 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1234,6 +1234,39 @@ static GLuint hash_key( const struct state_key *key )
 }
 
 
+/**
+ * Return a fragment program which implements the current
+ * fixed-function texture, fog and color-sum operations.
+ */
+struct gl_fragment_program *
+_mesa_get_fixed_func_fragment_program(GLcontext *ctx)
+{
+   struct gl_fragment_program *prog;
+   struct state_key key;
+   GLuint hash;
+	
+   make_state_key(ctx, &key);
+   hash = hash_key(&key);
+      
+   prog = search_cache(&ctx->Texture.env_fp_cache, hash, &key, sizeof(key));
+
+   if (!prog) {
+      if (0)
+         _mesa_printf("Building new texenv proggy for key %x\n", hash);
+
+      prog = (struct gl_fragment_program *) 
+         ctx->Driver.NewProgram(ctx, GL_FRAGMENT_PROGRAM_ARB, 0);
+
+      create_new_program(ctx, &key, prog);
+
+      cache_item(&ctx->Texture.env_fp_cache, hash, &key, prog);
+   }
+
+   return prog;
+}
+
+
+
 /**
  * If _MaintainTexEnvProgram is set we'll generate a fragment program that
  * implements the current texture env/combine mode.
@@ -1242,8 +1275,6 @@ static GLuint hash_key( const struct state_key *key )
 void
 _mesa_UpdateTexEnvProgram( GLcontext *ctx )
 {
-   struct state_key key;
-   GLuint hash;
    const struct gl_fragment_program *prev = ctx->FragmentProgram._Current;
 	
    ASSERT(ctx->FragmentProgram._MaintainTexEnvProgram);
@@ -1252,38 +1283,11 @@ _mesa_UpdateTexEnvProgram( GLcontext *ctx )
    if (!ctx->FragmentProgram._Enabled &&
        (!ctx->Shader.CurrentProgram ||
         !ctx->Shader.CurrentProgram->FragmentProgram) ) {
-      make_state_key(ctx, &key);
-      hash = hash_key(&key);
-      
-      ctx->FragmentProgram._Current =
-      ctx->FragmentProgram._TexEnvProgram =
-         search_cache(&ctx->Texture.env_fp_cache, hash, &key, sizeof(key));
-
-      if (!ctx->FragmentProgram._TexEnvProgram) {
-         if (0)
-            _mesa_printf("Building new texenv proggy for key %x\n", hash);
 
-         /* create new tex env program */
-	 ctx->FragmentProgram._Current =
-         ctx->FragmentProgram._TexEnvProgram =
-            (struct gl_fragment_program *) 
-            ctx->Driver.NewProgram(ctx, GL_FRAGMENT_PROGRAM_ARB, 0);
-
-         create_new_program(ctx, &key, ctx->FragmentProgram._TexEnvProgram);
-
-         cache_item(&ctx->Texture.env_fp_cache, hash, &key,
-                    ctx->FragmentProgram._TexEnvProgram);
-      }
-      else {
-         if (0)
-            _mesa_printf("Found existing texenv program for key %x\n", hash);
-      }
+      ctx->FragmentProgram._Current
+         = ctx->FragmentProgram._TexEnvProgram
+         = _mesa_get_fixed_func_fragment_program(ctx);
    } 
-#if 0
-   else {
-      ctx->FragmentProgram._Current = ctx->FragmentProgram.Current;
-   }
-#endif
 
    /* Tell the driver about the change.  Could define a new target for
     * this?
diff --git a/src/mesa/main/texenvprogram.h b/src/mesa/main/texenvprogram.h
index 6f017767c8..ac73910cb5 100644
--- a/src/mesa/main/texenvprogram.h
+++ b/src/mesa/main/texenvprogram.h
@@ -34,6 +34,9 @@
 
 #include "mtypes.h"
 
+extern struct gl_fragment_program *
+_mesa_get_fixed_func_fragment_program(GLcontext *ctx);
+
 extern void _mesa_UpdateTexEnvProgram( GLcontext *ctx );
 extern void _mesa_TexEnvProgramCacheInit( GLcontext *ctx );
 extern void _mesa_TexEnvProgramCacheDestroy( GLcontext *ctx );
-- 
cgit v1.2.3


From f18d4e058ed979c6e42e868c7febde4fa62c5810 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Mon, 29 Oct 2007 12:25:46 -0600
Subject: Remove ctx field from texenvprog_cache

---
 src/mesa/main/mtypes.h        |  1 -
 src/mesa/main/texenvprogram.c | 23 +++++++++++------------
 2 files changed, 11 insertions(+), 13 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 8adc4e3373..44587ae962 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -1566,7 +1566,6 @@ struct texenvprog_cache_item {
 struct texenvprog_cache {
    struct texenvprog_cache_item **items;
    GLuint size, n_items;
-   GLcontext *ctx;
 };
 
 /**
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index d8f3314c70..cf92503c34 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -399,9 +399,9 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 }
 
 
-static void release_temps( struct texenv_fragment_program *p )
+static void release_temps(GLcontext *ctx, struct texenv_fragment_program *p )
 {
-   GLuint max_temp = p->ctx->Const.FragmentProgram.MaxTemps;
+   GLuint max_temp = ctx->Const.FragmentProgram.MaxTemps;
 
    /* KW: To support tex_env_crossbar, don't release the registers in
     * temps_output.
@@ -1037,7 +1037,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    p.one = undef;
 
    p.last_tex_stage = 0;
-   release_temps(&p);
+   release_temps(ctx, &p);
 
    if (key->enabled_units) {
       /* First pass - to support texture_env_crossbar, first identify
@@ -1055,7 +1055,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
       for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++)
 	 if (key->enabled_units & (1<<unit)) {
 	    p.src_previous = emit_texenv( &p, unit );
-	    release_temps(&p);	/* release all temps */
+	    release_temps(ctx, &p);	/* release all temps */
 	 }
    }
 
@@ -1169,7 +1169,7 @@ static void rehash( struct texenvprog_cache *cache )
    cache->size = size;
 }
 
-static void clear_cache( struct texenvprog_cache *cache )
+static void clear_cache( GLcontext *ctx, struct texenvprog_cache *cache )
 {
    struct texenvprog_cache_item *c, *next;
    GLuint i;
@@ -1178,8 +1178,7 @@ static void clear_cache( struct texenvprog_cache *cache )
       for (c = cache->items[i]; c; c = next) {
 	 next = c->next;
 	 _mesa_free(c->key);
-	 cache->ctx->Driver.DeleteProgram(cache->ctx,
-                                          (struct gl_program *) c->data);
+	 ctx->Driver.DeleteProgram(ctx, (struct gl_program *) c->data);
 	 _mesa_free(c);
       }
       cache->items[i] = NULL;
@@ -1190,7 +1189,8 @@ static void clear_cache( struct texenvprog_cache *cache )
 }
 
 
-static void cache_item( struct texenvprog_cache *cache,
+static void cache_item( GLcontext *ctx,
+                        struct texenvprog_cache *cache,
 			GLuint hash,
 			const struct state_key *key,
 			void *data )
@@ -1208,7 +1208,7 @@ static void cache_item( struct texenvprog_cache *cache,
       if (cache->size < 1000)
 	 rehash(cache);
       else 
-	 clear_cache(cache);
+	 clear_cache(ctx, cache);
    }
 
    cache->n_items++;
@@ -1259,7 +1259,7 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
 
       create_new_program(ctx, &key, prog);
 
-      cache_item(&ctx->Texture.env_fp_cache, hash, &key, prog);
+      cache_item(ctx, &ctx->Texture.env_fp_cache, hash, &key, prog);
    }
 
    return prog;
@@ -1301,7 +1301,6 @@ _mesa_UpdateTexEnvProgram( GLcontext *ctx )
 
 void _mesa_TexEnvProgramCacheInit( GLcontext *ctx )
 {
-   ctx->Texture.env_fp_cache.ctx = ctx;
    ctx->Texture.env_fp_cache.size = 17;
    ctx->Texture.env_fp_cache.n_items = 0;
    ctx->Texture.env_fp_cache.items = (struct texenvprog_cache_item **)
@@ -1312,6 +1311,6 @@ void _mesa_TexEnvProgramCacheInit( GLcontext *ctx )
 
 void _mesa_TexEnvProgramCacheDestroy( GLcontext *ctx )
 {
-   clear_cache(&ctx->Texture.env_fp_cache);
+   clear_cache(ctx, &ctx->Texture.env_fp_cache);
    _mesa_free(ctx->Texture.env_fp_cache.items);
 }
-- 
cgit v1.2.3


From f4a5ea2ccb472958a4635c606e9510011bceaa3d Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Wed, 31 Oct 2007 12:45:32 -0600
Subject: Update texenvprogram.c code to use prog_cache.c routines.

---
 src/mesa/main/mtypes.h        |  14 -----
 src/mesa/main/texenvprogram.c | 135 +++---------------------------------------
 src/mesa/main/texenvprogram.h |  11 +---
 src/mesa/main/texstate.c      |   4 --
 4 files changed, 9 insertions(+), 155 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index b435c29793..8e49431a8f 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -1557,17 +1557,6 @@ struct gl_texture_unit
    /*@}*/
 };
 
-struct texenvprog_cache_item {
-   GLuint hash;
-   void *key;
-   struct gl_fragment_program *data;
-   struct texenvprog_cache_item *next;
-};
-
-struct texenvprog_cache {
-   struct texenvprog_cache_item **items;
-   GLuint size, n_items;
-};
 
 /**
  * Texture attribute group (GL_TEXTURE_BIT).
@@ -1599,9 +1588,6 @@ struct gl_texture_attrib
    /** GL_EXT_shared_texture_palette */
    GLboolean SharedPalette;
    struct gl_color_table Palette;
-   
-   /** Cached texenv fragment programs */
-   struct texenvprog_cache env_fp_cache;
 };
 
 
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index cf92503c34..efb3b35f6a 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1,6 +1,6 @@
 /**************************************************************************
  * 
- * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
  * All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -29,6 +29,7 @@
 #include "macros.h"
 #include "enums.h"
 #include "shader/prog_parameter.h"
+#include "shader/prog_cache.h"
 #include "shader/prog_instruction.h"
 #include "shader/prog_print.h"
 #include "shader/prog_statevars.h"
@@ -1131,109 +1132,6 @@ create_new_program(GLcontext *ctx, struct state_key *key,
 }
 
 
-static struct gl_fragment_program *
-search_cache(const struct texenvprog_cache *cache,
-             GLuint hash,
-             const void *key,
-             GLuint keysize)
-{
-   struct texenvprog_cache_item *c;
-
-   for (c = cache->items[hash % cache->size]; c; c = c->next) {
-      if (c->hash == hash && memcmp(c->key, key, keysize) == 0)
-	 return (struct gl_fragment_program *) c->data;
-   }
-
-   return NULL;
-}
-
-static void rehash( struct texenvprog_cache *cache )
-{
-   struct texenvprog_cache_item **items;
-   struct texenvprog_cache_item *c, *next;
-   GLuint size, i;
-
-   size = cache->size * 3;
-   items = (struct texenvprog_cache_item**) _mesa_malloc(size * sizeof(*items));
-   _mesa_memset(items, 0, size * sizeof(*items));
-
-   for (i = 0; i < cache->size; i++)
-      for (c = cache->items[i]; c; c = next) {
-	 next = c->next;
-	 c->next = items[c->hash % size];
-	 items[c->hash % size] = c;
-      }
-
-   _mesa_free(cache->items);
-   cache->items = items;
-   cache->size = size;
-}
-
-static void clear_cache( GLcontext *ctx, struct texenvprog_cache *cache )
-{
-   struct texenvprog_cache_item *c, *next;
-   GLuint i;
-
-   for (i = 0; i < cache->size; i++) {
-      for (c = cache->items[i]; c; c = next) {
-	 next = c->next;
-	 _mesa_free(c->key);
-	 ctx->Driver.DeleteProgram(ctx, (struct gl_program *) c->data);
-	 _mesa_free(c);
-      }
-      cache->items[i] = NULL;
-   }
-
-
-   cache->n_items = 0;
-}
-
-
-static void cache_item( GLcontext *ctx,
-                        struct texenvprog_cache *cache,
-			GLuint hash,
-			const struct state_key *key,
-			void *data )
-{
-   struct texenvprog_cache_item *c
-      = (struct texenvprog_cache_item *) MALLOC(sizeof(*c));
-   c->hash = hash;
-
-   c->key = _mesa_malloc(sizeof(*key));
-   memcpy(c->key, key, sizeof(*key));
-
-   c->data = (struct gl_fragment_program *) data;
-
-   if (cache->n_items > cache->size * 1.5) {
-      if (cache->size < 1000)
-	 rehash(cache);
-      else 
-	 clear_cache(ctx, cache);
-   }
-
-   cache->n_items++;
-   c->next = cache->items[hash % cache->size];
-   cache->items[hash % cache->size] = c;
-}
-
-static GLuint hash_key( const struct state_key *key )
-{
-   GLuint *ikey = (GLuint *)key;
-   GLuint hash = 0, i;
-
-   /* Make a slightly better attempt at a hash function:
-    */
-   for (i = 0; i < sizeof(*key)/sizeof(*ikey); i++)
-   {
-      hash += ikey[i];
-      hash += (hash << 10);
-      hash ^= (hash >> 6);
-   }
-
-   return hash;
-}
-
-
 /**
  * Return a fragment program which implements the current
  * fixed-function texture, fog and color-sum operations.
@@ -1243,23 +1141,21 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
 {
    struct gl_fragment_program *prog;
    struct state_key key;
-   GLuint hash;
 	
    make_state_key(ctx, &key);
-   hash = hash_key(&key);
       
-   prog = search_cache(&ctx->Texture.env_fp_cache, hash, &key, sizeof(key));
+   prog = (struct gl_fragment_program *)
+      _mesa_search_program_cache(ctx->FragmentProgram.Cache,
+                                 &key, sizeof(key));
 
    if (!prog) {
-      if (0)
-         _mesa_printf("Building new texenv proggy for key %x\n", hash);
-
       prog = (struct gl_fragment_program *) 
          ctx->Driver.NewProgram(ctx, GL_FRAGMENT_PROGRAM_ARB, 0);
 
       create_new_program(ctx, &key, prog);
 
-      cache_item(ctx, &ctx->Texture.env_fp_cache, hash, &key, prog);
+      _mesa_program_cache_insert(ctx, ctx->FragmentProgram.Cache,
+                                 &key, sizeof(key), &prog->Base);
    }
 
    return prog;
@@ -1297,20 +1193,3 @@ _mesa_UpdateTexEnvProgram( GLcontext *ctx )
                          (struct gl_program *) ctx->FragmentProgram._Current);
    }
 }
-
-
-void _mesa_TexEnvProgramCacheInit( GLcontext *ctx )
-{
-   ctx->Texture.env_fp_cache.size = 17;
-   ctx->Texture.env_fp_cache.n_items = 0;
-   ctx->Texture.env_fp_cache.items = (struct texenvprog_cache_item **)
-      _mesa_calloc(ctx->Texture.env_fp_cache.size * 
-		   sizeof(struct texenvprog_cache_item));
-}
-
-
-void _mesa_TexEnvProgramCacheDestroy( GLcontext *ctx )
-{
-   clear_cache(ctx, &ctx->Texture.env_fp_cache);
-   _mesa_free(ctx->Texture.env_fp_cache.items);
-}
diff --git a/src/mesa/main/texenvprogram.h b/src/mesa/main/texenvprogram.h
index ac73910cb5..a7aa60cf37 100644
--- a/src/mesa/main/texenvprogram.h
+++ b/src/mesa/main/texenvprogram.h
@@ -1,13 +1,8 @@
-/**
- * \file texenvprogram.h
- * Texture state management.
- */
-
 /*
  * Mesa 3-D graphics library
- * Version:  5.1
+ * Version:  7.1
  *
- * Copyright (C) 1999-2002  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -38,7 +33,5 @@ extern struct gl_fragment_program *
 _mesa_get_fixed_func_fragment_program(GLcontext *ctx);
 
 extern void _mesa_UpdateTexEnvProgram( GLcontext *ctx );
-extern void _mesa_TexEnvProgramCacheInit( GLcontext *ctx );
-extern void _mesa_TexEnvProgramCacheDestroy( GLcontext *ctx );
 
 #endif
diff --git a/src/mesa/main/texstate.c b/src/mesa/main/texstate.c
index c9f8a0656e..cb7da39b51 100644
--- a/src/mesa/main/texstate.c
+++ b/src/mesa/main/texstate.c
@@ -3210,8 +3210,6 @@ _mesa_init_texture(GLcontext *ctx)
    ctx->Texture.SharedPalette = GL_FALSE;
    _mesa_init_colortable(&ctx->Texture.Palette);
 
-   _mesa_TexEnvProgramCacheInit( ctx );
-
    /* Allocate proxy textures */
    if (!alloc_proxy_textures( ctx ))
       return GL_FALSE;
@@ -3239,6 +3237,4 @@ _mesa_free_texture_data(GLcontext *ctx)
 
    for (i = 0; i < MAX_TEXTURE_IMAGE_UNITS; i++)
       _mesa_free_colortable_data( &ctx->Texture.Unit[i].ColorTable );
-
-   _mesa_TexEnvProgramCacheDestroy( ctx );
 }
-- 
cgit v1.2.3


From 1e3b07f363e6bf512ab1a5c620656985aece40fd Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Fri, 14 Dec 2007 11:16:49 -0700
Subject: set program->SamplersUsed bit when using a texture instruction

---
 src/mesa/main/texenvprogram.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index efb3b35f6a..1e46d8c375 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -936,10 +936,12 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 			  
       /* TODO: Use D0_MASK_XY where possible.
        */
-      if (p->state->unit[unit].enabled) 
+      if (p->state->unit[unit].enabled) {
 	 p->src_texture[unit] = emit_texld( p, OPCODE_TXP,
 					    tmp, WRITEMASK_XYZW, 
 					    unit, dim, texcoord );
+         p->program->Base.SamplersUsed |= (1 << unit);
+      }
       else
 	 p->src_texture[unit] = get_zero(p);
    }
-- 
cgit v1.2.3


From fce4612f8a29ee1798c9326a431a139d856c7a04 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Fri, 14 Dec 2007 11:42:28 -0700
Subject: set SamplerUnit[] entry in load_texture() just to be safe

---
 src/mesa/main/texenvprogram.c | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 1e46d8c375..d866d10017 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -941,6 +941,10 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 					    tmp, WRITEMASK_XYZW, 
 					    unit, dim, texcoord );
          p->program->Base.SamplersUsed |= (1 << unit);
+         /* This identity mapping should already be in place
+          * (see _mesa_init_program_struct()) but let's be safe.
+          */
+         p->program->Base.SamplerUnits[unit] = unit;
       }
       else
 	 p->src_texture[unit] = get_zero(p);
-- 
cgit v1.2.3


From fdc8636bdc65deb0d95a62a51c8d9bca05bc6bb8 Mon Sep 17 00:00:00 2001
From: Brian <brian.paul@tungstengraphics.com>
Date: Fri, 18 Jan 2008 12:45:27 -0700
Subject: use PROGRAM_CONSTANT instead of PROGRAM_STATE_VAR when generating
 immediates/literals

---
 src/mesa/main/ffvertex_prog.c | 2 +-
 src/mesa/main/texenvprogram.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/ffvertex_prog.c b/src/mesa/main/ffvertex_prog.c
index 4a9a0cd975..04ece4cda8 100644
--- a/src/mesa/main/ffvertex_prog.c
+++ b/src/mesa/main/ffvertex_prog.c
@@ -436,7 +436,7 @@ static struct ureg register_const4f( struct tnl_program *p,
    idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                      &swizzle );
    ASSERT(swizzle == SWIZZLE_NOOP);
-   return make_ureg(PROGRAM_STATE_VAR, idx);
+   return make_ureg(PROGRAM_CONSTANT, idx);
 }
 
 #define register_const1f(p, s0)         register_const4f(p, s0, 0, 0, 1)
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index d866d10017..644b1f39c7 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -582,7 +582,7 @@ static struct ureg register_const4f( struct texenv_fragment_program *p,
    idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                      &swizzle );
    ASSERT(swizzle == SWIZZLE_NOOP);
-   return make_ureg(PROGRAM_STATE_VAR, idx);
+   return make_ureg(PROGRAM_CONSTANT, idx);
 }
 
 #define register_scalar_const(p, s0)    register_const4f(p, s0, s0, s0, s0)
-- 
cgit v1.2.3


From 53174afeeb68a79e471185cb463c13ff90af698f Mon Sep 17 00:00:00 2001
From: José Fonseca <jrfonseca@tungstengraphics.com>
Date: Sat, 31 May 2008 18:14:09 +0900
Subject: mesa: Apply MSVC portability fixes from Alan Hourihane.

---
 src/mesa/main/api_arrayelt.c                   | 96 +++++++++++++-------------
 src/mesa/main/api_validate.c                   |  2 +-
 src/mesa/main/context.c                        |  2 +-
 src/mesa/main/dlist.c                          |  2 +-
 src/mesa/main/drawpix.c                        |  2 +-
 src/mesa/main/framebuffer.c                    |  2 +-
 src/mesa/main/get.c                            |  4 +-
 src/mesa/main/image.c                          |  6 +-
 src/mesa/main/light.c                          |  2 +-
 src/mesa/main/mipmap.c                         |  2 +-
 src/mesa/main/pixel.c                          |  4 +-
 src/mesa/main/queryobj.c                       |  4 +-
 src/mesa/main/rastpos.c                        |  2 +-
 src/mesa/main/texenvprogram.c                  | 10 +--
 src/mesa/shader/prog_execute.c                 |  6 +-
 src/mesa/shader/prog_statevars.c               | 10 +--
 src/mesa/shader/prog_uniform.c                 |  2 +-
 src/mesa/shader/shader_api.c                   |  6 +-
 src/mesa/state_tracker/st_atom_pixeltransfer.c |  2 +-
 src/mesa/state_tracker/st_atom_scissor.c       |  8 +--
 src/mesa/state_tracker/st_atom_viewport.c      | 12 ++--
 src/mesa/state_tracker/st_cb_bitmap.c          | 38 +++++-----
 src/mesa/vbo/vbo_exec_array.c                  |  2 +-
 src/mesa/vbo/vbo_split_inplace.c               |  2 +-
 24 files changed, 114 insertions(+), 114 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/api_arrayelt.c b/src/mesa/main/api_arrayelt.c
index 72091b0789..d124c724c9 100644
--- a/src/mesa/main/api_arrayelt.c
+++ b/src/mesa/main/api_arrayelt.c
@@ -166,7 +166,7 @@ static void GLAPIENTRY VertexAttrib1NbvNV(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib1bvNV(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NbvNV(GLuint index, const GLbyte *v)
@@ -176,7 +176,7 @@ static void GLAPIENTRY VertexAttrib2NbvNV(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib2bvNV(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NbvNV(GLuint index, const GLbyte *v)
@@ -188,7 +188,7 @@ static void GLAPIENTRY VertexAttrib3NbvNV(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib3bvNV(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NbvNV(GLuint index, const GLbyte *v)
@@ -201,7 +201,7 @@ static void GLAPIENTRY VertexAttrib4NbvNV(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib4bvNV(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_BYTE attributes */
@@ -213,7 +213,7 @@ static void GLAPIENTRY VertexAttrib1NubvNV(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib1ubvNV(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NubvNV(GLuint index, const GLubyte *v)
@@ -224,7 +224,7 @@ static void GLAPIENTRY VertexAttrib2NubvNV(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib2ubvNV(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NubvNV(GLuint index, const GLubyte *v)
@@ -235,7 +235,7 @@ static void GLAPIENTRY VertexAttrib3NubvNV(GLuint index, const GLubyte *v)
 }
 static void GLAPIENTRY VertexAttrib3ubvNV(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NubvNV(GLuint index, const GLubyte *v)
@@ -248,7 +248,7 @@ static void GLAPIENTRY VertexAttrib4NubvNV(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib4ubvNV(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_SHORT attributes */
@@ -260,7 +260,7 @@ static void GLAPIENTRY VertexAttrib1NsvNV(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib1svNV(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NsvNV(GLuint index, const GLshort *v)
@@ -271,7 +271,7 @@ static void GLAPIENTRY VertexAttrib2NsvNV(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib2svNV(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NsvNV(GLuint index, const GLshort *v)
@@ -283,7 +283,7 @@ static void GLAPIENTRY VertexAttrib3NsvNV(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib3svNV(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NsvNV(GLuint index, const GLshort *v)
@@ -296,7 +296,7 @@ static void GLAPIENTRY VertexAttrib4NsvNV(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib4svNV(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_SHORT attributes */
@@ -308,7 +308,7 @@ static void GLAPIENTRY VertexAttrib1NusvNV(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib1usvNV(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NusvNV(GLuint index, const GLushort *v)
@@ -319,7 +319,7 @@ static void GLAPIENTRY VertexAttrib2NusvNV(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib2usvNV(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NusvNV(GLuint index, const GLushort *v)
@@ -331,7 +331,7 @@ static void GLAPIENTRY VertexAttrib3NusvNV(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib3usvNV(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NusvNV(GLuint index, const GLushort *v)
@@ -344,7 +344,7 @@ static void GLAPIENTRY VertexAttrib4NusvNV(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib4usvNV(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_INT attributes */
@@ -356,7 +356,7 @@ static void GLAPIENTRY VertexAttrib1NivNV(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib1ivNV(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NivNV(GLuint index, const GLint *v)
@@ -367,7 +367,7 @@ static void GLAPIENTRY VertexAttrib2NivNV(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib2ivNV(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NivNV(GLuint index, const GLint *v)
@@ -379,7 +379,7 @@ static void GLAPIENTRY VertexAttrib3NivNV(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib3ivNV(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NivNV(GLuint index, const GLint *v)
@@ -392,7 +392,7 @@ static void GLAPIENTRY VertexAttrib4NivNV(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib4ivNV(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_INT attributes */
@@ -404,7 +404,7 @@ static void GLAPIENTRY VertexAttrib1NuivNV(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib1uivNV(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fNV(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NuivNV(GLuint index, const GLuint *v)
@@ -415,7 +415,7 @@ static void GLAPIENTRY VertexAttrib2NuivNV(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib2uivNV(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NuivNV(GLuint index, const GLuint *v)
@@ -427,7 +427,7 @@ static void GLAPIENTRY VertexAttrib3NuivNV(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib3uivNV(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NuivNV(GLuint index, const GLuint *v)
@@ -440,7 +440,7 @@ static void GLAPIENTRY VertexAttrib4NuivNV(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib4uivNV(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fNV(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_FLOAT attributes */
@@ -602,7 +602,7 @@ static void GLAPIENTRY VertexAttrib1NbvARB(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib1bvARB(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NbvARB(GLuint index, const GLbyte *v)
@@ -612,7 +612,7 @@ static void GLAPIENTRY VertexAttrib2NbvARB(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib2bvARB(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NbvARB(GLuint index, const GLbyte *v)
@@ -624,7 +624,7 @@ static void GLAPIENTRY VertexAttrib3NbvARB(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib3bvARB(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NbvARB(GLuint index, const GLbyte *v)
@@ -637,7 +637,7 @@ static void GLAPIENTRY VertexAttrib4NbvARB(GLuint index, const GLbyte *v)
 
 static void GLAPIENTRY VertexAttrib4bvARB(GLuint index, const GLbyte *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_BYTE attributes */
@@ -649,7 +649,7 @@ static void GLAPIENTRY VertexAttrib1NubvARB(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib1ubvARB(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NubvARB(GLuint index, const GLubyte *v)
@@ -660,7 +660,7 @@ static void GLAPIENTRY VertexAttrib2NubvARB(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib2ubvARB(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NubvARB(GLuint index, const GLubyte *v)
@@ -671,7 +671,7 @@ static void GLAPIENTRY VertexAttrib3NubvARB(GLuint index, const GLubyte *v)
 }
 static void GLAPIENTRY VertexAttrib3ubvARB(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NubvARB(GLuint index, const GLubyte *v)
@@ -684,7 +684,7 @@ static void GLAPIENTRY VertexAttrib4NubvARB(GLuint index, const GLubyte *v)
 
 static void GLAPIENTRY VertexAttrib4ubvARB(GLuint index, const GLubyte *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_SHORT attributes */
@@ -696,7 +696,7 @@ static void GLAPIENTRY VertexAttrib1NsvARB(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib1svARB(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NsvARB(GLuint index, const GLshort *v)
@@ -707,7 +707,7 @@ static void GLAPIENTRY VertexAttrib2NsvARB(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib2svARB(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NsvARB(GLuint index, const GLshort *v)
@@ -719,7 +719,7 @@ static void GLAPIENTRY VertexAttrib3NsvARB(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib3svARB(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NsvARB(GLuint index, const GLshort *v)
@@ -732,7 +732,7 @@ static void GLAPIENTRY VertexAttrib4NsvARB(GLuint index, const GLshort *v)
 
 static void GLAPIENTRY VertexAttrib4svARB(GLuint index, const GLshort *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_SHORT attributes */
@@ -744,7 +744,7 @@ static void GLAPIENTRY VertexAttrib1NusvARB(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib1usvARB(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NusvARB(GLuint index, const GLushort *v)
@@ -755,7 +755,7 @@ static void GLAPIENTRY VertexAttrib2NusvARB(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib2usvARB(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NusvARB(GLuint index, const GLushort *v)
@@ -767,7 +767,7 @@ static void GLAPIENTRY VertexAttrib3NusvARB(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib3usvARB(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NusvARB(GLuint index, const GLushort *v)
@@ -780,7 +780,7 @@ static void GLAPIENTRY VertexAttrib4NusvARB(GLuint index, const GLushort *v)
 
 static void GLAPIENTRY VertexAttrib4usvARB(GLuint index, const GLushort *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_INT attributes */
@@ -792,7 +792,7 @@ static void GLAPIENTRY VertexAttrib1NivARB(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib1ivARB(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NivARB(GLuint index, const GLint *v)
@@ -803,7 +803,7 @@ static void GLAPIENTRY VertexAttrib2NivARB(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib2ivARB(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NivARB(GLuint index, const GLint *v)
@@ -815,7 +815,7 @@ static void GLAPIENTRY VertexAttrib3NivARB(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib3ivARB(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NivARB(GLuint index, const GLint *v)
@@ -828,7 +828,7 @@ static void GLAPIENTRY VertexAttrib4NivARB(GLuint index, const GLint *v)
 
 static void GLAPIENTRY VertexAttrib4ivARB(GLuint index, const GLint *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_UNSIGNED_INT attributes */
@@ -840,7 +840,7 @@ static void GLAPIENTRY VertexAttrib1NuivARB(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib1uivARB(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, v[0]));
+   CALL_VertexAttrib1fARB(GET_DISPATCH(), (index, (GLfloat)v[0]));
 }
 
 static void GLAPIENTRY VertexAttrib2NuivARB(GLuint index, const GLuint *v)
@@ -851,7 +851,7 @@ static void GLAPIENTRY VertexAttrib2NuivARB(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib2uivARB(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, v[0], v[1]));
+   CALL_VertexAttrib2fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1]));
 }
 
 static void GLAPIENTRY VertexAttrib3NuivARB(GLuint index, const GLuint *v)
@@ -863,7 +863,7 @@ static void GLAPIENTRY VertexAttrib3NuivARB(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib3uivARB(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, v[0], v[1], v[2]));
+   CALL_VertexAttrib3fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2]));
 }
 
 static void GLAPIENTRY VertexAttrib4NuivARB(GLuint index, const GLuint *v)
@@ -876,7 +876,7 @@ static void GLAPIENTRY VertexAttrib4NuivARB(GLuint index, const GLuint *v)
 
 static void GLAPIENTRY VertexAttrib4uivARB(GLuint index, const GLuint *v)
 {
-   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, v[0], v[1], v[2], v[3]));
+   CALL_VertexAttrib4fARB(GET_DISPATCH(), (index, (GLfloat)v[0], (GLfloat)v[1], (GLfloat)v[2], (GLfloat)v[3]));
 }
 
 /* GL_FLOAT attributes */
diff --git a/src/mesa/main/api_validate.c b/src/mesa/main/api_validate.c
index 64ab324af2..9144b4bc66 100644
--- a/src/mesa/main/api_validate.c
+++ b/src/mesa/main/api_validate.c
@@ -87,7 +87,7 @@ _mesa_validate_DrawElements(GLcontext *ctx,
          indexBytes = count * sizeof(GLushort);
       }
 
-      if (indexBytes > ctx->Array.ElementArrayBufferObj->Size) {
+      if (indexBytes > (GLuint) ctx->Array.ElementArrayBufferObj->Size) {
          _mesa_warning(ctx, "glDrawElements index out of buffer bounds");
          return GL_FALSE;
       }
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 2158eb6873..0053180000 100755
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -1532,7 +1532,7 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
           * if the DRIdrawable changes, and everything relies on them.
           * This is a bit messy (same as needed in _mesa_BindFramebufferEXT)
 	  */
-            int i;
+            unsigned int i;
             GLenum buffers[MAX_DRAW_BUFFERS];
 
             _mesa_reference_framebuffer(&newCtx->DrawBuffer, drawBuffer);
diff --git a/src/mesa/main/dlist.c b/src/mesa/main/dlist.c
index 13ebd4dd0d..f933580b2d 100644
--- a/src/mesa/main/dlist.c
+++ b/src/mesa/main/dlist.c
@@ -5069,7 +5069,7 @@ save_Indexfv(const GLfloat * v)
 static void GLAPIENTRY
 save_EdgeFlag(GLboolean x)
 {
-   save_Attr1fNV(VERT_ATTRIB_EDGEFLAG, x ? 1.0 : 0.0);
+   save_Attr1fNV(VERT_ATTRIB_EDGEFLAG, x ? (GLfloat)1.0 : (GLfloat)0.0);
 }
 
 static void GLAPIENTRY
diff --git a/src/mesa/main/drawpix.c b/src/mesa/main/drawpix.c
index fa422bb3c7..016ddd0a81 100644
--- a/src/mesa/main/drawpix.c
+++ b/src/mesa/main/drawpix.c
@@ -377,7 +377,7 @@ _mesa_Bitmap( GLsizei width, GLsizei height,
 
    if (ctx->RenderMode == GL_RENDER) {
       /* Truncate, to satisfy conformance tests (matches SGI's OpenGL). */
-      const GLfloat epsilon = 0.0001;
+      const GLfloat epsilon = (const GLfloat)0.0001;
       GLint x = IFLOOR(ctx->Current.RasterPos[0] + epsilon - xorig);
       GLint y = IFLOOR(ctx->Current.RasterPos[1] + epsilon - yorig);
 
diff --git a/src/mesa/main/framebuffer.c b/src/mesa/main/framebuffer.c
index c9b30d3252..894d99afd2 100644
--- a/src/mesa/main/framebuffer.c
+++ b/src/mesa/main/framebuffer.c
@@ -67,7 +67,7 @@ compute_depth_max(struct gl_framebuffer *fb)
    fb->_DepthMaxF = (GLfloat) fb->_DepthMax;
 
    /* Minimum resolvable depth value, for polygon offset */
-   fb->_MRD = 1.0 / fb->_DepthMaxF;
+   fb->_MRD = (GLfloat)1.0 / fb->_DepthMaxF;
 }
 
 
diff --git a/src/mesa/main/get.c b/src/mesa/main/get.c
index eb81ee4a52..ee48b78318 100644
--- a/src/mesa/main/get.c
+++ b/src/mesa/main/get.c
@@ -2124,7 +2124,7 @@ _mesa_GetFloatv( GLenum pname, GLfloat *params )
          params[0] = (GLfloat)(ctx->DrawBuffer->Visual.depthBits);
          break;
       case GL_DEPTH_CLEAR_VALUE:
-         params[0] = ctx->Depth.Clear;
+         params[0] = (GLfloat)ctx->Depth.Clear;
          break;
       case GL_DEPTH_FUNC:
          params[0] = ENUM_TO_FLOAT(ctx->Depth.Func);
@@ -2914,7 +2914,7 @@ _mesa_GetFloatv( GLenum pname, GLfloat *params )
          GLuint i, n = _mesa_get_compressed_formats(ctx, formats, GL_FALSE);
          ASSERT(n <= 100);
          for (i = 0; i < n; i++)
-            params[i] = ENUM_TO_INT(formats[i]);
+            params[i] = (GLfloat)(ENUM_TO_INT(formats[i]));
          }
          break;
       case GL_ARRAY_ELEMENT_LOCK_FIRST_EXT:
diff --git a/src/mesa/main/image.c b/src/mesa/main/image.c
index 76e105e65e..285c8346a5 100644
--- a/src/mesa/main/image.c
+++ b/src/mesa/main/image.c
@@ -1169,7 +1169,7 @@ _mesa_apply_stencil_transfer_ops(const GLcontext *ctx, GLuint n,
       GLuint mask = ctx->PixelMaps.StoS.Size - 1;
       GLuint i;
       for (i = 0; i < n; i++) {
-         stencil[i] = ctx->PixelMaps.StoS.Map[ stencil[i] & mask ];
+         stencil[i] = (GLstencil)ctx->PixelMaps.StoS.Map[ stencil[i] & mask ];
       }
    }
 }
@@ -3680,7 +3680,7 @@ _mesa_unpack_stencil_span( const GLcontext *ctx, GLuint n,
          const GLuint mask = ctx->PixelMaps.StoS.Size - 1;
          GLuint i;
          for (i = 0; i < n; i++) {
-            indexes[i] = ctx->PixelMaps.StoS.Map[ indexes[i] & mask ];
+            indexes[i] = (GLuint)ctx->PixelMaps.StoS.Map[ indexes[i] & mask ];
          }
       }
 
@@ -4035,7 +4035,7 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
    if (needClamp) {
       GLuint i;
       for (i = 0; i < n; i++) {
-         depthValues[i] = CLAMP(depthValues[i], 0.0, 1.0);
+         depthValues[i] = (GLfloat)CLAMP(depthValues[i], 0.0, 1.0);
       }
    }
 
diff --git a/src/mesa/main/light.c b/src/mesa/main/light.c
index f9715b4865..f37d1f216f 100644
--- a/src/mesa/main/light.c
+++ b/src/mesa/main/light.c
@@ -1119,7 +1119,7 @@ compute_light_positions( GLcontext *ctx )
       }
       else {
          /* positional light w/ homogeneous coordinate, divide by W */
-         GLfloat wInv = 1.0 / light->_Position[3];
+         GLfloat wInv = (GLfloat)1.0 / light->_Position[3];
          light->_Position[0] *= wInv;
          light->_Position[1] *= wInv;
          light->_Position[2] *= wInv;
diff --git a/src/mesa/main/mipmap.c b/src/mesa/main/mipmap.c
index d3d1958951..061378f3b7 100644
--- a/src/mesa/main/mipmap.c
+++ b/src/mesa/main/mipmap.c
@@ -292,7 +292,7 @@ do_row(GLenum datatype, GLuint comps, GLint srcWidth,
       GLfloat *dst = (GLfloat *) dstRow;
       for (i = j = 0, k = k0; i < (GLuint) dstWidth;
            i++, j += colStride, k += colStride) {
-         dst[i] = rowA[j] / 4 + rowA[k] / 4 + rowB[j] / 4 + rowB[k] / 4;
+         dst[i] = (GLfloat)(rowA[j] / 4 + rowA[k] / 4 + rowB[j] / 4 + rowB[k] / 4);
       }
    }
 
diff --git a/src/mesa/main/pixel.c b/src/mesa/main/pixel.c
index 0e9915dd38..7eeae05dbd 100644
--- a/src/mesa/main/pixel.c
+++ b/src/mesa/main/pixel.c
@@ -304,7 +304,7 @@ store_pixelmap(GLcontext *ctx, GLenum map, GLsizei mapsize,
       /* special case */
       ctx->PixelMaps.StoS.Size = mapsize;
       for (i = 0; i < mapsize; i++) {
-         ctx->PixelMaps.StoS.Map[i] = IROUND(values[i]);
+         ctx->PixelMaps.StoS.Map[i] = (GLfloat)IROUND(values[i]);
       }
       break;
    case GL_PIXEL_MAP_I_TO_I:
@@ -1142,7 +1142,7 @@ _mesa_lookup_rgba_ubyte(const struct gl_color_table *table,
                         GLuint n, GLubyte rgba[][4])
 {
    const GLubyte *lut = table->TableUB;
-   const GLfloat scale = (GLfloat) (table->Size - 1) / 255.0;
+   const GLfloat scale = (GLfloat) (table->Size - 1) / (GLfloat)255.0;
    GLuint i;
 
    if (!table->TableUB || table->Size == 0)
diff --git a/src/mesa/main/queryobj.c b/src/mesa/main/queryobj.c
index e30f5480da..a1e32e70ba 100644
--- a/src/mesa/main/queryobj.c
+++ b/src/mesa/main/queryobj.c
@@ -382,7 +382,7 @@ _mesa_GetQueryObjectivARB(GLuint id, GLenum pname, GLint *params)
             *params = 0x7fffffff;
          }
          else {
-            *params = q->Result;
+            *params = (GLint)q->Result;
          }
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
@@ -422,7 +422,7 @@ _mesa_GetQueryObjectuivARB(GLuint id, GLenum pname, GLuint *params)
             *params = 0xffffffff;
          }
          else {
-            *params = q->Result;
+            *params = (GLuint)q->Result;
          }
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
diff --git a/src/mesa/main/rastpos.c b/src/mesa/main/rastpos.c
index ee163e0c71..f0500083ec 100644
--- a/src/mesa/main/rastpos.c
+++ b/src/mesa/main/rastpos.c
@@ -63,7 +63,7 @@ rasterpos(GLfloat x, GLfloat y, GLfloat z, GLfloat w)
 void GLAPIENTRY
 _mesa_RasterPos2d(GLdouble x, GLdouble y)
 {
-   rasterpos(x, y, 0.0F, 1.0F);
+   rasterpos((GLfloat)x, (GLfloat)y, (GLfloat)0.0, (GLfloat)1.0);
 }
 
 void GLAPIENTRY
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 644b1f39c7..6877ef96f2 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -904,14 +904,14 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
     */
    if (alpha_shift || rgb_shift) {
       if (rgb_shift == alpha_shift) {
-	 shift = register_scalar_const(p, 1<<rgb_shift);
+	 shift = register_scalar_const(p, (GLfloat)(1<<rgb_shift));
       }
       else {
 	 shift = register_const4f(p, 
-				  1<<rgb_shift,
-				  1<<rgb_shift,
-				  1<<rgb_shift,
-				  1<<alpha_shift);
+				  (GLfloat)(1<<rgb_shift),
+				  (GLfloat)(1<<rgb_shift),
+				  (GLfloat)(1<<rgb_shift),
+				  (GLfloat)(1<<alpha_shift));
       }
       return emit_arith( p, OPCODE_MUL, dest, WRITEMASK_XYZW, 
 			 saturate, out, shift, undef );
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index 8ce2ca3964..abb143c4b9 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -92,7 +92,7 @@ get_register_pointer(const struct prog_src_register *source,
                 source->File == PROGRAM_CONSTANT ||
                 source->File == PROGRAM_STATE_VAR);
          params = machine->CurProgram->Parameters;
-         if (reg < 0 || reg >= params->NumParameters)
+         if (reg < 0 || reg >= (GLint)params->NumParameters)
             return ZeroVec;
          else
             return params->ParameterValues[reg];
@@ -227,7 +227,7 @@ fetch_vector4_deriv(GLcontext * ctx,
                     const struct gl_program_machine *machine,
                     char xOrY, GLfloat result[4])
 {
-   if (source->File == PROGRAM_INPUT && source->Index < 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;
@@ -506,7 +506,7 @@ _mesa_execute_program(GLcontext * ctx,
 {
    const GLuint numInst = program->NumInstructions;
    const GLuint maxExec = 10000;
-   GLint pc, numExec = 0;
+   GLuint pc, numExec = 0;
 
    machine->CurProgram = program;
 
diff --git a/src/mesa/shader/prog_statevars.c b/src/mesa/shader/prog_statevars.c
index 44fbfdcd04..8f48155825 100644
--- a/src/mesa/shader/prog_statevars.c
+++ b/src/mesa/shader/prog_statevars.c
@@ -250,7 +250,7 @@ _mesa_fetch_state(GLcontext *ctx, const gl_state_index state[],
       value[1] = ctx->Fog.Start;
       value[2] = ctx->Fog.End;
       value[3] = (ctx->Fog.End == ctx->Fog.Start)
-         ? 1.0 : 1.0F / (ctx->Fog.End - ctx->Fog.Start);
+         ? 1.0 : (GLfloat)(1.0 / (ctx->Fog.End - ctx->Fog.Start));
       return;
    case STATE_CLIPPLANE:
       {
@@ -411,7 +411,7 @@ _mesa_fetch_state(GLcontext *ctx, const gl_state_index state[],
             if (texObj) {
                struct gl_texture_image *texImage = texObj->Image[0][0];
                ASSIGN_4V(value, 1.0 / texImage->Width,
-                         1.0 / texImage->Height,
+                         (GLfloat)(1.0 / texImage->Height),
                          0.0, 1.0);
             }
          }
@@ -426,10 +426,10 @@ _mesa_fetch_state(GLcontext *ctx, const gl_state_index state[],
           * exp2: 2^-((density/(ln(2)^2) * fogcoord)^2)
           */
          value[0] = (ctx->Fog.End == ctx->Fog.Start)
-            ? 1.0 : -1.0F / (ctx->Fog.End - ctx->Fog.Start);
+            ? 1.0 : (GLfloat)(-1.0F / (ctx->Fog.End - ctx->Fog.Start));
          value[1] = ctx->Fog.End * -value[0];
-         value[2] = ctx->Fog.Density * ONE_DIV_LN2;
-         value[3] = ctx->Fog.Density * ONE_DIV_SQRT_LN2;
+         value[2] = (GLfloat)(ctx->Fog.Density * ONE_DIV_LN2);
+         value[3] = (GLfloat)(ctx->Fog.Density * ONE_DIV_SQRT_LN2);
          return;
 
       case STATE_LIGHT_SPOT_DIR_NORMALIZED: {
diff --git a/src/mesa/shader/prog_uniform.c b/src/mesa/shader/prog_uniform.c
index 20e004b350..d96a916533 100644
--- a/src/mesa/shader/prog_uniform.c
+++ b/src/mesa/shader/prog_uniform.c
@@ -135,7 +135,7 @@ _mesa_longest_uniform_name(const struct gl_uniform_list *list)
    GLuint i;
    for (i = 0; i < list->NumUniforms; i++) {
       GLuint len = _mesa_strlen(list->Uniforms[i].Name);
-      if (len > max)
+      if (len > (GLuint)max)
          max = len;
    }
    return max;
diff --git a/src/mesa/shader/shader_api.c b/src/mesa/shader/shader_api.c
index 24ab7568d6..856179e1d5 100644
--- a/src/mesa/shader/shader_api.c
+++ b/src/mesa/shader/shader_api.c
@@ -719,7 +719,7 @@ _mesa_get_attached_shaders(GLcontext *ctx, GLuint program, GLsizei maxCount,
    struct gl_shader_program *shProg
       = _mesa_lookup_shader_program(ctx, program);
    if (shProg) {
-      GLint i;
+      GLuint i;
       for (i = 0; i < maxCount && i < shProg->NumShaders; i++) {
          obj[i] = shProg->Shaders[i]->Name;
       }
@@ -893,7 +893,7 @@ _mesa_get_uniformfv(GLcontext *ctx, GLuint program, GLint location,
       = _mesa_lookup_shader_program(ctx, program);
    if (shProg) {
       if (location < shProg->Uniforms->NumUniforms) {
-         GLint progPos, i;
+         GLuint progPos, i;
          const struct gl_program *prog = NULL;
 
          progPos = shProg->Uniforms->Uniforms[location].VertPos;
@@ -1111,7 +1111,7 @@ set_program_uniform(GLcontext *ctx, struct gl_program *program, GLint location,
    }
    else {
       /* ordinary uniform variable */
-      GLint k, i;
+      GLuint k, i;
 
       if (count * elems > program->Parameters->Parameters[location].Size) {
          _mesa_error(ctx, GL_INVALID_OPERATION, "glUniform(count too large)");
diff --git a/src/mesa/state_tracker/st_atom_pixeltransfer.c b/src/mesa/state_tracker/st_atom_pixeltransfer.c
index e500ac8684..e4de875e8c 100644
--- a/src/mesa/state_tracker/st_atom_pixeltransfer.c
+++ b/src/mesa/state_tracker/st_atom_pixeltransfer.c
@@ -77,7 +77,7 @@ is_identity(const GLfloat m[16])
    GLuint i;
    for (i = 0; i < 16; i++) {
       const int row = i % 4, col = i / 4;
-      const float val = (row == col);
+      const float val = (GLfloat)(row == col);
       if (m[i] != val)
          return GL_FALSE;
    }
diff --git a/src/mesa/state_tracker/st_atom_scissor.c b/src/mesa/state_tracker/st_atom_scissor.c
index f5db492403..3fd59e1945 100644
--- a/src/mesa/state_tracker/st_atom_scissor.c
+++ b/src/mesa/state_tracker/st_atom_scissor.c
@@ -52,14 +52,14 @@ update_scissor( struct st_context *st )
    scissor.maxy = fb->Height;
 
    if (st->ctx->Scissor.Enabled) {
-      if (st->ctx->Scissor.X > scissor.minx)
+      if ((GLuint)st->ctx->Scissor.X > scissor.minx)
          scissor.minx = st->ctx->Scissor.X;
-      if (st->ctx->Scissor.Y > scissor.miny)
+      if ((GLuint)st->ctx->Scissor.Y > scissor.miny)
          scissor.miny = st->ctx->Scissor.Y;
 
-      if (st->ctx->Scissor.X + st->ctx->Scissor.Width < scissor.maxx)
+      if ((GLuint)st->ctx->Scissor.X + st->ctx->Scissor.Width < scissor.maxx)
          scissor.maxx = st->ctx->Scissor.X + st->ctx->Scissor.Width;
-      if (st->ctx->Scissor.Y + st->ctx->Scissor.Height < scissor.maxy)
+      if ((GLuint)st->ctx->Scissor.Y + st->ctx->Scissor.Height < scissor.maxy)
          scissor.maxy = st->ctx->Scissor.Y + st->ctx->Scissor.Height;
 
       /* check for null space */
diff --git a/src/mesa/state_tracker/st_atom_viewport.c b/src/mesa/state_tracker/st_atom_viewport.c
index 4b51521470..b105909e96 100644
--- a/src/mesa/state_tracker/st_atom_viewport.c
+++ b/src/mesa/state_tracker/st_atom_viewport.c
@@ -49,7 +49,7 @@ update_viewport( struct st_context *st )
     */
    if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) {
       yScale = -1;
-      yBias = ctx->DrawBuffer->Height ;
+      yBias = (GLfloat)ctx->DrawBuffer->Height;
    }
    else {
       yScale = 1.0;
@@ -59,12 +59,12 @@ update_viewport( struct st_context *st )
    /* _NEW_VIEWPORT 
     */
    {
-      GLfloat x = ctx->Viewport.X;
-      GLfloat y = ctx->Viewport.Y;
+      GLfloat x = (GLfloat)ctx->Viewport.X;
+      GLfloat y = (GLfloat)ctx->Viewport.Y;
       GLfloat z = ctx->Viewport.Near;
-      GLfloat half_width = ctx->Viewport.Width / 2.0;
-      GLfloat half_height = ctx->Viewport.Height / 2.0;
-      GLfloat half_depth = (ctx->Viewport.Far - ctx->Viewport.Near) / 2.0;
+      GLfloat half_width = (GLfloat)ctx->Viewport.Width / 2.0;
+      GLfloat half_height = (GLfloat)ctx->Viewport.Height / 2.0;
+      GLfloat half_depth = (GLfloat)(ctx->Viewport.Far - ctx->Viewport.Near) / 2.0;
       
       st->state.viewport.scale[0] = half_width;
       st->state.viewport.scale[1] = half_height * yScale;
diff --git a/src/mesa/state_tracker/st_cb_bitmap.c b/src/mesa/state_tracker/st_cb_bitmap.c
index 593938f8cf..9763eebe54 100644
--- a/src/mesa/state_tracker/st_cb_bitmap.c
+++ b/src/mesa/state_tracker/st_cb_bitmap.c
@@ -360,18 +360,18 @@ setup_bitmap_vertex_data(struct st_context *st,
 {
    struct pipe_context *pipe = st->pipe;
    const struct gl_framebuffer *fb = st->ctx->DrawBuffer;
-   const GLfloat fb_width = fb->Width;
-   const GLfloat fb_height = fb->Height;
-   const GLfloat x0 = x;
-   const GLfloat x1 = x + width;
-   const GLfloat y0 = y;
-   const GLfloat y1 = y + height;
-   const GLfloat sLeft = 0.0F, sRight = 1.0F;
-   const GLfloat tTop = 0.0, tBot = 1.0 - tTop;
-   const GLfloat clip_x0 = x0 / fb_width * 2.0 - 1.0;
-   const GLfloat clip_y0 = y0 / fb_height * 2.0 - 1.0;
-   const GLfloat clip_x1 = x1 / fb_width * 2.0 - 1.0;
-   const GLfloat clip_y1 = y1 / fb_height * 2.0 - 1.0;
+   const GLfloat fb_width = (GLfloat)fb->Width;
+   const GLfloat fb_height = (GLfloat)fb->Height;
+   const GLfloat x0 = (GLfloat)x;
+   const GLfloat x1 = (GLfloat)(x + width);
+   const GLfloat y0 = (GLfloat)y;
+   const GLfloat y1 = (GLfloat)(y + height);
+   const GLfloat sLeft = (GLfloat)0.0, sRight = (GLfloat)1.0;
+   const GLfloat tTop = (GLfloat)0.0, tBot = (GLfloat)1.0 - tTop;
+   const GLfloat clip_x0 = (GLfloat)(x0 / fb_width * 2.0 - 1.0);
+   const GLfloat clip_y0 = (GLfloat)(y0 / fb_height * 2.0 - 1.0);
+   const GLfloat clip_x1 = (GLfloat)(x1 / fb_width * 2.0 - 1.0);
+   const GLfloat clip_y1 = (GLfloat)(y1 / fb_height * 2.0 - 1.0);
    GLuint i;
    void *buf;
 
@@ -444,8 +444,8 @@ draw_bitmap_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
     * it up into chunks.
     */
    maxSize = 1 << (pipe->screen->get_param(pipe->screen, PIPE_CAP_MAX_TEXTURE_2D_LEVELS) - 1);
-   assert(width <= maxSize);
-   assert(height <= maxSize);
+   assert(width <= (GLsizei)maxSize);
+   assert(height <= (GLsizei)maxSize);
 
    cso_save_rasterizer(cso);
    cso_save_samplers(cso);
@@ -488,15 +488,15 @@ draw_bitmap_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
    {
       const struct gl_framebuffer *fb = st->ctx->DrawBuffer;
       const GLboolean invert = (st_fb_orientation(fb) == Y_0_TOP);
-      const float width = fb->Width;
-      const float height = fb->Height;
+      const GLfloat width = (GLfloat)fb->Width;
+      const GLfloat height = (GLfloat)fb->Height;
       struct pipe_viewport_state vp;
       vp.scale[0] =  0.5 * width;
-      vp.scale[1] = height * (invert ? -0.5 : 0.5);
+      vp.scale[1] = (GLfloat)(height * (invert ? -0.5 : 0.5));
       vp.scale[2] = 1.0;
       vp.scale[3] = 1.0;
-      vp.translate[0] = 0.5 * width;
-      vp.translate[1] = 0.5 * height;
+      vp.translate[0] = (GLfloat)(0.5 * width);
+      vp.translate[1] = (GLfloat)(0.5 * height);
       vp.translate[2] = 0.0;
       vp.translate[3] = 0.0;
       cso_set_viewport(cso, &vp);
diff --git a/src/mesa/vbo/vbo_exec_array.c b/src/mesa/vbo/vbo_exec_array.c
index a52521db64..dbee2188ee 100644
--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -41,7 +41,7 @@ static void get_minmax_index( GLuint count, GLuint type,
 			      GLuint *min_index,
 			      GLuint *max_index)
 {
-   GLint i;
+   GLuint i;
 
    switch(type) {
    case GL_UNSIGNED_INT: {
diff --git a/src/mesa/vbo/vbo_split_inplace.c b/src/mesa/vbo/vbo_split_inplace.c
index 958afccd0c..fbc856e93b 100644
--- a/src/mesa/vbo/vbo_split_inplace.c
+++ b/src/mesa/vbo/vbo_split_inplace.c
@@ -58,7 +58,7 @@ struct split_context {
 
 static void flush_vertex( struct split_context *split )
 {
-   GLint min_index, max_index;
+   GLuint min_index, max_index;
 
    if (!split->dstprim_nr) 
       return;
-- 
cgit v1.2.3


From af74abab6b9a1e32a1cc5cac7e547b953dcee0ab Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Fri, 12 Sep 2008 10:04:56 +0100
Subject: mesa: get fixed-function program generation working again

---
 src/mesa/main/state.c         | 62 +++++++++++++++++++++++++++++--------------
 src/mesa/main/texenvprogram.c |  3 +++
 2 files changed, 45 insertions(+), 20 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 5827f2211c..e340cd6686 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -219,23 +219,6 @@ update_program(GLcontext *ctx)
                                shProg->FragmentProgram);
    }
    else {
-      if (ctx->VertexProgram._Enabled) {
-         /* use user-defined vertex program */
-         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current,
-                                  ctx->VertexProgram.Current);
-      }
-      else if (ctx->VertexProgram._MaintainTnlProgram) {
-         /* Use vertex program generated from fixed-function state.
-          * The _Current pointer will get set in
-          * _tnl_UpdateFixedFunctionProgram() later if appropriate.
-          */
-         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current, NULL);
-      }
-      else {
-         /* no vertex program */
-         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current, NULL);
-      }
-
       if (ctx->FragmentProgram._Enabled) {
          /* use user-defined vertex program */
          _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._Current,
@@ -243,15 +226,38 @@ update_program(GLcontext *ctx)
       }
       else if (ctx->FragmentProgram._MaintainTexEnvProgram) {
          /* Use fragment program generated from fixed-function state.
-          * The _Current pointer will get set in _mesa_UpdateTexEnvProgram()
-          * later if appropriate.
           */
-         _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._Current, NULL);
+         _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._Current,
+                                  _mesa_get_fixed_func_fragment_program(ctx));
+         _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._TexEnvProgram,
+                                  ctx->FragmentProgram._Current);
       }
       else {
          /* no fragment program */
          _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._Current, NULL);
       }
+
+      /* Examine vertex program after fragment program as
+       * _mesa_get_fixed_func_vertex_program() needs to know active
+       * fragprog inputs.
+       */
+      if (ctx->VertexProgram._Enabled) {
+         /* use user-defined vertex program */
+         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current,
+                                  ctx->VertexProgram.Current);
+      }
+      else if (ctx->VertexProgram._MaintainTnlProgram) {
+         /* Use vertex program generated from fixed-function state.
+          */
+         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current,
+                                  _mesa_get_fixed_func_vertex_program(ctx));
+         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._TnlProgram,
+                                  ctx->VertexProgram._Current);
+      }
+      else {
+         /* no vertex program */
+         _mesa_reference_vertprog(ctx, &ctx->VertexProgram._Current, NULL);
+      }
    }
 
    if (ctx->VertexProgram._Current)
@@ -260,12 +266,28 @@ update_program(GLcontext *ctx)
       assert(ctx->FragmentProgram._Current->Base.Parameters);
 
 
+   /* XXX: get rid of _Active flag.
+    */
+#if 1
    ctx->FragmentProgram._Active = ctx->FragmentProgram._Enabled;
    if (ctx->FragmentProgram._MaintainTexEnvProgram &&
        !ctx->FragmentProgram._Enabled) {
       if (ctx->FragmentProgram._UseTexEnvProgram)
 	 ctx->FragmentProgram._Active = GL_TRUE;
    }
+#endif
+
+   /* Let the driver know what's happening:
+    */
+   if (ctx->FragmentProgram._Current != prevFP && ctx->Driver.BindProgram) {
+      ctx->Driver.BindProgram(ctx, GL_FRAGMENT_PROGRAM_ARB,
+                              (struct gl_program *) ctx->FragmentProgram._Current);
+   }
+   
+   if (ctx->VertexProgram._Current != prevVP && ctx->Driver.BindProgram) {
+      ctx->Driver.BindProgram(ctx, GL_VERTEX_PROGRAM_ARB,
+                              (struct gl_program *) ctx->VertexProgram._Current);
+   }
 }
 
 
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 713ff3d256..25e280bc3f 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1190,6 +1190,9 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
  * If _MaintainTexEnvProgram is set we'll generate a fragment program that
  * implements the current texture env/combine mode.
  * This function generates that program and puts it into effect.
+ *
+ * XXX: remove this function.  currently only called by some drivers,
+ * not by mesa core.  We now handle this properly from inside mesa.
  */
 void
 _mesa_UpdateTexEnvProgram( GLcontext *ctx )
-- 
cgit v1.2.3


From 006fb638188f083d64a2427cd28979b432622f3e Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Thu, 25 Sep 2008 18:27:22 -0600
Subject: mesa: fix swizzle failure, fix typo

---
 src/mesa/main/texenvprogram.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 6877ef96f2..c699c43429 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -375,7 +375,7 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 {
    int bit;
    
-   /* First try to find availble temp not previously used (to avoid
+   /* First try to find available temp not previously used (to avoid
     * starting a new texture indirection).  According to the spec, the
     * ~p->temps_output isn't necessary, but will keep it there for
     * now:
@@ -575,14 +575,16 @@ static struct ureg register_const4f( struct texenv_fragment_program *p,
 {
    GLfloat values[4];
    GLuint idx, swizzle;
+   struct ureg r;
    values[0] = s0;
    values[1] = s1;
    values[2] = s2;
    values[3] = s3;
    idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                      &swizzle );
-   ASSERT(swizzle == SWIZZLE_NOOP);
-   return make_ureg(PROGRAM_CONSTANT, idx);
+   r = make_ureg(PROGRAM_CONSTANT, idx);
+   r.swz = swizzle;
+   return r;
 }
 
 #define register_scalar_const(p, s0)    register_const4f(p, s0, s0, s0, s0)
-- 
cgit v1.2.3


From 3f99f501db2582e241851e63e432c18e2de415be Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Thu, 25 Sep 2008 18:40:16 -0600
Subject: mesa: increase MAX_INSTRUCTIONS

---
 src/mesa/main/texenvprogram.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index c699c43429..c64d88faf9 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -37,11 +37,9 @@
 #include "texenvprogram.h"
 
 /**
- * This MAX is probably a bit generous, but that's OK.  There can be
- * up to four instructions per texture unit (TEX + 3 for combine),
- * then there's fog and specular add.
+ * Up to nine instructions per tex unit, plus fog, specular color.
  */
-#define MAX_INSTRUCTIONS ((MAX_TEXTURE_UNITS * 4) + 12)
+#define MAX_INSTRUCTIONS ((MAX_TEXTURE_UNITS * 9) + 12)
 
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
-- 
cgit v1.2.3


From d01269a57f4cfdb859352c933bc546296545dd80 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Thu, 25 Sep 2008 18:27:22 -0600
Subject: mesa: fix swizzle failure, fix typo

---
 src/mesa/main/texenvprogram.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 25e280bc3f..4eb743736c 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -388,7 +388,7 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 {
    int bit;
    
-   /* First try to find availble temp not previously used (to avoid
+   /* First try to find available temp not previously used (to avoid
     * starting a new texture indirection).  According to the spec, the
     * ~p->temps_output isn't necessary, but will keep it there for
     * now:
@@ -588,14 +588,16 @@ static struct ureg register_const4f( struct texenv_fragment_program *p,
 {
    GLfloat values[4];
    GLuint idx, swizzle;
+   struct ureg r;
    values[0] = s0;
    values[1] = s1;
    values[2] = s2;
    values[3] = s3;
    idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                      &swizzle );
-   ASSERT(swizzle == SWIZZLE_NOOP);
-   return make_ureg(PROGRAM_CONSTANT, idx);
+   r = make_ureg(PROGRAM_CONSTANT, idx);
+   r.swz = swizzle;
+   return r;
 }
 
 #define register_scalar_const(p, s0)    register_const4f(p, s0, s0, s0, s0)
-- 
cgit v1.2.3


From b5e1a93036b22bd30738abccbb8a2645a515667f Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Thu, 25 Sep 2008 18:40:16 -0600
Subject: mesa: increase MAX_INSTRUCTIONS

---
 src/mesa/main/texenvprogram.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 4eb743736c..2bce93eef1 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -48,11 +48,9 @@ struct texenvprog_cache_item
 
 
 /**
- * This MAX is probably a bit generous, but that's OK.  There can be
- * up to four instructions per texture unit (TEX + 3 for combine),
- * then there's fog and specular add.
+ * Up to nine instructions per tex unit, plus fog, specular color.
  */
-#define MAX_INSTRUCTIONS ((MAX_TEXTURE_UNITS * 4) + 12)
+#define MAX_INSTRUCTIONS ((MAX_TEXTURE_UNITS * 9) + 12)
 
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
-- 
cgit v1.2.3


From 8fd329d04885eba7587bbe7604d3a1088e35de40 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Fri, 26 Sep 2008 11:18:06 -0600
Subject: mesa: fix temp register allocation problems.

Complex texcombine modes were running out of registers (>32 registers for 8 tex units).
---
 src/mesa/main/texenvprogram.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index c64d88faf9..97aa87e58c 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -398,6 +398,14 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 }
 
 
+/** Mark a temp reg as being no longer allocatable. */
+static void reserve_temp( struct texenv_fragment_program *p, struct ureg r )
+{
+   if (r.file == PROGRAM_TEMPORARY)
+      p->temps_output |= (1 << r.idx);
+}
+
+
 static void release_temps(GLcontext *ctx, struct texenv_fragment_program *p )
 {
    GLuint max_temp = ctx->Const.FragmentProgram.MaxTemps;
@@ -491,10 +499,12 @@ emit_op(struct texenv_fragment_program *p,
 
    emit_dst( &inst->DstReg, dest, mask );
 
+#if 0
    /* Accounting for indirection tracking:
     */
    if (dest.file == PROGRAM_TEMPORARY)
       p->temps_output |= 1 << dest.idx;
+#endif
 
    return inst;
 }
@@ -549,6 +559,10 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
 
    p->program->Base.NumTexInstructions++;
 
+   /* Accounting for indirection tracking:
+    */
+   reserve_temp(p, dest);
+
    /* Is this a texture indirection?
     */
    if ((coord.file == PROGRAM_TEMPORARY &&
@@ -1062,6 +1076,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
       for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++)
 	 if (key->enabled_units & (1<<unit)) {
 	    p.src_previous = emit_texenv( &p, unit );
+            reserve_temp(&p, p.src_previous); /* don't re-use this temp reg */
 	    release_temps(ctx, &p);	/* release all temps */
 	 }
    }
-- 
cgit v1.2.3


From 5620c20b24dc4f780a2246eb5270c4476b487e0a Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Fri, 26 Sep 2008 11:18:06 -0600
Subject: mesa: fix temp register allocation problems.

Complex texcombine modes were running out of registers (>32 registers for 8 tex units).
---
 src/mesa/main/texenvprogram.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 2bce93eef1..f6bbbcfaed 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -411,6 +411,14 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 }
 
 
+/** Mark a temp reg as being no longer allocatable. */
+static void reserve_temp( struct texenv_fragment_program *p, struct ureg r )
+{
+   if (r.file == PROGRAM_TEMPORARY)
+      p->temps_output |= (1 << r.idx);
+}
+
+
 static void release_temps(GLcontext *ctx, struct texenv_fragment_program *p )
 {
    GLuint max_temp = ctx->Const.FragmentProgram.MaxTemps;
@@ -504,10 +512,12 @@ emit_op(struct texenv_fragment_program *p,
 
    emit_dst( &inst->DstReg, dest, mask );
 
+#if 0
    /* Accounting for indirection tracking:
     */
    if (dest.file == PROGRAM_TEMPORARY)
       p->temps_output |= 1 << dest.idx;
+#endif
 
    return inst;
 }
@@ -562,6 +572,10 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
 
    p->program->Base.NumTexInstructions++;
 
+   /* Accounting for indirection tracking:
+    */
+   reserve_temp(p, dest);
+
    /* Is this a texture indirection?
     */
    if ((coord.file == PROGRAM_TEMPORARY &&
@@ -1079,6 +1093,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
       for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++)
 	 if (key->enabled_units & (1<<unit)) {
 	    p.src_previous = emit_texenv( &p, unit );
+            reserve_temp(&p, p.src_previous); /* don't re-use this temp reg */
 	    release_temps(ctx, &p);	/* release all temps */
 	 }
    }
-- 
cgit v1.2.3


From 6280e335706f95ed0ebb089d8f72aeede9b5a1ad Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Fri, 3 Oct 2008 13:51:56 +0100
Subject: mesa: shrink texenvprogram state key struct

---
 src/mesa/main/texenvprogram.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 97aa87e58c..ac49373604 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -44,15 +44,17 @@
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
 struct mode_opt {
-   GLuint Source:4;
-   GLuint Operand:3;
+   GLubyte Source:4;
+   GLubyte Operand:3;
 };
 
 struct state_key {
-   GLbitfield enabled_units;
+   GLuint nr_enabled_units:8;
+   GLuint enabled_units:8;
    GLuint separate_specular:1;
    GLuint fog_enabled:1;
    GLuint fog_mode:2;
+   GLuint inputs_available:12;
 
    struct {
       GLuint enabled:1;
@@ -62,10 +64,10 @@ struct state_key {
 
       GLuint NumArgsRGB:2;
       GLuint ModeRGB:4;
-      struct mode_opt OptRGB[3];
-
       GLuint NumArgsA:2;
       GLuint ModeA:4;
+
+      struct mode_opt OptRGB[3];
       struct mode_opt OptA[3];
    } unit[8];
 };
-- 
cgit v1.2.3


From 1680ef869625dc1fe9cf481b180382a34e0738e7 Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Fri, 3 Oct 2008 17:30:59 +0100
Subject: mesa: avoid generating constant vertex attributes in fixedfunc
 programs

Keep track of enabled/active vertex attributes.
Keep track of potential vertex program outputs.

When generating fragment program, replace references to fragment attributes
which are effectively non-varying and non-computed passthrough attributes with
references to the new CURRENT_ATTRIB tracked state value.

Only downside is slight ugliness in VBO code where we need to validate state
twice in succession.
---
 src/mesa/main/mtypes.h        |  2 +
 src/mesa/main/state.c         | 38 ++++++++++++++++-
 src/mesa/main/state.h         |  3 ++
 src/mesa/main/texenvprogram.c | 94 ++++++++++++++++++++++++++++++++++++++++---
 src/mesa/vbo/vbo_exec_array.c | 41 +++++++++++++++----
 src/mesa/vbo/vbo_exec_draw.c  |  8 ++++
 src/mesa/vbo/vbo_save_draw.c  |  4 ++
 7 files changed, 177 insertions(+), 13 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index bc099dabeb..ca1e369a35 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -3073,6 +3073,8 @@ struct __GLcontextRec
    GLenum RenderMode;        /**< either GL_RENDER, GL_SELECT, GL_FEEDBACK */
    GLbitfield NewState;      /**< bitwise-or of _NEW_* flags */
 
+   GLuint varying_vp_inputs;
+
    /** \name Derived state */
    /*@{*/
    GLbitfield _TriangleCaps;      /**< bitwise-or of DD_* flags */
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index eb8dc2a339..e0eb5f81e2 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -465,7 +465,8 @@ _mesa_update_state_locked( GLcontext *ctx )
       _mesa_update_tnl_spaces( ctx, new_state );
 
    if (ctx->FragmentProgram._MaintainTexEnvProgram) {
-      prog_flags |= (_NEW_TEXTURE | _NEW_FOG | _DD_NEW_SEPARATE_SPECULAR);
+      prog_flags |= (_NEW_ARRAY | _NEW_TEXTURE_MATRIX | _NEW_LIGHT |
+                     _NEW_TEXTURE | _NEW_FOG | _DD_NEW_SEPARATE_SPECULAR);
    }
    if (ctx->VertexProgram._MaintainTnlProgram) {
       prog_flags |= (_NEW_ARRAY | _NEW_TEXTURE | _NEW_TEXTURE_MATRIX |
@@ -504,3 +505,38 @@ _mesa_update_state( GLcontext *ctx )
    _mesa_update_state_locked(ctx);
    _mesa_unlock_context_textures(ctx);
 }
+
+
+
+
+/* Want to figure out which fragment program inputs are actually
+ * constant/current values from ctx->Current.  These should be
+ * referenced as a tracked state variable rather than a fragment
+ * program input, to save the overhead of putting a constant value in
+ * every submitted vertex, transferring it to hardware, interpolating
+ * it across the triangle, etc...
+ *
+ * When there is a VP bound, just use vp->outputs.  But when we're
+ * generating vp from fixed function state, basically want to
+ * calculate:
+ *
+ * vp_out_2_fp_in( vp_in_2_vp_out( varying_inputs ) | 
+ *                 potential_vp_outputs )
+ *
+ * Where potential_vp_outputs is calculated by looking at enabled
+ * texgen, etc.
+ * 
+ * The generated fragment program should then only declare inputs that
+ * may vary or otherwise differ from the ctx->Current values.
+ * Otherwise, the fp should track them as state values instead.
+ */
+void
+_mesa_set_varying_vp_inputs( GLcontext *ctx,
+                             unsigned varying_inputs )
+{
+   if (ctx->varying_vp_inputs != varying_inputs) {
+      ctx->varying_vp_inputs = varying_inputs;
+      ctx->NewState |= _NEW_ARRAY;
+      //_mesa_printf("%s %x\n", __FUNCTION__, varying_inputs);
+   }
+}
diff --git a/src/mesa/main/state.h b/src/mesa/main/state.h
index bb7cb8f32a..dc08043a76 100644
--- a/src/mesa/main/state.h
+++ b/src/mesa/main/state.h
@@ -37,5 +37,8 @@ _mesa_update_state( GLcontext *ctx );
 extern void
 _mesa_update_state_locked( GLcontext *ctx );
 
+void
+_mesa_set_varying_vp_inputs( GLcontext *ctx,
+                             unsigned varying_inputs );
 
 #endif
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index ac49373604..7cd82f98b0 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -189,6 +189,63 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
    }
 }
 
+#define VERT_BIT_TEX_ANY    (0xff << VERT_ATTRIB_TEX0)
+#define VERT_RESULT_TEX_ANY (0xff << VERT_RESULT_TEX0)
+
+/* Identify all possible varying inputs.  The fragment program will
+ * never reference non-varying inputs, but will track them via state
+ * constants instead.
+ *
+ * This function figures out all the inputs that the fragment program
+ * has access to.  The bitmask is later reduced to just those which
+ * are actually referenced.
+ */
+static GLuint get_fp_input_mask( GLcontext *ctx )
+{
+   GLuint fp_inputs = 0;
+
+   if (1) {
+      GLuint varying_inputs = ctx->varying_vp_inputs;
+
+      /* First look at what values may be computed by the generated
+       * vertex program:
+       */
+      if (ctx->Light.Enabled) {
+         fp_inputs |= FRAG_BIT_COL0;
+
+         if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR)
+            fp_inputs |= FRAG_BIT_COL1;
+      }
+
+      fp_inputs |= (ctx->Texture._TexGenEnabled |
+                    ctx->Texture._TexMatEnabled) << FRAG_ATTRIB_TEX0;
+
+      /* Then look at what might be varying as a result of enabled
+       * arrays, etc:
+       */
+      if (varying_inputs & VERT_BIT_COLOR0) fp_inputs |= FRAG_BIT_COL0;
+      if (varying_inputs & VERT_BIT_COLOR1) fp_inputs |= FRAG_BIT_COL1;
+
+      fp_inputs |= (((varying_inputs & VERT_BIT_TEX_ANY) >> VERT_ATTRIB_TEX0) 
+                    << FRAG_ATTRIB_TEX0);
+
+   }
+   else {
+      /* calculate from vp->outputs */
+      GLuint vp_outputs = 0;
+
+      if (vp_outputs & (1 << VERT_RESULT_COL0)) fp_inputs |= FRAG_BIT_COL0;
+      if (vp_outputs & (1 << VERT_RESULT_COL1)) fp_inputs |= FRAG_BIT_COL1;
+
+      fp_inputs |= (((vp_outputs & VERT_RESULT_TEX_ANY) 
+                   << VERT_RESULT_TEX0) 
+                  >> FRAG_ATTRIB_TEX0);
+   }
+   
+   return fp_inputs;
+}
+
+
 /**
  * Examine current texture environment state and generate a unique
  * key to identify it.
@@ -196,17 +253,21 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
 static void make_state_key( GLcontext *ctx,  struct state_key *key )
 {
    GLuint i, j;
-	
+   GLuint inputs_referenced = FRAG_BIT_COL0;
+   GLuint inputs_available = get_fp_input_mask( ctx );
+
    memset(key, 0, sizeof(*key));
 
    for (i=0;i<MAX_TEXTURE_UNITS;i++) {
       const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i];
 		
-      if (!texUnit->_ReallyEnabled)
+      if (!texUnit->_ReallyEnabled) 
          continue;
 
       key->unit[i].enabled = 1;
       key->enabled_units |= (1<<i);
+      key->nr_enabled_units = i+1;
+      inputs_referenced |= FRAG_BIT_TEX(i);
 
       key->unit[i].source_index = 
 	 translate_tex_src_bit(texUnit->_ReallyEnabled);		
@@ -234,13 +295,18 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
       }
    }
 	
-   if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR)
+   if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) {
       key->separate_specular = 1;
+      inputs_referenced |= FRAG_BIT_COL1;
+   }
 
    if (ctx->Fog.Enabled) {
       key->fog_enabled = 1;
       key->fog_mode = translate_fog_mode(ctx->Fog.Mode);
+      inputs_referenced |= FRAG_BIT_FOGC; /* maybe */
    }
+
+   key->inputs_available = (inputs_available & inputs_referenced);
 }
 
 /* Use uregs to represent registers internally, translate to Mesa's
@@ -446,11 +512,29 @@ static struct ureg register_param5( struct texenv_fragment_program *p,
 #define register_param3(p,s0,s1,s2)    register_param5(p,s0,s1,s2,0,0)
 #define register_param4(p,s0,s1,s2,s3) register_param5(p,s0,s1,s2,s3,0)
 
+static GLuint frag_to_vert_attrib( GLuint attrib )
+{
+   switch (attrib) {
+   case FRAG_ATTRIB_COL0: return VERT_ATTRIB_COLOR0;
+   case FRAG_ATTRIB_COL1: return VERT_ATTRIB_COLOR1;
+   default:
+      assert(attrib >= FRAG_ATTRIB_TEX0);
+      assert(attrib <= FRAG_ATTRIB_TEX7);
+      return attrib - FRAG_ATTRIB_TEX0 + VERT_ATTRIB_TEX0;
+   }
+}
+
 
 static struct ureg register_input( struct texenv_fragment_program *p, GLuint input )
 {
-   p->program->Base.InputsRead |= (1 << input);
-   return make_ureg(PROGRAM_INPUT, input);
+   if (p->state->inputs_available & (1<<input)) {
+      p->program->Base.InputsRead |= (1 << input);
+      return make_ureg(PROGRAM_INPUT, input);
+   }
+   else {
+      GLuint idx = frag_to_vert_attrib( input );
+      return register_param3( p, STATE_INTERNAL, STATE_CURRENT_ATTRIB, idx );
+   }
 }
 
 
diff --git a/src/mesa/vbo/vbo_exec_array.c b/src/mesa/vbo/vbo_exec_array.c
index 0f9d8da356..3d74f9f431 100644
--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -127,6 +127,7 @@ static void recalculate_input_bindings( GLcontext *ctx )
    struct vbo_context *vbo = vbo_context(ctx);
    struct vbo_exec_context *exec = &vbo->exec;
    const struct gl_client_array **inputs = &exec->array.inputs[0];
+   GLuint const_inputs = 0;
    GLuint i;
 
    exec->array.program_mode = get_program_mode(ctx);
@@ -141,19 +142,24 @@ static void recalculate_input_bindings( GLcontext *ctx )
       for (i = 0; i <= VERT_ATTRIB_TEX7; i++) {
 	 if (exec->array.legacy_array[i]->Enabled)
 	    inputs[i] = exec->array.legacy_array[i];
-	 else
+	 else {
 	    inputs[i] = &vbo->legacy_currval[i];
+            const_inputs |= 1 << i;
+         }
       }
 
       for (i = 0; i < MAT_ATTRIB_MAX; i++) {
 	 inputs[VERT_ATTRIB_GENERIC0 + i] = &vbo->mat_currval[i];
+         const_inputs |= 1 << (VERT_ATTRIB_GENERIC0 + i);
       }
 
       /* Could use just about anything, just to fill in the empty
        * slots:
        */
-      for (i = MAT_ATTRIB_MAX; i < VERT_ATTRIB_MAX - VERT_ATTRIB_GENERIC0; i++)
+      for (i = MAT_ATTRIB_MAX; i < VERT_ATTRIB_MAX - VERT_ATTRIB_GENERIC0; i++) {
 	 inputs[VERT_ATTRIB_GENERIC0 + i] = &vbo->generic_currval[i];
+         const_inputs |= 1 << (VERT_ATTRIB_GENERIC0 + i);
+      }
 
       break;
    case VP_NV:
@@ -166,15 +172,19 @@ static void recalculate_input_bindings( GLcontext *ctx )
 	    inputs[i] = exec->array.generic_array[i];
 	 else if (exec->array.legacy_array[i]->Enabled)
 	    inputs[i] = exec->array.legacy_array[i];
-	 else
+	 else {
 	    inputs[i] = &vbo->legacy_currval[i];
+            const_inputs |= 1 << i;
+         }
       }
 
       /* Could use just about anything, just to fill in the empty
        * slots:
        */
-      for (i = VERT_ATTRIB_GENERIC0; i < VERT_ATTRIB_MAX; i++)
+      for (i = VERT_ATTRIB_GENERIC0; i < VERT_ATTRIB_MAX; i++) {
 	 inputs[i] = &vbo->generic_currval[i - VERT_ATTRIB_GENERIC0];
+         const_inputs |= 1 << i;
+      }
 
       break;
    case VP_ARB:
@@ -189,25 +199,34 @@ static void recalculate_input_bindings( GLcontext *ctx )
 	 inputs[0] = exec->array.generic_array[0];
       else if (exec->array.legacy_array[0]->Enabled)
 	 inputs[0] = exec->array.legacy_array[0];
-      else
+      else {
 	 inputs[0] = &vbo->legacy_currval[0];
+         const_inputs |= 1 << 0;
+      }
 
 
       for (i = 1; i <= VERT_ATTRIB_TEX7; i++) {
 	 if (exec->array.legacy_array[i]->Enabled)
 	    inputs[i] = exec->array.legacy_array[i];
-	 else
+	 else {
 	    inputs[i] = &vbo->legacy_currval[i];
+            const_inputs |= 1 << i;
+         }
       }
 
       for (i = 0; i < 16; i++) {
 	 if (exec->array.generic_array[i]->Enabled)
 	    inputs[VERT_ATTRIB_GENERIC0 + i] = exec->array.generic_array[i];
-	 else
+	 else {
 	    inputs[VERT_ATTRIB_GENERIC0 + i] = &vbo->generic_currval[i];
+            const_inputs |= 1 << (VERT_ATTRIB_GENERIC0 + i);
+         }
+
       }
       break;
    }
+
+   _mesa_set_varying_vp_inputs( ctx, ~const_inputs );
 }
 
 static void bind_arrays( GLcontext *ctx )
@@ -257,6 +276,11 @@ vbo_exec_DrawArrays(GLenum mode, GLint start, GLsizei count)
 
    bind_arrays( ctx );
 
+   /* Again...
+    */
+   if (ctx->NewState)
+      _mesa_update_state( ctx );
+
    prim[0].begin = 1;
    prim[0].end = 1;
    prim[0].weak = 0;
@@ -297,6 +321,9 @@ vbo_exec_DrawRangeElements(GLenum mode,
 
    bind_arrays( ctx );
 
+   if (ctx->NewState)
+      _mesa_update_state( ctx );
+
    ib.count = count;
    ib.type = type; 
    ib.obj = ctx->Array.ElementArrayBufferObj;
diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index f497e9a5a5..ad60c9b05f 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -150,6 +150,7 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
    GLubyte *data = exec->vtx.buffer_map;
    const GLuint *map;
    GLuint attr;
+   GLuint varying_inputs = 0;
 
    /* Install the default (ie Current) attributes first, then overlay
     * all active ones.
@@ -211,8 +212,11 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
 	 arrays[attr]._MaxElement = count; /* ??? */
 
 	 data += exec->vtx.attrsz[src] * sizeof(GLfloat);
+         varying_inputs |= 1<<attr;
       }
    }
+
+   _mesa_set_varying_vp_inputs( ctx, varying_inputs );
 }
 
 
@@ -242,6 +246,10 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 	  */
 	 vbo_exec_bind_arrays( ctx );
 
+         if (ctx->NewState)
+            _mesa_update_state( ctx );
+
+
 	 ctx->Driver.UnmapBuffer(ctx, target, exec->vtx.bufferobj);
 	 exec->vtx.buffer_map = NULL;
 
diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
index 4c97acddb9..b015bf2786 100644
--- a/src/mesa/vbo/vbo_save_draw.c
+++ b/src/mesa/vbo/vbo_save_draw.c
@@ -118,6 +118,7 @@ static void vbo_bind_vertex_list( GLcontext *ctx,
    GLuint data = node->buffer_offset;
    const GLuint *map;
    GLuint attr;
+   GLuint varying_inputs = 0;
 
    /* Install the default (ie Current) attributes first, then overlay
     * all active ones.
@@ -167,8 +168,11 @@ static void vbo_bind_vertex_list( GLcontext *ctx,
 	 assert(arrays[attr].BufferObj->Name);
 
 	 data += node->attrsz[src] * sizeof(GLfloat);
+         varying_inputs |= 1<<attr;
       }
    }
+
+   _mesa_set_varying_vp_inputs( ctx, varying_inputs );
 }
 
 static void vbo_save_loopback_vertex_list( GLcontext *ctx,
-- 
cgit v1.2.3


From 0370d6b359016790c6b879c2a4b6661adac20dea Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Sat, 4 Oct 2008 12:41:56 +0100
Subject: mesa: handle vertex program enabled case also in texenvprogram.c

---
 src/mesa/main/texenvprogram.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 7cd82f98b0..ea2ee160e4 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -204,7 +204,7 @@ static GLuint get_fp_input_mask( GLcontext *ctx )
 {
    GLuint fp_inputs = 0;
 
-   if (1) {
+   if (!ctx->VertexProgram._Enabled) {
       GLuint varying_inputs = ctx->varying_vp_inputs;
 
       /* First look at what values may be computed by the generated
@@ -232,14 +232,13 @@ static GLuint get_fp_input_mask( GLcontext *ctx )
    }
    else {
       /* calculate from vp->outputs */
-      GLuint vp_outputs = 0;
+      GLuint vp_outputs = ctx->VertexProgram._Current->Base.OutputsWritten;
 
       if (vp_outputs & (1 << VERT_RESULT_COL0)) fp_inputs |= FRAG_BIT_COL0;
       if (vp_outputs & (1 << VERT_RESULT_COL1)) fp_inputs |= FRAG_BIT_COL1;
 
-      fp_inputs |= (((vp_outputs & VERT_RESULT_TEX_ANY) 
-                   << VERT_RESULT_TEX0) 
-                  >> FRAG_ATTRIB_TEX0);
+      fp_inputs |= (((vp_outputs & VERT_RESULT_TEX_ANY) >> VERT_RESULT_TEX0) 
+                    << FRAG_ATTRIB_TEX0);
    }
    
    return fp_inputs;
-- 
cgit v1.2.3


From 6ff1cf5b82488dc5a07513b0806c23e70f7a665e Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Tue, 7 Oct 2008 12:31:31 +0100
Subject: mesa: protect against segfault in get_fp_input_mask()

---
 src/mesa/main/texenvprogram.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index ea2ee160e4..7049467c22 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -204,7 +204,10 @@ static GLuint get_fp_input_mask( GLcontext *ctx )
 {
    GLuint fp_inputs = 0;
 
-   if (!ctx->VertexProgram._Enabled) {
+   if (!ctx->VertexProgram._Enabled ||
+       !ctx->VertexProgram._Current) {
+
+      /* Fixed function logic */
       GLuint varying_inputs = ctx->varying_vp_inputs;
 
       /* First look at what values may be computed by the generated
-- 
cgit v1.2.3


From 239617fbe22d4dd7b2794510a6665f09602b5adf Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Tue, 7 Oct 2008 11:22:47 -0600
Subject: mesa: replace GLuint with GLbitfield to be clearer about usage

Also, fix up some comments to be doxygen style.
---
 src/mesa/main/mtypes.h        |  2 +-
 src/mesa/main/state.c         |  2 +-
 src/mesa/main/state.h         |  2 +-
 src/mesa/main/texenvprogram.c | 30 ++++++++++++++++--------------
 src/mesa/vbo/vbo_exec_array.c |  2 +-
 src/mesa/vbo/vbo_exec_draw.c  |  2 +-
 src/mesa/vbo/vbo_save_draw.c  |  2 +-
 7 files changed, 22 insertions(+), 20 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index ca1e369a35..dff474d6d0 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -3073,7 +3073,7 @@ struct __GLcontextRec
    GLenum RenderMode;        /**< either GL_RENDER, GL_SELECT, GL_FEEDBACK */
    GLbitfield NewState;      /**< bitwise-or of _NEW_* flags */
 
-   GLuint varying_vp_inputs;
+   GLbitfield varying_vp_inputs;  /**< mask of VERT_BIT_* flags */
 
    /** \name Derived state */
    /*@{*/
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index e0eb5f81e2..b124d48269 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -532,7 +532,7 @@ _mesa_update_state( GLcontext *ctx )
  */
 void
 _mesa_set_varying_vp_inputs( GLcontext *ctx,
-                             unsigned varying_inputs )
+                             GLbitfield varying_inputs )
 {
    if (ctx->varying_vp_inputs != varying_inputs) {
       ctx->varying_vp_inputs = varying_inputs;
diff --git a/src/mesa/main/state.h b/src/mesa/main/state.h
index dc08043a76..79f2f6beb0 100644
--- a/src/mesa/main/state.h
+++ b/src/mesa/main/state.h
@@ -39,6 +39,6 @@ _mesa_update_state_locked( GLcontext *ctx );
 
 void
 _mesa_set_varying_vp_inputs( GLcontext *ctx,
-                             unsigned varying_inputs );
+                             GLbitfield varying_inputs );
 
 #endif
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 7049467c22..638d6be5ad 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -192,7 +192,8 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
 #define VERT_BIT_TEX_ANY    (0xff << VERT_ATTRIB_TEX0)
 #define VERT_RESULT_TEX_ANY (0xff << VERT_RESULT_TEX0)
 
-/* Identify all possible varying inputs.  The fragment program will
+/**
+ * Identify all possible varying inputs.  The fragment program will
  * never reference non-varying inputs, but will track them via state
  * constants instead.
  *
@@ -200,15 +201,15 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
  * has access to.  The bitmask is later reduced to just those which
  * are actually referenced.
  */
-static GLuint get_fp_input_mask( GLcontext *ctx )
+static GLbitfield get_fp_input_mask( GLcontext *ctx )
 {
-   GLuint fp_inputs = 0;
+   GLbitfield fp_inputs = 0x0;
 
    if (!ctx->VertexProgram._Enabled ||
        !ctx->VertexProgram._Current) {
 
       /* Fixed function logic */
-      GLuint varying_inputs = ctx->varying_vp_inputs;
+      GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
       /* First look at what values may be computed by the generated
        * vertex program:
@@ -235,7 +236,7 @@ static GLuint get_fp_input_mask( GLcontext *ctx )
    }
    else {
       /* calculate from vp->outputs */
-      GLuint vp_outputs = ctx->VertexProgram._Current->Base.OutputsWritten;
+      GLbitfield vp_outputs = ctx->VertexProgram._Current->Base.OutputsWritten;
 
       if (vp_outputs & (1 << VERT_RESULT_COL0)) fp_inputs |= FRAG_BIT_COL0;
       if (vp_outputs & (1 << VERT_RESULT_COL1)) fp_inputs |= FRAG_BIT_COL1;
@@ -255,8 +256,8 @@ static GLuint get_fp_input_mask( GLcontext *ctx )
 static void make_state_key( GLcontext *ctx,  struct state_key *key )
 {
    GLuint i, j;
-   GLuint inputs_referenced = FRAG_BIT_COL0;
-   GLuint inputs_available = get_fp_input_mask( ctx );
+   GLbitfield inputs_referenced = FRAG_BIT_COL0;
+   GLbitfield inputs_available = get_fp_input_mask( ctx );
 
    memset(key, 0, sizeof(*key));
 
@@ -311,7 +312,8 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
    key->inputs_available = (inputs_available & inputs_referenced);
 }
 
-/* Use uregs to represent registers internally, translate to Mesa's
+/**
+ * Use uregs to represent registers internally, translate to Mesa's
  * expected formats on emit.  
  *
  * NOTE: These are passed by value extensively in this file rather
@@ -344,16 +346,16 @@ static const struct ureg undef = {
 };
 
 
-/* State used to build the fragment program:
+/** State used to build the fragment program:
  */
 struct texenv_fragment_program {
    struct gl_fragment_program *program;
    GLcontext *ctx;
    struct state_key *state;
 
-   GLbitfield alu_temps;	/* Track texture indirections, see spec. */
-   GLbitfield temps_output;	/* Track texture indirections, see spec. */
-   GLbitfield temp_in_use;	/* Tracks temporary regs which are in use. */
+   GLbitfield alu_temps;	/**< Track texture indirections, see spec. */
+   GLbitfield temps_output;	/**< Track texture indirections, see spec. */
+   GLbitfield temp_in_use;	/**< Tracks temporary regs which are in use. */
    GLboolean error;
 
    struct ureg src_texture[MAX_TEXTURE_UNITS];   
@@ -361,11 +363,11 @@ struct texenv_fragment_program {
     * else undef.
     */
 
-   struct ureg src_previous;	/* Reg containing color from previous 
+   struct ureg src_previous;	/**< Reg containing color from previous 
 				 * stage.  May need to be decl'd.
 				 */
 
-   GLuint last_tex_stage;	/* Number of last enabled texture unit */
+   GLuint last_tex_stage;	/**< Number of last enabled texture unit */
 
    struct ureg half;
    struct ureg one;
diff --git a/src/mesa/vbo/vbo_exec_array.c b/src/mesa/vbo/vbo_exec_array.c
index 3d74f9f431..8871e10cf6 100644
--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -127,7 +127,7 @@ static void recalculate_input_bindings( GLcontext *ctx )
    struct vbo_context *vbo = vbo_context(ctx);
    struct vbo_exec_context *exec = &vbo->exec;
    const struct gl_client_array **inputs = &exec->array.inputs[0];
-   GLuint const_inputs = 0;
+   GLbitfield const_inputs = 0x0;
    GLuint i;
 
    exec->array.program_mode = get_program_mode(ctx);
diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index ad60c9b05f..ae43857c8a 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -150,7 +150,7 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
    GLubyte *data = exec->vtx.buffer_map;
    const GLuint *map;
    GLuint attr;
-   GLuint varying_inputs = 0;
+   GLbitfield varying_inputs = 0x0;
 
    /* Install the default (ie Current) attributes first, then overlay
     * all active ones.
diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
index 68f3a965a5..0488c5d718 100644
--- a/src/mesa/vbo/vbo_save_draw.c
+++ b/src/mesa/vbo/vbo_save_draw.c
@@ -118,7 +118,7 @@ static void vbo_bind_vertex_list( GLcontext *ctx,
    GLuint data = node->buffer_offset;
    const GLuint *map;
    GLuint attr;
-   GLuint varying_inputs = 0;
+   GLbitfield varying_inputs = 0x0;
 
    /* Install the default (ie Current) attributes first, then overlay
     * all active ones.
-- 
cgit v1.2.3


From 6d4d51d647c27288aa625560bc080231099c0b01 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Fri, 10 Oct 2008 13:39:14 -0600
Subject: mesa: new _mesa_set_vp_override() function for driver-override of
 vertex program

Patch provide by Keith.
Used in state tracker by DrawPixels to indicate that the state tracker (driver)
is using its own vertex program.  This prevents the texenvprogram code from
replacing conventional shader inputs with state vars.
Fixes glDraw/CopyPixels regressions.
---
 src/mesa/main/mtypes.h                    |  2 ++
 src/mesa/main/state.c                     | 29 ++++++++++++++++++++++++++---
 src/mesa/main/state.h                     | 17 +++++++++++------
 src/mesa/main/texenvprogram.c             | 14 ++++++++++++--
 src/mesa/state_tracker/st_cb_drawpixels.c |  6 ++++++
 5 files changed, 57 insertions(+), 11 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index dff474d6d0..fab6ad05ee 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -2007,6 +2007,8 @@ struct gl_vertex_program_state
    GLboolean CallbackEnabled;
    GLuint CurrentPosition;
 #endif
+
+   GLboolean _Overriden;
 };
 
 
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index b124d48269..a962f1cb41 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  7.1
+ * Version:  7.3
  *
  * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
  *
@@ -466,10 +466,12 @@ _mesa_update_state_locked( GLcontext *ctx )
 
    if (ctx->FragmentProgram._MaintainTexEnvProgram) {
       prog_flags |= (_NEW_ARRAY | _NEW_TEXTURE_MATRIX | _NEW_LIGHT |
+                     _NEW_RENDERMODE |
                      _NEW_TEXTURE | _NEW_FOG | _DD_NEW_SEPARATE_SPECULAR);
    }
    if (ctx->VertexProgram._MaintainTnlProgram) {
       prog_flags |= (_NEW_ARRAY | _NEW_TEXTURE | _NEW_TEXTURE_MATRIX |
+                     _NEW_RENDERMODE |
                      _NEW_TRANSFORM | _NEW_POINT |
                      _NEW_FOG | _NEW_LIGHT |
                      _MESA_NEW_NEED_EYE_COORDS);
@@ -509,7 +511,8 @@ _mesa_update_state( GLcontext *ctx )
 
 
 
-/* Want to figure out which fragment program inputs are actually
+/**
+ * Want to figure out which fragment program inputs are actually
  * constant/current values from ctx->Current.  These should be
  * referenced as a tracked state variable rather than a fragment
  * program input, to save the overhead of putting a constant value in
@@ -537,6 +540,26 @@ _mesa_set_varying_vp_inputs( GLcontext *ctx,
    if (ctx->varying_vp_inputs != varying_inputs) {
       ctx->varying_vp_inputs = varying_inputs;
       ctx->NewState |= _NEW_ARRAY;
-      //_mesa_printf("%s %x\n", __FUNCTION__, varying_inputs);
+      /*_mesa_printf("%s %x\n", __FUNCTION__, varying_inputs);*/
+   }
+}
+
+
+/**
+ * Used by drivers to tell core Mesa that the driver is going to
+ * install/ use its own vertex program.  In particular, this will
+ * prevent generated fragment programs from using state vars instead
+ * of ordinary varyings/inputs.
+ */
+void
+_mesa_set_vp_override(GLcontext *ctx, GLboolean flag)
+{
+   if (ctx->VertexProgram._Overriden != flag) {
+      ctx->VertexProgram._Overriden = flag;
+
+      /* Set one of the bits which will trigger fragment program
+       * regeneration:
+       */
+      ctx->NewState |= _NEW_ARRAY; 
    }
 }
diff --git a/src/mesa/main/state.h b/src/mesa/main/state.h
index 79f2f6beb0..29db08a0b9 100644
--- a/src/mesa/main/state.h
+++ b/src/mesa/main/state.h
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  7.1
+ * Version:  7.3
  *
  * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
  *
@@ -29,16 +29,21 @@
 #include "mtypes.h"
 
 extern void
-_mesa_update_state( GLcontext *ctx );
+_mesa_update_state(GLcontext *ctx);
 
 /* As above but can only be called between _mesa_lock_context_textures() and 
  * _mesa_unlock_context_textures().
  */
 extern void
-_mesa_update_state_locked( GLcontext *ctx );
+_mesa_update_state_locked(GLcontext *ctx);
+
+
+extern void
+_mesa_set_varying_vp_inputs(GLcontext *ctx, GLbitfield varying_inputs);
+
+
+extern void
+_mesa_set_vp_override(GLcontext *ctx, GLboolean flag);
 
-void
-_mesa_set_varying_vp_inputs( GLcontext *ctx,
-                             GLbitfield varying_inputs );
 
 #endif
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 638d6be5ad..f3bac86dfe 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -205,8 +205,18 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
 {
    GLbitfield fp_inputs = 0x0;
 
-   if (!ctx->VertexProgram._Enabled ||
-       !ctx->VertexProgram._Current) {
+   if (ctx->VertexProgram._Overriden) {
+      /* Somebody's messing with the vertex program and we don't have
+       * a clue what's happening.  Assume that it could be producing
+       * all possible outputs.
+       */
+      fp_inputs = ~0;
+   }
+   else if (ctx->RenderMode == GL_FEEDBACK) {
+      fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
+   }
+   else if (!ctx->VertexProgram._Enabled ||
+            !ctx->VertexProgram._Current) {
 
       /* Fixed function logic */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
diff --git a/src/mesa/state_tracker/st_cb_drawpixels.c b/src/mesa/state_tracker/st_cb_drawpixels.c
index 00bbcae32a..5b24b9f068 100644
--- a/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -35,6 +35,7 @@
 #include "main/bufferobj.h"
 #include "main/macros.h"
 #include "main/texformat.h"
+#include "main/state.h"
 #include "shader/program.h"
 #include "shader/prog_parameter.h"
 #include "shader/prog_print.h"
@@ -835,6 +836,9 @@ st_DrawPixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
       return;
    }
 
+   _mesa_set_vp_override( ctx, TRUE );
+   _mesa_update_state( ctx );
+
    st_validate_state(st);
 
    if (format == GL_DEPTH_COMPONENT) {
@@ -874,6 +878,8 @@ st_DrawPixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
       /* blit */
       draw_blit(st, width, height, format, type, pixels);
    }
+
+   _mesa_set_vp_override( ctx, FALSE );
 }
 
 
-- 
cgit v1.2.3


From 97e63437dc216c7fdb25220655ecbf26042cfec8 Mon Sep 17 00:00:00 2001
From: Keith Whitwell <keith@tungstengraphics.com>
Date: Mon, 20 Oct 2008 13:03:45 +0100
Subject: mesa: note that texcoords are generated by setup routines when
 pointsprite enabled

---
 src/mesa/main/texenvprogram.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index f3bac86dfe..c279956f2a 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -221,6 +221,12 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       /* Fixed function logic */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
+      /* These get generated in the setup routine regardless of the
+       * vertex program:
+       */
+      if (ctx->Point.PointSprite)
+         varying_inputs |= FRAG_BITS_TEX_ANY;
+
       /* First look at what values may be computed by the generated
        * vertex program:
        */
@@ -248,6 +254,12 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       /* calculate from vp->outputs */
       GLbitfield vp_outputs = ctx->VertexProgram._Current->Base.OutputsWritten;
 
+      /* These get generated in the setup routine regardless of the
+       * vertex program:
+       */
+      if (ctx->Point.PointSprite)
+         vp_outputs |= FRAG_BITS_TEX_ANY;
+
       if (vp_outputs & (1 << VERT_RESULT_COL0)) fp_inputs |= FRAG_BIT_COL0;
       if (vp_outputs & (1 << VERT_RESULT_COL1)) fp_inputs |= FRAG_BIT_COL1;
 
-- 
cgit v1.2.3


From 8dc88cb64305c591dfadded2b5acbb1e6b04cd7f Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Wed, 17 Dec 2008 10:55:40 -0700
Subject: mesa: fix vertex program test in get_fp_input_mask()

We were accidentally using the fixed-function logic when a vertex shader was
being used.
---
 src/mesa/main/texenvprogram.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 3cb103f51f..e160787eb9 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -227,9 +227,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
    else if (ctx->RenderMode == GL_FEEDBACK) {
       fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
    }
-   else if (!ctx->VertexProgram._Enabled ||
-            !ctx->VertexProgram._Current) {
-
+   else if (!ctx->VertexProgram._Current) {
       /* Fixed function logic */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
-- 
cgit v1.2.3


From d7296a1a8e846bc4d41ded1c2406b6f5c658188a Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Wed, 17 Dec 2008 11:29:06 -0700
Subject: Revert "mesa: fix vertex program test in get_fp_input_mask()"

This reverts commit 8dc88cb64305c591dfadded2b5acbb1e6b04cd7f.

This change broke other things...
---
 src/mesa/main/texenvprogram.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index e160787eb9..3cb103f51f 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -227,7 +227,9 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
    else if (ctx->RenderMode == GL_FEEDBACK) {
       fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
    }
-   else if (!ctx->VertexProgram._Current) {
+   else if (!ctx->VertexProgram._Enabled ||
+            !ctx->VertexProgram._Current) {
+
       /* Fixed function logic */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
-- 
cgit v1.2.3


From 1519b93b7bc519e187d98f99715a01ba866286b1 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Wed, 17 Dec 2008 13:17:15 -0700
Subject: mesa: add missing cases for texture array targets

---
 src/mesa/main/texenvprogram.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 3cb103f51f..1560e97c26 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -191,13 +191,17 @@ static GLuint translate_mode( GLenum mode )
 #define TEXTURE_UNKNOWN_INDEX 7
 static GLuint translate_tex_src_bit( GLbitfield bit )
 {
+   /* make sure number of switch cases is correct */
+   assert(NUM_TEXTURE_TARGETS == 7);
    switch (bit) {
-   case TEXTURE_1D_BIT:   return TEXTURE_1D_INDEX;
-   case TEXTURE_2D_BIT:   return TEXTURE_2D_INDEX;
-   case TEXTURE_RECT_BIT: return TEXTURE_RECT_INDEX;
-   case TEXTURE_3D_BIT:   return TEXTURE_3D_INDEX;
-   case TEXTURE_CUBE_BIT: return TEXTURE_CUBE_INDEX;
-   default: return TEXTURE_UNKNOWN_INDEX;
+   case TEXTURE_1D_BIT:         return TEXTURE_1D_INDEX;
+   case TEXTURE_2D_BIT:         return TEXTURE_2D_INDEX;
+   case TEXTURE_3D_BIT:         return TEXTURE_3D_INDEX;
+   case TEXTURE_CUBE_BIT:       return TEXTURE_CUBE_INDEX;
+   case TEXTURE_RECT_BIT:       return TEXTURE_RECT_INDEX;
+   case TEXTURE_1D_ARRAY_BIT:   return TEXTURE_1D_ARRAY_INDEX;
+   case TEXTURE_2D_ARRAY_BIT:   return TEXTURE_2D_ARRAY_INDEX;
+   default:                     return TEXTURE_UNKNOWN_INDEX;
    }
 }
 
-- 
cgit v1.2.3


From f0b0794b3885a2fdfb168ec4521c7b5e942d3228 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Wed, 17 Dec 2008 14:04:03 -0700
Subject: mesa: fix fixed-function test in get_fp_input_mask() - again.

The problem we're solving only occured when there was a user-defined
vertex shader but no fragment shader.  Check for that case now.
Fixes glean api2 vertex array failure.
---
 src/mesa/main/texenvprogram.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 1560e97c26..865ef8f9f8 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -219,6 +219,9 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
  */
 static GLbitfield get_fp_input_mask( GLcontext *ctx )
 {
+   const GLboolean vertexShader = (ctx->Shader.CurrentProgram &&
+                                   ctx->Shader.CurrentProgram->VertexProgram);
+   const GLboolean vertexProgram = ctx->VertexProgram._Enabled;
    GLbitfield fp_inputs = 0x0;
 
    if (ctx->VertexProgram._Overriden) {
@@ -231,10 +234,9 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
    else if (ctx->RenderMode == GL_FEEDBACK) {
       fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
    }
-   else if (!ctx->VertexProgram._Enabled ||
+   else if (!(vertexProgram || vertexShader) ||
             !ctx->VertexProgram._Current) {
-
-      /* Fixed function logic */
+      /* Fixed function vertex logic */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
       /* These get generated in the setup routine regardless of the
-- 
cgit v1.2.3


From 2389c055ed4c26ba5f3979c4a7871a333725dd88 Mon Sep 17 00:00:00 2001
From: Brian Paul <brian.paul@tungstengraphics.com>
Date: Wed, 17 Dec 2008 19:01:34 -0700
Subject: mesa: choose GLSL vertex shader over ARB/internal vertex program in
 get_fp_input_mask()

This is a work-around the for the fact that we do fragment shader state
validation before vertex shader validation (see comments in state.c) so in
get_fp_input_mask() we can't rely on ctx->VertexProgram._Current being up to
date yet.

This fixes a glean glsl1 test failure.
---
 src/mesa/main/texenvprogram.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 865ef8f9f8..2f90a34220 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -270,7 +270,19 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
    }
    else {
       /* calculate from vp->outputs */
-      GLbitfield vp_outputs = ctx->VertexProgram._Current->Base.OutputsWritten;
+      struct gl_vertex_program *vprog;
+      GLbitfield vp_outputs;
+
+      /* Choose GLSL vertex shader over ARB vertex program.  Need this
+       * since vertex shader state validation comes after fragment state
+       * validation (see additional comments in state.c).
+       */
+      if (vertexShader)
+         vprog = ctx->Shader.CurrentProgram->VertexProgram;
+      else
+         vprog = ctx->VertexProgram._Current;
+
+      vp_outputs = vprog->Base.OutputsWritten;
 
       /* These get generated in the setup routine regardless of the
        * vertex program:
-- 
cgit v1.2.3


From 9972d7147b1622074669111d72d85467c8c398cc Mon Sep 17 00:00:00 2001
From: José Fonseca <jfonseca@vmware.com>
Date: Tue, 30 Dec 2008 17:21:25 +0000
Subject: mesa: Do not specify types in bitfields.

As advised by gcc -pedantic.
---
 src/mesa/main/texenvprogram.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 2f90a34220..bd33cb4e05 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -55,8 +55,8 @@ struct texenvprog_cache_item
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
 struct mode_opt {
-   GLubyte Source:4;
-   GLubyte Operand:3;
+   GLuint Source:4;
+   GLuint Operand:3;
 };
 
 struct state_key {
-- 
cgit v1.2.3


From 6947f85cb5fbc433ba7763530285e470745b009b Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Fri, 23 Jan 2009 17:32:09 -0700
Subject: mesa: updated fragment shader codegen for GL_NV_texture_env_combine4

Expand struct fields, increase loops and handle new 4-term ADD modes.
Plus, some new assertions to catch switch/default cases that may be incorrect.
---
 src/mesa/main/texenvprogram.c | 121 +++++++++++++++++++++++++++++++-----------
 1 file changed, 90 insertions(+), 31 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 6a57d50193..066f3bd0be 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -2,6 +2,7 @@
  * 
  * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
  * All Rights Reserved.
+ * Copyright 2009 VMware, Inc.  All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the
@@ -38,6 +39,9 @@
 #include "texenvprogram.h"
 
 
+#define MAX_TERMS 4
+
+
 /*
  * Note on texture units:
  *
@@ -89,13 +93,13 @@ struct state_key {
       GLuint ScaleShiftRGB:2;
       GLuint ScaleShiftA:2;
 
-      GLuint NumArgsRGB:2;
+      GLuint NumArgsRGB:3;
       GLuint ModeRGB:4;
-      struct mode_opt OptRGB[3];
+      struct mode_opt OptRGB[MAX_TERMS];
 
-      GLuint NumArgsA:2;
+      GLuint NumArgsA:3;
       GLuint ModeA:4;
-      struct mode_opt OptA[3];
+      struct mode_opt OptA[MAX_TERMS];
    } unit[8];
 };
 
@@ -131,7 +135,9 @@ static GLuint translate_operand( GLenum operand )
    case GL_ONE_MINUS_SRC_ALPHA: return OPR_ONE_MINUS_SRC_ALPHA;
    case GL_ZERO: return OPR_ZERO;
    case GL_ONE: return OPR_ONE;
-   default:	return OPR_UNKNOWN;
+   default:
+      assert(0);
+      return OPR_UNKNOWN;
    }
 }
 
@@ -147,6 +153,7 @@ static GLuint translate_operand( GLenum operand )
 #define SRC_CONSTANT 9
 #define SRC_PRIMARY_COLOR 10
 #define SRC_PREVIOUS 11
+#define SRC_ZERO     12
 #define SRC_UNKNOWN  15
 
 static GLuint translate_source( GLenum src )
@@ -164,32 +171,49 @@ static GLuint translate_source( GLenum src )
    case GL_CONSTANT: return SRC_CONSTANT;
    case GL_PRIMARY_COLOR: return SRC_PRIMARY_COLOR;
    case GL_PREVIOUS: return SRC_PREVIOUS;
-   default: return SRC_UNKNOWN;
+   case GL_ZERO:
+      return SRC_ZERO;
+   default:
+      assert(0);
+      return SRC_UNKNOWN;
    }
 }
 
-#define MODE_REPLACE       0
-#define MODE_MODULATE      1
-#define MODE_ADD           2
-#define MODE_ADD_SIGNED    3
-#define MODE_INTERPOLATE   4
-#define MODE_SUBTRACT      5
-#define MODE_DOT3_RGB      6
-#define MODE_DOT3_RGB_EXT  7
-#define MODE_DOT3_RGBA     8
-#define MODE_DOT3_RGBA_EXT 9
-#define MODE_MODULATE_ADD_ATI           10
-#define MODE_MODULATE_SIGNED_ADD_ATI    11
-#define MODE_MODULATE_SUBTRACT_ATI      12
-#define MODE_UNKNOWN       15
-
-static GLuint translate_mode( GLenum mode )
+#define MODE_REPLACE                     0  /* r = a0 */
+#define MODE_MODULATE                    1  /* r = a0 * a1 */
+#define MODE_ADD                         2  /* r = a0 + a1 */
+#define MODE_ADD_SIGNED                  3  /* r = a0 + a1 - 0.5 */
+#define MODE_INTERPOLATE                 4  /* r = a0 * a2 + a1 * (1 - a2) */
+#define MODE_SUBTRACT                    5  /* r = a0 - a1 */
+#define MODE_DOT3_RGB                    6  /* r = a0 . a1 */
+#define MODE_DOT3_RGB_EXT                7  /* r = a0 . a1 */
+#define MODE_DOT3_RGBA                   8  /* r = a0 . a1 */
+#define MODE_DOT3_RGBA_EXT               9  /* r = a0 . a1 */
+#define MODE_MODULATE_ADD_ATI           10  /* r = a0 * a2 + a1 */
+#define MODE_MODULATE_SIGNED_ADD_ATI    11  /* r = a0 * a2 + a1 - 0.5 */
+#define MODE_MODULATE_SUBTRACT_ATI      12  /* r = a0 * a2 - a1 */
+#define MODE_ADD_PRODUCTS               13  /* r = a0 * a1 + a2 * a3 */
+#define MODE_ADD_PRODUCTS_SIGNED        14  /* r = a0 * a1 + a2 * a3 - 0.5 */
+#define MODE_UNKNOWN                    15
+
+/**
+ * Translate GL combiner state into a MODE_x value
+ */
+static GLuint translate_mode( GLenum envMode, GLenum mode )
 {
    switch (mode) {
    case GL_REPLACE: return MODE_REPLACE;
    case GL_MODULATE: return MODE_MODULATE;
-   case GL_ADD: return MODE_ADD;
-   case GL_ADD_SIGNED: return MODE_ADD_SIGNED;
+   case GL_ADD:
+      if (envMode == GL_COMBINE4_NV)
+         return MODE_ADD_PRODUCTS;
+      else
+         return MODE_ADD;
+   case GL_ADD_SIGNED:
+      if (envMode == GL_COMBINE4_NV)
+         return MODE_ADD_PRODUCTS_SIGNED;
+      else
+         return MODE_ADD_SIGNED;
    case GL_INTERPOLATE: return MODE_INTERPOLATE;
    case GL_SUBTRACT: return MODE_SUBTRACT;
    case GL_DOT3_RGB: return MODE_DOT3_RGB;
@@ -199,7 +223,9 @@ static GLuint translate_mode( GLenum mode )
    case GL_MODULATE_ADD_ATI: return MODE_MODULATE_ADD_ATI;
    case GL_MODULATE_SIGNED_ADD_ATI: return MODE_MODULATE_SIGNED_ADD_ATI;
    case GL_MODULATE_SUBTRACT_ATI: return MODE_MODULATE_SUBTRACT_ATI;
-   default: return MODE_UNKNOWN;
+   default:
+      assert(0);
+      return MODE_UNKNOWN;
    }
 }
 
@@ -212,7 +238,11 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
    case TEXTURE_RECT_BIT: return TEXTURE_RECT_INDEX;
    case TEXTURE_3D_BIT:   return TEXTURE_3D_INDEX;
    case TEXTURE_CUBE_BIT: return TEXTURE_CUBE_INDEX;
-   default: return TEXTURE_UNKNOWN_INDEX;
+   case TEXTURE_1D_ARRAY_BIT: return TEXTURE_1D_ARRAY_INDEX;
+   case TEXTURE_2D_ARRAY_BIT: return TEXTURE_2D_ARRAY_INDEX;
+   default:
+      assert(0);
+      return TEXTURE_UNKNOWN_INDEX;
    }
 }
 
@@ -248,14 +278,14 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
       key->unit[i].NumArgsA = texUnit->_CurrentCombine->_NumArgsA;
 
       key->unit[i].ModeRGB =
-	 translate_mode(texUnit->_CurrentCombine->ModeRGB);
+	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeRGB);
       key->unit[i].ModeA =
-	 translate_mode(texUnit->_CurrentCombine->ModeA);
+	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeA);
 		
       key->unit[i].ScaleShiftRGB = texUnit->_CurrentCombine->ScaleShiftRGB;
       key->unit[i].ScaleShiftA = texUnit->_CurrentCombine->ScaleShiftA;
 
-      for (j=0;j<3;j++) {
+      for (j = 0; j < MAX_TERMS; j++) {
          key->unit[i].OptRGB[j].Operand =
 	    translate_operand(texUnit->_CurrentCombine->OperandRGB[j]);
          key->unit[i].OptA[j].Operand =
@@ -677,12 +707,17 @@ static struct ureg get_source( struct texenv_fragment_program *p,
    case SRC_PRIMARY_COLOR:
       return register_input(p, FRAG_ATTRIB_COL0);
 
+   case SRC_ZERO:
+      return get_zero(p);
+
    case SRC_PREVIOUS:
-   default:
       if (is_undef(p->src_previous))
 	 return register_input(p, FRAG_ATTRIB_COL0);
       else
 	 return p->src_previous;
+
+   default:
+      assert(0);
    }
 }
 
@@ -723,7 +758,9 @@ static struct ureg emit_combine_source( struct texenv_fragment_program *p,
    case OPR_ONE:
       return get_one(p);
    case OPR_SRC_COLOR: 
+      return src;
    default:
+      assert(0);
       return src;
    }
 }
@@ -772,10 +809,12 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
 				 GLuint mode,
 				 const struct mode_opt *opt)
 {
-   struct ureg src[3];
+   struct ureg src[MAX_TERMS];
    struct ureg tmp, half;
    GLuint i;
 
+   assert(nr <= MAX_TERMS);
+
    tmp = undef; /* silence warning (bug 5318) */
 
    for (i = 0; i < nr; i++)
@@ -851,7 +890,26 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
       /* Arg0 * Arg2 - Arg1 */
       emit_arith( p, OPCODE_MAD, dest, mask, 0, src[0], src[2], negate(src[1]) );
       return dest;
+   case MODE_ADD_PRODUCTS:
+      /* Arg0 * Arg1 + Arg2 * Arg3 */
+      {
+         struct ureg tmp0 = get_temp(p);
+         emit_arith( p, OPCODE_MUL, tmp0, mask, 0, src[0], src[1], undef );
+         emit_arith( p, OPCODE_MAD, dest, mask, saturate, src[2], src[3], tmp0 );
+      }
+      return dest;
+   case MODE_ADD_PRODUCTS_SIGNED:
+      /* Arg0 * Arg1 + Arg2 * Arg3 - 0.5 */
+      {
+         struct ureg tmp0 = get_temp(p);
+         half = get_half(p);
+         emit_arith( p, OPCODE_MUL, tmp0, mask, 0, src[0], src[1], undef );
+         emit_arith( p, OPCODE_MAD, tmp0, mask, 0, src[2], src[3], tmp0 );
+         emit_arith( p, OPCODE_SUB, dest, mask, saturate, tmp0, half, undef );
+      }
+      return dest;
    default: 
+      assert(0);
       return src[0];
    }
 }
@@ -1007,6 +1065,7 @@ static GLboolean load_texenv_source( struct texenv_fragment_program *p,
       break;
       
    default:
+      /* not a texture src - do nothing */
       break;
    }
  
-- 
cgit v1.2.3


From 9fd26daec24f21dbe17afcb2e2ab272667ee9a69 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 10 Feb 2009 16:30:24 -0700
Subject: mesa: remove the unused _mesa_UpdateTexEnvProgram() function

---
 src/mesa/main/texenvprogram.c | 36 ------------------------------------
 src/mesa/main/texenvprogram.h |  2 --
 2 files changed, 38 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 066f3bd0be..d1f03405c1 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1261,39 +1261,3 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
 
    return prog;
 }
-
-
-
-/**
- * If _MaintainTexEnvProgram is set we'll generate a fragment program that
- * implements the current texture env/combine mode.
- * This function generates that program and puts it into effect.
- */
-void
-_mesa_UpdateTexEnvProgram( GLcontext *ctx )
-{
-   const struct gl_fragment_program *prev = ctx->FragmentProgram._Current;
-	
-   ASSERT(ctx->FragmentProgram._MaintainTexEnvProgram);
-
-   /* If a conventional fragment program/shader isn't in effect... */
-   if (!ctx->FragmentProgram._Enabled &&
-       (!ctx->Shader.CurrentProgram ||
-        !ctx->Shader.CurrentProgram->FragmentProgram) ) 
-   {
-      struct gl_fragment_program *newProg;
-
-      newProg = _mesa_get_fixed_func_fragment_program(ctx);
-
-      _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._Current, newProg);
-      _mesa_reference_fragprog(ctx, &ctx->FragmentProgram._TexEnvProgram, newProg);
-   } 
-
-   /* Tell the driver about the change.  Could define a new target for
-    * this?
-    */
-   if (ctx->FragmentProgram._Current != prev && ctx->Driver.BindProgram) {
-      ctx->Driver.BindProgram(ctx, GL_FRAGMENT_PROGRAM_ARB,
-                         (struct gl_program *) ctx->FragmentProgram._Current);
-   }
-}
diff --git a/src/mesa/main/texenvprogram.h b/src/mesa/main/texenvprogram.h
index a7aa60cf37..0a162d2e7a 100644
--- a/src/mesa/main/texenvprogram.h
+++ b/src/mesa/main/texenvprogram.h
@@ -32,6 +32,4 @@
 extern struct gl_fragment_program *
 _mesa_get_fixed_func_fragment_program(GLcontext *ctx);
 
-extern void _mesa_UpdateTexEnvProgram( GLcontext *ctx );
-
 #endif
-- 
cgit v1.2.3


From 44e018c09e7aeba9fd9f4c380da224bd6622c470 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Fri, 20 Feb 2009 13:42:08 -0700
Subject: mesa: add TexShadow field to prog_instruction

If the instruction is TEX/TXP/TXL/etc the TexShadow field will be true if
the instruction is a texture fetch with shadow compare.
---
 src/mesa/main/texenvprogram.c      | 19 +++++++++++++------
 src/mesa/shader/arbprogparse.c     |  3 +++
 src/mesa/shader/prog_instruction.h |  3 +++
 src/mesa/shader/prog_print.c       |  2 ++
 4 files changed, 21 insertions(+), 6 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index af51a206a5..d2a9e35dd5 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -736,6 +736,7 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
 			       GLuint destmask,
 			       GLuint tex_unit,
 			       GLuint tex_idx,
+                               GLuint tex_shadow,
 			       struct ureg coord )
 {
    struct prog_instruction *inst = emit_op( p, op, 
@@ -747,6 +748,7 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
    
    inst->TexSrcTarget = tex_idx;
    inst->TexSrcUnit = tex_unit;
+   inst->TexShadow = tex_shadow;
 
    p->program->Base.NumTexInstructions++;
 
@@ -1160,22 +1162,27 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 {
    if (is_undef(p->src_texture[unit])) {
-      GLuint dim = p->state->unit[unit].source_index;
+      GLuint texTarget = p->state->unit[unit].source_index;
       struct ureg texcoord = register_input(p, FRAG_ATTRIB_TEX0+unit);
       struct ureg tmp = get_tex_temp( p );
 
-      if (dim == TEXTURE_UNKNOWN_INDEX)
+      if (texTarget == TEXTURE_UNKNOWN_INDEX)
          program_error(p, "TexSrcBit");
 			  
       /* TODO: Use D0_MASK_XY where possible.
        */
       if (p->state->unit[unit].enabled) {
-	 p->src_texture[unit] = emit_texld( p, OPCODE_TXP,
-					    tmp, WRITEMASK_XYZW, 
-					    unit, dim, texcoord );
+         GLboolean shadow = GL_FALSE;
 
-	 if (p->state->unit[unit].shadow)
+	 if (p->state->unit[unit].shadow) {
 	    p->program->Base.ShadowSamplers |= 1 << unit;
+            shadow = GL_TRUE;
+         }
+
+	 p->src_texture[unit] = emit_texld( p, OPCODE_TXP,
+					    tmp, WRITEMASK_XYZW, 
+					    unit, texTarget, shadow,
+                                            texcoord );
 
          p->program->Base.SamplersUsed |= (1 << unit);
          /* This identity mapping should already be in place
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index 2e0fc3694a..62edb7f595 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -3104,6 +3104,9 @@ parse_fp_instruction (GLcontext * ctx, const GLubyte ** inst,
                break;
          }
 
+         if (shadow_tex)
+            fp->TexShadow = 1;
+
          /* Don't test the first time a particular sampler is seen.  Each time
           * after that, make sure the shadow state is the same.
           */
diff --git a/src/mesa/shader/prog_instruction.h b/src/mesa/shader/prog_instruction.h
index 84dd98b498..e3bb7ac01d 100644
--- a/src/mesa/shader/prog_instruction.h
+++ b/src/mesa/shader/prog_instruction.h
@@ -396,6 +396,9 @@ struct prog_instruction
    
    /** Source texture target, one of TEXTURE_{1D,2D,3D,CUBE,RECT}_INDEX */
    GLuint TexSrcTarget:3;
+
+   /** True if tex instruction should do shadow comparison */
+   GLuint TexShadow:1;
    /*@}*/
 
    /**
diff --git a/src/mesa/shader/prog_print.c b/src/mesa/shader/prog_print.c
index 516ea73301..80be51c3c5 100644
--- a/src/mesa/shader/prog_print.c
+++ b/src/mesa/shader/prog_print.c
@@ -592,6 +592,8 @@ _mesa_fprint_instruction_opt(FILE *f,
       default:
          ;
       }
+      if (inst->TexShadow)
+         _mesa_fprintf(f, " SHADOW");
       fprint_comment(f, inst);
       break;
 
-- 
cgit v1.2.3


From 8d475822e6e19fa79719c856a2db5b6a205db1b9 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Sat, 28 Feb 2009 11:49:46 -0700
Subject: mesa: rename, reorder FRAG_RESULT_x tokens

s/FRAG_RESULT_DEPR/FRAG_RESULT_DEPTH/
s/FRAG_RESULT_COLR/FRAG_RESULT/COLOR/
Remove FRAG_RESULT_COLH (NV half-precision) output since we never used it.
Next, we might merge the COLOR and DATA outputs (COLOR0, COLOR1, etc).
---
 src/mesa/drivers/dri/i915/i915_fragprog.c       |  4 +--
 src/mesa/drivers/dri/i965/brw_wm.c              |  2 +-
 src/mesa/drivers/dri/i965/brw_wm_fp.c           |  8 ++---
 src/mesa/drivers/dri/i965/brw_wm_state.c        |  2 +-
 src/mesa/drivers/dri/r300/r300_fragprog.c       |  4 +--
 src/mesa/drivers/dri/r300/r500_fragprog.c       |  4 +--
 src/mesa/drivers/dri/r300/radeon_nqssadce.c     |  2 +-
 src/mesa/drivers/dri/r300/radeon_program_pair.c |  4 +--
 src/mesa/main/mtypes.h                          |  7 ++---
 src/mesa/main/texenvprogram.c                   |  6 ++--
 src/mesa/shader/arbprogparse.c                  |  4 +--
 src/mesa/shader/nvfragparse.c                   | 40 ++++++++-----------------
 src/mesa/shader/nvfragparse.h                   |  5 ----
 src/mesa/shader/prog_debug.c                    | 12 +++-----
 src/mesa/shader/program.c                       |  6 ++--
 src/mesa/shader/programopt.c                    |  6 ++--
 src/mesa/shader/slang/slang_codegen.c           |  4 +--
 src/mesa/shader/slang/slang_link.c              |  2 +-
 src/mesa/state_tracker/st_atom_pixeltransfer.c  |  4 +--
 src/mesa/state_tracker/st_cb_drawpixels.c       | 10 +++----
 src/mesa/state_tracker/st_program.c             | 10 +++----
 src/mesa/swrast/s_context.c                     |  2 +-
 src/mesa/swrast/s_fragprog.c                    | 12 ++++----
 23 files changed, 68 insertions(+), 92 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/drivers/dri/i915/i915_fragprog.c b/src/mesa/drivers/dri/i915/i915_fragprog.c
index f091d600c3..52f09a4b1b 100644
--- a/src/mesa/drivers/dri/i915/i915_fragprog.c
+++ b/src/mesa/drivers/dri/i915/i915_fragprog.c
@@ -180,9 +180,9 @@ get_result_vector(struct i915_fragment_program *p,
    switch (inst->DstReg.File) {
    case PROGRAM_OUTPUT:
       switch (inst->DstReg.Index) {
-      case FRAG_RESULT_COLR:
+      case FRAG_RESULT_COLOR:
          return UREG(REG_TYPE_OC, 0);
-      case FRAG_RESULT_DEPR:
+      case FRAG_RESULT_DEPTH:
          p->depth_written = 1;
          return UREG(REG_TYPE_OD, 0);
       default:
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 10f38e02a4..06a6f3f0f4 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -198,7 +198,7 @@ static void brw_wm_populate_key( struct brw_context *brw,
        ctx->Color.AlphaEnabled)
       lookup |= IZ_PS_KILL_ALPHATEST_BIT;
 
-   if (fp->program.Base.OutputsWritten & (1<<FRAG_RESULT_DEPR))
+   if (fp->program.Base.OutputsWritten & (1<<FRAG_RESULT_DEPTH))
       lookup |= IZ_PS_COMPUTES_DEPTH_BIT;
 
    /* _NEW_DEPTH */
diff --git a/src/mesa/drivers/dri/i965/brw_wm_fp.c b/src/mesa/drivers/dri/i965/brw_wm_fp.c
index 977b9dd089..7e88ef0053 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_fp.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_fp.c
@@ -862,7 +862,7 @@ static void precalc_txp( struct brw_wm_compile *c,
 static void emit_fb_write( struct brw_wm_compile *c )
 {
    struct prog_src_register payload_r0_depth = src_reg(PROGRAM_PAYLOAD, PAYLOAD_DEPTH);
-   struct prog_src_register outdepth = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_DEPR);
+   struct prog_src_register outdepth = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_DEPTH);
    struct prog_src_register outcolor;
    GLuint i;
 
@@ -879,7 +879,7 @@ static void emit_fb_write( struct brw_wm_compile *c )
                                     outcolor, payload_r0_depth, outdepth);
          inst->Aux = (i<<1);
          if (c->fp_fragcolor_emitted) {
-            outcolor = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_COLR);
+            outcolor = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_COLOR);
             last_inst = inst = emit_op(c, WM_FB_WRITE, dst_mask(dst_undef(),0),
                                        0, outcolor, payload_r0_depth, outdepth);
             inst->Aux = (i<<1);
@@ -892,7 +892,7 @@ static void emit_fb_write( struct brw_wm_compile *c )
       if (c->fp->program.Base.OutputsWritten & (1 << FRAG_RESULT_DATA0))
          outcolor = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_DATA0);
       else 
-         outcolor = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_COLR);
+         outcolor = src_reg(PROGRAM_OUTPUT, FRAG_RESULT_COLOR);
 
       inst = emit_op(c, WM_FB_WRITE, dst_mask(dst_undef(),0),
                      0, outcolor, payload_r0_depth, outdepth);
@@ -928,7 +928,7 @@ static void validate_dst_regs( struct brw_wm_compile *c,
 {
    if (inst->DstReg.File == PROGRAM_OUTPUT) {
       GLuint idx = inst->DstReg.Index;
-      if (idx == FRAG_RESULT_COLR)
+      if (idx == FRAG_RESULT_COLOR)
          c->fp_fragcolor_emitted = 1;
    }
 }
diff --git a/src/mesa/drivers/dri/i965/brw_wm_state.c b/src/mesa/drivers/dri/i965/brw_wm_state.c
index 3c3b3473d6..f7ea756fb0 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_state.c
@@ -103,7 +103,7 @@ wm_unit_populate_key(struct brw_context *brw, struct brw_wm_unit_key *key)
 
    /* as far as we can tell */
    key->computes_depth =
-      (fp->Base.OutputsWritten & (1 << FRAG_RESULT_DEPR)) != 0;
+      (fp->Base.OutputsWritten & (1 << FRAG_RESULT_DEPTH)) != 0;
 
    /* _NEW_COLOR */
    key->uses_kill = fp->UsesKill || ctx->Color.AlphaEnabled;
diff --git a/src/mesa/drivers/dri/r300/r300_fragprog.c b/src/mesa/drivers/dri/r300/r300_fragprog.c
index 8d030c63fb..32182bb667 100644
--- a/src/mesa/drivers/dri/r300/r300_fragprog.c
+++ b/src/mesa/drivers/dri/r300/r300_fragprog.c
@@ -356,8 +356,8 @@ static void insert_WPOS_trailer(struct r300_fragment_program_compiler *compiler)
 
 static void nqssadce_init(struct nqssadce_state* s)
 {
-	s->Outputs[FRAG_RESULT_COLR].Sourced = WRITEMASK_XYZW;
-	s->Outputs[FRAG_RESULT_DEPR].Sourced = WRITEMASK_W;
+	s->Outputs[FRAG_RESULT_COLOR].Sourced = WRITEMASK_XYZW;
+	s->Outputs[FRAG_RESULT_DEPTH].Sourced = WRITEMASK_W;
 }
 
 
diff --git a/src/mesa/drivers/dri/r300/r500_fragprog.c b/src/mesa/drivers/dri/r300/r500_fragprog.c
index 926ddd5964..07a2a7b17c 100644
--- a/src/mesa/drivers/dri/r300/r500_fragprog.c
+++ b/src/mesa/drivers/dri/r300/r500_fragprog.c
@@ -298,8 +298,8 @@ static void insert_WPOS_trailer(struct r500_fragment_program_compiler *compiler)
 
 static void nqssadce_init(struct nqssadce_state* s)
 {
-	s->Outputs[FRAG_RESULT_COLR].Sourced = WRITEMASK_XYZW;
-	s->Outputs[FRAG_RESULT_DEPR].Sourced = WRITEMASK_W;
+	s->Outputs[FRAG_RESULT_COLOR].Sourced = WRITEMASK_XYZW;
+	s->Outputs[FRAG_RESULT_DEPTH].Sourced = WRITEMASK_W;
 }
 
 static GLboolean is_native_swizzle(GLuint opcode, struct prog_src_register reg)
diff --git a/src/mesa/drivers/dri/r300/radeon_nqssadce.c b/src/mesa/drivers/dri/r300/radeon_nqssadce.c
index 97ce016c99..a083c3d243 100644
--- a/src/mesa/drivers/dri/r300/radeon_nqssadce.c
+++ b/src/mesa/drivers/dri/r300/radeon_nqssadce.c
@@ -191,7 +191,7 @@ static void process_instruction(struct nqssadce_state* s)
 
 	if (inst->Opcode != OPCODE_KIL) {
 		if (s->Descr->RewriteDepthOut) {
-			if (inst->DstReg.File == PROGRAM_OUTPUT && inst->DstReg.Index == FRAG_RESULT_DEPR)
+			if (inst->DstReg.File == PROGRAM_OUTPUT && inst->DstReg.Index == FRAG_RESULT_DEPTH)
 				rewrite_depth_out(inst);
 		}
 
diff --git a/src/mesa/drivers/dri/r300/radeon_program_pair.c b/src/mesa/drivers/dri/r300/radeon_program_pair.c
index 58bc0d5843..365e7c1722 100644
--- a/src/mesa/drivers/dri/r300/radeon_program_pair.c
+++ b/src/mesa/drivers/dri/r300/radeon_program_pair.c
@@ -778,10 +778,10 @@ static void fill_dest_into_pair(struct pair_state *s, struct radeon_pair_instruc
 	struct prog_instruction *inst = s->Program->Instructions + ip;
 
 	if (inst->DstReg.File == PROGRAM_OUTPUT) {
-		if (inst->DstReg.Index == FRAG_RESULT_COLR) {
+		if (inst->DstReg.Index == FRAG_RESULT_COLOR) {
 			pair->RGB.OutputWriteMask |= inst->DstReg.WriteMask & WRITEMASK_XYZ;
 			pair->Alpha.OutputWriteMask |= GET_BIT(inst->DstReg.WriteMask, 3);
-		} else if (inst->DstReg.Index == FRAG_RESULT_DEPR) {
+		} else if (inst->DstReg.Index == FRAG_RESULT_DEPTH) {
 			pair->Alpha.DepthWriteMask |= GET_BIT(inst->DstReg.WriteMask, 3);
 		}
 	} else {
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 22add9d975..f17b9e1e71 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -268,10 +268,9 @@ typedef enum
  */
 typedef enum
 {
-   FRAG_RESULT_COLR = 0,
-   FRAG_RESULT_COLH = 1,
-   FRAG_RESULT_DEPR = 2,
-   FRAG_RESULT_DATA0 = 3,
+   FRAG_RESULT_DEPTH = 0,
+   FRAG_RESULT_COLOR = 1,
+   FRAG_RESULT_DATA0 = 2,
    FRAG_RESULT_MAX = (FRAG_RESULT_DATA0 + MAX_DRAW_BUFFERS)
 } gl_frag_result;
 
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index d2a9e35dd5..51c13a563d 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1098,7 +1098,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
        rgb_shift)
       dest = get_temp( p );
    else
-      dest = make_ureg(PROGRAM_OUTPUT, FRAG_RESULT_COLR);
+      dest = make_ureg(PROGRAM_OUTPUT, FRAG_RESULT_COLOR);
 
    /* Emit the RGB and A combine ops
     */
@@ -1278,7 +1278,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    p.program->Base.Parameters = _mesa_new_parameter_list();
 
    p.program->Base.InputsRead = 0;
-   p.program->Base.OutputsWritten = 1 << FRAG_RESULT_COLR;
+   p.program->Base.OutputsWritten = 1 << FRAG_RESULT_COLOR;
 
    for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
       p.src_texture[unit] = undef;
@@ -1313,7 +1313,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    }
 
    cf = get_source( &p, SRC_PREVIOUS, 0 );
-   out = make_ureg( PROGRAM_OUTPUT, FRAG_RESULT_COLR );
+   out = make_ureg( PROGRAM_OUTPUT, FRAG_RESULT_COLOR );
 
    if (key->separate_specular) {
       /* Emit specular add.
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index 62edb7f595..ccc0318a53 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -1688,7 +1688,7 @@ parse_result_binding(GLcontext *ctx, const GLubyte **inst,
              */
             parse_output_color_num(ctx, inst, Program, &out_color);
             ASSERT(out_color < MAX_DRAW_BUFFERS);
-            *outputReg = FRAG_RESULT_COLR;
+            *outputReg = FRAG_RESULT_COLOR;
          }
          else {
             /* for vtx programs, this is VERTEX_RESULT_POSITION */
@@ -1699,7 +1699,7 @@ parse_result_binding(GLcontext *ctx, const GLubyte **inst,
       case FRAGMENT_RESULT_DEPTH:
          if (Program->Base.Target == GL_FRAGMENT_PROGRAM_ARB) {
             /* for frag programs, this is FRAGMENT_RESULT_DEPTH */
-            *outputReg = FRAG_RESULT_DEPR;
+            *outputReg = FRAG_RESULT_DEPTH;
          }
          else {
             /* for vtx programs, this is VERTEX_RESULT_COLOR */
diff --git a/src/mesa/shader/nvfragparse.c b/src/mesa/shader/nvfragparse.c
index 20e4781372..37418ffb6e 100644
--- a/src/mesa/shader/nvfragparse.c
+++ b/src/mesa/shader/nvfragparse.c
@@ -384,18 +384,12 @@ static const char *InputRegisters[MAX_NV_FRAGMENT_PROGRAM_INPUTS + 1] = {
    "TEX0", "TEX1", "TEX2", "TEX3", "TEX4", "TEX5", "TEX6", "TEX7", NULL
 };
 
+
 static const char *OutputRegisters[MAX_NV_FRAGMENT_PROGRAM_OUTPUTS + 1] = {
-   "COLR", "COLH",
-   /* These are only allows for register combiners */
-   /*
-   "TEX0", "TEX1", "TEX2", "TEX3",
-   */
-   "DEPR", NULL
+   "DEPR", "COLR", "DATA0", NULL
 };
 
 
-
-
 /**********************************************************************/
 
 /**
@@ -828,7 +822,6 @@ static GLboolean
 Parse_OutputReg(struct parse_state *parseState, GLint *outputRegNum)
 {
    GLubyte token[100];
-   GLint j;
 
    /* Match "o[" */
    if (!Parse_String(parseState, "o["))
@@ -839,19 +832,19 @@ Parse_OutputReg(struct parse_state *parseState, GLint *outputRegNum)
       RETURN_ERROR;
 
    /* try to match an output register name */
-   for (j = 0; OutputRegisters[j]; j++) {
-      if (_mesa_strcmp((const char *) token, OutputRegisters[j]) == 0) {
-         static GLuint bothColors = (1 << FRAG_RESULT_COLR) | (1 << FRAG_RESULT_COLH);
-         *outputRegNum = j;
-         parseState->outputsWritten |= (1 << j);
-         if ((parseState->outputsWritten & bothColors) == bothColors) {
-            RETURN_ERROR1("Illegal to write to both o[COLR] and o[COLH]");
-         }
-         break;
-      }
+   if (_mesa_strcmp((char *) token, "COLR") == 0 ||
+       _mesa_strcmp((char *) token, "COLH") == 0) {
+      /* note that we don't distinguish between COLR and COLH */
+      *outputRegNum = FRAG_RESULT_COLOR;
+      parseState->outputsWritten |= (1 << FRAG_RESULT_COLOR);
    }
-   if (!OutputRegisters[j])
+   else if (_mesa_strcmp((char *) token, "DEPR") == 0) {
+      *outputRegNum = FRAG_RESULT_DEPTH;
+      parseState->outputsWritten |= (1 << FRAG_RESULT_DEPTH);
+   }
+   else {
       RETURN_ERROR1("Invalid output register name");
+   }
 
    /* Match ']' */
    if (!Parse_String(parseState, "]"))
@@ -1826,10 +1819,3 @@ _mesa_nv_fragment_input_register_name(GLuint i)
    return InputRegisters[i];
 }
 
-
-const char *
-_mesa_nv_fragment_output_register_name(GLuint i)
-{
-   ASSERT(i < MAX_NV_FRAGMENT_PROGRAM_OUTPUTS);
-   return OutputRegisters[i];
-}
diff --git a/src/mesa/shader/nvfragparse.h b/src/mesa/shader/nvfragparse.h
index de45cf543d..ac97921080 100644
--- a/src/mesa/shader/nvfragparse.h
+++ b/src/mesa/shader/nvfragparse.h
@@ -44,9 +44,4 @@ _mesa_print_nv_fragment_program(const struct gl_fragment_program *program);
 extern const char *
 _mesa_nv_fragment_input_register_name(GLuint i);
 
-
-extern const char *
-_mesa_nv_fragment_output_register_name(GLuint i);
-
-
 #endif
diff --git a/src/mesa/shader/prog_debug.c b/src/mesa/shader/prog_debug.c
index 7bcb2ef734..35ce37d5ce 100644
--- a/src/mesa/shader/prog_debug.c
+++ b/src/mesa/shader/prog_debug.c
@@ -222,20 +222,16 @@ _mesa_GetProgramRegisterfvMESA(GLenum target,
                         "glGetProgramRegisterfvMESA(registerName)");
             return;
          }
-         else if (_mesa_strcmp(reg, "o[COLR]") == 0) {
+         else if (_mesa_strcmp(reg, "o[COLR]") == 0 ||
+                  _mesa_strcmp(reg, "o[COLH]") == 0) {
             /* Fragment output color */
             ctx->Driver.GetProgramRegister(ctx, PROGRAM_OUTPUT,
-                                           FRAG_RESULT_COLR, v);
-         }
-         else if (_mesa_strcmp(reg, "o[COLH]") == 0) {
-            /* Fragment output color */
-            ctx->Driver.GetProgramRegister(ctx, PROGRAM_OUTPUT,
-                                           FRAG_RESULT_COLH, v);
+                                           FRAG_RESULT_COLOR, v);
          }
          else if (_mesa_strcmp(reg, "o[DEPR]") == 0) {
             /* Fragment output depth */
             ctx->Driver.GetProgramRegister(ctx, PROGRAM_OUTPUT,
-                                           FRAG_RESULT_DEPR, v);
+                                           FRAG_RESULT_DEPTH, v);
          }
          else {
             /* try user-defined identifiers */
diff --git a/src/mesa/shader/program.c b/src/mesa/shader/program.c
index 00655f0288..2e5632710e 100644
--- a/src/mesa/shader/program.c
+++ b/src/mesa/shader/program.c
@@ -738,7 +738,7 @@ _mesa_combine_programs(GLcontext *ctx,
       /* Connect color outputs of fprogA to color inputs of fprogB, via a
        * new temporary register.
        */
-      if ((progA->OutputsWritten & (1 << FRAG_RESULT_COLR)) &&
+      if ((progA->OutputsWritten & (1 << FRAG_RESULT_COLOR)) &&
           (progB_inputsRead & FRAG_BIT_COL0)) {
          GLint tempReg = _mesa_find_free_register(newProg, PROGRAM_TEMPORARY);
          if (tempReg < 0) {
@@ -748,7 +748,7 @@ _mesa_combine_programs(GLcontext *ctx,
          }
          /* replace writes to result.color[0] with tempReg */
          replace_registers(newInst, lenA,
-                           PROGRAM_OUTPUT, FRAG_RESULT_COLR,
+                           PROGRAM_OUTPUT, FRAG_RESULT_COLOR,
                            PROGRAM_TEMPORARY, tempReg);
          /* replace reads from the input color with tempReg */
          replace_registers(newInst + lenA, lenB,
@@ -758,7 +758,7 @@ _mesa_combine_programs(GLcontext *ctx,
 
       /* compute combined program's InputsRead */
       inputsB = progB_inputsRead;
-      if (progA->OutputsWritten & (1 << FRAG_RESULT_COLR)) {
+      if (progA->OutputsWritten & (1 << FRAG_RESULT_COLOR)) {
          inputsB &= ~(1 << FRAG_ATTRIB_COL0);
       }
       newProg->InputsRead = progA->InputsRead | inputsB;
diff --git a/src/mesa/shader/programopt.c b/src/mesa/shader/programopt.c
index 56f1eb832e..05d9bdf7ec 100644
--- a/src/mesa/shader/programopt.c
+++ b/src/mesa/shader/programopt.c
@@ -171,7 +171,7 @@ _mesa_append_fog_code(GLcontext *ctx, struct gl_fragment_program *fprog)
       if (inst->Opcode == OPCODE_END)
          break;
       if (inst->DstReg.File == PROGRAM_OUTPUT &&
-          inst->DstReg.Index == FRAG_RESULT_COLR) {
+          inst->DstReg.Index == FRAG_RESULT_COLOR) {
          /* change the instruction to write to colorTemp w/ clamping */
          inst->DstReg.File = PROGRAM_TEMPORARY;
          inst->DstReg.Index = colorTemp;
@@ -249,7 +249,7 @@ _mesa_append_fog_code(GLcontext *ctx, struct gl_fragment_program *fprog)
    /* LRP result.color.xyz, fogFactorTemp.xxxx, colorTemp, fogColorRef; */
    inst->Opcode = OPCODE_LRP;
    inst->DstReg.File = PROGRAM_OUTPUT;
-   inst->DstReg.Index = FRAG_RESULT_COLR;
+   inst->DstReg.Index = FRAG_RESULT_COLOR;
    inst->DstReg.WriteMask = WRITEMASK_XYZ;
    inst->SrcReg[0].File = PROGRAM_TEMPORARY;
    inst->SrcReg[0].Index = fogFactorTemp;
@@ -264,7 +264,7 @@ _mesa_append_fog_code(GLcontext *ctx, struct gl_fragment_program *fprog)
    /* MOV result.color.w, colorTemp.x;  # copy alpha */
    inst->Opcode = OPCODE_MOV;
    inst->DstReg.File = PROGRAM_OUTPUT;
-   inst->DstReg.Index = FRAG_RESULT_COLR;
+   inst->DstReg.Index = FRAG_RESULT_COLOR;
    inst->DstReg.WriteMask = WRITEMASK_W;
    inst->SrcReg[0].File = PROGRAM_TEMPORARY;
    inst->SrcReg[0].Index = colorTemp;
diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c
index 5e2ce0ce3a..ee24e2e138 100644
--- a/src/mesa/shader/slang/slang_codegen.c
+++ b/src/mesa/shader/slang/slang_codegen.c
@@ -450,8 +450,8 @@ _slang_output_index(const char *name, GLenum target)
       { NULL, 0 }
    };
    static const struct output_info fragOutputs[] = {
-      { "gl_FragColor", FRAG_RESULT_COLR },
-      { "gl_FragDepth", FRAG_RESULT_DEPR },
+      { "gl_FragColor", FRAG_RESULT_COLOR },
+      { "gl_FragDepth", FRAG_RESULT_DEPTH },
       { "gl_FragData", FRAG_RESULT_DATA0 },
       { NULL, 0 }
    };
diff --git a/src/mesa/shader/slang/slang_link.c b/src/mesa/shader/slang/slang_link.c
index bcabe71bfa..f98434892b 100644
--- a/src/mesa/shader/slang/slang_link.c
+++ b/src/mesa/shader/slang/slang_link.c
@@ -686,7 +686,7 @@ _slang_link(GLcontext *ctx,
    /* check that gl_FragColor and gl_FragData are not both written to */
    if (shProg->FragmentProgram) {
       GLbitfield outputsWritten = shProg->FragmentProgram->Base.OutputsWritten;
-      if ((outputsWritten & ((1 << FRAG_RESULT_COLR))) &&
+      if ((outputsWritten & ((1 << FRAG_RESULT_COLOR))) &&
           (outputsWritten >= (1 << FRAG_RESULT_DATA0))) {
          link_error(shProg, "Fragment program cannot write both gl_FragColor"
                     " and gl_FragData[].\n");
diff --git a/src/mesa/state_tracker/st_atom_pixeltransfer.c b/src/mesa/state_tracker/st_atom_pixeltransfer.c
index ce872458e3..05b69c9d00 100644
--- a/src/mesa/state_tracker/st_atom_pixeltransfer.c
+++ b/src/mesa/state_tracker/st_atom_pixeltransfer.c
@@ -212,7 +212,7 @@ get_pixel_transfer_program(GLcontext *ctx, const struct state_key *key)
    inst[ic].TexSrcTarget = TEXTURE_2D_INDEX;
    ic++;
    fp->Base.InputsRead = (1 << FRAG_ATTRIB_TEX0);
-   fp->Base.OutputsWritten = (1 << FRAG_RESULT_COLR);
+   fp->Base.OutputsWritten = (1 << FRAG_RESULT_COLOR);
    fp->Base.SamplersUsed = 0x1;  /* sampler 0 (bit 0) is used */
 
    if (key->scaleAndBias) {
@@ -400,7 +400,7 @@ get_pixel_transfer_program(GLcontext *ctx, const struct state_key *key)
    {
       struct prog_instruction *last = &inst[ic - 1];
       last->DstReg.File = PROGRAM_OUTPUT;
-      last->DstReg.Index = FRAG_RESULT_COLR;
+      last->DstReg.Index = FRAG_RESULT_COLOR;
    }
 
    /* END; */
diff --git a/src/mesa/state_tracker/st_cb_drawpixels.c b/src/mesa/state_tracker/st_cb_drawpixels.c
index c73d0080ca..cc7a9e7890 100644
--- a/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -75,7 +75,7 @@ is_passthrough_program(const struct gl_fragment_program *prog)
       if (inst[0].Opcode == OPCODE_MOV &&
           inst[1].Opcode == OPCODE_END &&
           inst[0].DstReg.File == PROGRAM_OUTPUT &&
-          inst[0].DstReg.Index == FRAG_RESULT_COLR &&
+          inst[0].DstReg.Index == FRAG_RESULT_COLOR &&
           inst[0].DstReg.WriteMask == WRITEMASK_XYZW &&
           inst[0].SrcReg[0].File == PROGRAM_INPUT &&
           inst[0].SrcReg[0].Index == FRAG_ATTRIB_COL0 &&
@@ -158,7 +158,7 @@ combined_drawpix_fragment_program(GLcontext *ctx)
 
 /**
  * Create fragment shader that does a TEX() instruction to get a Z
- * value, then writes to FRAG_RESULT_DEPR.
+ * value, then writes to FRAG_RESULT_DEPTH.
  * Pass fragment color through as-is.
  */
 static struct st_fragment_program *
@@ -191,7 +191,7 @@ make_fragment_shader_z(struct st_context *st)
    /* TEX result.depth, fragment.texcoord[0], texture[0], 2D; */
    p->Instructions[ic].Opcode = OPCODE_TEX;
    p->Instructions[ic].DstReg.File = PROGRAM_OUTPUT;
-   p->Instructions[ic].DstReg.Index = FRAG_RESULT_DEPR;
+   p->Instructions[ic].DstReg.Index = FRAG_RESULT_DEPTH;
    p->Instructions[ic].DstReg.WriteMask = WRITEMASK_Z;
    p->Instructions[ic].SrcReg[0].File = PROGRAM_INPUT;
    p->Instructions[ic].SrcReg[0].Index = FRAG_ATTRIB_TEX0;
@@ -202,7 +202,7 @@ make_fragment_shader_z(struct st_context *st)
    /* MOV result.color, fragment.color */
    p->Instructions[ic].Opcode = OPCODE_MOV;
    p->Instructions[ic].DstReg.File = PROGRAM_OUTPUT;
-   p->Instructions[ic].DstReg.Index = FRAG_RESULT_COLR;
+   p->Instructions[ic].DstReg.Index = FRAG_RESULT_COLOR;
    p->Instructions[ic].SrcReg[0].File = PROGRAM_INPUT;
    p->Instructions[ic].SrcReg[0].Index = FRAG_ATTRIB_COL0;
    ic++;
@@ -213,7 +213,7 @@ make_fragment_shader_z(struct st_context *st)
    assert(ic == p->NumInstructions);
 
    p->InputsRead = FRAG_BIT_TEX0 | FRAG_BIT_COL0;
-   p->OutputsWritten = (1 << FRAG_RESULT_COLR) | (1 << FRAG_RESULT_DEPR);
+   p->OutputsWritten = (1 << FRAG_RESULT_COLOR) | (1 << FRAG_RESULT_DEPTH);
    p->SamplersUsed = 0x1;  /* sampler 0 (bit 0) is used */
 
    st->drawpix.z_shader = (struct st_fragment_program *) p;
diff --git a/src/mesa/state_tracker/st_program.c b/src/mesa/state_tracker/st_program.c
index d2535b6a2f..f825204915 100644
--- a/src/mesa/state_tracker/st_program.c
+++ b/src/mesa/state_tracker/st_program.c
@@ -466,23 +466,23 @@ st_translate_fragment_program(struct st_context *st,
       GLbitfield outputsWritten = stfp->Base.Base.OutputsWritten;
 
       /* if z is written, emit that first */
-      if (outputsWritten & (1 << FRAG_RESULT_DEPR)) {
+      if (outputsWritten & (1 << FRAG_RESULT_DEPTH)) {
          fs_output_semantic_name[fs_num_outputs] = TGSI_SEMANTIC_POSITION;
          fs_output_semantic_index[fs_num_outputs] = 0;
-         outputMapping[FRAG_RESULT_DEPR] = fs_num_outputs;
+         outputMapping[FRAG_RESULT_DEPTH] = fs_num_outputs;
          fs_num_outputs++;
-         outputsWritten &= ~(1 << FRAG_RESULT_DEPR);
+         outputsWritten &= ~(1 << FRAG_RESULT_DEPTH);
       }
 
       /* handle remaning outputs (color) */
       for (attr = 0; attr < FRAG_RESULT_MAX; attr++) {
          if (outputsWritten & (1 << attr)) {
             switch (attr) {
-            case FRAG_RESULT_DEPR:
+            case FRAG_RESULT_DEPTH:
                /* handled above */
                assert(0);
                break;
-            case FRAG_RESULT_COLR:
+            case FRAG_RESULT_COLOR:
                fs_output_semantic_name[fs_num_outputs] = TGSI_SEMANTIC_COLOR;
                fs_output_semantic_index[fs_num_outputs] = numColors;
                outputMapping[attr] = fs_num_outputs;
diff --git a/src/mesa/swrast/s_context.c b/src/mesa/swrast/s_context.c
index 297940adbd..719e6b8296 100644
--- a/src/mesa/swrast/s_context.c
+++ b/src/mesa/swrast/s_context.c
@@ -206,7 +206,7 @@ _swrast_update_deferred_texture(GLcontext *ctx)
    else {
       const struct gl_fragment_program *fprog
          = ctx->FragmentProgram._Current;
-      if (fprog && (fprog->Base.OutputsWritten & (1 << FRAG_RESULT_DEPR))) {
+      if (fprog && (fprog->Base.OutputsWritten & (1 << FRAG_RESULT_DEPTH))) {
          /* Z comes from fragment program/shader */
          swrast->_DeferredTexture = GL_FALSE;
       }
diff --git a/src/mesa/swrast/s_fragprog.c b/src/mesa/swrast/s_fragprog.c
index c6601f5593..3735f041d6 100644
--- a/src/mesa/swrast/s_fragprog.c
+++ b/src/mesa/swrast/s_fragprog.c
@@ -202,9 +202,9 @@ run_program(GLcontext *ctx, SWspan *span, GLuint start, GLuint end)
          if (_mesa_execute_program(ctx, &program->Base, machine)) {
 
             /* Store result color */
-            if (outputsWritten & (1 << FRAG_RESULT_COLR)) {
+            if (outputsWritten & (1 << FRAG_RESULT_COLOR)) {
                COPY_4V(span->array->attribs[FRAG_ATTRIB_COL0][i],
-                       machine->Outputs[FRAG_RESULT_COLR]);
+                       machine->Outputs[FRAG_RESULT_COLOR]);
             }
             else {
                /* Multiple drawbuffers / render targets
@@ -221,8 +221,8 @@ run_program(GLcontext *ctx, SWspan *span, GLuint start, GLuint end)
             }
 
             /* Store result depth/z */
-            if (outputsWritten & (1 << FRAG_RESULT_DEPR)) {
-               const GLfloat depth = machine->Outputs[FRAG_RESULT_DEPR][2];
+            if (outputsWritten & (1 << FRAG_RESULT_DEPTH)) {
+               const GLfloat depth = machine->Outputs[FRAG_RESULT_DEPTH][2];
                if (depth <= 0.0)
                   span->array->z[i] = 0;
                else if (depth >= 1.0)
@@ -259,12 +259,12 @@ _swrast_exec_fragment_program( GLcontext *ctx, SWspan *span )
 
    run_program(ctx, span, 0, span->end);
 
-   if (program->Base.OutputsWritten & (1 << FRAG_RESULT_COLR)) {
+   if (program->Base.OutputsWritten & (1 << FRAG_RESULT_COLOR)) {
       span->interpMask &= ~SPAN_RGBA;
       span->arrayMask |= SPAN_RGBA;
    }
 
-   if (program->Base.OutputsWritten & (1 << FRAG_RESULT_DEPR)) {
+   if (program->Base.OutputsWritten & (1 << FRAG_RESULT_DEPTH)) {
       span->interpMask &= ~SPAN_Z;
       span->arrayMask |= SPAN_Z;
    }
-- 
cgit v1.2.3


From 114152e068ec919feb0a57a1259c2ada970b9f02 Mon Sep 17 00:00:00 2001
From: Roland Scheidegger <sroland@vmware.com>
Date: Thu, 12 Mar 2009 15:01:16 +0100
Subject: mesa: add support for ATI_envmap_bumpmap

add new entrypoints, new texture format, etc
translate in texenvprogram.c for drivers using the mesa-generated tex env
fragment program
also handled in swrast, but not tested (cannot work due to negative texel
results not handled correctly)
---
 src/mesa/glapi/gl_API.xml        |  50 +++++++++++
 src/mesa/main/api_exec.c         |   6 ++
 src/mesa/main/config.h           |   3 +-
 src/mesa/main/context.c          |   3 +
 src/mesa/main/dlist.c            |  46 +++++++++++
 src/mesa/main/extensions.c       |   2 +
 src/mesa/main/image.c            | 153 +++++++++++++++++++++++++++++++++-
 src/mesa/main/image.h            |   7 ++
 src/mesa/main/macros.h           |   7 +-
 src/mesa/main/mipmap.c           |  14 +++-
 src/mesa/main/mtypes.h           |   5 ++
 src/mesa/main/texenv.c           | 174 ++++++++++++++++++++++++++++++++++++++-
 src/mesa/main/texenv.h           |  11 +++
 src/mesa/main/texenvprogram.c    |  99 ++++++++++++++++++++--
 src/mesa/main/texformat.c        |  42 ++++++++++
 src/mesa/main/texformat.h        |  14 +++-
 src/mesa/main/texformat_tmp.h    |  17 +++-
 src/mesa/main/teximage.c         |  39 ++++++++-
 src/mesa/main/texstate.c         |  20 +++++
 src/mesa/main/texstore.c         |  91 +++++++++++++++++++-
 src/mesa/main/texstore.h         |   2 +-
 src/mesa/shader/prog_statevars.c |  28 +++++++
 src/mesa/shader/prog_statevars.h |   2 +
 src/mesa/swrast/s_texcombine.c   |  95 ++++++++++++++++++++-
 24 files changed, 906 insertions(+), 24 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/glapi/gl_API.xml b/src/mesa/glapi/gl_API.xml
index 8212fc2477..cc3e3ae6bf 100644
--- a/src/mesa/glapi/gl_API.xml
+++ b/src/mesa/glapi/gl_API.xml
@@ -11681,6 +11681,56 @@
     <enum name="DEPTH_STENCIL_TO_BGRA_NV"                 value="0x886F"/>
 </category>
 
+<category name="GL_ATI_envmap_bumpmap" number="244">
+    <enum name="BUMP_ROT_MATRIX_ATI"          count="4"   value="0x8775">
+        <size name="TexBumpParameterfv"/>
+        <size name="TexBumpParameteriv"/>
+        <size name="GetTexBumpParameterfv" mode="get"/>
+        <size name="GetTexBumpParameteriv" mode="get"/>
+    </enum>
+    <enum name="BUMP_ROT_MATRIX_SIZE_ATI"     count="1"   value="0x8776">
+        <size name="GetTexBumpParameterfv" mode="get"/>
+        <size name="GetTexBumpParameteriv" mode="get"/>
+    </enum>
+    <enum name="BUMP_NUM_TEX_UNITS_ATI"       count="1"   value="0x8777">
+        <size name="GetTexBumpParameterfv" mode="get"/>
+        <size name="GetTexBumpParameteriv" mode="get"/>
+    </enum>
+    <enum name="BUMP_TEX_UNITS_ATI"           count="-1"  value="0x8778">
+        <size name="GetTexBumpParameterfv" mode="get"/>
+        <size name="GetTexBumpParameteriv" mode="get"/>
+    </enum>
+    <enum name="DUDV_ATI"                                 value="0x8779"/>
+    <enum name="DU8DV8_ATI"                               value="0x877A"/>
+    <enum name="BUMP_ENVMAP_ATI"                          value="0x877B"/>
+    <enum name="BUMP_TARGET_ATI"              count="1"   value="0x877C">
+        <size name="TexEnviv"/>
+        <size name="TexEnvfv"/>
+        <size name="GetTexEnviv" mode="get"/>
+        <size name="GetTexEnvfv" mode="get"/>
+    </enum>
+    <function name="TexBumpParameterfvATI" offset="assign">
+      <param name="pname" type="GLenum"/>
+      <param name="param" type="const GLfloat *" variable_param="pname"/>
+      <glx ignore="true"/>
+    </function>
+    <function name="TexBumpParameterivATI" offset="assign">
+      <param name="pname" type="GLenum"/>
+      <param name="param" type="const GLint *" variable_param="pname"/>
+      <glx ignore="true"/>
+    </function>
+    <function name="GetTexBumpParameterfvATI" offset="assign">
+      <param name="pname" type="GLenum"/>
+      <param name="param" type="GLfloat *" variable_param="pname"/>
+      <glx ignore="true"/>
+    </function>
+    <function name="GetTexBumpParameterivATI" offset="assign">
+      <param name="pname" type="GLenum"/>
+      <param name="param" type="GLint *" variable_param="pname"/>
+      <glx ignore="true"/>
+    </function>
+</category>
+
 <category name="GL_ATI_fragment_shader" number="245">
     <function name="GenFragmentShadersATI" offset="assign">
       <return type="GLuint"/>
diff --git a/src/mesa/main/api_exec.c b/src/mesa/main/api_exec.c
index e0715817ad..6f66ff47a0 100644
--- a/src/mesa/main/api_exec.c
+++ b/src/mesa/main/api_exec.c
@@ -839,6 +839,12 @@ _mesa_init_exec_table(struct _glapi_table *exec)
    SET_SetFragmentShaderConstantATI(exec, _mesa_SetFragmentShaderConstantATI);
 #endif
 
+  /* GL_ATI_envmap_bumpmap */
+   SET_GetTexBumpParameterivATI(exec, _mesa_GetTexBumpParameterivATI);
+   SET_GetTexBumpParameterfvATI(exec, _mesa_GetTexBumpParameterfvATI);
+   SET_TexBumpParameterivATI(exec, _mesa_TexBumpParameterivATI);
+   SET_TexBumpParameterfvATI(exec, _mesa_TexBumpParameterfvATI);
+
 #if FEATURE_EXT_framebuffer_object
    SET_IsRenderbufferEXT(exec, _mesa_IsRenderbufferEXT);
    SET_BindRenderbufferEXT(exec, _mesa_BindRenderbufferEXT);
diff --git a/src/mesa/main/config.h b/src/mesa/main/config.h
index 282ad9514c..fc31155b35 100644
--- a/src/mesa/main/config.h
+++ b/src/mesa/main/config.h
@@ -232,7 +232,8 @@
 #define MAX_COLOR_ATTACHMENTS 8
 /*@}*/
 
-
+/** For GL_ATI_envmap_bump - support bump mapping on first 8 units */
+#define SUPPORTED_ATI_BUMP_UNITS 0xff
 
 /**
  * \name Mesa-specific parameters
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 9fd9e769d2..84bf0205a8 100644
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -589,6 +589,9 @@ _mesa_init_constants(GLcontext *ctx)
    /* GL_ARB_framebuffer_object */
    ctx->Const.MaxSamples = 0;
 
+   /* GL_ATI_envmap_bumpmap */
+   ctx->Const.SupportedBumpUnits = SUPPORTED_ATI_BUMP_UNITS;
+
    /* sanity checks */
    ASSERT(ctx->Const.MaxTextureUnits == MIN2(ctx->Const.MaxTextureImageUnits,
                                              ctx->Const.MaxTextureCoordUnits));
diff --git a/src/mesa/main/dlist.c b/src/mesa/main/dlist.c
index d4bd56be83..ebcef0268a 100644
--- a/src/mesa/main/dlist.c
+++ b/src/mesa/main/dlist.c
@@ -320,6 +320,8 @@ typedef enum
    /* GL_ARB_draw_buffers */
    OPCODE_DRAW_BUFFERS_ARB,
    /* GL_ATI_fragment_shader */
+   OPCODE_TEX_BUMP_PARAMETER_ATI,
+   /* GL_ATI_fragment_shader */
    OPCODE_BIND_FRAGMENT_SHADER_ATI,
    OPCODE_SET_FRAGMENT_SHADER_CONSTANTS_ATI,
    /* OpenGL 2.0 */
@@ -4803,6 +4805,36 @@ save_DrawBuffersARB(GLsizei count, const GLenum * buffers)
    }
 }
 
+static void GLAPIENTRY
+save_TexBumpParameterfvATI(GLenum pname, const GLfloat *param)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   Node *n;
+
+   n = ALLOC_INSTRUCTION(ctx, OPCODE_TEX_BUMP_PARAMETER_ATI, 5);
+   if (n) {
+      n[1].ui = pname;
+      n[2].f = param[0];
+      n[3].f = param[1];
+      n[4].f = param[2];
+      n[5].f = param[3];
+   }
+   if (ctx->ExecuteFlag) {
+      CALL_TexBumpParameterfvATI(ctx->Exec, (pname, param));
+   }
+}
+
+static void GLAPIENTRY
+save_TexBumpParameterivATI(GLenum pname, const GLint *param)
+{
+   GLfloat p[4];
+   p[0] = INT_TO_FLOAT(param[0]);
+   p[1] = INT_TO_FLOAT(param[1]);
+   p[2] = INT_TO_FLOAT(param[2]);
+   p[3] = INT_TO_FLOAT(param[3]);
+   save_TexBumpParameterfvATI(pname, p);
+}
+
 #if FEATURE_ATI_fragment_shader
 static void GLAPIENTRY
 save_BindFragmentShaderATI(GLuint id)
@@ -6505,6 +6537,16 @@ execute_list(GLcontext *ctx, GLuint list)
                                                 n[9].i, n[10].e));
 	    break;
 #endif
+         case OPCODE_TEX_BUMP_PARAMETER_ATI:
+            {
+               GLfloat values[4];
+               GLuint i, pname = n[1].ui;
+
+               for (i = 0; i < 4; i++)
+                  values[i] = n[1 + i].f;
+               CALL_TexBumpParameterfvATI(ctx->Exec, (pname, values));
+            }
+            break;
 #if FEATURE_ATI_fragment_shader
          case OPCODE_BIND_FRAGMENT_SHADER_ATI:
             CALL_BindFragmentShaderATI(ctx->Exec, (n[1].i));
@@ -8043,6 +8085,10 @@ _mesa_init_dlist_table(struct _glapi_table *table)
    SET_VertexAttribPointerNV(table, _mesa_VertexAttribPointerNV);
 #endif
 
+   /* 244. GL_ATI_envmap_bumpmap */
+   SET_TexBumpParameterivATI(table, save_TexBumpParameterivATI);
+   SET_TexBumpParameterfvATI(table, save_TexBumpParameterfvATI);
+
    /* 245. GL_ATI_fragment_shader */
 #if FEATURE_ATI_fragment_shader
    SET_BindFragmentShaderATI(table, save_BindFragmentShaderATI);
diff --git a/src/mesa/main/extensions.c b/src/mesa/main/extensions.c
index fbca924ad3..2d2bf51784 100644
--- a/src/mesa/main/extensions.c
+++ b/src/mesa/main/extensions.c
@@ -137,6 +137,7 @@ static const struct {
    { ON,  "GL_APPLE_packed_pixels",            F(APPLE_packed_pixels) },
    { OFF, "GL_APPLE_vertex_array_object",      F(APPLE_vertex_array_object) },
    { OFF, "GL_ATI_blend_equation_separate",    F(EXT_blend_equation_separate) },
+   { OFF, "GL_ATI_envmap_bumpmap",             F(ATI_envmap_bumpmap) },
    { OFF, "GL_ATI_texture_env_combine3",       F(ATI_texture_env_combine3)},
    { OFF, "GL_ATI_texture_mirror_once",        F(ATI_texture_mirror_once)},
    { OFF, "GL_ATI_fragment_shader",            F(ATI_fragment_shader)},
@@ -229,6 +230,7 @@ _mesa_enable_sw_extensions(GLcontext *ctx)
    /*ctx->Extensions.ARB_vertex_buffer_object = GL_TRUE;*/
 #endif
    ctx->Extensions.APPLE_vertex_array_object = GL_TRUE;
+   ctx->Extensions.ATI_envmap_bumpmap = GL_TRUE;
 #if FEATURE_ATI_fragment_shader
    ctx->Extensions.ATI_fragment_shader = GL_TRUE;
 #endif
diff --git a/src/mesa/main/image.c b/src/mesa/main/image.c
index 4d86c54777..ed0811660d 100644
--- a/src/mesa/main/image.c
+++ b/src/mesa/main/image.c
@@ -293,6 +293,8 @@ _mesa_components_in_format( GLenum format )
          return 2;
       case GL_DEPTH_STENCIL_EXT:
          return 2;
+      case GL_DUDV_ATI:
+         return 2;
       default:
          return -1;
    }
@@ -503,6 +505,20 @@ _mesa_is_legal_format_and_type( GLcontext *ctx, GLenum format, GLenum type )
             return GL_TRUE;
          else
             return GL_FALSE;
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            default:
+               return GL_FALSE;
+         }
       default:
          ; /* fall-through */
    }
@@ -1674,8 +1690,19 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
    GLfloat luminance[MAX_WIDTH];
    const GLint comps = _mesa_components_in_format(dstFormat);
    GLuint i;
-
-   if (dstType != GL_FLOAT || ctx->Color.ClampReadColor == GL_TRUE) {
+   /* clamping only applies to colors, not the dudv values, but still need
+      it if converting to unsigned values (which doesn't make much sense) */
+   if (dstFormat == GL_DUDV_ATI || dstFormat == GL_DU8DV8_ATI) {
+      switch (dstType) {
+         case GL_UNSIGNED_BYTE:
+         case GL_UNSIGNED_SHORT:
+         case GL_UNSIGNED_INT:
+            transferOps |= IMAGE_CLAMP_BIT;
+            break;
+            /* actually might want clamp to [-1,1] otherwise but shouldn't matter? */
+      }
+   }
+   else if (dstType != GL_FLOAT || ctx->Color.ClampReadColor == GL_TRUE) {
       /* need to clamp to [0, 1] */
       transferOps |= IMAGE_CLAMP_BIT;
    }
@@ -1774,6 +1801,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = FLOAT_TO_UBYTE(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_UBYTE(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_UBYTE(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -1847,6 +1881,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = FLOAT_TO_BYTE(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_BYTE(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_BYTE(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -1920,6 +1961,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      CLAMPED_FLOAT_TO_USHORT(dst[i*4+3], rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_USHORT(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_USHORT(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -1993,6 +2041,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = FLOAT_TO_SHORT(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_SHORT(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_SHORT(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -2066,6 +2121,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = FLOAT_TO_UINT(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_UINT(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_UINT(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -2139,6 +2201,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = FLOAT_TO_INT(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = FLOAT_TO_INT(rgba[i][RCOMP]);
+                     dst[i*2+1] = FLOAT_TO_INT(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -2212,6 +2281,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = rgba[i][RCOMP];
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = rgba[i][RCOMP];
+                     dst[i*2+1] = rgba[i][GCOMP];
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -2285,6 +2361,13 @@ _mesa_pack_rgba_span_float(GLcontext *ctx, GLuint n, GLfloat rgba[][4],
                      dst[i*4+3] = _mesa_float_to_half(rgba[i][RCOMP]);
                   }
                   break;
+               case GL_DUDV_ATI:
+               case GL_DU8DV8_ATI:
+                  for (i=0;i<n;i++) {
+                     dst[i*2+0] = _mesa_float_to_half(rgba[i][RCOMP]);
+                     dst[i*2+1] = _mesa_float_to_half(rgba[i][GCOMP]);
+                  }
+                  break;
                default:
                   _mesa_problem(ctx, "bad format in _mesa_pack_rgba_span\n");
             }
@@ -2834,7 +2917,8 @@ extract_float_rgba(GLuint n, GLfloat rgba[][4],
           srcFormat == GL_BGR ||
           srcFormat == GL_RGBA ||
           srcFormat == GL_BGRA ||
-          srcFormat == GL_ABGR_EXT);
+          srcFormat == GL_ABGR_EXT ||
+          srcFormat == GL_DUDV_ATI);
 
    ASSERT(srcType == GL_UNSIGNED_BYTE ||
           srcType == GL_BYTE ||
@@ -2949,6 +3033,13 @@ extract_float_rgba(GLuint n, GLfloat rgba[][4],
          aComp = 0;
          stride = 4;
          break;
+      case GL_DUDV_ATI:
+         redIndex = 0;
+         greenIndex = 1;
+         blueIndex = -1;
+         alphaIndex = -1;
+         stride = 2;
+         break;
       default:
          _mesa_problem(NULL, "bad srcFormat in extract float data");
          return;
@@ -3877,6 +3968,62 @@ _mesa_unpack_color_span_float( GLcontext *ctx,
    }
 }
 
+/**
+ * Similar to _mesa_unpack_color_span_float(), but for dudv data instead of rgba,
+ * directly return GLbyte data, no transfer ops apply.
+ */
+void
+_mesa_unpack_dudv_span_byte( GLcontext *ctx,
+                             GLuint n, GLenum dstFormat, GLbyte dest[],
+                             GLenum srcFormat, GLenum srcType,
+                             const GLvoid *source,
+                             const struct gl_pixelstore_attrib *srcPacking,
+                             GLbitfield transferOps )
+{
+   ASSERT(dstFormat == GL_DUDV_ATI);
+   ASSERT(srcFormat == GL_DUDV_ATI);
+
+   ASSERT(srcType == GL_UNSIGNED_BYTE ||
+          srcType == GL_BYTE ||
+          srcType == GL_UNSIGNED_SHORT ||
+          srcType == GL_SHORT ||
+          srcType == GL_UNSIGNED_INT ||
+          srcType == GL_INT ||
+          srcType == GL_HALF_FLOAT_ARB ||
+          srcType == GL_FLOAT);
+
+   /* general solution */
+   {
+      GLint dstComponents;
+      GLfloat rgba[MAX_WIDTH][4];
+
+      dstComponents = _mesa_components_in_format( dstFormat );
+      /* source & dest image formats should have been error checked by now */
+      assert(dstComponents > 0);
+
+      /*
+       * Extract image data and convert to RGBA floats
+       */
+      assert(n <= MAX_WIDTH);
+      extract_float_rgba(n, rgba, srcFormat, srcType, source,
+                         srcPacking->SwapBytes);
+
+
+      /* Now determine which color channels we need to produce.
+       * And determine the dest index (offset) within each color tuple.
+       */
+
+      /* Now pack results in the requested dstFormat */
+      GLbyte *dst = dest;
+      GLuint i;
+      for (i = 0; i < n; i++) {
+         /* not sure - need clamp[-1,1] here? */
+         dst[0] = FLOAT_TO_BYTE(rgba[i][RCOMP]);
+         dst[1] = FLOAT_TO_BYTE(rgba[i][GCOMP]);
+         dst += dstComponents;
+      }
+   }
+}
 
 /*
  * Unpack a row of color index data from a client buffer according to
diff --git a/src/mesa/main/image.h b/src/mesa/main/image.h
index 0e0bbd96d8..b26c27e5a8 100644
--- a/src/mesa/main/image.h
+++ b/src/mesa/main/image.h
@@ -198,6 +198,13 @@ _mesa_unpack_color_span_float( GLcontext *ctx,
                                const struct gl_pixelstore_attrib *srcPacking,
                                GLbitfield transferOps );
 
+extern void
+_mesa_unpack_dudv_span_byte( GLcontext *ctx,
+                             GLuint n, GLenum dstFormat, GLbyte dest[],
+                             GLenum srcFormat, GLenum srcType,
+                             const GLvoid *source,
+                             const struct gl_pixelstore_attrib *srcPacking,
+                             GLbitfield transferOps );
 
 extern void
 _mesa_unpack_index_span( const GLcontext *ctx, GLuint n,
diff --git a/src/mesa/main/macros.h b/src/mesa/main/macros.h
index 2630855a0e..bfd740870e 100644
--- a/src/mesa/main/macros.h
+++ b/src/mesa/main/macros.h
@@ -54,13 +54,16 @@ extern GLfloat _mesa_ubyte_to_float_color_tab[256];
 #define FLOAT_TO_BYTE(X)    ( (((GLint) (255.0F * (X))) - 1) / 2 )
 
 
-/** Convert GLushort in [0,65536] to GLfloat in [0.0,1.0] */
+/** Convert GLushort in [0,65535] to GLfloat in [0.0,1.0] */
 #define USHORT_TO_FLOAT(S)  ((GLfloat) (S) * (1.0F / 65535.0F))
 
+/** Convert GLfloat in [0.0,1.0] to GLushort in [0, 65535] */
+#define FLOAT_TO_USHORT(X)   ((GLuint) ((X) * 65535.0))
+
 /** Convert GLshort in [-32768,32767] to GLfloat in [-1.0,1.0] */
 #define SHORT_TO_FLOAT(S)   ((2.0F * (S) + 1.0F) * (1.0F/65535.0F))
 
-/** Convert GLfloat in [0.0,1.0] to GLshort in [-32768,32767] */
+/** Convert GLfloat in [-1.0,1.0] to GLshort in [-32768,32767] */
 #define FLOAT_TO_SHORT(X)   ( (((GLint) (65535.0F * (X))) - 1) / 2 )
 
 
diff --git a/src/mesa/main/mipmap.c b/src/mesa/main/mipmap.c
index 3dd4b3391b..af2bf8fe22 100644
--- a/src/mesa/main/mipmap.c
+++ b/src/mesa/main/mipmap.c
@@ -85,7 +85,7 @@ bytes_per_pixel(GLenum datatype, GLuint comps)
                                 rowC[j][e], rowC[k][e], \
                                 rowD[j][e], rowD[k][e]); \
    } while(0)
-   
+
 #define FILTER_F_3D(e) \
    do { \
       dst[i][e] = (rowA[j][e] + rowA[k][e] \
@@ -226,7 +226,6 @@ do_row(GLenum datatype, GLuint comps, GLint srcWidth,
          dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) / 4;
       }
    }
-
    else if (datatype == GL_FLOAT && comps == 4) {
       GLuint i, j, k;
       const GLfloat(*rowA)[4] = (const GLfloat(*)[4]) srcRowA;
@@ -471,6 +470,17 @@ do_row(GLenum datatype, GLuint comps, GLint srcWidth,
          dst[i] = (blue << 5) | (green << 2) | red;
       }
    }
+   else if (datatype == GL_BYTE && comps == 2) {
+      GLuint i, j, k;
+      const GLbyte(*rowA)[2] = (const GLbyte(*)[2]) srcRowA;
+      const GLbyte(*rowB)[2] = (const GLbyte(*)[2]) srcRowB;
+      GLbyte(*dst)[2] = (GLbyte(*)[2]) dstRow;
+      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
+           i++, j += colStride, k += colStride) {
+         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) >> 2;
+         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) >> 2;
+      }
+   }
    else {
       _mesa_problem(NULL, "bad format in do_row()");
    }
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 9080c83da4..169f1af057 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -1390,6 +1390,8 @@ struct gl_texture_unit
    GLbitfield _GenFlags;	/**< Bitwise-OR of Gen[STRQ]._ModeBit */
 
    GLfloat LodBias;		/**< for biasing mipmap levels */
+   GLenum BumpTarget;
+   GLfloat RotMatrix[4]; /* 2x2 matrix */
 
    /** 
     * \name GL_EXT_texture_env_combine 
@@ -2388,6 +2390,8 @@ struct gl_constants
    GLuint MaxSamples;            /**< GL_ARB_framebuffer_object */
 
    GLuint MaxVarying;  /**< Number of float[4] varying parameters */
+
+   GLbitfield SupportedBumpUnits; /**> units supporting GL_ATI_envmap_bumpmap as targets */
 };
 
 
@@ -2484,6 +2488,7 @@ struct gl_extensions
    GLboolean APPLE_client_storage;
    GLboolean APPLE_packed_pixels;
    GLboolean APPLE_vertex_array_object;
+   GLboolean ATI_envmap_bumpmap;
    GLboolean ATI_texture_mirror_once;
    GLboolean ATI_texture_env_combine3;
    GLboolean ATI_fragment_shader;
diff --git a/src/mesa/main/texenv.c b/src/mesa/main/texenv.c
index 95547a500e..c2960fc820 100644
--- a/src/mesa/main/texenv.c
+++ b/src/mesa/main/texenv.c
@@ -142,7 +142,11 @@ set_combiner_mode(GLcontext *ctx,
    case GL_MODULATE_ADD_ATI:
    case GL_MODULATE_SIGNED_ADD_ATI:
    case GL_MODULATE_SUBTRACT_ATI:
-      legal =ctx->Extensions.ATI_texture_env_combine3;
+      legal = ctx->Extensions.ATI_texture_env_combine3;
+      break;
+   case GL_BUMP_ENVMAP_ATI:
+      legal = (ctx->Extensions.ATI_envmap_bumpmap &&
+               pname == GL_COMBINE_RGB);
       break;
    default:
       legal = GL_FALSE;
@@ -500,6 +504,26 @@ _mesa_TexEnvfv( GLenum target, GLenum pname, const GLfloat *param )
       case GL_ALPHA_SCALE:
          set_combiner_scale(ctx, texUnit, pname, param[0]);
 	 break;
+      case GL_BUMP_TARGET_ATI:
+         if (!ctx->Extensions.ATI_envmap_bumpmap) {
+	    _mesa_error( ctx, GL_INVALID_ENUM, "glTexEnv(pname=0x%x)", pname );
+	    return;
+	 }
+	 if (((GLenum) (GLint) param[0] < GL_TEXTURE0) ||
+	 ((GLenum) (GLint) param[0] > GL_TEXTURE31)) {
+	    /* spec doesn't say this but it seems logical */
+	    _mesa_error( ctx, GL_INVALID_ENUM, "glTexEnv(param=0x%x)", (GLenum) (GLint) param[0]);
+	    return;
+	 }
+	 if (!((1 << ((GLenum) (GLint) param[0] - GL_TEXTURE0)) & ctx->Const.SupportedBumpUnits)) {
+	    _mesa_error( ctx, GL_INVALID_VALUE, "glTexEnv(param=0x%x)", (GLenum) (GLint) param[0]);
+	    return;
+	 }
+	 else {
+	    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+	    texUnit->BumpTarget = (GLenum) (GLint) param[0];
+	 }
+	 break;
       default:
 	 _mesa_error( ctx, GL_INVALID_ENUM, "glTexEnv(pname)" );
 	 return;
@@ -735,6 +759,16 @@ get_texenvi(GLcontext *ctx, const struct gl_texture_unit *texUnit,
          _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexEnvfv(pname)");
       }
       break;
+   case GL_BUMP_TARGET_ATI:
+      /* spec doesn't say so, but I think this should be queryable */
+      if (ctx->Extensions.ATI_envmap_bumpmap) {
+         return texUnit->BumpTarget;
+      }
+      else {
+         _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexEnvfv(pname)");
+      }
+      break;
+
    default:
       ;
    }
@@ -874,4 +908,142 @@ _mesa_GetTexEnviv( GLenum target, GLenum pname, GLint *params )
    }
 }
 
+/* why does ATI_envmap_bumpmap require new entrypoints? Should just
+   reuse TexEnv ones... */
+void GLAPIENTRY
+_mesa_TexBumpParameterivATI( GLenum pname, const GLint *param )
+{
+   GLfloat p[4];
+   if (pname == GL_BUMP_ROT_MATRIX_ATI) {
+      /* hope that conversion is correct here */
+      p[0] = INT_TO_FLOAT( param[0] );
+      p[1] = INT_TO_FLOAT( param[1] );
+      p[2] = INT_TO_FLOAT( param[2] );
+      p[3] = INT_TO_FLOAT( param[3] );
+   }
+   else {
+      p[0] = (GLfloat) param[0];
+      p[1] = p[2] = p[3] = 0;  /* init to zero, just to be safe */
+   }
+   _mesa_TexBumpParameterfvATI( pname, p );
+}
+
+void GLAPIENTRY
+_mesa_TexBumpParameterfvATI( GLenum pname, const GLfloat *param )
+{
+   struct gl_texture_unit *texUnit;
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   /* should return error if extension not supported? */
+   texUnit = &ctx->Texture.Unit[ctx->Texture.CurrentUnit];
+
+   if (pname == GL_BUMP_ROT_MATRIX_ATI) {
+      if (TEST_EQ_4V(param, texUnit->RotMatrix))
+         return;
+      FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+      COPY_4FV(texUnit->RotMatrix, param);
+   }
+   else {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glTexBumpParameter(pname)" );
+      return;
+   }
+   /* Drivers might want to know about this, instead of dedicated function
+      just shove it into TexEnv where it really belongs anyway */
+   if (ctx->Driver.TexEnv) {
+      (*ctx->Driver.TexEnv)( ctx, 0, pname, param );
+   }
+}
+
+void GLAPIENTRY
+_mesa_GetTexBumpParameterivATI( GLenum pname, GLint *param )
+{
+   const struct gl_texture_unit *texUnit;
+   GLint i;
+   GLint temp = 0;
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   /* should return error if extension not supported? */
+   texUnit = &ctx->Texture.Unit[ctx->Texture.CurrentUnit];
+
+   if (pname == GL_BUMP_ROT_MATRIX_SIZE_ATI) {
+      /* spec leaves open to support larger matrices.
+         Don't think anyone would ever want to use it
+         (and apps almost certainly would not understand it and
+         thus fail to submit matrices correctly) so hardcode this. */
+      *param = 4;
+   }
+   else if (pname == GL_BUMP_ROT_MATRIX_ATI) {
+      /* hope that conversion is correct here */
+      param[0] = FLOAT_TO_INT(texUnit->RotMatrix[0]);
+      param[1] = FLOAT_TO_INT(texUnit->RotMatrix[1]);
+      param[2] = FLOAT_TO_INT(texUnit->RotMatrix[2]);
+      param[3] = FLOAT_TO_INT(texUnit->RotMatrix[3]);
+   }
+   else if (pname == GL_BUMP_NUM_TEX_UNITS_ATI) {
+      for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
+         if (ctx->Const.SupportedBumpUnits & (1 << i)) {
+            temp++;
+         }
+      }
+      *param = temp;
+   }
+   else if (pname == GL_BUMP_TEX_UNITS_ATI) {
+      for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
+         if (ctx->Const.SupportedBumpUnits & (1 << i)) {
+            *param++ = i + GL_TEXTURE0;
+         }
+      }
+   }
+   else {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexBumpParameter(pname)" );
+      return;
+   }
+}
+
+void GLAPIENTRY
+_mesa_GetTexBumpParameterfvATI( GLenum pname, GLfloat *param )
+{
+   const struct gl_texture_unit *texUnit;
+   GLint i;
+   GLint temp = 0;
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   /* should return error if extension not supported? */
+   texUnit = &ctx->Texture.Unit[ctx->Texture.CurrentUnit];
+
+   if (pname == GL_BUMP_ROT_MATRIX_SIZE_ATI) {
+      /* spec leaves open to support larger matrices.
+         Don't think anyone would ever want to use it
+         (and apps might not understand it) so hardcode this. */
+      *param = (GLfloat) 4;
+   }
+   else if (pname == GL_BUMP_ROT_MATRIX_ATI) {
+      param[0] = texUnit->RotMatrix[0];
+      param[1] = texUnit->RotMatrix[1];
+      param[2] = texUnit->RotMatrix[2];
+      param[3] = texUnit->RotMatrix[3];
+   }
+   else if (pname == GL_BUMP_NUM_TEX_UNITS_ATI) {
+      for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
+         if (ctx->Const.SupportedBumpUnits & (1 << i)) {
+            temp++;
+         }
+      }
+      *param = (GLfloat) temp;
+   }
+   else if (pname == GL_BUMP_TEX_UNITS_ATI) {
+      for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
+         if (ctx->Const.SupportedBumpUnits & (1 << i)) {
+            *param++ = (GLfloat) (i + GL_TEXTURE0);
+         }
+      }
+   }
+   else {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexBumpParameter(pname)" );
+      return;
+   }
+}
 
diff --git a/src/mesa/main/texenv.h b/src/mesa/main/texenv.h
index bdff7fdb82..1e9c5faed7 100644
--- a/src/mesa/main/texenv.h
+++ b/src/mesa/main/texenv.h
@@ -48,5 +48,16 @@ _mesa_TexEnvi( GLenum target, GLenum pname, GLint param );
 extern void GLAPIENTRY
 _mesa_TexEnviv( GLenum target, GLenum pname, const GLint *param );
 
+extern void GLAPIENTRY
+_mesa_TexBumpParameterivATI( GLenum pname, const GLint *param );
+
+extern void GLAPIENTRY
+_mesa_TexBumpParameterfvATI( GLenum pname, const GLfloat *param );
+
+extern void GLAPIENTRY
+_mesa_GetTexBumpParameterivATI( GLenum pname, GLint *param );
+
+extern void GLAPIENTRY
+_mesa_GetTexBumpParameterfvATI( GLenum pname, GLfloat *param );
 
 #endif /* TEXENV_H */
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 51c13a563d..3fbd119b34 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -94,11 +94,11 @@ struct state_key {
       GLuint ScaleShiftA:2;
 
       GLuint NumArgsRGB:3;
-      GLuint ModeRGB:4;
+      GLuint ModeRGB:5;
       struct mode_opt OptRGB[MAX_TERMS];
 
       GLuint NumArgsA:3;
-      GLuint ModeA:4;
+      GLuint ModeA:5;
       struct mode_opt OptA[MAX_TERMS];
    } unit[8];
 };
@@ -194,7 +194,8 @@ static GLuint translate_source( GLenum src )
 #define MODE_MODULATE_SUBTRACT_ATI      12  /* r = a0 * a2 - a1 */
 #define MODE_ADD_PRODUCTS               13  /* r = a0 * a1 + a2 * a3 */
 #define MODE_ADD_PRODUCTS_SIGNED        14  /* r = a0 * a1 + a2 * a3 - 0.5 */
-#define MODE_UNKNOWN                    15
+#define MODE_BUMP_ENVMAP_ATI            15  /* special */
+#define MODE_UNKNOWN                    16
 
 /**
  * Translate GL combiner state into a MODE_x value
@@ -223,6 +224,7 @@ static GLuint translate_mode( GLenum envMode, GLenum mode )
    case GL_MODULATE_ADD_ATI: return MODE_MODULATE_ADD_ATI;
    case GL_MODULATE_SIGNED_ADD_ATI: return MODE_MODULATE_SIGNED_ADD_ATI;
    case GL_MODULATE_SUBTRACT_ATI: return MODE_MODULATE_SUBTRACT_ATI;
+   case GL_BUMP_ENVMAP_ATI: return MODE_BUMP_ENVMAP_ATI;
    default:
       assert(0);
       return MODE_UNKNOWN;
@@ -383,7 +385,7 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
 	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeRGB);
       key->unit[i].ModeA =
 	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeA);
-		
+
       key->unit[i].ScaleShiftRGB = texUnit->_CurrentCombine->ScaleShiftRGB;
       key->unit[i].ScaleShiftA = texUnit->_CurrentCombine->ScaleShiftA;
 
@@ -397,8 +399,18 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
          key->unit[i].OptA[j].Source =
 	    translate_source(texUnit->_CurrentCombine->SourceA[j]);
       }
+
+      if (key->unit[i].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
+         /* requires some special translation */
+         key->unit[i].NumArgsRGB = 2;
+         key->unit[i].ScaleShiftRGB = 0;
+         key->unit[i].OptRGB[0].Operand = OPR_SRC_COLOR;
+         key->unit[i].OptRGB[0].Source = SRC_TEXTURE;
+         key->unit[i].OptRGB[1].Operand = OPR_SRC_COLOR;
+         key->unit[i].OptRGB[1].Source = texUnit->BumpTarget - GL_TEXTURE0 + SRC_TEXTURE0;
+       }
    }
-	
+
    if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) {
       key->separate_specular = 1;
       inputs_referenced |= FRAG_BIT_COL1;
@@ -464,6 +476,11 @@ struct texenv_fragment_program {
     * else undef.
     */
 
+   struct ureg texcoord_tex[MAX_TEXTURE_COORD_UNITS];
+   /* Reg containing texcoord for a texture unit,
+    * needed for bump mapping, else undef.
+    */
+
    struct ureg src_previous;	/**< Reg containing color from previous 
 				 * stage.  May need to be decl'd.
 				 */
@@ -756,6 +773,7 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
     */
    reserve_temp(p, dest);
 
+#if 0
    /* Is this a texture indirection?
     */
    if ((coord.file == PROGRAM_TEMPORARY &&
@@ -767,6 +785,7 @@ static struct ureg emit_texld( struct texenv_fragment_program *p,
       p->alu_temps = 0;
       assert(0);		/* KW: texture env crossbar */
    }
+#endif
 
    return dest;
 }
@@ -1052,6 +1071,10 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
          emit_arith( p, OPCODE_SUB, dest, mask, saturate, tmp0, half, undef );
       }
       return dest;
+   case MODE_BUMP_ENVMAP_ATI:
+      /* special - not handled here */
+      assert(0);
+      return src[0];
    default: 
       assert(0);
       return src[0];
@@ -1074,6 +1097,10 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    if (!key->unit[unit].enabled) {
       return get_source(p, SRC_PREVIOUS, 0);
    }
+   if (key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
+      /* this isn't really a env stage delivering a color and handled elsewhere */
+      return get_source(p, SRC_PREVIOUS, 0);
+   }
    
    switch (key->unit[unit].ModeRGB) {
    case MODE_DOT3_RGB_EXT:
@@ -1163,9 +1190,17 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 {
    if (is_undef(p->src_texture[unit])) {
       GLuint texTarget = p->state->unit[unit].source_index;
-      struct ureg texcoord = register_input(p, FRAG_ATTRIB_TEX0+unit);
+      struct ureg texcoord;
       struct ureg tmp = get_tex_temp( p );
 
+      if (is_undef(p->texcoord_tex[unit])) {
+         texcoord = register_input(p, FRAG_ATTRIB_TEX0+unit);
+      }
+      else {
+         /* might want to reuse this reg for tex output actually */
+         texcoord = p->texcoord_tex[unit];
+      }
+
       if (texTarget == TEXTURE_UNKNOWN_INDEX)
          program_error(p, "TexSrcBit");
 			  
@@ -1233,7 +1268,7 @@ load_texunit_sources( struct texenv_fragment_program *p, int unit )
    GLuint i;
 
    for (i = 0; i < key->unit[unit].NumArgsRGB; i++) {
-      load_texenv_source( p, key->unit[unit].OptRGB[i].Source, unit);
+      load_texenv_source( p, key->unit[unit].OptRGB[i].Source, unit );
    }
 
    for (i = 0; i < key->unit[unit].NumArgsA; i++) {
@@ -1243,6 +1278,40 @@ load_texunit_sources( struct texenv_fragment_program *p, int unit )
    return GL_TRUE;
 }
 
+/**
+ * Generate instructions for loading bump map textures.
+ */
+static GLboolean
+load_texunit_bumpmap( struct texenv_fragment_program *p, int unit )
+{
+   struct state_key *key = p->state;
+   GLuint bumpedUnitNr = key->unit[unit].OptRGB[1].Source - SRC_TEXTURE0;
+   struct ureg texcDst, bumpMapRes;
+   struct ureg constdudvcolor = register_const4f(p, 0.0, 0.0, 0.0, 1.0);
+   struct ureg texcSrc = register_input(p, FRAG_ATTRIB_TEX0 + bumpedUnitNr);
+   struct ureg rotMat0 = register_param3( p, STATE_INTERNAL, STATE_ROT_MATRIX_0, unit );
+   struct ureg rotMat1 = register_param3( p, STATE_INTERNAL, STATE_ROT_MATRIX_1, unit );
+
+   load_texenv_source( p, unit + SRC_TEXTURE0, unit );
+
+   bumpMapRes = get_source(p, key->unit[unit].OptRGB[0].Source, unit);
+   texcDst = get_tex_temp( p );
+   p->texcoord_tex[bumpedUnitNr] = texcDst;
+
+   /* apply rot matrix and add coords to be available in next phase */
+   /* dest = (Arg0.xxxx * rotMat0 + Arg1) + (Arg0.yyyy * rotMat1) */
+   /* note only 2 coords are affected the rest are left unchanged (mul by 0) */
+   emit_arith( p, OPCODE_MAD, texcDst, WRITEMASK_XYZW, 0,
+               swizzle1(bumpMapRes, SWIZZLE_X), rotMat0, texcSrc );
+   emit_arith( p, OPCODE_MAD, texcDst, WRITEMASK_XYZW, 0,
+               swizzle1(bumpMapRes, SWIZZLE_Y), rotMat1, texcDst );
+
+   /* move 0,0,0,1 into bumpmap src if someone (crossbar) is foolish
+      enough to access this later, should optimize away */
+   emit_arith( p, OPCODE_MOV, bumpMapRes, WRITEMASK_XYZW, 0, constdudvcolor, undef, undef );
+
+   return GL_TRUE;
+}
 
 /**
  * Generate a new fragment program which implements the context's
@@ -1267,7 +1336,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
     */
    p.program->Base.Instructions = instBuffer;
    p.program->Base.Target = GL_FRAGMENT_PROGRAM_ARB;
-   p.program->Base.NumTexIndirections = 1;	/* correct? */
+   p.program->Base.NumTexIndirections = 1;
    p.program->Base.NumTexInstructions = 0;
    p.program->Base.NumAluInstructions = 0;
    p.program->Base.String = NULL;
@@ -1280,8 +1349,10 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    p.program->Base.InputsRead = 0;
    p.program->Base.OutputsWritten = 1 << FRAG_RESULT_COLOR;
 
-   for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
+   for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++) {
       p.src_texture[unit] = undef;
+      p.texcoord_tex[unit] = undef;
+   }
 
    p.src_previous = undef;
    p.half = undef;
@@ -1292,6 +1363,16 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    release_temps(ctx, &p);
 
    if (key->enabled_units) {
+       GLboolean needbumpstage = GL_FALSE;
+      /* Zeroth pass - bump map textures first */
+      for (unit = 0 ; unit < ctx->Const.MaxTextureUnits ; unit++)
+	 if (key->unit[unit].enabled && key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
+	    needbumpstage = GL_TRUE;
+	    load_texunit_bumpmap( &p, unit );
+	 }
+      if (needbumpstage)
+	 p.program->Base.NumTexIndirections++;
+
       /* First pass - to support texture_env_crossbar, first identify
        * all referenced texture sources and emit texld instructions
        * for each:
diff --git a/src/mesa/main/texformat.c b/src/mesa/main/texformat.c
index 16d05cc7d0..61a12493a6 100644
--- a/src/mesa/main/texformat.c
+++ b/src/mesa/main/texformat.c
@@ -696,6 +696,33 @@ const struct gl_texture_format _mesa_texformat_intensity_float16 = {
    store_texel_intensity_f16		/* StoreTexel */
 };
 
+const struct gl_texture_format _mesa_texformat_dudv8 = {
+   MESA_FORMAT_DUDV8,			/* MesaFormat */
+   GL_DUDV_ATI,				/* BaseFormat */
+   /* FIXME: spec doesn't say since that parameter didn't exist then,
+      but this should be something like SIGNED_NORMALIZED */
+   GL_UNSIGNED_NORMALIZED_ARB,		/* DataType */
+   /* maybe should add dudvBits field, but spec seems to be
+      lacking the ability to query with GetTexLevelParameter anyway */
+   0,					/* RedBits */
+   0,					/* GreenBits */
+   0,					/* BlueBits */
+   0,					/* AlphaBits */
+   0,					/* LuminanceBits */
+   0,					/* IntensityBits */
+   0,					/* IndexBits */
+   0,					/* DepthBits */
+   0,					/* StencilBits */
+   2,					/* TexelBytes */
+   _mesa_texstore_dudv8,		/* StoreTexImageFunc */
+   NULL,				/* FetchTexel1D */
+   NULL,		/* FetchTexel2D */
+   NULL,				/* FetchTexel3D */
+   NULL,				/* FetchTexel1Df */
+   fetch_texel_2d_dudv8,		/* FetchTexel2Df */
+   NULL,				/* FetchTexel3Df */
+   NULL					/* StoreTexel */
+};
 
 /*@}*/
 
@@ -1634,6 +1661,16 @@ _mesa_choose_tex_format( GLcontext *ctx, GLint internalFormat,
       }
    }
 
+   if (ctx->Extensions.ATI_envmap_bumpmap) {
+      switch (internalFormat) {
+         case GL_DUDV_ATI:
+         case GL_DU8DV8_ATI:
+            return &_mesa_texformat_dudv8;
+         default:
+            ; /* fallthrough */
+      }
+   }
+
 #if FEATURE_EXT_texture_sRGB
    if (ctx->Extensions.EXT_texture_sRGB) {
       switch (internalFormat) {
@@ -1778,6 +1815,11 @@ _mesa_format_to_type_and_comps(const struct gl_texture_format *format,
       *comps = 1;
       return;
 
+   case MESA_FORMAT_DUDV8:
+      *datatype = GL_BYTE;
+      *comps = 2;
+      return;
+
 #if FEATURE_EXT_texture_sRGB
    case MESA_FORMAT_SRGB8:
       *datatype = GL_UNSIGNED_BYTE;
diff --git a/src/mesa/main/texformat.h b/src/mesa/main/texformat.h
index 31364c36b1..7fa70ad4fe 100644
--- a/src/mesa/main/texformat.h
+++ b/src/mesa/main/texformat.h
@@ -161,7 +161,14 @@ enum _format {
    MESA_FORMAT_LUMINANCE_ALPHA_FLOAT32,
    MESA_FORMAT_LUMINANCE_ALPHA_FLOAT16,
    MESA_FORMAT_INTENSITY_FLOAT32,
-   MESA_FORMAT_INTENSITY_FLOAT16
+   MESA_FORMAT_INTENSITY_FLOAT16,
+   /*@}*/
+
+   /**
+    * \name Signed fixed point texture formats.
+    */
+   /*@{*/
+   MESA_FORMAT_DUDV8
    /*@}*/
 };
 
@@ -209,6 +216,11 @@ extern const struct gl_texture_format _mesa_texformat_intensity_float32;
 extern const struct gl_texture_format _mesa_texformat_intensity_float16;
 /*@}*/
 
+/** Signed fixed point texture formats */
+/*@{*/
+extern const struct gl_texture_format _mesa_texformat_dudv8;
+/*@}*/
+
 /** \name Assorted hardware-friendly formats */
 /*@{*/
 extern const struct gl_texture_format _mesa_texformat_rgba8888;
diff --git a/src/mesa/main/texformat_tmp.h b/src/mesa/main/texformat_tmp.h
index 275340cabd..f9f9d5b1f8 100644
--- a/src/mesa/main/texformat_tmp.h
+++ b/src/mesa/main/texformat_tmp.h
@@ -1269,7 +1269,7 @@ static void FETCH(sl8)(const struct gl_texture_image *texImage,
    texel[RCOMP] = 
    texel[GCOMP] = 
    texel[BCOMP] = nonlinear_to_linear(src[0]);
-   texel[ACOMP] = CHAN_MAX;
+   texel[ACOMP] = 1.0F;
 }
 
 #if DIM == 3
@@ -1308,7 +1308,22 @@ static void store_texel_sla8(struct gl_texture_image *texImage,
 
 #endif /* FEATURE_EXT_texture_sRGB */
 
+#if DIM == 2
+/* MESA_FORMAT_DUDV8 ********************************************************/
+
+/* this format by definition produces 0,0,0,1 as rgba values,
+   however we'll return the dudv values as rg and fix up elsewhere */
+static void FETCH(dudv8)(const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte *src = TEXEL_ADDR(GLbyte, texImage, i, j, k, 2);
+   texel[RCOMP] = BYTE_TO_FLOAT(src[0]);
+   texel[GCOMP] = BYTE_TO_FLOAT(src[1]);
+   texel[BCOMP] = 0;
+   texel[ACOMP] = 0;
 
+}
+#endif
 
 /* MESA_FORMAT_YCBCR *********************************************************/
 
diff --git a/src/mesa/main/teximage.c b/src/mesa/main/teximage.c
index e3d4404759..2ed72280ec 100644
--- a/src/mesa/main/teximage.c
+++ b/src/mesa/main/teximage.c
@@ -339,6 +339,17 @@ _mesa_base_tex_format( GLcontext *ctx, GLint internalFormat )
       }
    }
 
+   if (ctx->Extensions.ATI_envmap_bumpmap) {
+      switch (internalFormat) {
+         case GL_DUDV_ATI:
+         case GL_DU8DV8_ATI:
+            return GL_DUDV_ATI;
+         default:
+            ; /* fallthrough */
+      }
+   }
+
+
    if (ctx->Extensions.EXT_packed_depth_stencil) {
       switch (internalFormat) {
          case GL_DEPTH_STENCIL_EXT:
@@ -568,6 +579,20 @@ is_depthstencil_format(GLenum format)
    }
 }
 
+/**
+ * Test if the given image format is a dudv format.
+ */
+static GLboolean
+is_dudv_format(GLenum format)
+{
+   switch (format) {
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
 
 
 /**
@@ -1539,7 +1564,8 @@ texture_error_check( GLcontext *ctx, GLenum target,
        (is_index_format(internalFormat) && !indexFormat) ||
        (is_depth_format(internalFormat) != is_depth_format(format)) ||
        (is_ycbcr_format(internalFormat) != is_ycbcr_format(format)) ||
-       (is_depthstencil_format(internalFormat) != is_depthstencil_format(format))) {
+       (is_depthstencil_format(internalFormat) != is_depthstencil_format(format)) ||
+       (is_dudv_format(internalFormat) != is_dudv_format(format))) {
       if (!isProxy)
          _mesa_error(ctx, GL_INVALID_OPERATION,
                      "glTexImage(internalFormat/format)");
@@ -2273,6 +2299,12 @@ _mesa_GetTexImage( GLenum target, GLint level, GLenum format,
       return;
    }
 
+   if (!ctx->Extensions.ATI_envmap_bumpmap
+       && is_dudv_format(format)) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexImage(format)");
+      return;
+   }
+
    _mesa_lock_texture(ctx, texObj);
    {
       texImage = _mesa_select_tex_image(ctx, texObj, target, level);
@@ -2313,6 +2345,11 @@ _mesa_GetTexImage( GLenum target, GLint level, GLenum format,
 	 _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexImage(format mismatch)");
 	 goto out;
       }
+      else if (is_dudv_format(format)
+	       && !is_dudv_format(texImage->TexFormat->BaseFormat)) {
+	 _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexImage(format mismatch)");
+	 goto out;
+      }
 
       if (ctx->Pack.BufferObj->Name) {
 	 /* packing texture image into a PBO */
diff --git a/src/mesa/main/texstate.c b/src/mesa/main/texstate.c
index d71f9530cb..a5af622910 100644
--- a/src/mesa/main/texstate.c
+++ b/src/mesa/main/texstate.c
@@ -95,6 +95,11 @@ _mesa_copy_texture_state( const GLcontext *src, GLcontext *dst )
       /* GL_EXT_texture_env_combine */
       dst->Texture.Unit[u].Combine = src->Texture.Unit[u].Combine;
 
+      /* GL_ATI_envmap_bumpmap - need this? */
+      dst->Texture.Unit[u].BumpTarget = src->Texture.Unit[u].BumpTarget;
+      COPY_4V(dst->Texture.Unit[u].RotMatrix, src->Texture.Unit[u].RotMatrix);
+
+
       /* copy texture object bindings, not contents of texture objects */
       _mesa_lock_context_textures(dst);
 
@@ -411,6 +416,10 @@ update_tex_combine(GLcontext *ctx, struct gl_texture_unit *texUnit)
    case GL_MODULATE_SUBTRACT_ATI:
       combine->_NumArgsRGB = 3;
       break;
+   case GL_BUMP_ENVMAP_ATI:
+      /* no real arguments for this case */
+      combine->_NumArgsRGB = 0;
+      break;
    default:
       combine->_NumArgsRGB = 0;
       _mesa_problem(ctx, "invalid RGB combine mode in update_texture_state");
@@ -682,6 +691,7 @@ init_texture_unit( GLcontext *ctx, GLuint unit )
    texUnit->Combine = default_combine_state;
    texUnit->_EnvMode = default_combine_state;
    texUnit->_CurrentCombine = & texUnit->_EnvMode;
+   texUnit->BumpTarget = GL_TEXTURE0;
 
    texUnit->TexGenEnabled = 0x0;
    texUnit->GenS.Mode = GL_EYE_LINEAR;
@@ -702,6 +712,16 @@ init_texture_unit( GLcontext *ctx, GLuint unit )
    ASSIGN_4V( texUnit->GenT.EyePlane, 0.0, 1.0, 0.0, 0.0 );
    ASSIGN_4V( texUnit->GenR.EyePlane, 0.0, 0.0, 0.0, 0.0 );
    ASSIGN_4V( texUnit->GenQ.EyePlane, 0.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenS.ObjectPlane, 1.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenT.ObjectPlane, 0.0, 1.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenR.ObjectPlane, 0.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenQ.ObjectPlane, 0.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenS.EyePlane, 1.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenT.EyePlane, 0.0, 1.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenR.EyePlane, 0.0, 0.0, 0.0, 0.0 );
+   ASSIGN_4V( texUnit->GenQ.EyePlane, 0.0, 0.0, 0.0, 0.0 );
+   /* no mention of this in spec, but maybe id matrix expected? */
+   ASSIGN_4V( texUnit->RotMatrix, 1.0, 0.0, 0.0, 1.0 );
 
    /* initialize current texture object ptrs to the shared default objects */
    for (tex = 0; tex < NUM_TEXTURE_TARGETS; tex++) {
diff --git a/src/mesa/main/texstore.c b/src/mesa/main/texstore.c
index 6360ca15f8..cc3c6958c7 100644
--- a/src/mesa/main/texstore.c
+++ b/src/mesa/main/texstore.c
@@ -2471,6 +2471,95 @@ _mesa_texstore_ycbcr(TEXSTORE_PARAMS)
    return GL_TRUE;
 }
 
+GLboolean
+_mesa_texstore_dudv8(TEXSTORE_PARAMS)
+{
+   const GLboolean littleEndian = _mesa_little_endian();
+
+   ASSERT(dstFormat == &_mesa_texformat_dudv8);
+   ASSERT(dstFormat->TexelBytes == 2);
+   ASSERT(ctx->Extensions.ATI_envmap_bumpmap);
+   ASSERT((srcFormat == GL_DU8DV8_ATI) ||
+	  (srcFormat == GL_DUDV_ATI));
+   ASSERT(baseInternalFormat == GL_DUDV_ATI);
+
+   if (!srcPacking->SwapBytes && srcType == GL_BYTE &&
+       littleEndian) {
+      /* simple memcpy path */
+      memcpy_texture(ctx, dims,
+                     dstFormat, dstAddr, dstXoffset, dstYoffset, dstZoffset,
+                     dstRowStride,
+                     dstImageOffsets,
+                     srcWidth, srcHeight, srcDepth, srcFormat, srcType,
+                     srcAddr, srcPacking);
+   }
+   else if (srcType == GL_BYTE) {
+
+      GLubyte dstmap[4];
+
+      /* dstmap - how to swizzle from RGBA to dst format:
+       */
+      if (littleEndian) {
+	 dstmap[0] = 0;
+	 dstmap[1] = 3;
+      }
+      else {
+	 dstmap[0] = 3;
+	 dstmap[1] = 0;
+      }
+      dstmap[2] = ZERO;		/* ? */
+      dstmap[3] = ONE;		/* ? */
+      
+      _mesa_swizzle_ubyte_image(ctx, dims,
+				GL_LUMINANCE_ALPHA, /* hack */
+				GL_UNSIGNED_BYTE, /* hack */
+				GL_LUMINANCE_ALPHA, /* hack */
+				dstmap, 2,
+				dstAddr, dstXoffset, dstYoffset, dstZoffset,
+				dstRowStride, dstImageOffsets,
+				srcWidth, srcHeight, srcDepth, srcAddr,
+				srcPacking);      
+   }   
+   else {
+      /* general path - note this is defined for 2d textures only */
+      const GLint components = _mesa_components_in_format(baseInternalFormat);
+      const GLint srcStride = _mesa_image_row_stride(srcPacking,
+                                                 srcWidth, srcFormat, srcType);
+      GLbyte *tempImage, *dst, *src;
+      GLint row;
+
+      tempImage = (GLbyte *) _mesa_malloc(srcWidth * srcHeight * srcDepth
+                                          * components * sizeof(GLbyte));
+      if (!tempImage)
+         return GL_FALSE;
+
+      src = (GLbyte *) _mesa_image_address(dims, srcPacking, srcAddr,
+                                           srcWidth, srcHeight,
+                                           srcFormat, srcType,
+                                           0, 0, 0);
+
+      dst = tempImage;
+      for (row = 0; row < srcHeight; row++) {
+         _mesa_unpack_dudv_span_byte(ctx, srcWidth, baseInternalFormat,
+                                     dst, srcFormat, srcType, src,
+                                     srcPacking, 0);
+         dst += srcWidth * components;
+         src += srcStride;
+      }
+ 
+      src = tempImage;
+      dst = (GLbyte *) dstAddr
+            + dstYoffset * dstRowStride
+            + dstXoffset * dstFormat->TexelBytes;
+      for (row = 0; row < srcHeight; row++) {
+         memcpy(dst, src, srcWidth * dstFormat->TexelBytes);
+         dst += dstRowStride;
+         src += srcWidth * dstFormat->TexelBytes;
+      }
+      _mesa_free((void *) tempImage);
+   }
+   return GL_TRUE;
+}
 
 
 /**
@@ -3882,7 +3971,7 @@ _mesa_get_teximage(GLcontext *ctx, GLenum target, GLint level,
                GLint col;
                GLbitfield transferOps = 0x0;
 
-               if (type == GL_FLOAT && 
+               if (type == GL_FLOAT && texImage->TexFormat->BaseFormat != GL_DUDV_ATI &&
                    ((ctx->Color.ClampReadColor == GL_TRUE) ||
                     (ctx->Color.ClampReadColor == GL_FIXED_ONLY_ARB &&
                      texImage->TexFormat->DataType != GL_FLOAT)))
diff --git a/src/mesa/main/texstore.h b/src/mesa/main/texstore.h
index b03386b2ac..c9e639be4e 100644
--- a/src/mesa/main/texstore.h
+++ b/src/mesa/main/texstore.h
@@ -78,7 +78,7 @@ extern GLboolean _mesa_texstore_sargb8(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_sl8(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_sla8(TEXSTORE_PARAMS);
 #endif
-
+extern GLboolean _mesa_texstore_dudv8(TEXSTORE_PARAMS);
 
 extern GLchan *
 _mesa_make_temp_chan_image(GLcontext *ctx, GLuint dims,
diff --git a/src/mesa/shader/prog_statevars.c b/src/mesa/shader/prog_statevars.c
index 1f7d87ccc9..f51d9e2651 100644
--- a/src/mesa/shader/prog_statevars.c
+++ b/src/mesa/shader/prog_statevars.c
@@ -506,6 +506,26 @@ _mesa_fetch_state(GLcontext *ctx, const gl_state_index state[],
             }
          }
          return;
+      case STATE_ROT_MATRIX_0:
+         {
+            const int unit = (int) state[2];
+            GLfloat *rotMat22 = ctx->Texture.Unit[unit].RotMatrix;
+            value[0] = rotMat22[0]; 
+            value[1] = rotMat22[2];
+            value[2] = 0.0;
+            value[3] = 0.0;
+         }
+         break;
+      case STATE_ROT_MATRIX_1:
+         {
+            const int unit = (int) state[2];
+            GLfloat *rotMat22 = ctx->Texture.Unit[unit].RotMatrix;
+            value[0] = rotMat22[1];
+            value[1] = rotMat22[3];
+            value[2] = 0.0;
+            value[3] = 0.0;
+         }
+         break;
 
          /* XXX: make sure new tokens added here are also handled in the 
           * _mesa_program_state_flags() switch, below.
@@ -591,6 +611,8 @@ _mesa_program_state_flags(const gl_state_index state[STATE_LENGTH])
 
       case STATE_TEXRECT_SCALE:
       case STATE_SHADOW_AMBIENT:
+      case STATE_ROT_MATRIX_0:
+      case STATE_ROT_MATRIX_1:
 	 return _NEW_TEXTURE;
       case STATE_FOG_PARAMS_OPTIMIZED:
 	 return _NEW_FOG;
@@ -806,6 +828,12 @@ append_token(char *dst, gl_state_index k)
    case STATE_SHADOW_AMBIENT:
       append(dst, "CompareFailValue");
       break;
+   case STATE_ROT_MATRIX_0:
+      append(dst, "rotMatrixRow0");
+      break;
+   case STATE_ROT_MATRIX_1:
+      append(dst, "rotMatrixRow1");
+      break;
    default:
       /* probably STATE_INTERNAL_DRIVER+i (driver private state) */
       append(dst, "driverState");
diff --git a/src/mesa/shader/prog_statevars.h b/src/mesa/shader/prog_statevars.h
index d5358a1d04..d563080db1 100644
--- a/src/mesa/shader/prog_statevars.h
+++ b/src/mesa/shader/prog_statevars.h
@@ -117,6 +117,8 @@ typedef enum gl_state_index_ {
    STATE_PCM_SCALE,             /**< Post color matrix RGBA scale */
    STATE_PCM_BIAS,              /**< Post color matrix RGBA bias */
    STATE_SHADOW_AMBIENT,        /**< ARB_shadow_ambient fail value; token[2] is texture unit index */
+   STATE_ROT_MATRIX_0,          /**< ATI_envmap_bumpmap, rot matrix row 0 */
+   STATE_ROT_MATRIX_1,          /**< ATI_envmap_bumpmap, rot matrix row 1 */
    STATE_INTERNAL_DRIVER	/* first available state index for drivers (must be last) */
 } gl_state_index;
 
diff --git a/src/mesa/swrast/s_texcombine.c b/src/mesa/swrast/s_texcombine.c
index 38b5633887..aa28311672 100644
--- a/src/mesa/swrast/s_texcombine.c
+++ b/src/mesa/swrast/s_texcombine.c
@@ -591,6 +591,25 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
             }
 	 }
          break;
+      case GL_BUMP_ENVMAP_ATI:
+         {
+            /* this produces a fixed rgba color, and the coord calc is done elsewhere */
+            for (i = 0; i < n; i++) {
+            /* rgba result is 0,0,0,1 */
+#if CHAN_TYPE == GL_FLOAT
+               rgba[i][RCOMP] = 0.0;
+               rgba[i][GCOMP] = 0.0;
+               rgba[i][BCOMP] = 0.0;
+               rgba[i][ACOMP] = 1.0;
+#else
+               rgba[i][RCOMP] = 0;
+               rgba[i][GCOMP] = 0;
+               rgba[i][BCOMP] = 0;
+               rgba[i][ACOMP] = CHAN_MAX;
+#endif
+            }
+	 }
+         return; /* no alpha processing */
       default:
          _mesa_problem(ctx, "invalid combine mode");
    }
@@ -1218,12 +1237,86 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span )
    if (swrast->_AnyTextureCombine)
       MEMCPY(primary_rgba, span->array->rgba, 4 * span->end * sizeof(GLchan));
 
+   /* First must sample all bump maps */
+   for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++) {
+      if (ctx->Texture.Unit[unit]._ReallyEnabled &&
+         ctx->Texture.Unit[unit]._CurrentCombine->ModeRGB == GL_BUMP_ENVMAP_ATI) {
+         const GLfloat (*texcoords)[4]
+            = (const GLfloat (*)[4])
+            span->array->attribs[FRAG_ATTRIB_TEX0 + unit];
+         GLfloat (*targetcoords)[4]
+            = (GLfloat (*)[4])
+            span->array->attribs[FRAG_ATTRIB_TEX0 +
+               ctx->Texture.Unit[unit].BumpTarget - GL_TEXTURE0];
+
+         const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
+         const struct gl_texture_object *curObj = texUnit->_Current;
+         GLfloat *lambda = span->array->lambda[unit];
+         GLchan (*texels)[4] = (GLchan (*)[4])
+            (swrast->TexelBuffer + unit * (span->end * 4 * sizeof(GLchan)));
+         GLuint i;
+         GLfloat rotMatrix00 = ctx->Texture.Unit[unit].RotMatrix[0];
+         GLfloat rotMatrix01 = ctx->Texture.Unit[unit].RotMatrix[1];
+         GLfloat rotMatrix10 = ctx->Texture.Unit[unit].RotMatrix[2];
+         GLfloat rotMatrix11 = ctx->Texture.Unit[unit].RotMatrix[3];
+
+         /* adjust texture lod (lambda) */
+         if (span->arrayMask & SPAN_LAMBDA) {
+            if (texUnit->LodBias + curObj->LodBias != 0.0F) {
+               /* apply LOD bias, but don't clamp yet */
+               const GLfloat bias = CLAMP(texUnit->LodBias + curObj->LodBias,
+                                          -ctx->Const.MaxTextureLodBias,
+                                          ctx->Const.MaxTextureLodBias);
+               GLuint i;
+               for (i = 0; i < span->end; i++) {
+                  lambda[i] += bias;
+               }
+            }
+
+            if (curObj->MinLod != -1000.0 || curObj->MaxLod != 1000.0) {
+               /* apply LOD clamping to lambda */
+               const GLfloat min = curObj->MinLod;
+               const GLfloat max = curObj->MaxLod;
+               GLuint i;
+               for (i = 0; i < span->end; i++) {
+                  GLfloat l = lambda[i];
+                  lambda[i] = CLAMP(l, min, max);
+               }
+            }
+         }
+
+         /* Sample the texture (span->end = number of fragments) */
+         swrast->TextureSample[unit]( ctx, texUnit->_Current, span->end,
+                                      texcoords, lambda, texels );
+
+         /* manipulate the span values of the bump target
+            not sure this can work correctly even ignoring
+            the problem that channel is unsigned */
+         for (i = 0; i < span->end; i++) {
+#if CHAN_TYPE == GL_FLOAT
+            targetcoords[i][0] += (texels[i][0] * rotMatrix00 + texels[i][1] *
+                                  rotMatrix01) / targetcoords[i][3];
+            targetcoords[i][1] += (texels[i][0] * rotMatrix10 + texels[i][1] *
+                                  rotMatrix11) / targetcoords[i][3];
+#else
+            targetcoords[i][0] += (CHAN_TO_FLOAT(texels[i][1]) * rotMatrix00 +
+                                  CHAN_TO_FLOAT(texels[i][1]) * rotMatrix01) /
+                                  targetcoords[i][3];
+            targetcoords[i][1] += (CHAN_TO_FLOAT(texels[i][0]) * rotMatrix10 + 
+                                  CHAN_TO_FLOAT(texels[i][1]) * rotMatrix11) /
+                                  targetcoords[i][3];
+#endif
+         }
+      }
+   }
+
    /*
     * Must do all texture sampling before combining in order to
     * accomodate GL_ARB_texture_env_crossbar.
     */
    for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++) {
-      if (ctx->Texture.Unit[unit]._ReallyEnabled) {
+      if (ctx->Texture.Unit[unit]._ReallyEnabled &&
+         ctx->Texture.Unit[unit]._CurrentCombine->ModeRGB != GL_BUMP_ENVMAP_ATI) {
          const GLfloat (*texcoords)[4]
             = (const GLfloat (*)[4])
             span->array->attribs[FRAG_ATTRIB_TEX0 + unit];
-- 
cgit v1.2.3


From bd9b2be8284fda3f8aac235908ded118b5648a38 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Sun, 8 Mar 2009 17:58:54 -0600
Subject: mesa: texture combine clean-ups

Use MAX_COMBINER_TERMS instead of 4.
Rename some vars.
Update comments.
---
 src/mesa/main/mtypes.h         |  15 +++--
 src/mesa/main/texenv.c         |  50 +++++++--------
 src/mesa/main/texenvprogram.c  |  13 ++--
 src/mesa/swrast/s_fragprog.c   |   2 -
 src/mesa/swrast/s_texcombine.c | 140 ++++++++++++++++++++---------------------
 5 files changed, 108 insertions(+), 112 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index e77dd1d226..c1f06885ec 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -1338,18 +1338,23 @@ struct gl_texture_object
 };
 
 
+/** Up to four combiner sources are possible with GL_NV_texture_env_combine4 */
+#define MAX_COMBINER_TERMS 4
+
+
 /**
  * Texture combine environment state.
- * Up to four combiner sources are possible with GL_NV_texture_env_combine4.
  */
 struct gl_tex_env_combine_state
 {
    GLenum ModeRGB;       /**< GL_REPLACE, GL_DECAL, GL_ADD, etc. */
    GLenum ModeA;         /**< GL_REPLACE, GL_DECAL, GL_ADD, etc. */
-   GLenum SourceRGB[4];  /**< GL_PRIMARY_COLOR, GL_TEXTURE, etc. */
-   GLenum SourceA[4];    /**< GL_PRIMARY_COLOR, GL_TEXTURE, etc. */
-   GLenum OperandRGB[4]; /**< SRC_COLOR, ONE_MINUS_SRC_COLOR, etc */
-   GLenum OperandA[4];   /**< SRC_ALPHA, ONE_MINUS_SRC_ALPHA, etc */
+   /** Source terms: GL_PRIMARY_COLOR, GL_TEXTURE, etc */
+   GLenum SourceRGB[MAX_COMBINER_TERMS];
+   GLenum SourceA[MAX_COMBINER_TERMS];
+   /** Source operands: GL_SRC_COLOR, GL_ONE_MINUS_SRC_COLOR, etc */
+   GLenum OperandRGB[MAX_COMBINER_TERMS];
+   GLenum OperandA[MAX_COMBINER_TERMS];
    GLuint ScaleShiftRGB; /**< 0, 1 or 2 */
    GLuint ScaleShiftA;   /**< 0, 1 or 2 */
    GLuint _NumArgsRGB;   /**< Number of inputs used for the RGB combiner */
diff --git a/src/mesa/main/texenv.c b/src/mesa/main/texenv.c
index c2960fc820..4d511f2f7e 100644
--- a/src/mesa/main/texenv.c
+++ b/src/mesa/main/texenv.c
@@ -184,7 +184,7 @@ set_combiner_source(GLcontext *ctx,
                     struct gl_texture_unit *texUnit,
                     GLenum pname, GLenum param)
 {
-   GLuint src;
+   GLuint term;
    GLboolean alpha, legal;
 
    if (!ctx->Extensions.EXT_texture_env_combine &&
@@ -194,24 +194,24 @@ set_combiner_source(GLcontext *ctx,
    }
 
    /*
-    * Translate pname to (src, alpha).
+    * Translate pname to (term, alpha).
     */
    switch (pname) {
    case GL_SOURCE0_RGB:
-      src = 0;
+      term = 0;
       alpha = GL_FALSE;
       break;
    case GL_SOURCE1_RGB:
-      src = 1;
+      term = 1;
       alpha = GL_FALSE;
       break;
    case GL_SOURCE2_RGB:
-      src = 2;
+      term = 2;
       alpha = GL_FALSE;
       break;
    case GL_SOURCE3_RGB_NV:
       if (ctx->Extensions.NV_texture_env_combine4) {
-         src = 3;
+         term = 3;
          alpha = GL_FALSE;
       }
       else {
@@ -220,20 +220,20 @@ set_combiner_source(GLcontext *ctx,
       }
       break;
    case GL_SOURCE0_ALPHA:
-      src = 0;
+      term = 0;
       alpha = GL_TRUE;
       break;
    case GL_SOURCE1_ALPHA:
-      src = 1;
+      term = 1;
       alpha = GL_TRUE;
       break;
    case GL_SOURCE2_ALPHA:
-      src = 2;
+      term = 2;
       alpha = GL_TRUE;
       break;
    case GL_SOURCE3_ALPHA_NV:
       if (ctx->Extensions.NV_texture_env_combine4) {
-         src = 3;
+         term = 3;
          alpha = GL_TRUE;
       }
       else {
@@ -246,7 +246,7 @@ set_combiner_source(GLcontext *ctx,
       return;
    }
 
-   assert(src < 4);
+   assert(term < MAX_COMBINER_TERMS);
 
    /*
     * Error-check param (the source term)
@@ -288,9 +288,9 @@ set_combiner_source(GLcontext *ctx,
    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
 
    if (alpha)
-      texUnit->Combine.SourceA[src] = param;
+      texUnit->Combine.SourceA[term] = param;
    else
-      texUnit->Combine.SourceRGB[src] = param;
+      texUnit->Combine.SourceRGB[term] = param;
 }
 
 
@@ -300,7 +300,7 @@ set_combiner_operand(GLcontext *ctx,
                      struct gl_texture_unit *texUnit,
                      GLenum pname, GLenum param)
 {
-   GLuint op;
+   GLuint term;
    GLboolean alpha, legal;
 
    if (!ctx->Extensions.EXT_texture_env_combine &&
@@ -311,16 +311,16 @@ set_combiner_operand(GLcontext *ctx,
 
    switch (pname) {
    case GL_OPERAND0_RGB:
-      op = 0;
+      term = 0;
       alpha = GL_FALSE;
       break;
    case GL_OPERAND1_RGB:
-      op = 1;
+      term = 1;
       alpha = GL_FALSE;
       break;
    case GL_OPERAND2_RGB:
       if (ctx->Extensions.ARB_texture_env_combine) {
-         op = 2;
+         term = 2;
          alpha = GL_FALSE;
       }
       else {
@@ -330,7 +330,7 @@ set_combiner_operand(GLcontext *ctx,
       break;
    case GL_OPERAND3_RGB_NV:
       if (ctx->Extensions.NV_texture_env_combine4) {
-         op = 3;
+         term = 3;
          alpha = GL_FALSE;
       }
       else {
@@ -339,16 +339,16 @@ set_combiner_operand(GLcontext *ctx,
       }
       break;
    case GL_OPERAND0_ALPHA:
-      op = 0;
+      term = 0;
       alpha = GL_TRUE;
       break;
    case GL_OPERAND1_ALPHA:
-      op = 1;
+      term = 1;
       alpha = GL_TRUE;
       break;
    case GL_OPERAND2_ALPHA:
       if (ctx->Extensions.ARB_texture_env_combine) {
-         op = 2;
+         term = 2;
          alpha = GL_TRUE;
       }
       else {
@@ -358,7 +358,7 @@ set_combiner_operand(GLcontext *ctx,
       break;
    case GL_OPERAND3_ALPHA_NV:
       if (ctx->Extensions.NV_texture_env_combine4) {
-         op = 3;
+         term = 3;
          alpha = GL_TRUE;
       }
       else {
@@ -371,7 +371,7 @@ set_combiner_operand(GLcontext *ctx,
       return;
    }
 
-   assert(op < 4);
+   assert(term < MAX_COMBINER_TERMS);
 
    /*
     * Error-check param (the source operand)
@@ -397,9 +397,9 @@ set_combiner_operand(GLcontext *ctx,
    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
 
    if (alpha)
-      texUnit->Combine.OperandA[op] = param;
+      texUnit->Combine.OperandA[term] = param;
    else
-      texUnit->Combine.OperandRGB[op] = param;
+      texUnit->Combine.OperandRGB[term] = param;
 }
 
 
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 3fbd119b34..4a124bf27e 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -39,9 +39,6 @@
 #include "texenvprogram.h"
 
 
-#define MAX_TERMS 4
-
-
 /*
  * Note on texture units:
  *
@@ -95,11 +92,11 @@ struct state_key {
 
       GLuint NumArgsRGB:3;
       GLuint ModeRGB:5;
-      struct mode_opt OptRGB[MAX_TERMS];
+      struct mode_opt OptRGB[MAX_COMBINER_TERMS];
 
       GLuint NumArgsA:3;
       GLuint ModeA:5;
-      struct mode_opt OptA[MAX_TERMS];
+      struct mode_opt OptA[MAX_COMBINER_TERMS];
    } unit[8];
 };
 
@@ -389,7 +386,7 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
       key->unit[i].ScaleShiftRGB = texUnit->_CurrentCombine->ScaleShiftRGB;
       key->unit[i].ScaleShiftA = texUnit->_CurrentCombine->ScaleShiftA;
 
-      for (j = 0; j < MAX_TERMS; j++) {
+      for (j = 0; j < MAX_COMBINER_TERMS; j++) {
          key->unit[i].OptRGB[j].Operand =
 	    translate_operand(texUnit->_CurrentCombine->OperandRGB[j]);
          key->unit[i].OptA[j].Operand =
@@ -972,11 +969,11 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
 				 GLuint mode,
 				 const struct mode_opt *opt)
 {
-   struct ureg src[MAX_TERMS];
+   struct ureg src[MAX_COMBINER_TERMS];
    struct ureg tmp, half;
    GLuint i;
 
-   assert(nr <= MAX_TERMS);
+   assert(nr <= MAX_COMBINER_TERMS);
 
    tmp = undef; /* silence warning (bug 5318) */
 
diff --git a/src/mesa/swrast/s_fragprog.c b/src/mesa/swrast/s_fragprog.c
index 5f032bbd69..b71fb9eae9 100644
--- a/src/mesa/swrast/s_fragprog.c
+++ b/src/mesa/swrast/s_fragprog.c
@@ -74,7 +74,6 @@ fetch_texel_lod( GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
 
       lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
 
-      /* XXX use a float-valued TextureSample routine here!!! */
       swrast->TextureSample[unit](ctx, texObj, 1,
                                   (const GLfloat (*)[4]) texcoord,
                                   &lambda, &rgba);
@@ -116,7 +115,6 @@ fetch_texel_deriv( GLcontext *ctx, const GLfloat texcoord[4],
 
       lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
 
-      /* XXX use a float-valued TextureSample routine here!!! */
       swrast->TextureSample[unit](ctx, texObj, 1,
                                   (const GLfloat (*)[4]) texcoord,
                                   &lambda, &rgba);
diff --git a/src/mesa/swrast/s_texcombine.c b/src/mesa/swrast/s_texcombine.c
index 27c5c15cf1..76b95e8ae7 100644
--- a/src/mesa/swrast/s_texcombine.c
+++ b/src/mesa/swrast/s_texcombine.c
@@ -37,23 +37,23 @@
 #include "s_texcombine.h"
 
 
-#define MAX_COMBINER_TERMS 4
-
-
 /**
- * Do texture application for GL_ARB/EXT_texture_env_combine.
- * This function also supports GL_{EXT,ARB}_texture_env_dot3 and
- * GL_ATI_texture_env_combine3.  Since "classic" texture environments are
- * implemented using GL_ARB_texture_env_combine-like state, this same function
- * is used for classic texture environment application as well.
+ * Do texture application for:
+ *  GL_EXT_texture_env_combine
+ *  GL_ARB_texture_env_combine
+ *  GL_EXT_texture_env_dot3
+ *  GL_ARB_texture_env_dot3
+ *  GL_ATI_texture_env_combine3
+ *  GL_NV_texture_env_combine4
+ *  conventional GL texture env modes
  *
  * \param ctx          rendering context
- * \param textureUnit  the texture unit to apply
+ * \param unit         the texture combiner unit
  * \param n            number of fragments to process (span width)
  * \param primary_rgba incoming fragment color array
  * \param texelBuffer  pointer to texel colors for all texture units
  * 
- * \param rgba         incoming colors, which get modified here
+ * \param rgba         incoming/result fragment colors
  */
 static void
 texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
@@ -65,16 +65,14 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
    const struct gl_tex_env_combine_state *combine = textureUnit->_CurrentCombine;
    const GLfloat (*argRGB [MAX_COMBINER_TERMS])[4];
    const GLfloat (*argA [MAX_COMBINER_TERMS])[4];
-   const GLfloat RGBmult = (GLfloat) (1 << combine->ScaleShiftRGB);
-   const GLfloat Amult = (GLfloat) (1 << combine->ScaleShiftA);
-   const GLuint numColorArgs = combine->_NumArgsRGB;
-   const GLuint numAlphaArgs = combine->_NumArgsA;
+   const GLfloat scaleRGB = (GLfloat) (1 << combine->ScaleShiftRGB);
+   const GLfloat scaleA = (GLfloat) (1 << combine->ScaleShiftA);
+   const GLuint numArgsRGB = combine->_NumArgsRGB;
+   const GLuint numArgsA = combine->_NumArgsA;
    GLfloat ccolor[MAX_COMBINER_TERMS][MAX_WIDTH][4]; /* temp color buffers */
    GLfloat rgba[MAX_WIDTH][4];
    GLuint i, term;
 
-   ASSERT(ctx->Extensions.EXT_texture_env_combine ||
-          ctx->Extensions.ARB_texture_env_combine);
    ASSERT(CONST_SWRAST_CONTEXT(ctx)->_AnyTextureCombine);
 
    for (i = 0; i < n; i++) {
@@ -97,7 +95,7 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
    /*
     * Do operand setup for up to 4 operands.  Loop over the terms.
     */
-   for (term = 0; term < numColorArgs; term++) {
+   for (term = 0; term < numArgsRGB; term++) {
       const GLenum srcRGB = combine->SourceRGB[term];
       const GLenum operandRGB = combine->OperandRGB[term];
 
@@ -193,9 +191,9 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
    }
 
    /*
-    * Set up the argA[i] pointers
+    * Set up the argA[term] pointers
     */
-   for (term = 0; term < numAlphaArgs; term++) {
+   for (term = 0; term < numArgsA; term++) {
       const GLenum srcA = combine->SourceA[term];
       const GLenum operandA = combine->OperandA[term];
 
@@ -269,16 +267,16 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
       switch (combine->ModeRGB) {
       case GL_REPLACE:
          for (i = 0; i < n; i++) {
-            rgba[i][RCOMP] = arg0[i][RCOMP] * RGBmult;
-            rgba[i][GCOMP] = arg0[i][GCOMP] * RGBmult;
-            rgba[i][BCOMP] = arg0[i][BCOMP] * RGBmult;
+            rgba[i][RCOMP] = arg0[i][RCOMP] * scaleRGB;
+            rgba[i][GCOMP] = arg0[i][GCOMP] * scaleRGB;
+            rgba[i][BCOMP] = arg0[i][BCOMP] * scaleRGB;
          }
          break;
       case GL_MODULATE:
          for (i = 0; i < n; i++) {
-            rgba[i][RCOMP] = arg0[i][RCOMP] * arg1[i][RCOMP] * RGBmult;
-            rgba[i][GCOMP] = arg0[i][GCOMP] * arg1[i][GCOMP] * RGBmult;
-            rgba[i][BCOMP] = arg0[i][BCOMP] * arg1[i][BCOMP] * RGBmult;
+            rgba[i][RCOMP] = arg0[i][RCOMP] * arg1[i][RCOMP] * scaleRGB;
+            rgba[i][GCOMP] = arg0[i][GCOMP] * arg1[i][GCOMP] * scaleRGB;
+            rgba[i][BCOMP] = arg0[i][BCOMP] * arg1[i][BCOMP] * scaleRGB;
          }
          break;
       case GL_ADD:
@@ -286,19 +284,19 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
             /* (a * b) + (c * d) */
             for (i = 0; i < n; i++) {
                rgba[i][RCOMP] = (arg0[i][RCOMP] * arg1[i][RCOMP] +
-                                 arg2[i][RCOMP] * arg3[i][RCOMP]) * RGBmult;
+                                 arg2[i][RCOMP] * arg3[i][RCOMP]) * scaleRGB;
                rgba[i][GCOMP] = (arg0[i][GCOMP] * arg1[i][GCOMP] +
-                                 arg2[i][GCOMP] * arg3[i][GCOMP]) * RGBmult;
+                                 arg2[i][GCOMP] * arg3[i][GCOMP]) * scaleRGB;
                rgba[i][BCOMP] = (arg0[i][BCOMP] * arg1[i][BCOMP] +
-                                 arg2[i][BCOMP] * arg3[i][BCOMP]) * RGBmult;
+                                 arg2[i][BCOMP] * arg3[i][BCOMP]) * scaleRGB;
             }
          }
          else {
             /* 2-term addition */
             for (i = 0; i < n; i++) {
-               rgba[i][RCOMP] = (arg0[i][RCOMP] + arg1[i][RCOMP]) * RGBmult;
-               rgba[i][GCOMP] = (arg0[i][GCOMP] + arg1[i][GCOMP]) * RGBmult;
-               rgba[i][BCOMP] = (arg0[i][BCOMP] + arg1[i][BCOMP]) * RGBmult;
+               rgba[i][RCOMP] = (arg0[i][RCOMP] + arg1[i][RCOMP]) * scaleRGB;
+               rgba[i][GCOMP] = (arg0[i][GCOMP] + arg1[i][GCOMP]) * scaleRGB;
+               rgba[i][BCOMP] = (arg0[i][BCOMP] + arg1[i][BCOMP]) * scaleRGB;
             }
          }
          break;
@@ -307,45 +305,45 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
             /* (a * b) + (c * d) - 0.5 */
             for (i = 0; i < n; i++) {
                rgba[i][RCOMP] = (arg0[i][RCOMP] + arg1[i][RCOMP] *
-                                 arg2[i][RCOMP] + arg3[i][RCOMP] - 0.5) * RGBmult;
+                                 arg2[i][RCOMP] + arg3[i][RCOMP] - 0.5) * scaleRGB;
                rgba[i][GCOMP] = (arg0[i][GCOMP] + arg1[i][GCOMP] *
-                                 arg2[i][GCOMP] + arg3[i][GCOMP] - 0.5) * RGBmult;
+                                 arg2[i][GCOMP] + arg3[i][GCOMP] - 0.5) * scaleRGB;
                rgba[i][BCOMP] = (arg0[i][BCOMP] + arg1[i][BCOMP] *
-                                 arg2[i][BCOMP] + arg3[i][BCOMP] - 0.5) * RGBmult;
+                                 arg2[i][BCOMP] + arg3[i][BCOMP] - 0.5) * scaleRGB;
             }
          }
          else {
             for (i = 0; i < n; i++) {
-               rgba[i][RCOMP] = (arg0[i][RCOMP] + arg1[i][RCOMP] - 0.5) * RGBmult;
-               rgba[i][GCOMP] = (arg0[i][GCOMP] + arg1[i][GCOMP] - 0.5) * RGBmult;
-               rgba[i][BCOMP] = (arg0[i][BCOMP] + arg1[i][BCOMP] - 0.5) * RGBmult;
+               rgba[i][RCOMP] = (arg0[i][RCOMP] + arg1[i][RCOMP] - 0.5) * scaleRGB;
+               rgba[i][GCOMP] = (arg0[i][GCOMP] + arg1[i][GCOMP] - 0.5) * scaleRGB;
+               rgba[i][BCOMP] = (arg0[i][BCOMP] + arg1[i][BCOMP] - 0.5) * scaleRGB;
             }
          }
          break;
       case GL_INTERPOLATE:
          for (i = 0; i < n; i++) {
             rgba[i][RCOMP] = (arg0[i][RCOMP] * arg2[i][RCOMP] +
-                          arg1[i][RCOMP] * (1.0F - arg2[i][RCOMP])) * RGBmult;
+                          arg1[i][RCOMP] * (1.0F - arg2[i][RCOMP])) * scaleRGB;
             rgba[i][GCOMP] = (arg0[i][GCOMP] * arg2[i][GCOMP] +
-                          arg1[i][GCOMP] * (1.0F - arg2[i][GCOMP])) * RGBmult;
+                          arg1[i][GCOMP] * (1.0F - arg2[i][GCOMP])) * scaleRGB;
             rgba[i][BCOMP] = (arg0[i][BCOMP] * arg2[i][BCOMP] +
-                          arg1[i][BCOMP] * (1.0F - arg2[i][BCOMP])) * RGBmult;
+                          arg1[i][BCOMP] * (1.0F - arg2[i][BCOMP])) * scaleRGB;
          }
          break;
       case GL_SUBTRACT:
          for (i = 0; i < n; i++) {
-            rgba[i][RCOMP] = (arg0[i][RCOMP] - arg1[i][RCOMP]) * RGBmult;
-            rgba[i][GCOMP] = (arg0[i][GCOMP] - arg1[i][GCOMP]) * RGBmult;
-            rgba[i][BCOMP] = (arg0[i][BCOMP] - arg1[i][BCOMP]) * RGBmult;
+            rgba[i][RCOMP] = (arg0[i][RCOMP] - arg1[i][RCOMP]) * scaleRGB;
+            rgba[i][GCOMP] = (arg0[i][GCOMP] - arg1[i][GCOMP]) * scaleRGB;
+            rgba[i][BCOMP] = (arg0[i][BCOMP] - arg1[i][BCOMP]) * scaleRGB;
          }
          break;
       case GL_DOT3_RGB_EXT:
       case GL_DOT3_RGBA_EXT:
          /* Do not scale the result by 1 2 or 4 */
          for (i = 0; i < n; i++) {
-            GLfloat dot = ((arg0[i][RCOMP]-0.5F) * (arg1[i][RCOMP]-0.5F) +
-                           (arg0[i][GCOMP]-0.5F) * (arg1[i][GCOMP]-0.5F) +
-                           (arg0[i][BCOMP]-0.5F) * (arg1[i][BCOMP]-0.5F))
+            GLfloat dot = ((arg0[i][RCOMP] - 0.5F) * (arg1[i][RCOMP] - 0.5F) +
+                           (arg0[i][GCOMP] - 0.5F) * (arg1[i][GCOMP] - 0.5F) +
+                           (arg0[i][BCOMP] - 0.5F) * (arg1[i][BCOMP] - 0.5F))
                * 4.0F;
             dot = CLAMP(dot, 0.0F, 1.0F);
             rgba[i][RCOMP] = rgba[i][GCOMP] = rgba[i][BCOMP] = (GLfloat) dot;
@@ -355,10 +353,10 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
       case GL_DOT3_RGBA:
          /* DO scale the result by 1 2 or 4 */
          for (i = 0; i < n; i++) {
-            GLfloat dot = ((arg0[i][RCOMP]-0.5F) * (arg1[i][RCOMP]-0.5F) +
-                           (arg0[i][GCOMP]-0.5F) * (arg1[i][GCOMP]-0.5F) +
-                           (arg0[i][BCOMP]-0.5F) * (arg1[i][BCOMP]-0.5F))
-               * 4.0F * RGBmult;
+            GLfloat dot = ((arg0[i][RCOMP] - 0.5F) * (arg1[i][RCOMP] - 0.5F) +
+                           (arg0[i][GCOMP] - 0.5F) * (arg1[i][GCOMP] - 0.5F) +
+                           (arg0[i][BCOMP] - 0.5F) * (arg1[i][BCOMP] - 0.5F))
+               * 4.0F * scaleRGB;
             dot = CLAMP(dot, 0.0, 1.0F);
             rgba[i][RCOMP] = rgba[i][GCOMP] = rgba[i][BCOMP] = (GLfloat) dot;
          }
@@ -366,31 +364,31 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
       case GL_MODULATE_ADD_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][RCOMP] = ((arg0[i][RCOMP] * arg2[i][RCOMP]) +
-                              arg1[i][RCOMP]) * RGBmult;
+                              arg1[i][RCOMP]) * scaleRGB;
             rgba[i][GCOMP] = ((arg0[i][GCOMP] * arg2[i][GCOMP]) +
-                              arg1[i][GCOMP]) * RGBmult;
+                              arg1[i][GCOMP]) * scaleRGB;
             rgba[i][BCOMP] = ((arg0[i][BCOMP] * arg2[i][BCOMP]) +
-                              arg1[i][BCOMP]) * RGBmult;
+                              arg1[i][BCOMP]) * scaleRGB;
 	 }
          break;
       case GL_MODULATE_SIGNED_ADD_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][RCOMP] = ((arg0[i][RCOMP] * arg2[i][RCOMP]) +
-                              arg1[i][RCOMP] - 0.5) * RGBmult;
+                              arg1[i][RCOMP] - 0.5) * scaleRGB;
             rgba[i][GCOMP] = ((arg0[i][GCOMP] * arg2[i][GCOMP]) +
-                              arg1[i][GCOMP] - 0.5) * RGBmult;
+                              arg1[i][GCOMP] - 0.5) * scaleRGB;
             rgba[i][BCOMP] = ((arg0[i][BCOMP] * arg2[i][BCOMP]) +
-                              arg1[i][BCOMP] - 0.5) * RGBmult;
+                              arg1[i][BCOMP] - 0.5) * scaleRGB;
 	 }
          break;
       case GL_MODULATE_SUBTRACT_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][RCOMP] = ((arg0[i][RCOMP] * arg2[i][RCOMP]) -
-                              arg1[i][RCOMP]) * RGBmult;
+                              arg1[i][RCOMP]) * scaleRGB;
             rgba[i][GCOMP] = ((arg0[i][GCOMP] * arg2[i][GCOMP]) -
-                              arg1[i][GCOMP]) * RGBmult;
+                              arg1[i][GCOMP]) * scaleRGB;
             rgba[i][BCOMP] = ((arg0[i][BCOMP] * arg2[i][BCOMP]) -
-                              arg1[i][BCOMP]) * RGBmult;
+                              arg1[i][BCOMP]) * scaleRGB;
 	 }
          break;
       case GL_BUMP_ENVMAP_ATI:
@@ -427,13 +425,13 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
       switch (combine->ModeA) {
       case GL_REPLACE:
          for (i = 0; i < n; i++) {
-            GLfloat a = arg0[i][ACOMP] * Amult;
+            GLfloat a = arg0[i][ACOMP] * scaleA;
             rgba[i][ACOMP] = (GLfloat) MIN2(a, 1.0F);
          }
          break;
       case GL_MODULATE:
          for (i = 0; i < n; i++) {
-            rgba[i][ACOMP] = arg0[i][ACOMP] * arg1[i][ACOMP] * Amult;
+            rgba[i][ACOMP] = arg0[i][ACOMP] * arg1[i][ACOMP] * scaleA;
          }
          break;
       case GL_ADD:
@@ -441,13 +439,13 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
             /* (a * b) + (c * d) */
             for (i = 0; i < n; i++) {
                rgba[i][ACOMP] = (arg0[i][ACOMP] * arg1[i][ACOMP] +
-                                 arg2[i][ACOMP] * arg3[i][ACOMP]) * Amult;
+                                 arg2[i][ACOMP] * arg3[i][ACOMP]) * scaleA;
             }
          }
          else {
             /* two-term add */
             for (i = 0; i < n; i++) {
-               rgba[i][ACOMP] = (arg0[i][ACOMP] + arg1[i][ACOMP]) * Amult;
+               rgba[i][ACOMP] = (arg0[i][ACOMP] + arg1[i][ACOMP]) * scaleA;
             }
          }
          break;
@@ -457,13 +455,13 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
             for (i = 0; i < n; i++) {
                rgba[i][ACOMP] = (arg0[i][ACOMP] * arg1[i][ACOMP] +
                                  arg2[i][ACOMP] * arg3[i][ACOMP] -
-                                 0.5) * Amult;
+                                 0.5) * scaleA;
             }
          }
          else {
             /* a + b - 0.5 */
             for (i = 0; i < n; i++) {
-               rgba[i][ACOMP] = (arg0[i][ACOMP] + arg1[i][ACOMP] - 0.5F) * Amult;
+               rgba[i][ACOMP] = (arg0[i][ACOMP] + arg1[i][ACOMP] - 0.5F) * scaleA;
             }
          }
          break;
@@ -471,30 +469,30 @@ texture_combine( const GLcontext *ctx, GLuint unit, GLuint n,
          for (i=0; i<n; i++) {
             rgba[i][ACOMP] = (arg0[i][ACOMP] * arg2[i][ACOMP] +
                               arg1[i][ACOMP] * (1.0F - arg2[i][ACOMP]))
-               * Amult;
+               * scaleA;
          }
          break;
       case GL_SUBTRACT:
          for (i = 0; i < n; i++) {
-            rgba[i][ACOMP] = (arg0[i][ACOMP] - arg1[i][ACOMP]) * Amult;
+            rgba[i][ACOMP] = (arg0[i][ACOMP] - arg1[i][ACOMP]) * scaleA;
          }
          break;
       case GL_MODULATE_ADD_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][ACOMP] = ((arg0[i][ACOMP] * arg2[i][ACOMP])
-                              + arg1[i][ACOMP]) * Amult;
+                              + arg1[i][ACOMP]) * scaleA;
          }
          break;
       case GL_MODULATE_SIGNED_ADD_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][ACOMP] = ((arg0[i][ACOMP] * arg2[i][ACOMP]) +
-                              arg1[i][ACOMP] - 0.5F) * Amult;
+                              arg1[i][ACOMP] - 0.5F) * scaleA;
          }
          break;
       case GL_MODULATE_SUBTRACT_ATI:
          for (i = 0; i < n; i++) {
             rgba[i][ACOMP] = ((arg0[i][ACOMP] * arg2[i][ACOMP])
-                              - arg1[i][ACOMP]) * Amult;
+                              - arg1[i][ACOMP]) * scaleA;
          }
          break;
       default:
@@ -841,8 +839,6 @@ texture_apply( const GLcontext *ctx,
 	 }
 	 break;
 
-     /* XXX don't clamp results if GLfloat is float??? */
-
       case GL_ADD:  /* GL_EXT_texture_add_env */
          switch (format) {
             case GL_ALPHA:
-- 
cgit v1.2.3


From 7db7ff878d3e5a6b345228e6eaee4797bb68b360 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 14 Apr 2009 22:14:30 -0600
Subject: mesa: merge the prog_src_register::NegateBase and NegateAbs fields

There's really no need for two negation fields.  This came from the
GL_NV_fragment_program extension.  The new, unified Negate bitfield applies
after the absolute value step.
---
 src/mesa/drivers/dri/i915/i915_fragprog.c         | 10 ++---
 src/mesa/drivers/dri/i965/brw_vs_constval.c       |  2 +-
 src/mesa/drivers/dri/i965/brw_vs_emit.c           |  8 ++--
 src/mesa/drivers/dri/i965/brw_wm_fp.c             | 13 +++---
 src/mesa/drivers/dri/i965/brw_wm_glsl.c           |  6 +--
 src/mesa/drivers/dri/i965/brw_wm_pass0.c          |  2 +-
 src/mesa/drivers/dri/r200/r200_vertprog.c         | 34 ++++++++--------
 src/mesa/drivers/dri/r300/r300_fragprog.c         |  4 +-
 src/mesa/drivers/dri/r300/r300_fragprog_swizzle.c | 17 ++++----
 src/mesa/drivers/dri/r300/r300_vertprog.c         | 49 +++++++++--------------
 src/mesa/drivers/dri/r300/r500_fragprog.c         | 19 +++++----
 src/mesa/drivers/dri/r300/radeon_nqssadce.c       |  7 ++--
 src/mesa/drivers/dri/r300/radeon_program_alu.c    | 20 +++++----
 src/mesa/drivers/dri/r300/radeon_program_pair.c   | 11 +++--
 src/mesa/main/ffvertex_prog.c                     |  3 +-
 src/mesa/main/texenvprogram.c                     |  3 +-
 src/mesa/shader/arbprogparse.c                    |  8 ++--
 src/mesa/shader/nvfragparse.c                     | 33 ++++++++-------
 src/mesa/shader/nvvertparse.c                     | 10 ++---
 src/mesa/shader/prog_execute.c                    | 27 ++++---------
 src/mesa/shader/prog_instruction.h                | 32 +++------------
 src/mesa/shader/prog_print.c                      | 20 ++++-----
 src/mesa/shader/programopt.c                      |  2 +-
 src/mesa/shader/slang/slang_emit.c                |  2 +-
 src/mesa/state_tracker/st_cb_bitmap.c             |  2 +-
 src/mesa/state_tracker/st_mesa_to_tgsi.c          | 20 ++++-----
 26 files changed, 156 insertions(+), 208 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/drivers/dri/i915/i915_fragprog.c b/src/mesa/drivers/dri/i915/i915_fragprog.c
index 52f09a4b1b..a5158de945 100644
--- a/src/mesa/drivers/dri/i915/i915_fragprog.c
+++ b/src/mesa/drivers/dri/i915/i915_fragprog.c
@@ -162,12 +162,12 @@ src_vector(struct i915_fragment_program *p,
                  GET_SWZ(source->Swizzle, 1),
                  GET_SWZ(source->Swizzle, 2), GET_SWZ(source->Swizzle, 3));
 
-   if (source->NegateBase)
+   if (source->Negate)
       src = negate(src,
-                   GET_BIT(source->NegateBase, 0),
-                   GET_BIT(source->NegateBase, 1),
-                   GET_BIT(source->NegateBase, 2),
-                   GET_BIT(source->NegateBase, 3));
+                   GET_BIT(source->Negate, 0),
+                   GET_BIT(source->Negate, 1),
+                   GET_BIT(source->Negate, 2),
+                   GET_BIT(source->Negate, 3));
 
    return src;
 }
diff --git a/src/mesa/drivers/dri/i965/brw_vs_constval.c b/src/mesa/drivers/dri/i965/brw_vs_constval.c
index d29eb17f8c..2637344b48 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_constval.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_constval.c
@@ -96,7 +96,7 @@ static GLubyte get_active( struct tracker *t,
 			   struct prog_src_register src )
 {
    GLuint i;
-   GLubyte active = src.NegateBase; /* NOTE! */
+   GLubyte active = src.Negate; /* NOTE! */
 
    if (src.RelAddr)
       return 0xf;
diff --git a/src/mesa/drivers/dri/i965/brw_vs_emit.c b/src/mesa/drivers/dri/i965/brw_vs_emit.c
index 2ee63129bc..42f6a99142 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_emit.c
@@ -899,7 +899,7 @@ static struct brw_reg get_arg( struct brw_vs_compile *c,
 
    /* Note this is ok for non-swizzle instructions: 
     */
-   reg.negate = src->NegateBase ? 1 : 0;   
+   reg.negate = src->Negate ? 1 : 0;   
 
    return reg;
 }
@@ -945,7 +945,7 @@ static void emit_swz( struct brw_vs_compile *c,
    GLuint ones_mask = 0;
    GLuint src_mask = 0;
    GLubyte src_swz[4];
-   GLboolean need_tmp = (src.NegateBase &&
+   GLboolean need_tmp = (src.Negate &&
 			 dst.file != BRW_GENERAL_REGISTER_FILE);
    struct brw_reg tmp = dst;
    GLuint i;
@@ -997,8 +997,8 @@ static void emit_swz( struct brw_vs_compile *c,
    if (ones_mask) 
       brw_MOV(p, brw_writemask(tmp, ones_mask), brw_imm_f(1));
 
-   if (src.NegateBase)
-      brw_MOV(p, brw_writemask(tmp, src.NegateBase), negate(tmp));
+   if (src.Negate)
+      brw_MOV(p, brw_writemask(tmp, src.Negate), negate(tmp));
    
    if (need_tmp) {
       brw_MOV(p, dst, tmp);
diff --git a/src/mesa/drivers/dri/i965/brw_wm_fp.c b/src/mesa/drivers/dri/i965/brw_wm_fp.c
index a7f5f1b9a2..1798d842c7 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_fp.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_fp.c
@@ -80,9 +80,8 @@ static struct prog_src_register src_reg(GLuint file, GLuint idx)
    reg.Index = idx;
    reg.Swizzle = SWIZZLE_NOOP;
    reg.RelAddr = 0;
-   reg.NegateBase = 0;
+   reg.Negate = NEGATE_NONE;
    reg.Abs = 0;
-   reg.NegateAbs = 0;
    return reg;
 }
 
@@ -569,7 +568,7 @@ static void precalc_dst( struct brw_wm_compile *c,
 		    src_undef(),
 		    src_undef());
       /* Avoid letting negation flag of src0 affect our 1 constant. */
-      swz->SrcReg[0].NegateBase &= ~NEGATE_X;
+      swz->SrcReg[0].Negate &= ~NEGATE_X;
    }
    if (dst.WriteMask & WRITEMASK_W) {
       /* dst.w = mov src1.w
@@ -604,7 +603,7 @@ static void precalc_lit( struct brw_wm_compile *c,
 		    src_undef(),
 		    src_undef());
       /* Avoid letting the negation flag of src0 affect our 1 constant. */
-      swz->SrcReg[0].NegateBase = 0;
+      swz->SrcReg[0].Negate = NEGATE_NONE;
    }
 
    if (dst.WriteMask & WRITEMASK_YZ) {
@@ -651,7 +650,7 @@ static void precalc_tex( struct brw_wm_compile *c,
                      src0,
                      src_undef(),
                      src_undef());
-       out->SrcReg[0].NegateBase = 0;
+       out->SrcReg[0].Negate = NEGATE_NONE;
        out->SrcReg[0].Abs = 1;
 
        /* tmp0 = MAX(coord.X, coord.Y) */
@@ -1050,14 +1049,14 @@ void brw_wm_pass_fp( struct brw_wm_compile *c )
       case OPCODE_ABS:
 	 out = emit_insn(c, inst);
 	 out->Opcode = OPCODE_MOV;
-	 out->SrcReg[0].NegateBase = 0;
+	 out->SrcReg[0].Negate = NEGATE_NONE;
 	 out->SrcReg[0].Abs = 1;
 	 break;
 
       case OPCODE_SUB: 
 	 out = emit_insn(c, inst);
 	 out->Opcode = OPCODE_ADD;
-	 out->SrcReg[1].NegateBase ^= 0xf;
+	 out->SrcReg[1].Negate ^= NEGATE_XYZW;
 	 break;
 
       case OPCODE_SCS: 
diff --git a/src/mesa/drivers/dri/i965/brw_wm_glsl.c b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
index 49fea2e41a..385efd2dd3 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_glsl.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
@@ -340,7 +340,7 @@ get_src_reg_const(struct brw_wm_compile *c,
    const_reg = stride(const_reg, 0, 1, 0);
    const_reg.subnr = component * 4;
 
-   if (src->NegateBase)
+   if (src->Negate & (1 << component))
       const_reg = negate(const_reg);
    if (src->Abs)
       const_reg = brw_abs(const_reg);
@@ -377,7 +377,7 @@ static struct brw_reg get_src_reg(struct brw_wm_compile *c,
     else {
        /* other type of source register */
        return get_reg(c, src->File, src->Index, component, nr, 
-                      src->NegateBase, src->Abs);
+                      src->Negate, src->Abs);
     }
 }
 
@@ -402,7 +402,7 @@ static struct brw_reg get_src_reg_imm(struct brw_wm_compile *c,
        const GLfloat *param =
           c->fp->program.Base.Parameters->ParameterValues[src->Index];
        GLfloat value = param[component];
-       if (src->NegateBase)
+       if (src->Negate & (1 << channel))
           value = -value;
        if (src->Abs)
           value = FABSF(value);
diff --git a/src/mesa/drivers/dri/i965/brw_wm_pass0.c b/src/mesa/drivers/dri/i965/brw_wm_pass0.c
index 2debd0678a..92142764f5 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_pass0.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_pass0.c
@@ -322,7 +322,7 @@ static struct brw_wm_ref *get_new_ref( struct brw_wm_compile *c,
       newref->value->lastuse = newref;
    }
 
-   if (src.NegateBase & (1<<i))
+   if (src.Negate & (1 << i))
       newref->hw_reg.negate ^= 1;
 
    if (src.Abs) {
diff --git a/src/mesa/drivers/dri/r200/r200_vertprog.c b/src/mesa/drivers/dri/r200/r200_vertprog.c
index a2561df579..4ce93b5145 100644
--- a/src/mesa/drivers/dri/r200/r200_vertprog.c
+++ b/src/mesa/drivers/dri/r200/r200_vertprog.c
@@ -290,7 +290,7 @@ static unsigned long t_src(struct r200_vertex_program *vp, struct prog_src_regis
 			t_swizzle(GET_SWZ(src->Swizzle, 2)),
 			t_swizzle(GET_SWZ(src->Swizzle, 3)),
 			t_src_class(src->File),
-			src->NegateBase) | (src->RelAddr << 4);
+			src->Negate) | (src->RelAddr << 4);
 }
 
 static unsigned long t_src_scalar(struct r200_vertex_program *vp, struct prog_src_register *src)
@@ -302,7 +302,7 @@ static unsigned long t_src_scalar(struct r200_vertex_program *vp, struct prog_sr
 			t_swizzle(GET_SWZ(src->Swizzle, 0)),
 			t_swizzle(GET_SWZ(src->Swizzle, 0)),
 			t_src_class(src->File),
-			src->NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src->RelAddr << 4);
+			src->Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src->RelAddr << 4);
 }
 
 static unsigned long t_opcode(enum prog_opcode opcode)
@@ -700,7 +700,7 @@ static GLboolean r200_translate_vertex_program(GLcontext *ctx, struct r200_verte
 		   t_swizzle(GET_SWZ(src[1].Swizzle, 0)),
 		   SWIZZLE_ZERO,
 		   t_src_class(src[0].File),
-		   src[0].NegateBase) | (src[0].RelAddr << 4);
+		   src[0].Negate) | (src[0].RelAddr << 4);
 	    o_inst->src1 = UNUSED_SRC_0;
 	    o_inst->src2 = UNUSED_SRC_0;
 	 }
@@ -712,12 +712,12 @@ static GLboolean r200_translate_vertex_program(GLcontext *ctx, struct r200_verte
 		   t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
 		   SWIZZLE_ZERO, SWIZZLE_ZERO, SWIZZLE_ZERO,
 		   t_src_class(src[0].File),
-		   src[0].NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+		   src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4);
 	    o_inst->src1 = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
 		   SWIZZLE_ZERO, SWIZZLE_ZERO,
 		   t_swizzle(GET_SWZ(src[1].Swizzle, 0)), SWIZZLE_ZERO,
 		   t_src_class(src[1].File),
-		   src[1].NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
+		   src[1].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
 	    o_inst->src2 = UNUSED_SRC_1;
 	    o_inst++;
 
@@ -766,11 +766,11 @@ if ((o_inst - vp->instr) == 31) {
 o_inst->src1 = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
 			SWIZZLE_X, SWIZZLE_X, SWIZZLE_X, SWIZZLE_X,
 			t_src_class(src[1].File),
-			src[1].NegateBase) | (src[1].RelAddr << 4);
+			src[1].Negate) | (src[1].RelAddr << 4);
 o_inst->src2 = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
 			SWIZZLE_Y, SWIZZLE_Y, SWIZZLE_Y, SWIZZLE_Y,
 			t_src_class(src[1].File),
-			src[1].NegateBase) | (src[1].RelAddr << 4);
+			src[1].Negate) | (src[1].RelAddr << 4);
 }
 else {
 	 o_inst->src1 = t_src(vp, &src[1]);
@@ -792,7 +792,7 @@ else {
 		t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
 		SWIZZLE_ZERO,
 		t_src_class(src[0].File),
-		src[0].NegateBase) | (src[0].RelAddr << 4);
+		src[0].Negate) | (src[0].RelAddr << 4);
 
 	 o_inst->src1 = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
 		t_swizzle(GET_SWZ(src[1].Swizzle, 0)),
@@ -800,7 +800,7 @@ else {
 		t_swizzle(GET_SWZ(src[1].Swizzle, 2)),
 		SWIZZLE_ZERO,
 		t_src_class(src[1].File),
-		src[1].NegateBase) | (src[1].RelAddr << 4);
+		src[1].Negate) | (src[1].RelAddr << 4);
 
 	 o_inst->src2 = UNUSED_SRC_1;
 	 goto next;
@@ -815,7 +815,7 @@ else {
 		t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
 		VSF_IN_COMPONENT_ONE,
 		t_src_class(src[0].File),
-		src[0].NegateBase) | (src[0].RelAddr << 4);
+		src[0].Negate) | (src[0].RelAddr << 4);
 	 o_inst->src1 = t_src(vp, &src[1]);
 	 o_inst->src2 = UNUSED_SRC_1;
 	 goto next;
@@ -831,7 +831,7 @@ else {
 		t_swizzle(GET_SWZ(src[1].Swizzle, 2)),
 		t_swizzle(GET_SWZ(src[1].Swizzle, 3)),
 		t_src_class(src[1].File),
-		(!src[1].NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
+		(!src[1].Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
 	 o_inst->src2 = UNUSED_SRC_1;
 	 goto next;
 
@@ -846,7 +846,7 @@ else {
 		t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
 		t_swizzle(GET_SWZ(src[0].Swizzle, 3)),
 		t_src_class(src[0].File),
-		(!src[0].NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+		(!src[0].Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4);
 	 o_inst->src2 = UNUSED_SRC_1;
 	 goto next;
 
@@ -874,7 +874,7 @@ else {
 		VSF_IN_COMPONENT_W,
 		VSF_IN_CLASS_TMP,
 		/* Not 100% sure about this */
-		(!src[0].NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE/*VSF_FLAG_ALL*/);
+		(!src[0].Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE/*VSF_FLAG_ALL*/);
 
 	 o_inst->src2 = UNUSED_SRC_0;
 	 u_temp_i--;
@@ -899,7 +899,7 @@ else {
 		t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x
 		t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w
 		t_src_class(src[0].File),
-		src[0].NegateBase) | (src[0].RelAddr << 4);
+		src[0].Negate) | (src[0].RelAddr << 4);
 
 	 o_inst->src1 = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
 		t_swizzle(GET_SWZ(src[1].Swizzle, 2)), // z
@@ -907,7 +907,7 @@ else {
 		t_swizzle(GET_SWZ(src[1].Swizzle, 1)), // y
 		t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w
 		t_src_class(src[1].File),
-		src[1].NegateBase) | (src[1].RelAddr << 4);
+		src[1].Negate) | (src[1].RelAddr << 4);
 
 	 o_inst->src2 = UNUSED_SRC_1;
 	 o_inst++;
@@ -922,7 +922,7 @@ else {
 		t_swizzle(GET_SWZ(src[1].Swizzle, 0)), // x
 		t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w
 		t_src_class(src[1].File),
-		(!src[1].NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
+		(!src[1].Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4);
 
 	 o_inst->src1 = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
 		t_swizzle(GET_SWZ(src[0].Swizzle, 2)), // z
@@ -930,7 +930,7 @@ else {
 		t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y
 		t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w
 		t_src_class(src[0].File),
-		src[0].NegateBase) | (src[0].RelAddr << 4);
+		src[0].Negate) | (src[0].RelAddr << 4);
 
 	 o_inst->src2 = MAKE_VSF_SOURCE(u_temp_i+1,
 		VSF_IN_COMPONENT_X,
diff --git a/src/mesa/drivers/dri/r300/r300_fragprog.c b/src/mesa/drivers/dri/r300/r300_fragprog.c
index 32182bb667..873cde4414 100644
--- a/src/mesa/drivers/dri/r300/r300_fragprog.c
+++ b/src/mesa/drivers/dri/r300/r300_fragprog.c
@@ -214,9 +214,9 @@ static GLboolean transform_TEX(
 		 *   r  < tex  <=>      -tex+r < 0
 		 *   r >= tex  <=> not (-tex+r < 0 */
 		if (comparefunc == GL_LESS || comparefunc == GL_GEQUAL)
-			tgt[1].SrcReg[2].NegateBase = tgt[0].SrcReg[2].NegateBase ^ NEGATE_XYZW;
+			tgt[1].SrcReg[2].Negate = tgt[0].SrcReg[2].Negate ^ NEGATE_XYZW;
 		else
-			tgt[1].SrcReg[0].NegateBase = tgt[0].SrcReg[0].NegateBase ^ NEGATE_XYZW;
+			tgt[1].SrcReg[0].Negate = tgt[0].SrcReg[0].Negate ^ NEGATE_XYZW;
 
 		tgt[2].Opcode = OPCODE_CMP;
 		tgt[2].DstReg = orig_inst->DstReg;
diff --git a/src/mesa/drivers/dri/r300/r300_fragprog_swizzle.c b/src/mesa/drivers/dri/r300/r300_fragprog_swizzle.c
index a86d2bd471..191853ac1f 100644
--- a/src/mesa/drivers/dri/r300/r300_fragprog_swizzle.c
+++ b/src/mesa/drivers/dri/r300/r300_fragprog_swizzle.c
@@ -92,7 +92,7 @@ static const struct swizzle_data* lookup_native_swizzle(GLuint swizzle)
 GLboolean r300FPIsNativeSwizzle(GLuint opcode, struct prog_src_register reg)
 {
 	if (reg.Abs)
-		reg.NegateBase = 0;
+		reg.Negate = NEGATE_NONE;
 
 	if (opcode == OPCODE_KIL ||
 	    opcode == OPCODE_TEX ||
@@ -100,7 +100,8 @@ GLboolean r300FPIsNativeSwizzle(GLuint opcode, struct prog_src_register reg)
 	    opcode == OPCODE_TXP) {
 		int j;
 
-		if (reg.Abs || reg.NegateBase != (15*reg.NegateAbs))
+		if (reg.Abs || (reg.Negate != NEGATE_XYZW &&
+                                reg.Negate != NEGATE_NONE))
 			return GL_FALSE;
 
 		for(j = 0; j < 4; ++j) {
@@ -121,7 +122,7 @@ GLboolean r300FPIsNativeSwizzle(GLuint opcode, struct prog_src_register reg)
 		if (GET_SWZ(reg.Swizzle, j) != SWIZZLE_NIL)
 			relevant |= 1 << j;
 
-	if ((reg.NegateBase & relevant) && (reg.NegateBase & relevant) != relevant)
+	if ((reg.Negate & relevant) && (reg.Negate & relevant) != relevant)
 		return GL_FALSE;
 
 	if (!lookup_native_swizzle(reg.Swizzle))
@@ -137,7 +138,7 @@ GLboolean r300FPIsNativeSwizzle(GLuint opcode, struct prog_src_register reg)
 void r300FPBuildSwizzle(struct nqssadce_state *s, struct prog_dst_register dst, struct prog_src_register src)
 {
 	if (src.Abs)
-		src.NegateBase = 0;
+		src.Negate = NEGATE_NONE;
 
 	while(dst.WriteMask) {
 		const struct swizzle_data *best_swizzle = 0;
@@ -170,11 +171,11 @@ void r300FPBuildSwizzle(struct nqssadce_state *s, struct prog_dst_register dst,
 			}
 		}
 
-		if ((src.NegateBase & best_matchmask) != 0) {
-			best_matchmask &= src.NegateBase;
-			rgbnegate = !src.NegateAbs;
+		if ((src.Negate & best_matchmask) != 0) {
+			best_matchmask &= src.Negate;
+			rgbnegate = !src.Negate;
 		} else {
-			rgbnegate = src.NegateAbs;
+			rgbnegate = src.Negate;
 		}
 
 		struct prog_instruction *inst;
diff --git a/src/mesa/drivers/dri/r300/r300_vertprog.c b/src/mesa/drivers/dri/r300/r300_vertprog.c
index 50806575ce..146daa367c 100644
--- a/src/mesa/drivers/dri/r300/r300_vertprog.c
+++ b/src/mesa/drivers/dri/r300/r300_vertprog.c
@@ -245,7 +245,7 @@ static unsigned long t_src_index(struct r300_vertex_program *vp,
 static unsigned long t_src(struct r300_vertex_program *vp,
 			   struct prog_src_register *src)
 {
-	/* src->NegateBase uses the NEGATE_ flags from program_instruction.h,
+	/* src->Negate uses the NEGATE_ flags from program_instruction.h,
 	 * which equal our VSF_FLAGS_ values, so it's safe to just pass it here.
 	 */
 	return PVS_SRC_OPERAND(t_src_index(vp, src),
@@ -254,13 +254,13 @@ static unsigned long t_src(struct r300_vertex_program *vp,
 			       t_swizzle(GET_SWZ(src->Swizzle, 2)),
 			       t_swizzle(GET_SWZ(src->Swizzle, 3)),
 			       t_src_class(src->File),
-			       src->NegateBase) | (src->RelAddr << 4);
+			       src->Negate) | (src->RelAddr << 4);
 }
 
 static unsigned long t_src_scalar(struct r300_vertex_program *vp,
 				  struct prog_src_register *src)
 {
-	/* src->NegateBase uses the NEGATE_ flags from program_instruction.h,
+	/* src->Negate uses the NEGATE_ flags from program_instruction.h,
 	 * which equal our VSF_FLAGS_ values, so it's safe to just pass it here.
 	 */
 	return PVS_SRC_OPERAND(t_src_index(vp, src),
@@ -269,8 +269,7 @@ static unsigned long t_src_scalar(struct r300_vertex_program *vp,
 			       t_swizzle(GET_SWZ(src->Swizzle, 0)),
 			       t_swizzle(GET_SWZ(src->Swizzle, 0)),
 			       t_src_class(src->File),
-			       src->
-			       NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+			       src->Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src->RelAddr << 4);
 }
 
@@ -307,7 +306,7 @@ static GLuint *r300TranslateOpcodeABS(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),
 				  t_src_class(src[0].File),
 				  (!src[0].
-				   NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				   Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[3] = 0;
 
@@ -369,8 +368,7 @@ static GLuint *r300TranslateOpcodeDP3(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
 				  SWIZZLE_ZERO,
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[2] =
 	    PVS_SRC_OPERAND(t_src_index(vp, &src[1]),
@@ -378,8 +376,7 @@ static GLuint *r300TranslateOpcodeDP3(struct r300_vertex_program *vp,
 			    t_swizzle(GET_SWZ(src[1].Swizzle, 1)),
 			    t_swizzle(GET_SWZ(src[1].Swizzle, 2)), SWIZZLE_ZERO,
 			    t_src_class(src[1].File),
-			    src[1].
-			    NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+			    src[1].Negate ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
 	    (src[1].RelAddr << 4);
 	inst[3] = __CONST(1, SWIZZLE_ZERO);
 
@@ -422,8 +419,7 @@ static GLuint *r300TranslateOpcodeDPH(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
 				  PVS_SRC_SELECT_FORCE_1,
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[2] = t_src(vp, &src[1]);
 	inst[3] = __CONST(1, SWIZZLE_ZERO);
@@ -519,7 +515,7 @@ static GLuint *r300TranslateOpcodeFLR(struct r300_vertex_program *vp,
 				  PVS_SRC_SELECT_W, PVS_SRC_REG_TEMPORARY,
 				  /* Not 100% sure about this */
 				  (!src[0].
-				   NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE
+				   Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE
 				  /*VSF_FLAG_ALL */ );
 	inst[3] = __CONST(0, SWIZZLE_ZERO);
 	(*u_temp_i)--;
@@ -564,8 +560,7 @@ static GLuint *r300TranslateOpcodeLG2(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[2] = __CONST(0, SWIZZLE_ZERO);
 	inst[3] = __CONST(0, SWIZZLE_ZERO);
@@ -592,24 +587,21 @@ static GLuint *r300TranslateOpcodeLIT(struct r300_vertex_program *vp,
 				  PVS_SRC_SELECT_FORCE_0,	// Z
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// Y
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[2] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// Y
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// W
 				  PVS_SRC_SELECT_FORCE_0,	// Z
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// X
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[3] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// Y
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// X
 				  PVS_SRC_SELECT_FORCE_0,	// Z
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// W
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 
 	return inst;
@@ -837,7 +829,7 @@ static GLuint *r300TranslateOpcodeSUB(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),
 				  t_src_class(src[1].File),
 				  (!src[1].
-				   NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				   Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[1].RelAddr << 4);
 	inst[3] = 0;
 #else
@@ -857,7 +849,7 @@ static GLuint *r300TranslateOpcodeSUB(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),
 				  t_src_class(src[1].File),
 				  (!src[1].
-				   NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				   Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[1].RelAddr << 4);
 #endif
 
@@ -905,16 +897,14 @@ static GLuint *r300TranslateOpcodeXPD(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// X
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// W
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[2] = PVS_SRC_OPERAND(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 2)),	// Z
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 0)),	// X
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 1)),	// Y
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// W
 				  t_src_class(src[1].File),
-				  src[1].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[1].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[1].RelAddr << 4);
 	inst[3] = __CONST(1, SWIZZLE_ZERO);
 	inst += 4;
@@ -931,15 +921,14 @@ static GLuint *r300TranslateOpcodeXPD(struct r300_vertex_program *vp,
 				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// W
 				  t_src_class(src[1].File),
 				  (!src[1].
-				   NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				   Negate) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[1].RelAddr << 4);
 	inst[2] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 2)),	// Z
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// X
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// Y
 				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// W
 				  t_src_class(src[0].File),
-				  src[0].
-				  NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+				  src[0].Negate ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
 	    (src[0].RelAddr << 4);
 	inst[3] =
 	    PVS_SRC_OPERAND(*u_temp_i, PVS_SRC_SELECT_X, PVS_SRC_SELECT_Y,
diff --git a/src/mesa/drivers/dri/r300/r500_fragprog.c b/src/mesa/drivers/dri/r300/r500_fragprog.c
index 07a2a7b17c..292573de89 100644
--- a/src/mesa/drivers/dri/r300/r500_fragprog.c
+++ b/src/mesa/drivers/dri/r300/r500_fragprog.c
@@ -156,9 +156,9 @@ static GLboolean transform_TEX(
 		 *   r  < tex  <=>      -tex+r < 0
 		 *   r >= tex  <=> not (-tex+r < 0 */
 		if (comparefunc == GL_LESS || comparefunc == GL_GEQUAL)
-			tgt[1].SrcReg[2].NegateBase = tgt[0].SrcReg[2].NegateBase ^ NEGATE_XYZW;
+			tgt[1].SrcReg[2].Negate = tgt[0].SrcReg[2].Negate ^ NEGATE_XYZW;
 		else
-			tgt[1].SrcReg[0].NegateBase = tgt[0].SrcReg[0].NegateBase ^ NEGATE_XYZW;
+			tgt[1].SrcReg[0].Negate = tgt[0].SrcReg[0].Negate ^ NEGATE_XYZW;
 
 		tgt[2].Opcode = OPCODE_CMP;
 		tgt[2].DstReg = orig_inst->DstReg;
@@ -314,8 +314,8 @@ static GLboolean is_native_swizzle(GLuint opcode, struct prog_src_register reg)
 		if (reg.Abs)
 			return GL_FALSE;
 
-		if (reg.NegateAbs)
-			reg.NegateBase ^= 15;
+		if (reg.Negate)
+			reg.Negate ^= NEGATE_XYZW;
 
 		if (opcode == OPCODE_KIL) {
 			if (reg.Swizzle != SWIZZLE_NOOP)
@@ -324,7 +324,7 @@ static GLboolean is_native_swizzle(GLuint opcode, struct prog_src_register reg)
 			for(i = 0; i < 4; ++i) {
 				GLuint swz = GET_SWZ(reg.Swizzle, i);
 				if (swz == SWIZZLE_NIL) {
-					reg.NegateBase &= ~(1 << i);
+					reg.Negate &= ~(1 << i);
 					continue;
 				}
 				if (swz >= 4)
@@ -332,15 +332,14 @@ static GLboolean is_native_swizzle(GLuint opcode, struct prog_src_register reg)
 			}
 		}
 
-		if (reg.NegateBase)
+		if (reg.Negate)
 			return GL_FALSE;
 
 		return GL_TRUE;
 	} else if (opcode == OPCODE_DDX || opcode == OPCODE_DDY) {
 		/* DDX/MDH and DDY/MDV explicitly ignore incoming swizzles;
 		 * if it doesn't fit perfectly into a .xyzw case... */
-		if (reg.Swizzle == SWIZZLE_NOOP && !reg.Abs
-				&& !reg.NegateBase && !reg.NegateAbs)
+		if (reg.Swizzle == SWIZZLE_NOOP && !reg.Abs && !reg.Negate)
 			return GL_TRUE;
 
 		return GL_FALSE;
@@ -355,7 +354,7 @@ static GLboolean is_native_swizzle(GLuint opcode, struct prog_src_register reg)
 			if (swz != SWIZZLE_NIL && swz != SWIZZLE_ZERO)
 				relevant |= 1 << i;
 		}
-		if ((reg.NegateBase & relevant) && ((reg.NegateBase & relevant) != relevant))
+		if ((reg.Negate & relevant) && ((reg.Negate & relevant) != relevant))
 			return GL_FALSE;
 
 		return GL_TRUE;
@@ -379,7 +378,7 @@ static void nqssadce_build_swizzle(struct nqssadce_state *s,
 		GLuint swz = GET_SWZ(src.Swizzle, i);
 		if (swz == SWIZZLE_NIL)
 			continue;
-		negatebase[GET_BIT(src.NegateBase, i)] |= 1 << i;
+		negatebase[GET_BIT(src.Negate, i)] |= 1 << i;
 	}
 
 	_mesa_insert_instructions(s->Program, s->IP, (negatebase[0] ? 1 : 0) + (negatebase[1] ? 1 : 0));
diff --git a/src/mesa/drivers/dri/r300/radeon_nqssadce.c b/src/mesa/drivers/dri/r300/radeon_nqssadce.c
index a083c3d243..4a2e1cba40 100644
--- a/src/mesa/drivers/dri/r300/radeon_nqssadce.c
+++ b/src/mesa/drivers/dri/r300/radeon_nqssadce.c
@@ -61,12 +61,12 @@ static struct prog_src_register lmul_swizzle(GLuint swizzle, struct prog_src_reg
 	struct prog_src_register tmp = srcreg;
 	int i;
 	tmp.Swizzle = 0;
-	tmp.NegateBase = 0;
+	tmp.Negate = NEGATE_NONE;
 	for(i = 0; i < 4; ++i) {
 		GLuint swz = GET_SWZ(swizzle, i);
 		if (swz < 4) {
 			tmp.Swizzle |= GET_SWZ(srcreg.Swizzle, swz) << (i*3);
-			tmp.NegateBase |= GET_BIT(srcreg.NegateBase, swz) << i;
+			tmp.Negate |= GET_BIT(srcreg.Negate, swz) << i;
 		} else {
 			tmp.Swizzle |= swz << (i*3);
 		}
@@ -103,9 +103,8 @@ static struct prog_instruction* track_used_srcreg(struct nqssadce_state* s,
 		inst->SrcReg[src].File = PROGRAM_TEMPORARY;
 		inst->SrcReg[src].Index = dstreg.Index;
 		inst->SrcReg[src].Swizzle = 0;
-		inst->SrcReg[src].NegateBase = 0;
+		inst->SrcReg[src].Negate = NEGATE_NONE;
 		inst->SrcReg[src].Abs = 0;
-		inst->SrcReg[src].NegateAbs = 0;
 		for(i = 0; i < 4; ++i) {
 			if (GET_BIT(sourced, i))
 				inst->SrcReg[src].Swizzle |= i << (3*i);
diff --git a/src/mesa/drivers/dri/r300/radeon_program_alu.c b/src/mesa/drivers/dri/r300/radeon_program_alu.c
index 1ef71e74dc..ebc5c913b2 100644
--- a/src/mesa/drivers/dri/r300/radeon_program_alu.c
+++ b/src/mesa/drivers/dri/r300/radeon_program_alu.c
@@ -89,8 +89,9 @@ static void set_swizzle(struct prog_src_register *SrcReg, int coordinate, int sw
 
 static void set_negate_base(struct prog_src_register *SrcReg, int coordinate, int negate)
 {
-	SrcReg->NegateBase &= ~(1 << coordinate);
-	SrcReg->NegateBase |= (negate << coordinate);
+	/* XXX note sure about this negation logic here */
+	SrcReg->Negate &= ~(1 << coordinate);
+	SrcReg->Negate |= (negate << coordinate);
 }
 
 static struct prog_dst_register dstreg(int file, int index)
@@ -156,15 +157,14 @@ static struct prog_src_register absolute(struct prog_src_register reg)
 {
 	struct prog_src_register newreg = reg;
 	newreg.Abs = 1;
-	newreg.NegateBase = 0;
-	newreg.NegateAbs = 0;
+	newreg.Negate = NEGATE_NONE;
 	return newreg;
 }
 
 static struct prog_src_register negate(struct prog_src_register reg)
 {
 	struct prog_src_register newreg = reg;
-	newreg.NegateAbs = !newreg.NegateAbs;
+	newreg.Negate = newreg.Negate ^ NEGATE_XYZW;
 	return newreg;
 }
 
@@ -189,8 +189,7 @@ static void transform_ABS(struct radeon_transform_context* t,
 {
 	struct prog_src_register src = inst->SrcReg[0];
 	src.Abs = 1;
-	src.NegateBase = 0;
-	src.NegateAbs = 0;
+	src.Negate = NEGATE_NONE;
 	emit1(t->Program, OPCODE_MOV, inst->SaturateMode, inst->DstReg, src);
 }
 
@@ -198,14 +197,13 @@ static void transform_DPH(struct radeon_transform_context* t,
 	struct prog_instruction* inst)
 {
 	struct prog_src_register src0 = inst->SrcReg[0];
-	if (src0.NegateAbs) {
+	if (src0.Negate) {
 		if (src0.Abs) {
 			int tempreg = radeonFindFreeTemporary(t);
 			emit1(t->Program, OPCODE_MOV, 0, dstreg(PROGRAM_TEMPORARY, tempreg), src0);
 			src0 = srcreg(src0.File, src0.Index);
 		} else {
-			src0.NegateAbs = 0;
-			src0.NegateBase ^= NEGATE_XYZW;
+			src0.Negate ^= NEGATE_XYZW;
 		}
 	}
 	set_swizzle(&src0, 3, SWIZZLE_ONE);
@@ -649,7 +647,7 @@ GLboolean radeonTransformDeriv(struct radeon_transform_context* t,
 
 	B.Swizzle = MAKE_SWIZZLE4(SWIZZLE_ONE, SWIZZLE_ONE,
 						SWIZZLE_ONE, SWIZZLE_ONE);
-	B.NegateBase = NEGATE_XYZW;
+	B.Negate = NEGATE_XYZW;
 
 	emit2(t->Program, inst->Opcode, inst->SaturateMode, inst->DstReg,
 		inst->SrcReg[0], B);
diff --git a/src/mesa/drivers/dri/r300/radeon_program_pair.c b/src/mesa/drivers/dri/r300/radeon_program_pair.c
index f398404f9f..ecc82ff8a8 100644
--- a/src/mesa/drivers/dri/r300/radeon_program_pair.c
+++ b/src/mesa/drivers/dri/r300/radeon_program_pair.c
@@ -255,8 +255,7 @@ static void final_rewrite(struct pair_state *s, struct prog_instruction *inst)
 		inst->SrcReg[2] = inst->SrcReg[1];
 		inst->SrcReg[1].File = PROGRAM_BUILTIN;
 		inst->SrcReg[1].Swizzle = SWIZZLE_1111;
-		inst->SrcReg[1].NegateBase = 0;
-		inst->SrcReg[1].NegateAbs = 0;
+		inst->SrcReg[1].Negate = NEGATE_NONE;
 		inst->Opcode = OPCODE_MAD;
 		break;
 	case OPCODE_CMP:
@@ -730,7 +729,7 @@ static GLboolean fill_instruction_into_pair(struct pair_state *s, struct radeon_
 					srcrgb = GL_TRUE;
 				else if (swz < 4)
 					srcalpha = GL_TRUE;
-				if (swz != SWIZZLE_NIL && GET_BIT(inst->SrcReg[i].NegateBase, j))
+				if (swz != SWIZZLE_NIL && GET_BIT(inst->SrcReg[i].Negate, j))
 					negatebase = 1;
 			}
 			source = alloc_pair_source(s, pair, inst->SrcReg[i], srcrgb, srcalpha);
@@ -739,12 +738,12 @@ static GLboolean fill_instruction_into_pair(struct pair_state *s, struct radeon_
 			pair->RGB.Arg[i].Source = source;
 			pair->RGB.Arg[i].Swizzle = inst->SrcReg[i].Swizzle & 0x1ff;
 			pair->RGB.Arg[i].Abs = inst->SrcReg[i].Abs;
-			pair->RGB.Arg[i].Negate = (negatebase & ~pair->RGB.Arg[i].Abs) ^ inst->SrcReg[i].NegateAbs;
+			pair->RGB.Arg[i].Negate = (negatebase & ~pair->RGB.Arg[i].Abs) ^ inst->SrcReg[i].Negate;
 		}
 		if (pairinst->NeedAlpha) {
 			GLboolean srcrgb = GL_FALSE;
 			GLboolean srcalpha = GL_FALSE;
-			GLuint negatebase = GET_BIT(inst->SrcReg[i].NegateBase, pairinst->IsTranscendent ? 0 : 3);
+			GLuint negatebase = GET_BIT(inst->SrcReg[i].Negate, pairinst->IsTranscendent ? 0 : 3);
 			GLuint swz = GET_SWZ(inst->SrcReg[i].Swizzle, pairinst->IsTranscendent ? 0 : 3);
 			if (swz < 3)
 				srcrgb = GL_TRUE;
@@ -756,7 +755,7 @@ static GLboolean fill_instruction_into_pair(struct pair_state *s, struct radeon_
 			pair->Alpha.Arg[i].Source = source;
 			pair->Alpha.Arg[i].Swizzle = swz;
 			pair->Alpha.Arg[i].Abs = inst->SrcReg[i].Abs;
-			pair->Alpha.Arg[i].Negate = (negatebase & ~pair->RGB.Arg[i].Abs) ^ inst->SrcReg[i].NegateAbs;
+			pair->Alpha.Arg[i].Negate = (negatebase & ~pair->RGB.Arg[i].Abs) ^ inst->SrcReg[i].Negate;
 		}
 	}
 
diff --git a/src/mesa/main/ffvertex_prog.c b/src/mesa/main/ffvertex_prog.c
index 03f42704a7..1ce5685af4 100644
--- a/src/mesa/main/ffvertex_prog.c
+++ b/src/mesa/main/ffvertex_prog.c
@@ -570,9 +570,8 @@ static void emit_arg( struct prog_src_register *src,
    src->File = reg.file;
    src->Index = reg.idx;
    src->Swizzle = reg.swz;
-   src->NegateBase = reg.negate ? NEGATE_XYZW : 0;
+   src->Negate = reg.negate ? NEGATE_XYZW : NEGATE_NONE;
    src->Abs = 0;
-   src->NegateAbs = 0;
    src->RelAddr = 0;
    /* Check that bitfield sizes aren't exceeded */
    ASSERT(src->Index == reg.idx);
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 4a124bf27e..a70d069bd9 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -663,9 +663,8 @@ static void emit_arg( struct prog_src_register *reg,
    reg->File = ureg.file;
    reg->Index = ureg.idx;
    reg->Swizzle = ureg.swz;
-   reg->NegateBase = ureg.negatebase ? 0xf : 0x0;
+   reg->Negate = ureg.negatebase ? NEGATE_XYZW : NEGATE_NONE;
    reg->Abs = ureg.abs;
-   reg->NegateAbs = ureg.negateabs;
 }
 
 static void emit_dst( struct prog_dst_register *dst,
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index 35253daa2e..b47bf360cf 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -2669,7 +2669,7 @@ parse_vector_src_reg(GLcontext *ctx, const GLubyte **inst,
    reg->File = file;
    reg->Index = index;
    reg->Swizzle = MAKE_SWIZZLE4(swizzle[0], swizzle[1], swizzle[2], swizzle[3]);
-   reg->NegateBase = negateMask;
+   reg->Negate = negateMask;
    reg->RelAddr = isRelOffset;
    return 0;
 }
@@ -2703,7 +2703,7 @@ parse_scalar_src_reg(GLcontext *ctx, const GLubyte **inst,
    reg->File = file;
    reg->Index = index;
    reg->Swizzle = (swizzle[0] << 0);
-   reg->NegateBase = negateMask;
+   reg->Negate = negateMask;
    reg->RelAddr = isRelOffset;
    return 0;
 }
@@ -3019,7 +3019,7 @@ parse_fp_instruction (GLcontext * ctx, const GLubyte ** inst,
 	    parse_extended_swizzle_mask(inst, swizzle, &negateMask);
 	    fp->SrcReg[0].File = file;
 	    fp->SrcReg[0].Index = index;
-	    fp->SrcReg[0].NegateBase = negateMask;
+	    fp->SrcReg[0].Negate = negateMask;
 	    fp->SrcReg[0].Swizzle = MAKE_SWIZZLE4(swizzle[0],
                                                   swizzle[1],
                                                   swizzle[2],
@@ -3363,7 +3363,7 @@ parse_vp_instruction (GLcontext * ctx, const GLubyte ** inst,
 	    parse_extended_swizzle_mask (inst, swizzle, &negateMask);
 	    vp->SrcReg[0].File = file;
 	    vp->SrcReg[0].Index = index;
-	    vp->SrcReg[0].NegateBase = negateMask;
+	    vp->SrcReg[0].Negate = negateMask;
 	    vp->SrcReg[0].Swizzle = MAKE_SWIZZLE4(swizzle[0],
                                                   swizzle[1],
                                                   swizzle[2],
diff --git a/src/mesa/shader/nvfragparse.c b/src/mesa/shader/nvfragparse.c
index 56b7c29bea..0fd55524ab 100644
--- a/src/mesa/shader/nvfragparse.c
+++ b/src/mesa/shader/nvfragparse.c
@@ -957,6 +957,7 @@ Parse_VectorSrc(struct parse_state *parseState,
    GLfloat sign = 1.0F;
    GLubyte token[100];
    GLint idx;
+   GLuint negateBase, negateAbs;
 
    /*
     * First, take care of +/- and absolute value stuff.
@@ -968,21 +969,23 @@ Parse_VectorSrc(struct parse_state *parseState,
 
    if (Parse_String(parseState, "|")) {
       srcReg->Abs = GL_TRUE;
-      srcReg->NegateAbs = (sign < 0.0F) ? GL_TRUE : GL_FALSE;
+      negateAbs = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
 
       if (Parse_String(parseState, "-"))
-         srcReg->NegateBase = NEGATE_XYZW;
+         negateBase = NEGATE_XYZW;
       else if (Parse_String(parseState, "+"))
-         srcReg->NegateBase = NEGATE_NONE;
+         negateBase = NEGATE_NONE;
       else
-         srcReg->NegateBase = NEGATE_NONE;
+         negateBase = NEGATE_NONE;
    }
    else {
       srcReg->Abs = GL_FALSE;
-      srcReg->NegateAbs = GL_FALSE;
-      srcReg->NegateBase = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
+      negateAbs = NEGATE_NONE;
+      negateBase = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
    }
 
+   srcReg->Negate = srcReg->Abs ? negateAbs : negateBase;
+
    /* This should be the real src vector/register name */
    if (!Peek_Token(parseState, token))
       RETURN_ERROR;
@@ -1083,6 +1086,7 @@ Parse_ScalarSrcReg(struct parse_state *parseState,
    GLfloat sign = 1.0F;
    GLboolean needSuffix = GL_TRUE;
    GLint idx;
+   GLuint negateBase, negateAbs;
 
    /*
     * First, take care of +/- and absolute value stuff.
@@ -1094,21 +1098,23 @@ Parse_ScalarSrcReg(struct parse_state *parseState,
 
    if (Parse_String(parseState, "|")) {
       srcReg->Abs = GL_TRUE;
-      srcReg->NegateAbs = (sign < 0.0F) ? GL_TRUE : GL_FALSE;
+      negateAbs = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
 
       if (Parse_String(parseState, "-"))
-         srcReg->NegateBase = NEGATE_XYZW;
+         negateBase = NEGATE_XYZW;
       else if (Parse_String(parseState, "+"))
-         srcReg->NegateBase = NEGATE_NONE;
+         negateBase = NEGATE_NONE;
       else
-         srcReg->NegateBase = NEGATE_NONE;
+         negateBase = NEGATE_NONE;
    }
    else {
       srcReg->Abs = GL_FALSE;
-      srcReg->NegateAbs = GL_FALSE;
-      srcReg->NegateBase = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
+      negateAbs = NEGATE_NONE;
+      negateBase = (sign < 0.0F) ? NEGATE_XYZW : NEGATE_NONE;
    }
 
+   srcReg->Negate = srcReg->Abs ? negateAbs : negateBase;
+
    if (!Peek_Token(parseState, token))
       RETURN_ERROR;
 
@@ -1247,9 +1253,8 @@ Parse_PrintInstruction(struct parse_state *parseState,
    }
 
    inst->SrcReg[0].Swizzle = SWIZZLE_NOOP;
-   inst->SrcReg[0].NegateBase = NEGATE_NONE;
    inst->SrcReg[0].Abs = GL_FALSE;
-   inst->SrcReg[0].NegateAbs = GL_FALSE;
+   inst->SrcReg[0].Negate = NEGATE_NONE;
 
    return GL_TRUE;
 }
diff --git a/src/mesa/shader/nvvertparse.c b/src/mesa/shader/nvvertparse.c
index 624262395b..f5e2df2670 100644
--- a/src/mesa/shader/nvvertparse.c
+++ b/src/mesa/shader/nvvertparse.c
@@ -641,12 +641,12 @@ Parse_SwizzleSrcReg(struct parse_state *parseState, struct prog_src_register *sr
       RETURN_ERROR;
    if (token[0] == '-') {
       (void) Parse_String(parseState, "-");
-      srcReg->NegateBase = NEGATE_XYZW;
+      srcReg->Negate = NEGATE_XYZW;
       if (!Peek_Token(parseState, token))
          RETURN_ERROR;
    }
    else {
-      srcReg->NegateBase = NEGATE_NONE;
+      srcReg->Negate = NEGATE_NONE;
    }
 
    /* Src reg can be R<n>, c[n], c[n +/- offset], or a named vertex attrib */
@@ -734,13 +734,13 @@ Parse_ScalarSrcReg(struct parse_state *parseState, struct prog_src_register *src
    if (!Peek_Token(parseState, token))
       RETURN_ERROR;
    if (token[0] == '-') {
-      srcReg->NegateBase = NEGATE_XYZW;
+      srcReg->Negate = NEGATE_XYZW;
       (void) Parse_String(parseState, "-"); /* consume '-' */
       if (!Peek_Token(parseState, token))
          RETURN_ERROR;
    }
    else {
-      srcReg->NegateBase = NEGATE_NONE;
+      srcReg->Negate = NEGATE_NONE;
    }
 
    /* Src reg can be R<n>, c[n], c[n +/- offset], or a named vertex attrib */
@@ -1062,7 +1062,7 @@ Parse_PrintInstruction(struct parse_state *parseState, struct prog_instruction *
          RETURN_ERROR;
 
       srcReg->RelAddr = GL_FALSE;
-      srcReg->NegateBase = NEGATE_NONE;
+      srcReg->Negate = NEGATE_NONE;
       srcReg->Swizzle = SWIZZLE_NOOP;
 
       /* Register can be R<n>, c[n], c[n +/- offset], a named vertex attrib,
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index bdac1d4f8a..68a59350a1 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -212,19 +212,14 @@ fetch_vector4(const struct prog_src_register *source,
       result[3] = src[GET_SWZ(source->Swizzle, 3)];
    }
 
-   if (source->NegateBase) {
-      result[0] = -result[0];
-      result[1] = -result[1];
-      result[2] = -result[2];
-      result[3] = -result[3];
-   }
    if (source->Abs) {
       result[0] = FABSF(result[0]);
       result[1] = FABSF(result[1]);
       result[2] = FABSF(result[2]);
       result[3] = FABSF(result[3]);
    }
-   if (source->NegateAbs) {
+   if (source->Negate) {
+      ASSERT(source->Negate == NEGATE_XYZW);
       result[0] = -result[0];
       result[1] = -result[1];
       result[2] = -result[2];
@@ -259,7 +254,7 @@ fetch_vector4ui(const struct prog_src_register *source,
       result[3] = src[GET_SWZ(source->Swizzle, 3)];
    }
 
-   /* Note: no NegateBase, Abs, NegateAbs here */
+   /* Note: no Negate or Abs here */
 }
 
 
@@ -299,19 +294,14 @@ fetch_vector4_deriv(GLcontext * ctx,
       result[2] = deriv[GET_SWZ(source->Swizzle, 2)];
       result[3] = deriv[GET_SWZ(source->Swizzle, 3)];
       
-      if (source->NegateBase) {
-         result[0] = -result[0];
-         result[1] = -result[1];
-         result[2] = -result[2];
-         result[3] = -result[3];
-      }
       if (source->Abs) {
          result[0] = FABSF(result[0]);
          result[1] = FABSF(result[1]);
          result[2] = FABSF(result[2]);
          result[3] = FABSF(result[3]);
       }
-      if (source->NegateAbs) {
+      if (source->Negate) {
+         ASSERT(source->Negate == NEGATE_XYZW);
          result[0] = -result[0];
          result[1] = -result[1];
          result[2] = -result[2];
@@ -336,13 +326,10 @@ fetch_vector1(const struct prog_src_register *source,
 
    result[0] = src[GET_SWZ(source->Swizzle, 0)];
 
-   if (source->NegateBase) {
-      result[0] = -result[0];
-   }
    if (source->Abs) {
       result[0] = FABSF(result[0]);
    }
-   if (source->NegateAbs) {
+   if (source->Negate) {
       result[0] = -result[0];
    }
 }
@@ -1514,7 +1501,7 @@ _mesa_execute_program(GLcontext * ctx,
                   ASSERT(swz <= 3);
                   result[i] = src[swz];
                }
-               if (source->NegateBase & (1 << i))
+               if (source->Negate & (1 << i))
                   result[i] = -result[i];
             }
             store_vector4(inst, machine, result);
diff --git a/src/mesa/shader/prog_instruction.h b/src/mesa/shader/prog_instruction.h
index 4adce11f95..3109f6cbae 100644
--- a/src/mesa/shader/prog_instruction.h
+++ b/src/mesa/shader/prog_instruction.h
@@ -261,37 +261,15 @@ struct prog_src_register
    GLuint Swizzle:12;
    GLuint RelAddr:1;
 
-   /**
-    * \name Source register "sign" control.
-    *
-    * The ARB and NV extensions allow varrying degrees of control over the
-    * sign of the source vector components.  These values allow enough control
-    * for all flavors of the extensions.
-    */
-   /*@{*/
-   /**
-    * Per-component negation for the SWZ instruction.  For non-SWZ
-    * instructions the only possible values are NEGATE_XYZW and NEGATE_NONE.
-    *
-    * \since
-    * ARB_vertex_program, ARB_fragment_program
-    */
-   GLuint NegateBase:4;
-
-   /**
-    * Take the component-wise absolute value.
-    *
-    * \since
-    * NV_fragment_program, NV_fragment_program_option, NV_vertex_program2,
-    * NV_vertex_program2_option.
-    */
+   /** Take the component-wise absolute value */
    GLuint Abs:1;
 
    /**
-    * Post-absolute value negation (all components).
+    * Post-Abs negation.
+    * This will either be NEGATE_NONE or NEGATE_XYZW, except for the SWZ
+    * instruction which allows per-component negation.
     */
-   GLuint NegateAbs:1;
-   /*@}*/
+   GLuint Negate:4;
 };
 
 
diff --git a/src/mesa/shader/prog_print.c b/src/mesa/shader/prog_print.c
index b832ddb477..d73c619fea 100644
--- a/src/mesa/shader/prog_print.c
+++ b/src/mesa/shader/prog_print.c
@@ -325,19 +325,19 @@ reg_string(gl_register_file f, GLint index, gl_prog_print_mode mode,
  * \param extended  if true, also allow 0, 1 values
  */
 const char *
-_mesa_swizzle_string(GLuint swizzle, GLuint negateBase, GLboolean extended)
+_mesa_swizzle_string(GLuint swizzle, GLuint negateMask, GLboolean extended)
 {
    static const char swz[] = "xyzw01!?";  /* See SWIZZLE_x definitions */
    static char s[20];
    GLuint i = 0;
 
-   if (!extended && swizzle == SWIZZLE_NOOP && negateBase == 0)
+   if (!extended && swizzle == SWIZZLE_NOOP && negateMask == 0)
       return ""; /* no swizzle/negation */
 
    if (!extended)
       s[i++] = '.';
 
-   if (negateBase & NEGATE_X)
+   if (negateMask & NEGATE_X)
       s[i++] = '-';
    s[i++] = swz[GET_SWZ(swizzle, 0)];
 
@@ -345,7 +345,7 @@ _mesa_swizzle_string(GLuint swizzle, GLuint negateBase, GLboolean extended)
       s[i++] = ',';
    }
 
-   if (negateBase & NEGATE_Y)
+   if (negateMask & NEGATE_Y)
       s[i++] = '-';
    s[i++] = swz[GET_SWZ(swizzle, 1)];
 
@@ -353,7 +353,7 @@ _mesa_swizzle_string(GLuint swizzle, GLuint negateBase, GLboolean extended)
       s[i++] = ',';
    }
 
-   if (negateBase & NEGATE_Z)
+   if (negateMask & NEGATE_Z)
       s[i++] = '-';
    s[i++] = swz[GET_SWZ(swizzle, 2)];
 
@@ -361,7 +361,7 @@ _mesa_swizzle_string(GLuint swizzle, GLuint negateBase, GLboolean extended)
       s[i++] = ',';
    }
 
-   if (negateBase & NEGATE_W)
+   if (negateMask & NEGATE_W)
       s[i++] = '-';
    s[i++] = swz[GET_SWZ(swizzle, 3)];
 
@@ -465,14 +465,14 @@ fprint_src_reg(FILE *f,
                  reg_string((gl_register_file) srcReg->File,
                             srcReg->Index, mode, srcReg->RelAddr, prog),
                  _mesa_swizzle_string(srcReg->Swizzle,
-                                      srcReg->NegateBase, GL_FALSE),
+                                      srcReg->Negate, GL_FALSE),
                  abs);
 #if 0
    _mesa_fprintf(f, "%s[%d]%s",
                  file_string((gl_register_file) srcReg->File, mode),
                  srcReg->Index,
                  _mesa_swizzle_string(srcReg->Swizzle,
-                                      srcReg->NegateBase, GL_FALSE));
+                                      srcReg->Negate, GL_FALSE));
 #endif
 }
 
@@ -566,7 +566,7 @@ _mesa_fprint_instruction_opt(FILE *f,
                                   mode),
                       inst->SrcReg[0].Index,
                       _mesa_swizzle_string(inst->SrcReg[0].Swizzle,
-                                           inst->SrcReg[0].NegateBase, GL_FALSE));
+                                           inst->SrcReg[0].Negate, GL_FALSE));
       }
       if (inst->Comment)
          _mesa_fprintf(f, "  # %s", inst->Comment);
@@ -583,7 +583,7 @@ _mesa_fprint_instruction_opt(FILE *f,
                                mode),
                    inst->SrcReg[0].Index,
                    _mesa_swizzle_string(inst->SrcReg[0].Swizzle,
-                                        inst->SrcReg[0].NegateBase, GL_TRUE));
+                                        inst->SrcReg[0].Negate, GL_TRUE));
       fprint_comment(f, inst);
       break;
    case OPCODE_TEX:
diff --git a/src/mesa/shader/programopt.c b/src/mesa/shader/programopt.c
index e283f8933b..ecd98dc85c 100644
--- a/src/mesa/shader/programopt.c
+++ b/src/mesa/shader/programopt.c
@@ -241,7 +241,7 @@ _mesa_append_fog_code(GLcontext *ctx, struct gl_fragment_program *fprog)
       inst->DstReg.WriteMask = WRITEMASK_X;
       inst->SrcReg[0].File = PROGRAM_TEMPORARY;
       inst->SrcReg[0].Index = fogFactorTemp;
-      inst->SrcReg[0].NegateBase = NEGATE_XYZW;
+      inst->SrcReg[0].Negate = NEGATE_XYZW;
       inst->SrcReg[0].Swizzle = SWIZZLE_XXXX;
       inst->SaturateMode = SATURATE_ZERO_ONE;
       inst++;
diff --git a/src/mesa/shader/slang/slang_emit.c b/src/mesa/shader/slang/slang_emit.c
index 8493c490fb..3f455e0640 100644
--- a/src/mesa/shader/slang/slang_emit.c
+++ b/src/mesa/shader/slang/slang_emit.c
@@ -1135,7 +1135,7 @@ emit_negation(slang_emit_info *emitInfo, slang_ir_node *n)
                            n->Children[0]->Store,
                            NULL,
                            NULL);
-   inst->SrcReg[0].NegateBase = NEGATE_XYZW;
+   inst->SrcReg[0].Negate = NEGATE_XYZW;
    return inst;
 }
 
diff --git a/src/mesa/state_tracker/st_cb_bitmap.c b/src/mesa/state_tracker/st_cb_bitmap.c
index 3b2ad00f5c..fa4f4082a7 100644
--- a/src/mesa/state_tracker/st_cb_bitmap.c
+++ b/src/mesa/state_tracker/st_cb_bitmap.c
@@ -147,7 +147,7 @@ make_bitmap_fragment_program(GLcontext *ctx, GLuint samplerIndex)
       p->Instructions[ic].SrcReg[0].Swizzle = SWIZZLE_XXXX;
 
    p->Instructions[ic].SrcReg[0].Index = 0;
-   p->Instructions[ic].SrcReg[0].NegateBase = NEGATE_XYZW;
+   p->Instructions[ic].SrcReg[0].Negate = NEGATE_XYZW;
    ic++;
 
    /* END; */
diff --git a/src/mesa/state_tracker/st_mesa_to_tgsi.c b/src/mesa/state_tracker/st_mesa_to_tgsi.c
index ffa607dd87..43c9afccc3 100644
--- a/src/mesa/state_tracker/st_mesa_to_tgsi.c
+++ b/src/mesa/state_tracker/st_mesa_to_tgsi.c
@@ -275,8 +275,8 @@ compile_instruction(
       /* swizzle (ext swizzle also depends on negation) */
       {
          GLuint swz[4];
-         GLboolean extended = (inst->SrcReg[i].NegateBase != NEGATE_NONE &&
-                               inst->SrcReg[i].NegateBase != NEGATE_XYZW);
+         GLboolean extended = (inst->SrcReg[i].Negate != NEGATE_NONE &&
+                               inst->SrcReg[i].Negate != NEGATE_XYZW);
          for( j = 0; j < 4; j++ ) {
             swz[j] = GET_SWZ( inst->SrcReg[i].Swizzle, j );
             if (swz[j] > SWIZZLE_W)
@@ -296,20 +296,20 @@ compile_instruction(
          }
       }
 
-      if( inst->SrcReg[i].NegateBase == NEGATE_XYZW ) {
+      if( inst->SrcReg[i].Negate == NEGATE_XYZW ) {
          fullsrc->SrcRegister.Negate = 1;
       }
-      else if( inst->SrcReg[i].NegateBase != NEGATE_NONE ) {
-         if( inst->SrcReg[i].NegateBase & NEGATE_X ) {
+      else if( inst->SrcReg[i].Negate != NEGATE_NONE ) {
+         if( inst->SrcReg[i].Negate & NEGATE_X ) {
             fullsrc->SrcRegisterExtSwz.NegateX = 1;
          }
-         if( inst->SrcReg[i].NegateBase & NEGATE_Y ) {
+         if( inst->SrcReg[i].Negate & NEGATE_Y ) {
             fullsrc->SrcRegisterExtSwz.NegateY = 1;
          }
-         if( inst->SrcReg[i].NegateBase & NEGATE_Z ) {
+         if( inst->SrcReg[i].Negate & NEGATE_Z ) {
             fullsrc->SrcRegisterExtSwz.NegateZ = 1;
          }
-         if( inst->SrcReg[i].NegateBase & NEGATE_W ) {
+         if( inst->SrcReg[i].Negate & NEGATE_W ) {
             fullsrc->SrcRegisterExtSwz.NegateW = 1;
          }
       }
@@ -318,10 +318,6 @@ compile_instruction(
          fullsrc->SrcRegisterExtMod.Absolute = 1;
       }
 
-      if( inst->SrcReg[i].NegateAbs ) {
-         fullsrc->SrcRegisterExtMod.Negate = 1;
-      }
-
       if( inst->SrcReg[i].RelAddr ) {
          fullsrc->SrcRegister.Indirect = 1;
 
-- 
cgit v1.2.3


From c6af9b29476e4e445623e7a2f737ba95003bbe13 Mon Sep 17 00:00:00 2001
From: José Fonseca <jfonseca@vmware.com>
Date: Mon, 15 Jun 2009 19:20:25 +0100
Subject: mesa: Always return a value.

---
 src/mesa/main/texenvprogram.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index a70d069bd9..b92ba2542d 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -877,6 +877,7 @@ static struct ureg get_source( struct texenv_fragment_program *p,
 
    default:
       assert(0);
+      return undef;
    }
 }
 
-- 
cgit v1.2.3


From e5f63c403b767f9974e8eb5d412c012b8a69287f Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Tue, 2 Jun 2009 07:47:20 -0700
Subject: Fix state flag dependencies for fixed function fragment program
 updates.

I started looking into why _NEW_ARRAY punishes us, and while annotating
dependencies noticed that a bunch of dependencies were missing.
---
 src/mesa/main/state.c         |  2 +-
 src/mesa/main/texenvprogram.c | 10 ++++++++++
 2 files changed, 11 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 3b2c6ec618..9ba131bee5 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -547,7 +547,7 @@ _mesa_update_state_locked( GLcontext *ctx )
    /* Determine which state flags effect vertex/fragment program state */
    if (ctx->FragmentProgram._MaintainTexEnvProgram) {
       prog_flags |= (_NEW_TEXTURE | _NEW_FOG | _DD_NEW_SEPARATE_SPECULAR |
-		     _NEW_ARRAY);
+		     _NEW_ARRAY | _NEW_LIGHT | _NEW_POINT | _NEW_RENDERMODE);
    }
    if (ctx->VertexProgram._MaintainTnlProgram) {
       prog_flags |= (_NEW_ARRAY | _NEW_TEXTURE | _NEW_TEXTURE_MATRIX |
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index b92ba2542d..a3f1246c98 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -261,6 +261,7 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
  */
 static GLbitfield get_fp_input_mask( GLcontext *ctx )
 {
+   /* _NEW_PROGRAM */
    const GLboolean vertexShader = (ctx->Shader.CurrentProgram &&
                                    ctx->Shader.CurrentProgram->VertexProgram);
    const GLboolean vertexProgram = ctx->VertexProgram._Enabled;
@@ -274,22 +275,26 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       fp_inputs = ~0;
    }
    else if (ctx->RenderMode == GL_FEEDBACK) {
+      /* _NEW_RENDERMODE */
       fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
    }
    else if (!(vertexProgram || vertexShader) ||
             !ctx->VertexProgram._Current) {
       /* Fixed function vertex logic */
+      /* _NEW_ARRAY */
       GLbitfield varying_inputs = ctx->varying_vp_inputs;
 
       /* These get generated in the setup routine regardless of the
        * vertex program:
        */
+      /* _NEW_POINT */
       if (ctx->Point.PointSprite)
          varying_inputs |= FRAG_BITS_TEX_ANY;
 
       /* First look at what values may be computed by the generated
        * vertex program:
        */
+      /* _NEW_LIGHT */
       if (ctx->Light.Enabled) {
          fp_inputs |= FRAG_BIT_COL0;
 
@@ -297,6 +302,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
             fp_inputs |= FRAG_BIT_COL1;
       }
 
+      /* _NEW_TEXTURE */
       fp_inputs |= (ctx->Texture._TexGenEnabled |
                     ctx->Texture._TexMatEnabled) << FRAG_ATTRIB_TEX0;
 
@@ -329,6 +335,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       /* These get generated in the setup routine regardless of the
        * vertex program:
        */
+      /* _NEW_POINT */
       if (ctx->Point.PointSprite)
          vp_outputs |= FRAG_BITS_TEX_ANY;
 
@@ -355,6 +362,7 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
 
    memset(key, 0, sizeof(*key));
 
+   /* _NEW_TEXTURE */
    for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
       const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i];
       GLenum format;
@@ -408,11 +416,13 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
        }
    }
 
+   /* _DD_NEW_SEPARATE_SPECULAR */
    if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) {
       key->separate_specular = 1;
       inputs_referenced |= FRAG_BIT_COL1;
    }
 
+   /* _NEW_FOG */
    if (ctx->Fog.Enabled) {
       key->fog_enabled = 1;
       key->fog_mode = translate_fog_mode(ctx->Fog.Mode);
-- 
cgit v1.2.3


From 9cea84b6b5f96bf3bb42546e49b76024d7126e30 Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Thu, 16 Jul 2009 18:41:03 -0700
Subject: texenv: Calculate whether we need to do secondary color on our own.

The _TriangleCaps bit is deprecated, not updated when we require, and
is set based on state that hasn't been updated at that point in
_mesa_update_state_locked().

Fixes incorrect clear color in glsl/twoside.c with meta_clear_tris.
---
 src/mesa/main/state.c         |  2 +-
 src/mesa/main/texenvprogram.c | 18 +++++++++++++++---
 2 files changed, 16 insertions(+), 4 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 9ba131bee5..d8191ab518 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -546,7 +546,7 @@ _mesa_update_state_locked( GLcontext *ctx )
 
    /* Determine which state flags effect vertex/fragment program state */
    if (ctx->FragmentProgram._MaintainTexEnvProgram) {
-      prog_flags |= (_NEW_TEXTURE | _NEW_FOG | _DD_NEW_SEPARATE_SPECULAR |
+      prog_flags |= (_NEW_TEXTURE | _NEW_FOG |
 		     _NEW_ARRAY | _NEW_LIGHT | _NEW_POINT | _NEW_RENDERMODE);
    }
    if (ctx->VertexProgram._MaintainTnlProgram) {
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index a3f1246c98..6b090ff399 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -62,6 +62,18 @@ struct texenvprog_cache_item
    struct texenvprog_cache_item *next;
 };
 
+static GLboolean
+texenv_doing_secondary_color(GLcontext *ctx)
+{
+   if (ctx->Light.Enabled &&
+       (ctx->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR))
+      return GL_TRUE;
+
+   if (ctx->Fog.ColorSumEnabled)
+      return GL_TRUE;
+
+   return GL_FALSE;
+}
 
 /**
  * Up to nine instructions per tex unit, plus fog, specular color.
@@ -298,7 +310,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       if (ctx->Light.Enabled) {
          fp_inputs |= FRAG_BIT_COL0;
 
-         if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR)
+         if (texenv_doing_secondary_color(ctx))
             fp_inputs |= FRAG_BIT_COL1;
       }
 
@@ -416,8 +428,8 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
        }
    }
 
-   /* _DD_NEW_SEPARATE_SPECULAR */
-   if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) {
+   /* _NEW_LIGHT | _NEW_FOG */
+   if (texenv_doing_secondary_color(ctx)) {
       key->separate_specular = 1;
       inputs_referenced |= FRAG_BIT_COL1;
    }
-- 
cgit v1.2.3


From 40990d9dfb20b69585859b2a45596aa46c20140a Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Mon, 3 Aug 2009 12:36:52 -0700
Subject: texenv: Match state.c in deciding whether we'll be using a vertex
 shader.

---
 src/mesa/main/texenvprogram.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 6b090ff399..2eefae8255 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -275,6 +275,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
 {
    /* _NEW_PROGRAM */
    const GLboolean vertexShader = (ctx->Shader.CurrentProgram &&
+				   ctx->Shader.CurrentProgram->LinkStatus &&
                                    ctx->Shader.CurrentProgram->VertexProgram);
    const GLboolean vertexProgram = ctx->VertexProgram._Enabled;
    GLbitfield fp_inputs = 0x0;
-- 
cgit v1.2.3


From a9ba1bfeb3a2852c6eda718e73c46c972a286648 Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Mon, 3 Aug 2009 12:38:56 -0700
Subject: texenv: Use VP->Current, since _Current isn't updated at this point.

---
 src/mesa/main/texenvprogram.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 2eefae8255..1ee78ffa16 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -341,7 +341,7 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       if (vertexShader)
          vprog = ctx->Shader.CurrentProgram->VertexProgram;
       else
-         vprog = ctx->VertexProgram._Current;
+         vprog = ctx->VertexProgram.Current;
 
       vp_outputs = vprog->Base.OutputsWritten;
 
-- 
cgit v1.2.3


From 6807d96f8efeecd9d71e1e1bff856e7e04f5f364 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Fri, 7 Aug 2009 09:42:28 -0600
Subject: mesa: reformat code to allow setting breakpoints on the
 true-statement

---
 src/mesa/main/texenvprogram.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 1ee78ffa16..3736138b9e 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -322,8 +322,10 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       /* Then look at what might be varying as a result of enabled
        * arrays, etc:
        */
-      if (varying_inputs & VERT_BIT_COLOR0) fp_inputs |= FRAG_BIT_COL0;
-      if (varying_inputs & VERT_BIT_COLOR1) fp_inputs |= FRAG_BIT_COL1;
+      if (varying_inputs & VERT_BIT_COLOR0)
+         fp_inputs |= FRAG_BIT_COL0;
+      if (varying_inputs & VERT_BIT_COLOR1)
+         fp_inputs |= FRAG_BIT_COL1;
 
       fp_inputs |= (((varying_inputs & VERT_BIT_TEX_ANY) >> VERT_ATTRIB_TEX0) 
                     << FRAG_ATTRIB_TEX0);
@@ -352,8 +354,10 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
       if (ctx->Point.PointSprite)
          vp_outputs |= FRAG_BITS_TEX_ANY;
 
-      if (vp_outputs & (1 << VERT_RESULT_COL0)) fp_inputs |= FRAG_BIT_COL0;
-      if (vp_outputs & (1 << VERT_RESULT_COL1)) fp_inputs |= FRAG_BIT_COL1;
+      if (vp_outputs & (1 << VERT_RESULT_COL0))
+         fp_inputs |= FRAG_BIT_COL0;
+      if (vp_outputs & (1 << VERT_RESULT_COL1))
+         fp_inputs |= FRAG_BIT_COL1;
 
       fp_inputs |= (((vp_outputs & VERT_RESULT_TEX_ANY) >> VERT_RESULT_TEX0) 
                     << FRAG_ATTRIB_TEX0);
-- 
cgit v1.2.3


From 956e6c3978abe918348377cf05e5c92971e50d3f Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Mon, 31 Aug 2009 11:14:16 -0600
Subject: mesa: fix saturation logic in emit_texenv()

We need to clamp/saturate after each texenv stage, not just the last one.
Fixes glean texEnv failure for softpipe (and probably other fragment program-
based drivers).
---
 src/mesa/main/texenvprogram.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 3736138b9e..847db5cfb6 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1113,7 +1113,7 @@ static struct ureg
 emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 {
    struct state_key *key = p->state;
-   GLboolean saturate = (unit < p->last_tex_stage);
+   GLboolean saturate;
    GLuint rgb_shift, alpha_shift;
    struct ureg out, shift;
    struct ureg dest;
@@ -1141,6 +1141,11 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
       break;
    }
    
+   /* If we'll do rgb/alpha shifting don't saturate in emit_combine().
+    * We don't want to clamp twice.
+    */
+   saturate = !(rgb_shift || alpha_shift);
+
    /* If this is the very last calculation, emit direct to output reg:
     */
    if (key->separate_specular ||
@@ -1189,6 +1194,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    /* Deal with the final shift:
     */
    if (alpha_shift || rgb_shift) {
+      saturate = GL_TRUE;  /* always saturate at this point */
       if (rgb_shift == alpha_shift) {
 	 shift = register_scalar_const(p, (GLfloat)(1<<rgb_shift));
       }
-- 
cgit v1.2.3


From a5e7003565a021ceb82a8b16f4a0ecc7cc6c02e0 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Mon, 31 Aug 2009 11:17:59 -0600
Subject: mesa: added const qualifiers, move local var

---
 src/mesa/main/texenvprogram.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 847db5cfb6..3e5d65c875 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -952,7 +952,7 @@ static struct ureg emit_combine_source( struct texenv_fragment_program *p,
    }
 }
 
-static GLboolean args_match( struct state_key *key, GLuint unit )
+static GLboolean args_match( const struct state_key *key, GLuint unit )
 {
    GLuint i, nr = key->unit[unit].NumArgsRGB;
 
@@ -1112,11 +1112,10 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
 static struct ureg
 emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 {
-   struct state_key *key = p->state;
+   const struct state_key *key = p->state;
    GLboolean saturate;
    GLuint rgb_shift, alpha_shift;
-   struct ureg out, shift;
-   struct ureg dest;
+   struct ureg out, dest;
 
    if (!key->unit[unit].enabled) {
       return get_source(p, SRC_PREVIOUS, 0);
@@ -1168,7 +1167,6 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    }
    else if (key->unit[unit].ModeRGB == MODE_DOT3_RGBA_EXT ||
 	    key->unit[unit].ModeRGB == MODE_DOT3_RGBA) {
-
       out = emit_combine( p, dest, WRITEMASK_XYZW, saturate,
 			  unit,
 			  key->unit[unit].NumArgsRGB,
@@ -1194,7 +1192,10 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    /* Deal with the final shift:
     */
    if (alpha_shift || rgb_shift) {
+      struct ureg shift;
+
       saturate = GL_TRUE;  /* always saturate at this point */
+
       if (rgb_shift == alpha_shift) {
 	 shift = register_scalar_const(p, (GLfloat)(1<<rgb_shift));
       }
@@ -1294,7 +1295,7 @@ static GLboolean load_texenv_source( struct texenv_fragment_program *p,
 static GLboolean
 load_texunit_sources( struct texenv_fragment_program *p, int unit )
 {
-   struct state_key *key = p->state;
+   const struct state_key *key = p->state;
    GLuint i;
 
    for (i = 0; i < key->unit[unit].NumArgsRGB; i++) {
@@ -1314,7 +1315,7 @@ load_texunit_sources( struct texenv_fragment_program *p, int unit )
 static GLboolean
 load_texunit_bumpmap( struct texenv_fragment_program *p, int unit )
 {
-   struct state_key *key = p->state;
+   const struct state_key *key = p->state;
    GLuint bumpedUnitNr = key->unit[unit].OptRGB[1].Source - SRC_TEXTURE0;
    struct ureg texcDst, bumpMapRes;
    struct ureg constdudvcolor = register_const4f(p, 0.0, 0.0, 0.0, 1.0);
-- 
cgit v1.2.3


From d55a28e3dc712342e4a2f25aa9b661add7461e82 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 15:34:16 -0600
Subject: mesa: replace 8 with MAX_TEXTURE_UNITS

---
 src/mesa/main/texenvprogram.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 3e5d65c875..f19bb99300 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -109,7 +109,7 @@ struct state_key {
       GLuint NumArgsA:3;
       GLuint ModeA:5;
       struct mode_opt OptA[MAX_COMBINER_TERMS];
-   } unit[8];
+   } unit[MAX_TEXTURE_UNITS];
 };
 
 #define FOG_LINEAR  0
-- 
cgit v1.2.3


From f337e2c4954b4779dcf2c2b4e0e068cb3ed7b841 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 15:49:55 -0600
Subject: mesa: minor code clean-ups, comments

---
 src/mesa/main/texenvprogram.c | 58 +++++++++++++++++++++++++------------------
 1 file changed, 34 insertions(+), 24 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index f19bb99300..0898581485 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -83,8 +83,8 @@ texenv_doing_secondary_color(GLcontext *ctx)
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
 struct mode_opt {
-   GLuint Source:4;
-   GLuint Operand:3;
+   GLuint Source:4;  /**< SRC_x */
+   GLuint Operand:3; /**< OPR_x */
 };
 
 struct state_key {
@@ -103,11 +103,11 @@ struct state_key {
       GLuint ScaleShiftA:2;
 
       GLuint NumArgsRGB:3;
-      GLuint ModeRGB:5;
+      GLuint ModeRGB:5;     /**< MODE_x */
       struct mode_opt OptRGB[MAX_COMBINER_TERMS];
 
       GLuint NumArgsA:3;
-      GLuint ModeA:5;
+      GLuint ModeA:5;     /**< MODE_x */
       struct mode_opt OptA[MAX_COMBINER_TERMS];
    } unit[MAX_TEXTURE_UNITS];
 };
@@ -240,7 +240,12 @@ static GLuint translate_mode( GLenum envMode, GLenum mode )
    }
 }
 
+
 #define TEXTURE_UNKNOWN_INDEX 7
+
+/**
+ * Translate TEXTURE_x_BIT to TEXTURE_x_INDEX.
+ */
 static GLuint translate_tex_src_bit( GLbitfield bit )
 {
    /* make sure number of switch cases is correct */
@@ -259,6 +264,7 @@ static GLuint translate_tex_src_bit( GLbitfield bit )
    }
 }
 
+
 #define VERT_BIT_TEX_ANY    (0xff << VERT_ATTRIB_TEX0)
 #define VERT_RESULT_TEX_ANY (0xff << VERT_RESULT_TEX0)
 
@@ -375,51 +381,50 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
 {
    GLuint i, j;
    GLbitfield inputs_referenced = FRAG_BIT_COL0;
-   GLbitfield inputs_available = get_fp_input_mask( ctx );
+   const GLbitfield inputs_available = get_fp_input_mask( ctx );
 
    memset(key, 0, sizeof(*key));
 
    /* _NEW_TEXTURE */
    for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
       const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i];
+      const struct gl_texture_object *texObj = texUnit->_Current;
+      const struct gl_tex_env_combine_state *comb = texUnit->_CurrentCombine;
       GLenum format;
 
       if (!texUnit->_ReallyEnabled || !texUnit->Enabled)
          continue;
 
-      format = texUnit->_Current->Image[0][texUnit->_Current->BaseLevel]->_BaseFormat;
+      format = texObj->Image[0][texObj->BaseLevel]->_BaseFormat;
 
       key->unit[i].enabled = 1;
       key->enabled_units |= (1<<i);
       key->nr_enabled_units = i+1;
       inputs_referenced |= FRAG_BIT_TEX(i);
 
-      key->unit[i].source_index = 
-	 translate_tex_src_bit(texUnit->_ReallyEnabled);		
-      key->unit[i].shadow = ((texUnit->_Current->CompareMode == GL_COMPARE_R_TO_TEXTURE) && 
+      key->unit[i].source_index =
+         translate_tex_src_bit(texUnit->_ReallyEnabled);
+
+      key->unit[i].shadow = ((texObj->CompareMode == GL_COMPARE_R_TO_TEXTURE) &&
                              ((format == GL_DEPTH_COMPONENT) || 
                               (format == GL_DEPTH_STENCIL_EXT)));
 
-      key->unit[i].NumArgsRGB = texUnit->_CurrentCombine->_NumArgsRGB;
-      key->unit[i].NumArgsA = texUnit->_CurrentCombine->_NumArgsA;
+      key->unit[i].NumArgsRGB = comb->_NumArgsRGB;
+      key->unit[i].NumArgsA = comb->_NumArgsA;
 
       key->unit[i].ModeRGB =
-	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeRGB);
+	 translate_mode(texUnit->EnvMode, comb->ModeRGB);
       key->unit[i].ModeA =
-	 translate_mode(texUnit->EnvMode, texUnit->_CurrentCombine->ModeA);
+	 translate_mode(texUnit->EnvMode, comb->ModeA);
 
-      key->unit[i].ScaleShiftRGB = texUnit->_CurrentCombine->ScaleShiftRGB;
-      key->unit[i].ScaleShiftA = texUnit->_CurrentCombine->ScaleShiftA;
+      key->unit[i].ScaleShiftRGB = comb->ScaleShiftRGB;
+      key->unit[i].ScaleShiftA = comb->ScaleShiftA;
 
       for (j = 0; j < MAX_COMBINER_TERMS; j++) {
-         key->unit[i].OptRGB[j].Operand =
-	    translate_operand(texUnit->_CurrentCombine->OperandRGB[j]);
-         key->unit[i].OptA[j].Operand =
-	    translate_operand(texUnit->_CurrentCombine->OperandA[j]);
-         key->unit[i].OptRGB[j].Source =
-	    translate_source(texUnit->_CurrentCombine->SourceRGB[j]);
-         key->unit[i].OptA[j].Source =
-	    translate_source(texUnit->_CurrentCombine->SourceA[j]);
+         key->unit[i].OptRGB[j].Operand = translate_operand(comb->OperandRGB[j]);
+         key->unit[i].OptA[j].Operand = translate_operand(comb->OperandA[j]);
+         key->unit[i].OptRGB[j].Source = translate_source(comb->SourceRGB[j]);
+         key->unit[i].OptA[j].Source = translate_source(comb->SourceA[j]);
       }
 
       if (key->unit[i].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
@@ -952,6 +957,11 @@ static struct ureg emit_combine_source( struct texenv_fragment_program *p,
    }
 }
 
+/**
+ * Check if the RGB and Alpha sources and operands match for the given
+ * texture unit's combinder state.  When the RGB and A sources and
+ * operands match, we can emit fewer instructions.
+ */
 static GLboolean args_match( const struct state_key *key, GLuint unit )
 {
    GLuint i, nr = key->unit[unit].NumArgsRGB;
@@ -1234,7 +1244,7 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 
       if (texTarget == TEXTURE_UNKNOWN_INDEX)
          program_error(p, "TexSrcBit");
-			  
+
       /* TODO: Use D0_MASK_XY where possible.
        */
       if (p->state->unit[unit].enabled) {
-- 
cgit v1.2.3


From b5ec0a68589c5d8f0d4be2556ea6d6b01fd4e37d Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 15:51:23 -0600
Subject: mesa: simplify translate_tex_src_bit()

---
 src/mesa/main/texenvprogram.c | 23 +++--------------------
 1 file changed, 3 insertions(+), 20 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 0898581485..ea0aa448e1 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -241,27 +241,13 @@ static GLuint translate_mode( GLenum envMode, GLenum mode )
 }
 
 
-#define TEXTURE_UNKNOWN_INDEX 7
-
 /**
  * Translate TEXTURE_x_BIT to TEXTURE_x_INDEX.
  */
 static GLuint translate_tex_src_bit( GLbitfield bit )
 {
-   /* make sure number of switch cases is correct */
-   assert(NUM_TEXTURE_TARGETS == 7);
-   switch (bit) {
-   case TEXTURE_1D_BIT:   return TEXTURE_1D_INDEX;
-   case TEXTURE_2D_BIT:   return TEXTURE_2D_INDEX;
-   case TEXTURE_RECT_BIT: return TEXTURE_RECT_INDEX;
-   case TEXTURE_3D_BIT:   return TEXTURE_3D_INDEX;
-   case TEXTURE_CUBE_BIT: return TEXTURE_CUBE_INDEX;
-   case TEXTURE_1D_ARRAY_BIT: return TEXTURE_1D_ARRAY_INDEX;
-   case TEXTURE_2D_ARRAY_BIT: return TEXTURE_2D_ARRAY_INDEX;
-   default:
-      assert(0);
-      return TEXTURE_UNKNOWN_INDEX;
-   }
+   ASSERT(bit);
+   return _mesa_ffs(bit) - 1;
 }
 
 
@@ -1230,7 +1216,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 static void load_texture( struct texenv_fragment_program *p, GLuint unit )
 {
    if (is_undef(p->src_texture[unit])) {
-      GLuint texTarget = p->state->unit[unit].source_index;
+      const GLuint texTarget = p->state->unit[unit].source_index;
       struct ureg texcoord;
       struct ureg tmp = get_tex_temp( p );
 
@@ -1242,9 +1228,6 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit )
          texcoord = p->texcoord_tex[unit];
       }
 
-      if (texTarget == TEXTURE_UNKNOWN_INDEX)
-         program_error(p, "TexSrcBit");
-
       /* TODO: Use D0_MASK_XY where possible.
        */
       if (p->state->unit[unit].enabled) {
-- 
cgit v1.2.3


From 9ed0315271c38f5fe9cdd65020a54cbb7f61ed2b Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 15:57:36 -0600
Subject: mesa: more comments, clean-ups

---
 src/mesa/main/texenvprogram.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index ea0aa448e1..54960a1fa1 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -92,21 +92,21 @@ struct state_key {
    GLuint enabled_units:8;
    GLuint separate_specular:1;
    GLuint fog_enabled:1;
-   GLuint fog_mode:2;
+   GLuint fog_mode:2;          /**< FOG_x */
    GLuint inputs_available:12;
 
    struct {
       GLuint enabled:1;
-      GLuint source_index:3;   /* one of TEXTURE_1D/2D/3D/CUBE/RECT_INDEX */
+      GLuint source_index:3;   /**< TEXTURE_x_INDEX */
       GLuint shadow:1;
       GLuint ScaleShiftRGB:2;
       GLuint ScaleShiftA:2;
 
-      GLuint NumArgsRGB:3;
+      GLuint NumArgsRGB:3;  /**< up to MAX_COMBINER_TERMS */
       GLuint ModeRGB:5;     /**< MODE_x */
       struct mode_opt OptRGB[MAX_COMBINER_TERMS];
 
-      GLuint NumArgsA:3;
+      GLuint NumArgsA:3;  /**< up to MAX_COMBINER_TERMS */
       GLuint ModeA:5;     /**< MODE_x */
       struct mode_opt OptA[MAX_COMBINER_TERMS];
    } unit[MAX_TEXTURE_UNITS];
@@ -705,7 +705,7 @@ emit_op(struct texenv_fragment_program *p,
 	struct ureg src1,
 	struct ureg src2 )
 {
-   GLuint nr = p->program->Base.NumInstructions++;
+   const GLuint nr = p->program->Base.NumInstructions++;
    struct prog_instruction *inst = &p->program->Base.Instructions[nr];
 
    assert(nr < MAX_INSTRUCTIONS);
@@ -950,15 +950,15 @@ static struct ureg emit_combine_source( struct texenv_fragment_program *p,
  */
 static GLboolean args_match( const struct state_key *key, GLuint unit )
 {
-   GLuint i, nr = key->unit[unit].NumArgsRGB;
+   GLuint i, numArgs = key->unit[unit].NumArgsRGB;
 
-   for (i = 0 ; i < nr ; i++) {
+   for (i = 0 ; i < numArgs ; i++) {
       if (key->unit[unit].OptA[i].Source != key->unit[unit].OptRGB[i].Source) 
 	 return GL_FALSE;
 
-      switch(key->unit[unit].OptA[i].Operand) {
+      switch (key->unit[unit].OptA[i].Operand) {
       case OPR_SRC_ALPHA: 
-	 switch(key->unit[unit].OptRGB[i].Operand) {
+	 switch (key->unit[unit].OptRGB[i].Operand) {
 	 case OPR_SRC_COLOR: 
 	 case OPR_SRC_ALPHA: 
 	    break;
@@ -967,7 +967,7 @@ static GLboolean args_match( const struct state_key *key, GLuint unit )
 	 }
 	 break;
       case OPR_ONE_MINUS_SRC_ALPHA: 
-	 switch(key->unit[unit].OptRGB[i].Operand) {
+	 switch (key->unit[unit].OptRGB[i].Operand) {
 	 case OPR_ONE_MINUS_SRC_COLOR: 
 	 case OPR_ONE_MINUS_SRC_ALPHA: 
 	    break;
-- 
cgit v1.2.3


From c7ba2b8622d8a924b3ac4c7d94a3c19d12e88e03 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 15:58:49 -0600
Subject: mesa: remove unused ureg:negateabs field

---
 src/mesa/main/texenvprogram.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 54960a1fa1..869d8d026b 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -458,9 +458,8 @@ struct ureg {
    GLuint idx:8;
    GLuint negatebase:1;
    GLuint abs:1;
-   GLuint negateabs:1;
    GLuint swz:12;
-   GLuint pad:5;
+   GLuint pad:6;
 };
 
 static const struct ureg undef = { 
@@ -469,7 +468,6 @@ static const struct ureg undef = {
    0,
    0,
    0,
-   0,
    0
 };
 
@@ -516,7 +514,6 @@ static struct ureg make_ureg(GLuint file, GLuint idx)
    reg.idx = idx;
    reg.negatebase = 0;
    reg.abs = 0;
-   reg.negateabs = 0;
    reg.swz = SWIZZLE_NOOP;
    reg.pad = 0;
    return reg;
-- 
cgit v1.2.3


From 1480bcacfec8f215c5827278a5f984c6bf68ba69 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 16:01:12 -0600
Subject: mesa: remove unused ureg::abs field

---
 src/mesa/main/texenvprogram.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 869d8d026b..8c588538b1 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -457,9 +457,8 @@ struct ureg {
    GLuint file:4;
    GLuint idx:8;
    GLuint negatebase:1;
-   GLuint abs:1;
    GLuint swz:12;
-   GLuint pad:6;
+   GLuint pad:7;
 };
 
 static const struct ureg undef = { 
@@ -467,7 +466,6 @@ static const struct ureg undef = {
    ~0,
    0,
    0,
-   0,
    0
 };
 
@@ -513,7 +511,6 @@ static struct ureg make_ureg(GLuint file, GLuint idx)
    reg.file = file;
    reg.idx = idx;
    reg.negatebase = 0;
-   reg.abs = 0;
    reg.swz = SWIZZLE_NOOP;
    reg.pad = 0;
    return reg;
@@ -679,7 +676,7 @@ static void emit_arg( struct prog_src_register *reg,
    reg->Index = ureg.idx;
    reg->Swizzle = ureg.swz;
    reg->Negate = ureg.negatebase ? NEGATE_XYZW : NEGATE_NONE;
-   reg->Abs = ureg.abs;
+   reg->Abs = GL_FALSE;
 }
 
 static void emit_dst( struct prog_dst_register *dst,
-- 
cgit v1.2.3


From 95136694ab28829f75c0a9f955a79836a0c433c7 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 16:03:28 -0600
Subject: mesa: remove unused texenv_fragment_program::ctx field

---
 src/mesa/main/texenvprogram.c | 2 --
 1 file changed, 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 8c588538b1..5c61c97354 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -474,7 +474,6 @@ static const struct ureg undef = {
  */
 struct texenv_fragment_program {
    struct gl_fragment_program *program;
-   GLcontext *ctx;
    struct state_key *state;
 
    GLbitfield alu_temps;	/**< Track texture indirections, see spec. */
@@ -1345,7 +1344,6 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    struct ureg cf, out;
 
    _mesa_memset(&p, 0, sizeof(p));
-   p.ctx = ctx;
    p.state = key;
    p.program = program;
 
-- 
cgit v1.2.3


From 25e5a6f279cec0863dcc341805aa7ca71189b626 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 16:05:07 -0600
Subject: mesa: updated #includes

---
 src/mesa/main/texenvprogram.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 5c61c97354..7efb10c5a0 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -27,8 +27,7 @@
  **************************************************************************/
 
 #include "glheader.h"
-#include "macros.h"
-#include "enums.h"
+#include "imports.h"
 #include "shader/program.h"
 #include "shader/prog_parameter.h"
 #include "shader/prog_cache.h"
-- 
cgit v1.2.3


From 84d6bed4d6844110af23a995e2fad0280a92d46c Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 16:10:57 -0600
Subject: mesa: change conditional to match the previous one

---
 src/mesa/main/texenvprogram.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 7efb10c5a0..b316c25809 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1401,7 +1401,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
       /* Second pass - emit combine instructions to build final color:
        */
       for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++)
-	 if (key->enabled_units & (1<<unit)) {
+	 if (key->unit[unit].enabled) {
 	    p.src_previous = emit_texenv( &p, unit );
             reserve_temp(&p, p.src_previous); /* don't re-use this temp reg */
 	    release_temps(ctx, &p);	/* release all temps */
-- 
cgit v1.2.3


From 457e427e32324884acb064b756e327dc131ff135 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Tue, 1 Sep 2009 16:22:02 -0600
Subject: mesa: more clean-ups

---
 src/mesa/main/texenvprogram.c | 55 ++++++++++++++++++++++++-------------------
 1 file changed, 31 insertions(+), 24 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index b316c25809..5913957969 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -944,7 +944,7 @@ static GLboolean args_match( const struct state_key *key, GLuint unit )
 {
    GLuint i, numArgs = key->unit[unit].NumArgsRGB;
 
-   for (i = 0 ; i < numArgs ; i++) {
+   for (i = 0; i < numArgs; i++) {
       if (key->unit[unit].OptA[i].Source != key->unit[unit].OptRGB[i].Source) 
 	 return GL_FALSE;
 
@@ -1278,7 +1278,7 @@ static GLboolean load_texenv_source( struct texenv_fragment_program *p,
  * Generate instructions for loading all texture source terms.
  */
 static GLboolean
-load_texunit_sources( struct texenv_fragment_program *p, int unit )
+load_texunit_sources( struct texenv_fragment_program *p, GLuint unit )
 {
    const struct state_key *key = p->state;
    GLuint i;
@@ -1298,7 +1298,7 @@ load_texunit_sources( struct texenv_fragment_program *p, int unit )
  * Generate instructions for loading bump map textures.
  */
 static GLboolean
-load_texunit_bumpmap( struct texenv_fragment_program *p, int unit )
+load_texunit_bumpmap( struct texenv_fragment_program *p, GLuint unit )
 {
    const struct state_key *key = p->state;
    GLuint bumpedUnitNr = key->unit[unit].OptRGB[1].Source - SRC_TEXTURE0;
@@ -1314,17 +1314,20 @@ load_texunit_bumpmap( struct texenv_fragment_program *p, int unit )
    texcDst = get_tex_temp( p );
    p->texcoord_tex[bumpedUnitNr] = texcDst;
 
-   /* apply rot matrix and add coords to be available in next phase */
-   /* dest = (Arg0.xxxx * rotMat0 + Arg1) + (Arg0.yyyy * rotMat1) */
-   /* note only 2 coords are affected the rest are left unchanged (mul by 0) */
+   /* Apply rot matrix and add coords to be available in next phase.
+    * dest = (Arg0.xxxx * rotMat0 + Arg1) + (Arg0.yyyy * rotMat1)
+    * note only 2 coords are affected the rest are left unchanged (mul by 0)
+    */
    emit_arith( p, OPCODE_MAD, texcDst, WRITEMASK_XYZW, 0,
                swizzle1(bumpMapRes, SWIZZLE_X), rotMat0, texcSrc );
    emit_arith( p, OPCODE_MAD, texcDst, WRITEMASK_XYZW, 0,
                swizzle1(bumpMapRes, SWIZZLE_Y), rotMat1, texcDst );
 
-   /* move 0,0,0,1 into bumpmap src if someone (crossbar) is foolish
-      enough to access this later, should optimize away */
-   emit_arith( p, OPCODE_MOV, bumpMapRes, WRITEMASK_XYZW, 0, constdudvcolor, undef, undef );
+   /* Move 0,0,0,1 into bumpmap src if someone (crossbar) is foolish
+    * enough to access this later, should optimize away.
+    */
+   emit_arith( p, OPCODE_MOV, bumpMapRes, WRITEMASK_XYZW, 0,
+               constdudvcolor, undef, undef );
 
    return GL_TRUE;
 }
@@ -1351,17 +1354,17 @@ create_new_program(GLcontext *ctx, struct state_key *key,
     */
    p.program->Base.Instructions = instBuffer;
    p.program->Base.Target = GL_FRAGMENT_PROGRAM_ARB;
-   p.program->Base.NumTexIndirections = 1;
+   p.program->Base.String = NULL;
+   p.program->Base.NumTexIndirections = 1; /* is this right? */
    p.program->Base.NumTexInstructions = 0;
    p.program->Base.NumAluInstructions = 0;
-   p.program->Base.String = NULL;
-   p.program->Base.NumInstructions =
-   p.program->Base.NumTemporaries =
-   p.program->Base.NumParameters =
-   p.program->Base.NumAttributes = p.program->Base.NumAddressRegs = 0;
+   p.program->Base.NumInstructions = 0;
+   p.program->Base.NumTemporaries = 0;
+   p.program->Base.NumParameters = 0;
+   p.program->Base.NumAttributes = 0;
+   p.program->Base.NumAddressRegs = 0;
    p.program->Base.Parameters = _mesa_new_parameter_list();
-
-   p.program->Base.InputsRead = 0;
+   p.program->Base.InputsRead = 0x0;
    p.program->Base.OutputsWritten = 1 << FRAG_RESULT_COLOR;
 
    for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++) {
@@ -1378,10 +1381,12 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    release_temps(ctx, &p);
 
    if (key->enabled_units) {
-       GLboolean needbumpstage = GL_FALSE;
+      GLboolean needbumpstage = GL_FALSE;
+
       /* Zeroth pass - bump map textures first */
-      for (unit = 0 ; unit < ctx->Const.MaxTextureUnits ; unit++)
-	 if (key->unit[unit].enabled && key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
+      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
+	 if (key->unit[unit].enabled &&
+             key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
 	    needbumpstage = GL_TRUE;
 	    load_texunit_bumpmap( &p, unit );
 	 }
@@ -1392,7 +1397,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
        * all referenced texture sources and emit texld instructions
        * for each:
        */
-      for (unit = 0 ; unit < ctx->Const.MaxTextureUnits ; unit++)
+      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
 	 if (key->unit[unit].enabled) {
 	    load_texunit_sources( &p, unit );
 	    p.last_tex_stage = unit;
@@ -1400,7 +1405,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
 
       /* Second pass - emit combine instructions to build final color:
        */
-      for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++)
+      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
 	 if (key->unit[unit].enabled) {
 	    p.src_previous = emit_texenv( &p, unit );
             reserve_temp(&p, p.src_previous); /* don't re-use this temp reg */
@@ -1434,9 +1439,11 @@ create_new_program(GLcontext *ctx, struct state_key *key,
        * a reduced value and not what is expected in FogOption
        */
       p.program->FogOption = ctx->Fog.Mode;
-      p.program->Base.InputsRead |= FRAG_BIT_FOGC; /* XXX new */
-   } else
+      p.program->Base.InputsRead |= FRAG_BIT_FOGC;
+   }
+   else {
       p.program->FogOption = GL_NONE;
+   }
 
    if (p.program->Base.NumTexIndirections > ctx->Const.FragmentProgram.MaxTexIndirections) 
       program_error(&p, "Exceeded max nr indirect texture lookups");
-- 
cgit v1.2.3


From a46e32787d98192d35bf2834011cfa5537e0e703 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Wed, 2 Sep 2009 05:11:25 -0700
Subject: mesa: Compact state key for TexEnv program cache

By rearranging the bitfields within the key we can reduce the size
of the key from 644 to 196 bytes, reducing the cost of both the
hashing and equality tests.
---
 src/mesa/main/texenvprogram.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 5913957969..b16b8dd36b 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -82,8 +82,8 @@ texenv_doing_secondary_color(GLcontext *ctx)
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
 struct mode_opt {
-   GLuint Source:4;  /**< SRC_x */
-   GLuint Operand:3; /**< OPR_x */
+   GLubyte Source:4;  /**< SRC_x */
+   GLubyte Operand:3; /**< OPR_x */
 };
 
 struct state_key {
@@ -103,10 +103,11 @@ struct state_key {
 
       GLuint NumArgsRGB:3;  /**< up to MAX_COMBINER_TERMS */
       GLuint ModeRGB:5;     /**< MODE_x */
-      struct mode_opt OptRGB[MAX_COMBINER_TERMS];
 
       GLuint NumArgsA:3;  /**< up to MAX_COMBINER_TERMS */
       GLuint ModeA:5;     /**< MODE_x */
+
+      struct mode_opt OptRGB[MAX_COMBINER_TERMS];
       struct mode_opt OptA[MAX_COMBINER_TERMS];
    } unit[MAX_TEXTURE_UNITS];
 };
@@ -1497,6 +1498,7 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
    struct gl_fragment_program *prog;
    struct state_key key;
 	
+   printf("SIZE OF KEY %d\n", sizeof(key));
    make_state_key(ctx, &key);
       
    prog = (struct gl_fragment_program *)
-- 
cgit v1.2.3


From 83e4470d9a3367883b3fe699629f1d8648459a89 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Wed, 2 Sep 2009 08:45:34 -0600
Subject: mesa: added #ifdef __GNUC__ around GLubyte bitfield usage
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

It would be nice if there were a #pragma or something to disable the warnings:

main/texenvprogram.c:87: warning: type of bit-field ‘Source’ is a GCC extension
main/texenvprogram.c:88: warning: type of bit-field ‘Operand’ is a GCC extension

but there doesn't appear to be a way to do that.
---
 src/mesa/main/texenvprogram.c | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index b16b8dd36b..770a2c26f7 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -82,8 +82,13 @@ texenv_doing_secondary_color(GLcontext *ctx)
 #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
 
 struct mode_opt {
+#ifdef __GNUC__
    GLubyte Source:4;  /**< SRC_x */
    GLubyte Operand:3; /**< OPR_x */
+#else
+   GLubyte Source;  /**< SRC_x */
+   GLubyte Operand; /**< OPR_x */
+#endif
 };
 
 struct state_key {
-- 
cgit v1.2.3


From 8cb651578de8dc4c956eaa8c0e5f1d442a3d4a92 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Wed, 2 Sep 2009 09:25:28 -0600
Subject: mesa: remove accidentally committed printf

---
 src/mesa/main/texenvprogram.c | 1 -
 1 file changed, 1 deletion(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 770a2c26f7..d5343d44db 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -1503,7 +1503,6 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
    struct gl_fragment_program *prog;
    struct state_key key;
 	
-   printf("SIZE OF KEY %d\n", sizeof(key));
    make_state_key(ctx, &key);
       
    prog = (struct gl_fragment_program *)
-- 
cgit v1.2.3


From d98b9f4edbbee1654b66c8df7b1deeb401b3dc5f Mon Sep 17 00:00:00 2001
From: Gary Wong <gtw@gnu.org>
Date: Wed, 2 Sep 2009 10:11:15 -0600
Subject: mesa: silence gcc bit-field warning

---
 src/mesa/main/texenvprogram.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index d5343d44db..b4923e904d 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -83,8 +83,8 @@ texenv_doing_secondary_color(GLcontext *ctx)
 
 struct mode_opt {
 #ifdef __GNUC__
-   GLubyte Source:4;  /**< SRC_x */
-   GLubyte Operand:3; /**< OPR_x */
+   __extension__ GLubyte Source:4;  /**< SRC_x */
+   __extension__ GLubyte Operand:3; /**< OPR_x */
 #else
    GLubyte Source;  /**< SRC_x */
    GLubyte Operand; /**< OPR_x */
-- 
cgit v1.2.3


From 4b08e7498230eac30eea1721f33994b30999acd4 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Wed, 2 Sep 2009 10:12:15 -0600
Subject: mesa: fix the non-GNU C bit-field case

---
 src/mesa/main/texenvprogram.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index b4923e904d..165d464e3b 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -86,8 +86,8 @@ struct mode_opt {
    __extension__ GLubyte Source:4;  /**< SRC_x */
    __extension__ GLubyte Operand:3; /**< OPR_x */
 #else
-   GLubyte Source;  /**< SRC_x */
-   GLubyte Operand; /**< OPR_x */
+   GLuint Source;  /**< SRC_x */
+   GLuint Operand; /**< OPR_x */
 #endif
 };
 
-- 
cgit v1.2.3


From 5e809216a8d2e7ad093a1d25f3f6495f29e57572 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Wed, 2 Sep 2009 10:38:46 -0600
Subject: Revert "mesa: fix the non-GNU C bit-field case"

This reverts commit 4b08e7498230eac30eea1721f33994b30999acd4.

Don't know what I was thinking there.
---
 src/mesa/main/texenvprogram.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 165d464e3b..b4923e904d 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -86,8 +86,8 @@ struct mode_opt {
    __extension__ GLubyte Source:4;  /**< SRC_x */
    __extension__ GLubyte Operand:3; /**< OPR_x */
 #else
-   GLuint Source;  /**< SRC_x */
-   GLuint Operand; /**< OPR_x */
+   GLubyte Source;  /**< SRC_x */
+   GLubyte Operand; /**< OPR_x */
 #endif
 };
 
-- 
cgit v1.2.3


From 7a7d5879e28607c151132a8b6a07cf4be36b5b3e Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Wed, 2 Sep 2009 15:39:13 -0600
Subject: mesa: compute tex env program key size based on nr_enabled_units

When no textures are enabled, the key size is 4 bytes.  If unit[0] is
enabled, the key size is 16 bytes.

Inspired by a patch from Chris Wilson.
---
 src/mesa/main/texenvprogram.c | 26 ++++++++++++++++++--------
 1 file changed, 18 insertions(+), 8 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index b4923e904d..870970a355 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -99,6 +99,7 @@ struct state_key {
    GLuint fog_mode:2;          /**< FOG_x */
    GLuint inputs_available:12;
 
+   /* NOTE: This array of structs must be last! (see "keySize" below) */
    struct {
       GLuint enabled:1;
       GLuint source_index:3;   /**< TEXTURE_x_INDEX */
@@ -368,11 +369,12 @@ static GLbitfield get_fp_input_mask( GLcontext *ctx )
  * Examine current texture environment state and generate a unique
  * key to identify it.
  */
-static void make_state_key( GLcontext *ctx,  struct state_key *key )
+static GLuint make_state_key( GLcontext *ctx,  struct state_key *key )
 {
    GLuint i, j;
    GLbitfield inputs_referenced = FRAG_BIT_COL0;
    const GLbitfield inputs_available = get_fp_input_mask( ctx );
+   GLuint keySize;
 
    memset(key, 0, sizeof(*key));
 
@@ -390,7 +392,7 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
 
       key->unit[i].enabled = 1;
       key->enabled_units |= (1<<i);
-      key->nr_enabled_units = i+1;
+      key->nr_enabled_units = i + 1;
       inputs_referenced |= FRAG_BIT_TEX(i);
 
       key->unit[i].source_index =
@@ -443,8 +445,15 @@ static void make_state_key( GLcontext *ctx,  struct state_key *key )
    }
 
    key->inputs_available = (inputs_available & inputs_referenced);
+
+   /* compute size of state key, ignoring unused texture units */
+   keySize = sizeof(*key) - sizeof(key->unit)
+      + key->nr_enabled_units * sizeof(key->unit[0]);
+
+   return keySize;
 }
 
+
 /**
  * Use uregs to represent registers internally, translate to Mesa's
  * expected formats on emit.  
@@ -1390,7 +1399,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
       GLboolean needbumpstage = GL_FALSE;
 
       /* Zeroth pass - bump map textures first */
-      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
+      for (unit = 0; unit < key->nr_enabled_units; unit++)
 	 if (key->unit[unit].enabled &&
              key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) {
 	    needbumpstage = GL_TRUE;
@@ -1403,7 +1412,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
        * all referenced texture sources and emit texld instructions
        * for each:
        */
-      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
+      for (unit = 0; unit < key->nr_enabled_units; unit++)
 	 if (key->unit[unit].enabled) {
 	    load_texunit_sources( &p, unit );
 	    p.last_tex_stage = unit;
@@ -1411,7 +1420,7 @@ create_new_program(GLcontext *ctx, struct state_key *key,
 
       /* Second pass - emit combine instructions to build final color:
        */
-      for (unit = 0; unit < ctx->Const.MaxTextureUnits; unit++)
+      for (unit = 0; unit < key->nr_enabled_units; unit++)
 	 if (key->unit[unit].enabled) {
 	    p.src_previous = emit_texenv( &p, unit );
             reserve_temp(&p, p.src_previous); /* don't re-use this temp reg */
@@ -1502,12 +1511,13 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
 {
    struct gl_fragment_program *prog;
    struct state_key key;
+   GLuint keySize;
 	
-   make_state_key(ctx, &key);
+   keySize = make_state_key(ctx, &key);
       
    prog = (struct gl_fragment_program *)
       _mesa_search_program_cache(ctx->FragmentProgram.Cache,
-                                 &key, sizeof(key));
+                                 &key, keySize);
 
    if (!prog) {
       prog = (struct gl_fragment_program *) 
@@ -1516,7 +1526,7 @@ _mesa_get_fixed_func_fragment_program(GLcontext *ctx)
       create_new_program(ctx, &key, prog);
 
       _mesa_program_cache_insert(ctx, ctx->FragmentProgram.Cache,
-                                 &key, sizeof(key), &prog->Base);
+                                 &key, keySize, &prog->Base);
    }
 
    return prog;
-- 
cgit v1.2.3


From e9ba9ffc9aa5dcc2de02dc3e58279ffda2318c79 Mon Sep 17 00:00:00 2001
From: Brian Paul <brianp@vmware.com>
Date: Thu, 10 Sep 2009 08:41:12 -0600
Subject: mesa: in texenvprogram code, only do saturation when really needed.

For some env modes (like modulate or replace) we don't have to clamp
because we know the results will be in [0,1].
---
 src/mesa/main/texenvprogram.c | 61 +++++++++++++++++++++++++++++++++++++------
 1 file changed, 53 insertions(+), 8 deletions(-)

(limited to 'src/mesa/main/texenvprogram.c')

diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 870970a355..2f3e47e69e 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -247,6 +247,40 @@ static GLuint translate_mode( GLenum envMode, GLenum mode )
 }
 
 
+/**
+ * Do we need to clamp the results of the given texture env/combine mode?
+ * If the inputs to the mode are in [0,1] we don't always have to clamp
+ * the results.
+ */
+static GLboolean
+need_saturate( GLuint mode )
+{
+   switch (mode) {
+   case MODE_REPLACE:
+   case MODE_MODULATE:
+   case MODE_INTERPOLATE:
+      return GL_FALSE;
+   case MODE_ADD:
+   case MODE_ADD_SIGNED:
+   case MODE_SUBTRACT:
+   case MODE_DOT3_RGB:
+   case MODE_DOT3_RGB_EXT:
+   case MODE_DOT3_RGBA:
+   case MODE_DOT3_RGBA_EXT:
+   case MODE_MODULATE_ADD_ATI:
+   case MODE_MODULATE_SIGNED_ADD_ATI:
+   case MODE_MODULATE_SUBTRACT_ATI:
+   case MODE_ADD_PRODUCTS:
+   case MODE_ADD_PRODUCTS_SIGNED:
+   case MODE_BUMP_ENVMAP_ATI:
+      return GL_TRUE;
+   default:
+      assert(0);
+   }
+}
+
+
+
 /**
  * Translate TEXTURE_x_BIT to TEXTURE_x_INDEX.
  */
@@ -1116,7 +1150,7 @@ static struct ureg
 emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 {
    const struct state_key *key = p->state;
-   GLboolean saturate;
+   GLboolean rgb_saturate, alpha_saturate;
    GLuint rgb_shift, alpha_shift;
    struct ureg out, dest;
 
@@ -1146,7 +1180,19 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    /* If we'll do rgb/alpha shifting don't saturate in emit_combine().
     * We don't want to clamp twice.
     */
-   saturate = !(rgb_shift || alpha_shift);
+   if (rgb_shift)
+      rgb_saturate = GL_FALSE;  /* saturate after rgb shift */
+   else if (need_saturate(key->unit[unit].ModeRGB))
+      rgb_saturate = GL_TRUE;
+   else
+      rgb_saturate = GL_FALSE;
+
+   if (alpha_shift)
+      alpha_saturate = GL_FALSE;  /* saturate after alpha shift */
+   else if (need_saturate(key->unit[unit].ModeA))
+      alpha_saturate = GL_TRUE;
+   else
+      alpha_saturate = GL_FALSE;
 
    /* If this is the very last calculation, emit direct to output reg:
     */
@@ -1162,7 +1208,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
     */
    if (key->unit[unit].ModeRGB == key->unit[unit].ModeA &&
        args_match(key, unit)) {
-      out = emit_combine( p, dest, WRITEMASK_XYZW, saturate,
+      out = emit_combine( p, dest, WRITEMASK_XYZW, rgb_saturate,
 			  unit,
 			  key->unit[unit].NumArgsRGB,
 			  key->unit[unit].ModeRGB,
@@ -1170,7 +1216,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    }
    else if (key->unit[unit].ModeRGB == MODE_DOT3_RGBA_EXT ||
 	    key->unit[unit].ModeRGB == MODE_DOT3_RGBA) {
-      out = emit_combine( p, dest, WRITEMASK_XYZW, saturate,
+      out = emit_combine( p, dest, WRITEMASK_XYZW, rgb_saturate,
 			  unit,
 			  key->unit[unit].NumArgsRGB,
 			  key->unit[unit].ModeRGB,
@@ -1180,12 +1226,12 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
       /* Need to do something to stop from re-emitting identical
        * argument calculations here:
        */
-      out = emit_combine( p, dest, WRITEMASK_XYZ, saturate,
+      out = emit_combine( p, dest, WRITEMASK_XYZ, rgb_saturate,
 			  unit,
 			  key->unit[unit].NumArgsRGB,
 			  key->unit[unit].ModeRGB,
 			  key->unit[unit].OptRGB);
-      out = emit_combine( p, dest, WRITEMASK_W, saturate,
+      out = emit_combine( p, dest, WRITEMASK_W, alpha_saturate,
 			  unit,
 			  key->unit[unit].NumArgsA,
 			  key->unit[unit].ModeA,
@@ -1196,8 +1242,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
     */
    if (alpha_shift || rgb_shift) {
       struct ureg shift;
-
-      saturate = GL_TRUE;  /* always saturate at this point */
+      GLboolean saturate = GL_TRUE;  /* always saturate at this point */
 
       if (rgb_shift == alpha_shift) {
 	 shift = register_scalar_const(p, (GLfloat)(1<<rgb_shift));
-- 
cgit v1.2.3