Merge remote-tracking branch 'origin/released'

Conflicts:
	mesalib/src/mesa/swrast/s_readpix.c

45 files changed, 4484 insertions, 5215 deletions

diff --git a/mesalib/src/gallium/auxiliary/util/u_draw.c b/mesalib/src/gallium/auxiliary/util/u_draw.c
index 4f2565edb..d16575b73 100644
--- a/mesalib/src/gallium/auxiliary/util/u_draw.c
+++ b/mesalib/src/gallium/auxiliary/util/u_draw.c
@@ -1,99 +1,115 @@
-/**************************************************************************

- *

- * Copyright 2011 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

- * "Software"), to deal in the Software without restriction, including

- * without limitation the rights to use, copy, modify, merge, publish,

- * distribute, sub license, and/or sell copies of the Software, and to

- * permit persons to whom the Software is furnished to do so, subject to

- * the following conditions:

- *

- * The above copyright notice and this permission notice (including the

- * next paragraph) shall be included in all copies or substantial portions

- * of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF

- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.

- * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR

- * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,

- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE

- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- *

- **************************************************************************/

-

-

-#include "util/u_debug.h"

-#include "util/u_math.h"

-#include "util/u_format.h"

-#include "util/u_draw.h"

-

-

-/**

- * Returns the largest legal index value for the current set of bound vertex

- * buffers.  Regardless of any other consideration, all vertex lookups need to

- * be clamped to 0..max_index to prevent an out-of-bound access.

- */

-unsigned

-util_draw_max_index(

-      const struct pipe_vertex_buffer *vertex_buffers,

-      unsigned nr_vertex_buffers,

-      const struct pipe_vertex_element *vertex_elements,

-      unsigned nr_vertex_elements,

-      const struct pipe_draw_info *info)

-{

-   unsigned max_index;

-   unsigned i;

-

-   max_index = ~0;

-   for (i = 0; i < nr_vertex_elements; i++) {

-      const struct pipe_vertex_element *element =

-         &vertex_elements[i];

-      const struct pipe_vertex_buffer *buffer =

-         &vertex_buffers[element->vertex_buffer_index];

-      unsigned buffer_size;

-      const struct util_format_description *format_desc;

-      unsigned format_size;

-

-      assert(buffer->buffer->height0 == 1);

-      assert(buffer->buffer->depth0 == 1);

-      buffer_size = buffer->buffer->width0;

-

-      format_desc = util_format_description(element->src_format);

-      assert(format_desc->block.width == 1);

-      assert(format_desc->block.height == 1);

-      assert(format_desc->block.bits % 8 == 0);

-      format_size = format_desc->block.bits/8;

-

-      assert(buffer_size - buffer->buffer_offset <= buffer_size);

-      buffer_size -= buffer->buffer_offset;

-

-      assert(buffer_size - element->src_offset <= buffer_size);

-      buffer_size -= element->src_offset;

-

-      assert(buffer_size - format_size <= buffer_size);

-      buffer_size -= format_size;

-

-      if (buffer->stride != 0) {

-         unsigned buffer_max_index;

-

-         buffer_max_index = buffer_size / buffer->stride;

-

-         if (element->instance_divisor == 0) {

-            /* Per-vertex data */

-            max_index = MIN2(max_index, buffer_max_index);

-         }

-         else {

-            /* Per-instance data. Simply make sure the state tracker didn't

-             * request more instances than those that fit in the buffer */

-            assert((info->start_instance + info->instance_count)/element->instance_divisor

-                   <= (buffer_max_index + 1));

-         }

-      }

-   }

-

-   return max_index;

-}

+/**************************************************************************
+ *
+ * Copyright 2011 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
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+
+#include "util/u_debug.h"
+#include "util/u_math.h"
+#include "util/u_format.h"
+#include "util/u_draw.h"
+
+
+/**
+ * Returns the largest legal index value plus one for the current set
+ * of bound vertex buffers.  Regardless of any other consideration,
+ * all vertex lookups need to be clamped to 0..max_index-1 to prevent
+ * an out-of-bound access.
+ *
+ * Note that if zero is returned it means that one or more buffers is
+ * too small to contain any valid vertex data.
+ */
+unsigned
+util_draw_max_index(
+      const struct pipe_vertex_buffer *vertex_buffers,
+      unsigned nr_vertex_buffers,
+      const struct pipe_vertex_element *vertex_elements,
+      unsigned nr_vertex_elements,
+      const struct pipe_draw_info *info)
+{
+   unsigned max_index;
+   unsigned i;
+
+   max_index = ~0U - 1;
+   for (i = 0; i < nr_vertex_elements; i++) {
+      const struct pipe_vertex_element *element =
+         &vertex_elements[i];
+      const struct pipe_vertex_buffer *buffer =
+         &vertex_buffers[element->vertex_buffer_index];
+      unsigned buffer_size;
+      const struct util_format_description *format_desc;
+      unsigned format_size;
+
+      assert(buffer->buffer->height0 == 1);
+      assert(buffer->buffer->depth0 == 1);
+      buffer_size = buffer->buffer->width0;
+
+      format_desc = util_format_description(element->src_format);
+      assert(format_desc->block.width == 1);
+      assert(format_desc->block.height == 1);
+      assert(format_desc->block.bits % 8 == 0);
+      format_size = format_desc->block.bits/8;
+
+      if (buffer->buffer_offset >= buffer_size) {
+         /* buffer is too small */
+         return 0;
+      }
+
+      buffer_size -= buffer->buffer_offset;
+
+      if (element->src_offset >= buffer_size) {
+         /* buffer is too small */
+         return 0;
+      }
+
+      buffer_size -= element->src_offset;
+
+      if (format_size > buffer_size) {
+         /* buffer is too small */
+         return 0;
+      }
+
+      buffer_size -= format_size;
+
+      if (buffer->stride != 0) {
+         unsigned buffer_max_index;
+
+         buffer_max_index = buffer_size / buffer->stride;
+
+         if (element->instance_divisor == 0) {
+            /* Per-vertex data */
+            max_index = MIN2(max_index, buffer_max_index);
+         }
+         else {
+            /* Per-instance data. Simply make sure the state tracker didn't
+             * request more instances than those that fit in the buffer */
+            assert((info->start_instance + info->instance_count)/element->instance_divisor
+                   <= (buffer_max_index + 1));
+         }
+      }
+   }
+
+   return max_index + 1;
+}
diff --git a/mesalib/src/gallium/auxiliary/util/u_prim.h b/mesalib/src/gallium/auxiliary/util/u_prim.h
index ca7c67d7c..070df643d 100644
--- a/mesalib/src/gallium/auxiliary/util/u_prim.h
+++ b/mesalib/src/gallium/auxiliary/util/u_prim.h
@@ -78,7 +78,7 @@ static INLINE boolean u_validate_pipe_prim( unsigned pipe_prim, unsigned nr )
 static INLINE boolean u_trim_pipe_prim( unsigned pipe_prim, unsigned *nr )
 {
    boolean ok = TRUE;
-   const static int values[][2] = {
+   const static unsigned values[][2] = {
       { 1, 0 }, /* PIPE_PRIM_POINTS */
       { 2, 2 }, /* PIPE_PRIM_LINES */
       { 2, 0 }, /* PIPE_PRIM_LINE_LOOP */
diff --git a/mesalib/src/glsl/ast_function.cpp b/mesalib/src/glsl/ast_function.cpp
index fc0d7497d..126b610d1 100644
--- a/mesalib/src/glsl/ast_function.cpp
+++ b/mesalib/src/glsl/ast_function.cpp
@@ -93,214 +93,256 @@ prototype_string(const glsl_type *return_type, const char *name,
    return str;
 }
 
-
 static ir_rvalue *
-match_function_by_name(exec_list *instructions, const char *name,
-		       YYLTYPE *loc, exec_list *actual_parameters,
-		       struct _mesa_glsl_parse_state *state)
+generate_call(exec_list *instructions, ir_function_signature *sig,
+	      YYLTYPE *loc, exec_list *actual_parameters,
+	      struct _mesa_glsl_parse_state *state)
 {
    void *ctx = state;
-   ir_function *f = state->symbols->get_function(name);
-   ir_function_signature *sig;
-
-   sig = f ? f->matching_signature(actual_parameters) : NULL;
+   exec_list post_call_conversions;
 
-   /* FINISHME: This doesn't handle the case where shader X contains a
-    * FINISHME: matching signature but shader X + N contains an _exact_
-    * FINISHME: matching signature.
+   /* Verify that 'out' and 'inout' actual parameters are lvalues.  This
+    * isn't done in ir_function::matching_signature because that function
+    * cannot generate the necessary diagnostics.
+    *
+    * Also, validate that 'const_in' formal parameters (an extension of our
+    * IR) correspond to ir_constant actual parameters.
+    *
+    * Also, perform implicit conversion of arguments.  Note: to implicitly
+    * convert out parameters, we need to place them in a temporary
+    * variable, and do the conversion after the call takes place.  Since we
+    * haven't emitted the call yet, we'll place the post-call conversions
+    * in a temporary exec_list, and emit them later.
     */
-   if (sig == NULL
-       && (f == NULL || state->es_shader || !f->has_user_signature())
-       && state->symbols->get_type(name) == NULL
-       && (state->language_version == 110
-	   || state->symbols->get_variable(name) == NULL)) {
-      /* The current shader doesn't contain a matching function or signature.
-       * Before giving up, look for the prototype in the built-in functions.
-       */
-      _mesa_glsl_initialize_functions(state);
-      for (unsigned i = 0; i < state->num_builtins_to_link; i++) {
-	 ir_function *builtin;
-	 builtin = state->builtins_to_link[i]->symbols->get_function(name);
-	 sig = builtin ? builtin->matching_signature(actual_parameters) : NULL;
-	 if (sig != NULL) {
-	    if (f == NULL) {
-	       f = new(ctx) ir_function(name);
-	       state->symbols->add_global_function(f);
-	       emit_function(state, f);
-	    }
+   exec_list_iterator actual_iter = actual_parameters->iterator();
+   exec_list_iterator formal_iter = sig->parameters.iterator();
+
+   while (actual_iter.has_next()) {
+      ir_rvalue *actual = (ir_rvalue *) actual_iter.get();
+      ir_variable *formal = (ir_variable *) formal_iter.get();
 
-	    f->add_signature(sig->clone_prototype(f, NULL));
+      assert(actual != NULL);
+      assert(formal != NULL);
+
+      if (formal->mode == ir_var_const_in && !actual->as_constant()) {
+	 _mesa_glsl_error(loc, state,
+			  "parameter `%s' must be a constant expression",
+			  formal->name);
+	 return ir_call::get_error_instruction(ctx);
+      }
+
+      if ((formal->mode == ir_var_out)
+	  || (formal->mode == ir_var_inout)) {
+	 const char *mode = NULL;
+	 switch (formal->mode) {
+	 case ir_var_out:   mode = "out";   break;
+	 case ir_var_inout: mode = "inout"; break;
+	 default:           assert(false);  break;
+	 }
+	 /* FIXME: 'loc' is incorrect (as of 2011-01-21). It is always
+	  * FIXME: 0:0(0).
+	  */
+	 if (actual->variable_referenced()
+	     && actual->variable_referenced()->read_only) {
+	    _mesa_glsl_error(loc, state,
+			     "function parameter '%s %s' references the "
+			     "read-only variable '%s'",
+			     mode, formal->name,
+			     actual->variable_referenced()->name);
+
+	 } else if (!actual->is_lvalue()) {
+	    _mesa_glsl_error(loc, state,
+			     "function parameter '%s %s' is not an lvalue",
+			     mode, formal->name);
+	 }
+      }
+
+      if (formal->type->is_numeric() || formal->type->is_boolean()) {
+	 switch (formal->mode) {
+	 case ir_var_const_in:
+	 case ir_var_in: {
+	    ir_rvalue *converted
+	       = convert_component(actual, formal->type);
+	    actual->replace_with(converted);
+	    break;
+	 }
+	 case ir_var_out:
+	    if (actual->type != formal->type) {
+	       /* To convert an out parameter, we need to create a
+		* temporary variable to hold the value before conversion,
+		* and then perform the conversion after the function call
+		* returns.
+		*
+		* This has the effect of transforming code like this:
+		*
+		*   void f(out int x);
+		*   float value;
+		*   f(value);
+		*
+		* Into IR that's equivalent to this:
+		*
+		*   void f(out int x);
+		*   float value;
+		*   int out_parameter_conversion;
+		*   f(out_parameter_conversion);
+		*   value = float(out_parameter_conversion);
+		*/
+	       ir_variable *tmp =
+		  new(ctx) ir_variable(formal->type,
+				       "out_parameter_conversion",
+				       ir_var_temporary);
+	       instructions->push_tail(tmp);
+	       ir_dereference_variable *deref_tmp_1
+		  = new(ctx) ir_dereference_variable(tmp);
+	       ir_dereference_variable *deref_tmp_2
+		  = new(ctx) ir_dereference_variable(tmp);
+	       ir_rvalue *converted_tmp
+		  = convert_component(deref_tmp_1, actual->type);
+	       ir_assignment *assignment
+		  = new(ctx) ir_assignment(actual, converted_tmp);
+	       post_call_conversions.push_tail(assignment);
+	       actual->replace_with(deref_tmp_2);
+	    }
+	    break;
+	 case ir_var_inout:
+	    /* Inout parameters should never require conversion, since that
+	     * would require an implicit conversion to exist both to and
+	     * from the formal parameter type, and there are no
+	     * bidirectional implicit conversions.
+	     */
+	    assert (actual->type == formal->type);
+	    break;
+	 default:
+	    assert (!"Illegal formal parameter mode");
 	    break;
 	 }
       }
-   }
 
-   exec_list post_call_conversions;
+      actual_iter.next();
+      formal_iter.next();
+   }
 
-   if (sig != NULL) {
-      /* Verify that 'out' and 'inout' actual parameters are lvalues.  This
-       * isn't done in ir_function::matching_signature because that function
-       * cannot generate the necessary diagnostics.
-       *
-       * Also, validate that 'const_in' formal parameters (an extension of our
-       * IR) correspond to ir_constant actual parameters.
+   /* Always insert the call in the instruction stream, and return a deref
+    * of its return val if it returns a value, since we don't know if
+    * the rvalue is going to be assigned to anything or not.
+    *
+    * Also insert any out parameter conversions after the call.
+    */
+   ir_call *call = new(ctx) ir_call(sig, actual_parameters);
+   ir_dereference_variable *deref;
+   if (!sig->return_type->is_void()) {
+      /* If the function call is a constant expression, don't
+       * generate the instructions to call it; just generate an
+       * ir_constant representing the constant value.
        *
-       * Also, perform implicit conversion of arguments.  Note: to implicitly
-       * convert out parameters, we need to place them in a temporary
-       * variable, and do the conversion after the call takes place.  Since we
-       * haven't emitted the call yet, we'll place the post-call conversions
-       * in a temporary exec_list, and emit them later.
+       * Function calls can only be constant expressions starting
+       * in GLSL 1.20.
        */
-      exec_list_iterator actual_iter = actual_parameters->iterator();
-      exec_list_iterator formal_iter = sig->parameters.iterator();
+      if (state->language_version >= 120) {
+	 ir_constant *const_val = call->constant_expression_value();
+	 if (const_val) {
+	    return const_val;
+	 }
+      }
 
-      while (actual_iter.has_next()) {
-	 ir_rvalue *actual = (ir_rvalue *) actual_iter.get();
-	 ir_variable *formal = (ir_variable *) formal_iter.get();
+      ir_variable *var;
 
-	 assert(actual != NULL);
-	 assert(formal != NULL);
+      var = new(ctx) ir_variable(sig->return_type,
+				 ralloc_asprintf(ctx, "%s_retval",
+						 sig->function_name()),
+				 ir_var_temporary);
+      instructions->push_tail(var);
 
-	 if (formal->mode == ir_var_const_in && !actual->as_constant()) {
-	    _mesa_glsl_error(loc, state,
-			     "parameter `%s' must be a constant expression",
-			     formal->name);
-	    return ir_call::get_error_instruction(ctx);
-	 }
+      deref = new(ctx) ir_dereference_variable(var);
+      ir_assignment *assign = new(ctx) ir_assignment(deref, call, NULL);
+      instructions->push_tail(assign);
 
-	 if ((formal->mode == ir_var_out)
-	     || (formal->mode == ir_var_inout)) {
-	    const char *mode = NULL;
-	    switch (formal->mode) {
-	    case ir_var_out:   mode = "out";   break;
-	    case ir_var_inout: mode = "inout"; break;
-	    default:           assert(false);  break;
-	    }
-            /* FIXME: 'loc' is incorrect (as of 2011-01-21). It is always
-             * FIXME: 0:0(0).
-             */
-	    if (actual->variable_referenced()
-	        && actual->variable_referenced()->read_only) {
-	       _mesa_glsl_error(loc, state,
-	                        "function parameter '%s %s' references the "
-	                        "read-only variable '%s'",
-	                        mode, formal->name,
-	                        actual->variable_referenced()->name);
-
-	    } else if (!actual->is_lvalue()) {
-               _mesa_glsl_error(loc, state,
-                                "function parameter '%s %s' is not an lvalue",
-                                mode, formal->name);
-	    }
-	 }
+      deref = new(ctx) ir_dereference_variable(var);
+   } else {
+      instructions->push_tail(call);
+      deref = NULL;
+   }
+   instructions->append_list(&post_call_conversions);
+   return deref;
+}
 
-	 if (formal->type->is_numeric() || formal->type->is_boolean()) {
-            switch (formal->mode) {
-            case ir_var_const_in:
-            case ir_var_in: {
-               ir_rvalue *converted
-                  = convert_component(actual, formal->type);
-               actual->replace_with(converted);
-               break;
-            }
-            case ir_var_out:
-               if (actual->type != formal->type) {
-                  /* To convert an out parameter, we need to create a
-                   * temporary variable to hold the value before conversion,
-                   * and then perform the conversion after the function call
-                   * returns.
-                   *
-                   * This has the effect of transforming code like this:
-                   *
-                   *   void f(out int x);
-                   *   float value;
-                   *   f(value);
-                   *
-                   * Into IR that's equivalent to this:
-                   *
-                   *   void f(out int x);
-                   *   float value;
-                   *   int out_parameter_conversion;
-                   *   f(out_parameter_conversion);
-                   *   value = float(out_parameter_conversion);
-                   */
-                  ir_variable *tmp =
-                     new(ctx) ir_variable(formal->type,
-                                          "out_parameter_conversion",
-                                          ir_var_temporary);
-                  instructions->push_tail(tmp);
-                  ir_dereference_variable *deref_tmp_1
-                     = new(ctx) ir_dereference_variable(tmp);
-                  ir_dereference_variable *deref_tmp_2
-                     = new(ctx) ir_dereference_variable(tmp);
-                  ir_rvalue *converted_tmp
-                     = convert_component(deref_tmp_1, actual->type);
-                  ir_assignment *assignment
-                     = new(ctx) ir_assignment(actual, converted_tmp);
-                  post_call_conversions.push_tail(assignment);
-                  actual->replace_with(deref_tmp_2);
-               }
-               break;
-            case ir_var_inout:
-               /* Inout parameters should never require conversion, since that
-                * would require an implicit conversion to exist both to and
-                * from the formal parameter type, and there are no
-                * bidirectional implicit conversions.
-                */
-               assert (actual->type == formal->type);
-               break;
-            default:
-               assert (!"Illegal formal parameter mode");
-               break;
-            }
-	 }
+static ir_rvalue *
+match_function_by_name(exec_list *instructions, const char *name,
+		       YYLTYPE *loc, exec_list *actual_parameters,
+		       struct _mesa_glsl_parse_state *state)
+{
+   void *ctx = state;
+   ir_function *f = state->symbols->get_function(name);
+   ir_function_signature *local_sig = NULL;
+   ir_function_signature *sig = NULL;
+
+   /* Is the function hidden by a record type constructor? */
+   if (state->symbols->get_type(name))
+      goto done; /* no match */
+
+   /* Is the function hidden by a variable (impossible in 1.10)? */
+   if (state->language_version != 110 && state->symbols->get_variable(name))
+      goto done; /* no match */
+
+   if (f != NULL) {
+      /* Look for a match in the local shader.  If exact, we're done. */
+      bool is_exact = false;
+      sig = local_sig = f->matching_signature(actual_parameters, &is_exact);
+      if (is_exact)
+	 goto done;
+
+      if (!state->es_shader && f->has_user_signature()) {
+	 /* In desktop GL, the presence of a user-defined signature hides any
+	  * built-in signatures, so we must ignore them.  In contrast, in ES2
+	  * user-defined signatures add new overloads, so we must proceed.
+	  */
+	 goto done;
+      }
+   }
 
-	 actual_iter.next();
-	 formal_iter.next();
+   /* Local shader has no exact candidates; check the built-ins. */
+   _mesa_glsl_initialize_functions(state);
+   for (unsigned i = 0; i < state->num_builtins_to_link; i++) {
+      ir_function *builtin =
+	 state->builtins_to_link[i]->symbols->get_function(name);
+      if (builtin == NULL)
+	 continue;
+
+      bool is_exact = false;
+      ir_function_signature *builtin_sig =
+	 builtin->matching_signature(actual_parameters, &is_exact);
+
+      if (builtin_sig == NULL)
+	 continue;
+
+      /* If the built-in signature is exact, we can stop. */
+      if (is_exact) {
+	 sig = builtin_sig;
+	 goto done;
       }
 
-      /* Always insert the call in the instruction stream, and return a deref
-       * of its return val if it returns a value, since we don't know if
-       * the rvalue is going to be assigned to anything or not.
-       *
-       * Also insert any out parameter conversions after the call.
-       */
-      ir_call *call = new(ctx) ir_call(sig, actual_parameters);
-      ir_dereference_variable *deref;
-      if (!sig->return_type->is_void()) {
-         /* If the function call is a constant expression, don't
-          * generate the instructions to call it; just generate an
-          * ir_constant representing the constant value.
-          *
-          * Function calls can only be constant expressions starting
-          * in GLSL 1.20.
-          */
-         if (state->language_version >= 120) {
-            ir_constant *const_val = call->constant_expression_value();
-            if (const_val) {
-               return const_val;
-            }
-         }
-
-	 ir_variable *var;
-
-	 var = new(ctx) ir_variable(sig->return_type,
-				    ralloc_asprintf(ctx, "%s_retval",
-						    sig->function_name()),
-				    ir_var_temporary);
-	 instructions->push_tail(var);
-
-	 deref = new(ctx) ir_dereference_variable(var);
-	 ir_assignment *assign = new(ctx) ir_assignment(deref, call, NULL);
-	 instructions->push_tail(assign);
-
-	 deref = new(ctx) ir_dereference_variable(var);
-      } else {
-	 instructions->push_tail(call);
-	 deref = NULL;
+      if (sig == NULL) {
+	 /* We found an inexact match, which is better than nothing.  However,
+	  * we should keep searching for an exact match.
+	  */
+	 sig = builtin_sig;
+      }
+   }
+
+done:
+   if (sig != NULL) {
+      /* If the match is from a linked built-in shader, import the prototype. */
+      if (sig != local_sig) {
+	 if (f == NULL) {
+	    f = new(ctx) ir_function(name);
+	    state->symbols->add_global_function(f);
+	    emit_function(state, f);
+	 }
+	 f->add_signature(sig->clone_prototype(f, NULL));
       }
-      instructions->append_list(&post_call_conversions);
-      return deref;
+
+      /* Finally, generate a call instruction. */
+      return generate_call(instructions, sig, loc, actual_parameters, state);
    } else {
       char *str = prototype_string(NULL, name, actual_parameters);
 
diff --git a/mesalib/src/glsl/glsl_parser.yy b/mesalib/src/glsl/glsl_parser.yy
index f3e873800..836390453 100644
--- a/mesalib/src/glsl/glsl_parser.yy
+++ b/mesalib/src/glsl/glsl_parser.yy
@@ -1706,7 +1706,7 @@ case_statement:
 	   ast_case_statement *stmts = new(state) ast_case_statement($1);
 
 	   stmts->stmts.push_tail(& $2->link);
-	   $$ = stmts
+	   $$ = stmts;
 	}
 	| case_statement statement
 	{
diff --git a/mesalib/src/glsl/ir.h b/mesalib/src/glsl/ir.h
index 5878c051b..1faae3c72 100644
--- a/mesalib/src/glsl/ir.h
+++ b/mesalib/src/glsl/ir.h
@@ -566,6 +566,13 @@ public:
 
    /**
     * Find a signature that matches a set of actual parameters, taking implicit
+    * conversions into account.  Also flags whether the match was exact.
+    */
+   ir_function_signature *matching_signature(const exec_list *actual_param,
+					     bool *match_is_exact);
+
+   /**
+    * Find a signature that matches a set of actual parameters, taking implicit
     * conversions into account.
     */
    ir_function_signature *matching_signature(const exec_list *actual_param);
diff --git a/mesalib/src/glsl/ir_function.cpp b/mesalib/src/glsl/ir_function.cpp
index 51d32b46f..b34a50081 100644
--- a/mesalib/src/glsl/ir_function.cpp
+++ b/mesalib/src/glsl/ir_function.cpp
@@ -118,6 +118,14 @@ parameter_lists_match(const exec_list *list_a, const exec_list *list_b)
 ir_function_signature *
 ir_function::matching_signature(const exec_list *actual_parameters)
 {
+   bool is_exact;
+   return matching_signature(actual_parameters, &is_exact);
+}
+
+ir_function_signature *
+ir_function::matching_signature(const exec_list *actual_parameters,
+			        bool *is_exact)
+{
    ir_function_signature *match = NULL;
    bool multiple_inexact_matches = false;
 
@@ -137,6 +145,7 @@ ir_function::matching_signature(const exec_list *actual_parameters)
 
       switch (parameter_lists_match(& sig->parameters, actual_parameters)) {
       case PARAMETER_LIST_EXACT_MATCH:
+	 *is_exact = true;
 	 return sig;
       case PARAMETER_LIST_INEXACT_MATCH:
 	 if (match == NULL)
@@ -159,6 +168,8 @@ ir_function::matching_signature(const exec_list *actual_parameters)
     * FINISHME: a "no matching signature" error; it should report that the
     * FINISHME: call is ambiguous.  But reporting errors from here is hard.
     */
+   *is_exact = false;
+
    if (multiple_inexact_matches)
       return NULL;
 
diff --git a/mesalib/src/glsl/link_uniforms.cpp b/mesalib/src/glsl/link_uniforms.cpp
index 1f860440c..ef8e436a5 100644
--- a/mesalib/src/glsl/link_uniforms.cpp
+++ b/mesalib/src/glsl/link_uniforms.cpp
@@ -113,11 +113,18 @@ uniform_field_visitor::recursion(const glsl_type *t, char **name,
 class count_uniform_size : public uniform_field_visitor {
 public:
    count_uniform_size(struct string_to_uint_map *map)
-      : num_active_uniforms(0), num_values(0), map(map)
+      : num_active_uniforms(0), num_values(0), num_shader_samplers(0),
+	num_shader_uniforms(0), map(map)
    {
       /* empty */
    }
 
+   void start_shader()
+   {
+      this->num_shader_samplers = 0;
+      this->num_shader_uniforms = 0;
+   }
+
    /**
     * Total number of active uniforms counted
     */
@@ -128,12 +135,39 @@ public:
     */
    unsigned num_values;
 
+   /**
+    * Number of samplers used
+    */
+   unsigned num_shader_samplers;
+
+   /**
+    * Number of uniforms used in the current shader
+    */
+   unsigned num_shader_uniforms;
+
 private:
    virtual void visit_field(const glsl_type *type, const char *name)
    {
       assert(!type->is_record());
       assert(!(type->is_array() && type->fields.array->is_record()));
 
+      /* Count the number of samplers regardless of whether the uniform is
+       * already in the hash table.  The hash table prevents adding the same
+       * uniform for multiple shader targets, but in this case we want to
+       * count it for each shader target.
+       */
+      const unsigned values = values_for_type(type);
+      if (type->contains_sampler()) {
+	 this->num_shader_samplers +=
+	    type->is_array() ? type->array_size() : 1;
+      } else {
+	 /* Accumulate the total number of uniform slots used by this shader.
+	  * Note that samplers do not count against this limit because they
+	  * don't use any storage on current hardware.
+	  */
+	 this->num_shader_uniforms += values;
+      }
+
       /* If the uniform is already in the map, there's nothing more to do.
        */
       unsigned id;
@@ -147,7 +181,7 @@ private:
        * uniforms.
        */
       this->num_active_uniforms++;
-      this->num_values += values_for_type(type);
+      this->num_values += values;
    }
 
    struct string_to_uint_map *map;
@@ -267,6 +301,10 @@ link_assign_uniform_locations(struct gl_shader_program *prog)
       if (prog->_LinkedShaders[i] == NULL)
 	 continue;
 
+      /* Reset various per-shader target counts.
+       */
+      uniform_size.start_shader();
+
       foreach_list(node, prog->_LinkedShaders[i]->ir) {
 	 ir_variable *const var = ((ir_instruction *) node)->as_variable();
 
@@ -280,6 +318,10 @@ link_assign_uniform_locations(struct gl_shader_program *prog)
 
 	 uniform_size.process(var);
       }
+
+      prog->_LinkedShaders[i]->num_samplers = uniform_size.num_shader_samplers;
+      prog->_LinkedShaders[i]->num_uniform_components =
+	 uniform_size.num_shader_uniforms * 4;
    }
 
    const unsigned num_user_uniforms = uniform_size.num_active_uniforms;
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index 351680d43..0ec773d6c 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -1875,6 +1875,47 @@ store_tfeedback_info(struct gl_context *ctx, struct gl_shader_program *prog,
    return true;
 }
 
+/**
+ * Validate the resources used by a program versus the implementation limits
+ */
+static bool
+check_resources(struct gl_context *ctx, struct gl_shader_program *prog)
+{
+   static const char *const shader_names[MESA_SHADER_TYPES] = {
+      "vertex", "fragment", "geometry"
+   };
+
+   const unsigned max_samplers[MESA_SHADER_TYPES] = {
+      ctx->Const.MaxVertexTextureImageUnits,
+      ctx->Const.MaxTextureImageUnits,
+      ctx->Const.MaxGeometryTextureImageUnits
+   };
+
+   const unsigned max_uniform_components[MESA_SHADER_TYPES] = {
+      ctx->Const.VertexProgram.MaxUniformComponents,
+      ctx->Const.FragmentProgram.MaxUniformComponents,
+      0          /* FINISHME: Geometry shaders. */
+   };
+
+   for (unsigned i = 0; i < MESA_SHADER_TYPES; i++) {
+      struct gl_shader *sh = prog->_LinkedShaders[i];
+
+      if (sh == NULL)
+	 continue;
+
+      if (sh->num_samplers > max_samplers[i]) {
+	 linker_error(prog, "Too many %s shader texture samplers",
+		      shader_names[i]);
+      }
+
+      if (sh->num_uniform_components > max_uniform_components[i]) {
+         linker_error(prog, "Too many %s shader uniform components",
+		      shader_names[i]);
+      }
+   }
+
+   return prog->LinkStatus;
+}
 
 void
 link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
@@ -2137,6 +2178,9 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
    update_array_sizes(prog);
    link_assign_uniform_locations(prog);
 
+   if (!check_resources(ctx, prog))
+      goto done;
+
    /* OpenGL ES requires that a vertex shader and a fragment shader both be
     * present in a linked program.  By checking for use of shading language
     * version 1.00, we also catch the GL_ARB_ES2_compatibility case.
diff --git a/mesalib/src/mesa/SConscript b/mesalib/src/mesa/SConscript
index 3e42bfd8b..807608636 100644
--- a/mesalib/src/mesa/SConscript
+++ b/mesalib/src/mesa/SConscript
@@ -169,7 +169,6 @@ swrast_sources = [
     'swrast/s_logic.c',
     'swrast/s_masking.c',
     'swrast/s_points.c',
-    'swrast/s_readpix.c',
     'swrast/s_span.c',
     'swrast/s_stencil.c',
     'swrast/s_texcombine.c',
diff --git a/mesalib/src/mesa/drivers/common/driverfuncs.c b/mesalib/src/mesa/drivers/common/driverfuncs.c
index 27ea559b3..5e25d7fda 100644
--- a/mesalib/src/mesa/drivers/common/driverfuncs.c
+++ b/mesalib/src/mesa/drivers/common/driverfuncs.c
@@ -30,6 +30,7 @@
 #include "main/framebuffer.h"
 #include "main/mipmap.h"
 #include "main/queryobj.h"
+#include "main/readpix.h"
 #include "main/renderbuffer.h"
 #include "main/shaderobj.h"
 #include "main/texcompress.h"
@@ -82,7 +83,7 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->Accum = _swrast_Accum;
    driver->RasterPos = _tnl_RasterPos;
    driver->DrawPixels = _swrast_DrawPixels;
-   driver->ReadPixels = _swrast_ReadPixels;
+   driver->ReadPixels = _mesa_readpixels;
    driver->CopyPixels = _swrast_CopyPixels;
    driver->Bitmap = _swrast_Bitmap;
 
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index 8d589e4c3..259041f67 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -879,40 +879,25 @@ _mesa_meta_end(struct gl_context *ctx)
 
       /* restore texture objects for unit[0] only */
       for (tgt = 0; tgt < NUM_TEXTURE_TARGETS; tgt++) {
-         _mesa_reference_texobj(&ctx->Texture.Unit[0].CurrentTex[tgt],
-                                save->CurrentTexture[tgt]);
+	 if (ctx->Texture.Unit[0].CurrentTex[tgt] != save->CurrentTexture[tgt]) {
+	    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+	    _mesa_reference_texobj(&ctx->Texture.Unit[0].CurrentTex[tgt],
+				   save->CurrentTexture[tgt]);
+	 }
          _mesa_reference_texobj(&save->CurrentTexture[tgt], NULL);
       }
 
-      /* Re-enable textures, texgen */
+      /* Restore fixed function texture enables, texgen */
       for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
-         if (save->TexEnabled[u]) {
-            _mesa_ActiveTextureARB(GL_TEXTURE0 + u);
-
-            if (save->TexEnabled[u] & TEXTURE_1D_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_1D, GL_TRUE);
-            if (save->TexEnabled[u] & TEXTURE_2D_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_2D, GL_TRUE);
-            if (save->TexEnabled[u] & TEXTURE_3D_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_3D, GL_TRUE);
-            if (save->TexEnabled[u] & TEXTURE_CUBE_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_CUBE_MAP, GL_TRUE);
-            if (save->TexEnabled[u] & TEXTURE_RECT_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_RECTANGLE, GL_TRUE);
-         }
-
-         if (save->TexGenEnabled[u]) {
-            _mesa_ActiveTextureARB(GL_TEXTURE0 + u);
+	 if (ctx->Texture.Unit[u].Enabled != save->TexEnabled[u]) {
+	    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+	    ctx->Texture.Unit[u].Enabled = save->TexEnabled[u];
+	 }
 
-            if (save->TexGenEnabled[u] & S_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_GEN_S, GL_TRUE);
-            if (save->TexGenEnabled[u] & T_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_GEN_T, GL_TRUE);
-            if (save->TexGenEnabled[u] & R_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_GEN_R, GL_TRUE);
-            if (save->TexGenEnabled[u] & Q_BIT)
-               _mesa_set_enable(ctx, GL_TEXTURE_GEN_Q, GL_TRUE);
-         }
+	 if (ctx->Texture.Unit[u].TexGenEnabled != save->TexGenEnabled[u]) {
+	    FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+	    ctx->Texture.Unit[u].TexGenEnabled = save->TexGenEnabled[u];
+	 }
       }
 
       /* restore current unit state */
diff --git a/mesalib/src/mesa/main/ffvertex_prog.c b/mesalib/src/mesa/main/ffvertex_prog.c
index 846907842..2c937386a 100644
--- a/mesalib/src/mesa/main/ffvertex_prog.c
+++ b/mesalib/src/mesa/main/ffvertex_prog.c
@@ -949,7 +949,7 @@ static struct ureg calculate_light_attenuation( struct tnl_program *p,
 {
    struct ureg attenuation = register_param3(p, STATE_LIGHT, i,
 					     STATE_ATTENUATION);
-   struct ureg att = get_temp(p);
+   struct ureg att = undef;
 
    /* Calculate spot attenuation:
     */
@@ -959,6 +959,8 @@ static struct ureg calculate_light_attenuation( struct tnl_program *p,
       struct ureg spot = get_temp(p);
       struct ureg slt = get_temp(p);
 
+      att = get_temp(p);
+
       emit_op2(p, OPCODE_DP3, spot, 0, negate(VPpli), spot_dir_norm);
       emit_op2(p, OPCODE_SLT, slt, 0, swizzle1(spot_dir_norm,W), spot);
       emit_op2(p, OPCODE_POW, spot, 0, spot, swizzle1(attenuation, W));
@@ -968,9 +970,13 @@ static struct ureg calculate_light_attenuation( struct tnl_program *p,
       release_temp(p, slt);
    }
 
-   /* Calculate distance attenuation:
+   /* Calculate distance attenuation(See formula (2.4) at glspec 2.1 page 62):
+    *
+    * Skip the calucation when _dist_ is undefined(light_eyepos3_is_zero)
     */
-   if (p->state->unit[i].light_attenuated) {
+   if (p->state->unit[i].light_attenuated && !is_undef(dist)) {
+      if (is_undef(att))
+         att = get_temp(p);
       /* 1/d,d,d,1/d */
       emit_op1(p, OPCODE_RCP, dist, WRITEMASK_YZ, dist);
       /* 1,d,d*d,1/d */
@@ -1113,73 +1119,54 @@ static void build_lighting( struct tnl_program *p )
       if (p->state->unit[i].light_enabled) {
 	 struct ureg half = undef;
 	 struct ureg att = undef, VPpli = undef;
+	 struct ureg dist = undef;
 
 	 count++;
+         if (p->state->unit[i].light_eyepos3_is_zero) {
+             VPpli = register_param3(p, STATE_INTERNAL,
+                                     STATE_LIGHT_POSITION_NORMALIZED, i);
+         } else {
+            struct ureg Ppli = register_param3(p, STATE_INTERNAL,
+                                               STATE_LIGHT_POSITION, i);
+            struct ureg V = get_eye_position(p);
+
+            VPpli = get_temp(p);
+            dist = get_temp(p);
+
+            /* Calculate VPpli vector
+             */
+            emit_op2(p, OPCODE_SUB, VPpli, 0, Ppli, V);
 
-	 if (p->state->unit[i].light_eyepos3_is_zero) {
-	    /* Can used precomputed constants in this case.
-	     * Attenuation never applies to infinite lights.
-	     */
-	    VPpli = register_param3(p, STATE_INTERNAL,
-				    STATE_LIGHT_POSITION_NORMALIZED, i);
-
-            if (!p->state->material_shininess_is_zero) {
-               if (p->state->light_local_viewer) {
-                  struct ureg eye_hat = get_eye_position_normalized(p);
-                  half = get_temp(p);
-                  emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
-                  emit_normalize_vec3(p, half, half);
-               }
-               else {
-                  half = register_param3(p, STATE_INTERNAL,
-                                         STATE_LIGHT_HALF_VECTOR, i);
-               }
-            }
-	 }
-	 else {
-	    struct ureg Ppli = register_param3(p, STATE_INTERNAL,
-					       STATE_LIGHT_POSITION, i);
-	    struct ureg V = get_eye_position(p);
-	    struct ureg dist = get_temp(p);
-
-	    VPpli = get_temp(p);
-
-	    /* Calculate VPpli vector
-	     */
-	    emit_op2(p, OPCODE_SUB, VPpli, 0, Ppli, V);
-
-	    /* Normalize VPpli.  The dist value also used in
-	     * attenuation below.
-	     */
-	    emit_op2(p, OPCODE_DP3, dist, 0, VPpli, VPpli);
-	    emit_op1(p, OPCODE_RSQ, dist, 0, dist);
-	    emit_op2(p, OPCODE_MUL, VPpli, 0, VPpli, dist);
-
-	    /* Calculate attenuation:
-	     */
-	    if (!p->state->unit[i].light_spotcutoff_is_180 ||
-		p->state->unit[i].light_attenuated) {
-	       att = calculate_light_attenuation(p, i, VPpli, dist);
-	    }
-
-	    /* Calculate viewer direction, or use infinite viewer:
-	     */
-            if (!p->state->material_shininess_is_zero) {
-               half = get_temp(p);
+            /* Normalize VPpli.  The dist value also used in
+             * attenuation below.
+             */
+            emit_op2(p, OPCODE_DP3, dist, 0, VPpli, VPpli);
+            emit_op1(p, OPCODE_RSQ, dist, 0, dist);
+            emit_op2(p, OPCODE_MUL, VPpli, 0, VPpli, dist);
+         }
 
-               if (p->state->light_local_viewer) {
-                  struct ureg eye_hat = get_eye_position_normalized(p);
-                  emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
-               }
-               else {
-                  struct ureg z_dir = swizzle(get_identity_param(p),X,Y,W,Z);
-                  emit_op2(p, OPCODE_ADD, half, 0, VPpli, z_dir);
-               }
+         /* Calculate attenuation:
+          */
+         att = calculate_light_attenuation(p, i, VPpli, dist);
+         release_temp(p, dist);
 
+	 /* Calculate viewer direction, or use infinite viewer:
+	  */
+         if (!p->state->material_shininess_is_zero) {
+            if (p->state->light_local_viewer) {
+               struct ureg eye_hat = get_eye_position_normalized(p);
+               half = get_temp(p);
+               emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
+               emit_normalize_vec3(p, half, half);
+            } else if (p->state->unit[i].light_eyepos3_is_zero) {
+               half = register_param3(p, STATE_INTERNAL,
+                                      STATE_LIGHT_HALF_VECTOR, i);
+            } else {
+               struct ureg z_dir = swizzle(get_identity_param(p),X,Y,W,Z);
+               half = get_temp(p);
+               emit_op2(p, OPCODE_ADD, half, 0, VPpli, z_dir);
                emit_normalize_vec3(p, half, half);
             }
-
-	    release_temp(p, dist);
 	 }
 
 	 /* Calculate dot products:
diff --git a/mesalib/src/mesa/main/format_unpack.c b/mesalib/src/mesa/main/format_unpack.c
index 525bbcb1c..6e2ce7a05 100644
--- a/mesalib/src/mesa/main/format_unpack.c
+++ b/mesalib/src/mesa/main/format_unpack.c
@@ -57,1049 +57,1335 @@ nonlinear_to_linear(GLubyte cs8)
 }
 
 
-typedef void (*unpack_rgba_func)(const void *src, GLfloat dst[4]);
+typedef void (*unpack_rgba_func)(const void *src, GLfloat dst[][4], GLuint n);
 
 
 static void
-unpack_RGBA8888(const void *src, GLfloat dst[4])
+unpack_RGBA8888(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i] >> 24)        );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( (s[i]      ) & 0xff );
+   }
 }
 
 static void
-unpack_RGBA8888_REV(const void *src, GLfloat dst[4])
+unpack_RGBA8888_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i]      ) & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( (s[i] >> 24)        );
+   }
 }
 
 static void
-unpack_ARGB8888(const void *src, GLfloat dst[4])
+unpack_ARGB8888(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i]      ) & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( (s[i] >> 24)        );
+   }
 }
 
 static void
-unpack_ARGB8888_REV(const void *src, GLfloat dst[4])
+unpack_ARGB8888_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i] >> 24)        );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( (s[i]      ) & 0xff );
+   }
 }
 
 static void
-unpack_XRGB8888(const void *src, GLfloat dst[4])
+unpack_XRGB8888(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
-   dst[ACOMP] = 1.0f;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i]      ) & 0xff );
+      dst[i][ACOMP] = 1.0f;
+   }
 }
 
 static void
-unpack_XRGB8888_REV(const void *src, GLfloat dst[4])
+unpack_XRGB8888_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
-   dst[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
-   dst[ACOMP] = 1.0f;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( (s[i] >>  8) & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( (s[i] >> 16) & 0xff );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( (s[i] >> 24)        );
+      dst[i][ACOMP] = 1.0f;
+   }
 }
 
 static void
-unpack_RGB888(const void *src, GLfloat dst[4])
+unpack_RGB888(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLubyte *s = (const GLubyte *) src;
-   dst[RCOMP] = UBYTE_TO_FLOAT( s[2] );
-   dst[GCOMP] = UBYTE_TO_FLOAT( s[1] );
-   dst[BCOMP] = UBYTE_TO_FLOAT( s[0] );
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( s[i*3+2] );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( s[i*3+1] );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( s[i*3+0] );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_BGR888(const void *src, GLfloat dst[4])
+unpack_BGR888(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLubyte *s = (const GLubyte *) src;
-   dst[RCOMP] = UBYTE_TO_FLOAT( s[0] );
-   dst[GCOMP] = UBYTE_TO_FLOAT( s[1] );
-   dst[BCOMP] = UBYTE_TO_FLOAT( s[2] );
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( s[i*3+0] );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( s[i*3+1] );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( s[i*3+2] );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_RGB565(const void *src, GLfloat dst[4])
+unpack_RGB565(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);
-   dst[GCOMP] = ((s >> 5 ) & 0x3f) * (1.0F / 63.0F);
-   dst[BCOMP] = ((s      ) & 0x1f) * (1.0F / 31.0F);
-   dst[ACOMP] = 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >> 11) & 0x1f) * (1.0F / 31.0F);
+      dst[i][GCOMP] = ((s[i] >> 5 ) & 0x3f) * (1.0F / 63.0F);
+      dst[i][BCOMP] = ((s[i]      ) & 0x1f) * (1.0F / 31.0F);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_RGB565_REV(const void *src, GLfloat dst[4])
+unpack_RGB565_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   GLushort s = *((const GLushort *) src);
-   s = (s >> 8) | (s << 8); /* byte swap */
-   dst[RCOMP] = UBYTE_TO_FLOAT( ((s >> 8) & 0xf8) | ((s >> 13) & 0x7) );
-   dst[GCOMP] = UBYTE_TO_FLOAT( ((s >> 3) & 0xfc) | ((s >>  9) & 0x3) );
-   dst[BCOMP] = UBYTE_TO_FLOAT( ((s << 3) & 0xf8) | ((s >>  2) & 0x7) );
-   dst[ACOMP] = 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      GLuint t = (s[i] >> 8) | (s[i] << 8); /* byte swap */
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( ((t >> 8) & 0xf8) | ((t >> 13) & 0x7) );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( ((t >> 3) & 0xfc) | ((t >>  9) & 0x3) );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( ((t << 3) & 0xf8) | ((t >>  2) & 0x7) );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_ARGB4444(const void *src, GLfloat dst[4])
+unpack_ARGB4444(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);
-   dst[GCOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);
-   dst[BCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);
-   dst[ACOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >>  8) & 0xf) * (1.0F / 15.0F);
+      dst[i][GCOMP] = ((s[i] >>  4) & 0xf) * (1.0F / 15.0F);
+      dst[i][BCOMP] = ((s[i]      ) & 0xf) * (1.0F / 15.0F);
+      dst[i][ACOMP] = ((s[i] >> 12) & 0xf) * (1.0F / 15.0F);
+   }
 }
 
 static void
-unpack_ARGB4444_REV(const void *src, GLfloat dst[4])
+unpack_ARGB4444_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);
-   dst[GCOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);
-   dst[BCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);
-   dst[ACOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i]      ) & 0xf) * (1.0F / 15.0F);
+      dst[i][GCOMP] = ((s[i] >> 12) & 0xf) * (1.0F / 15.0F);
+      dst[i][BCOMP] = ((s[i] >>  8) & 0xf) * (1.0F / 15.0F);
+      dst[i][ACOMP] = ((s[i] >>  4) & 0xf) * (1.0F / 15.0F);
+   }
 }
 
 static void
-unpack_RGBA5551(const void *src, GLfloat dst[4])
+unpack_RGBA5551(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);
-   dst[GCOMP] = ((s >>  6) & 0x1f) * (1.0F / 31.0F);
-   dst[BCOMP] = ((s >>  1) & 0x1f) * (1.0F / 31.0F);
-   dst[ACOMP] = ((s      ) & 0x01) * 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >> 11) & 0x1f) * (1.0F / 31.0F);
+      dst[i][GCOMP] = ((s[i] >>  6) & 0x1f) * (1.0F / 31.0F);
+      dst[i][BCOMP] = ((s[i] >>  1) & 0x1f) * (1.0F / 31.0F);
+      dst[i][ACOMP] = ((s[i]      ) & 0x01) * 1.0F;
+   }
 }
 
 static void
-unpack_ARGB1555(const void *src, GLfloat dst[4])
+unpack_ARGB1555(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = ((s >> 10) & 0x1f) * (1.0F / 31.0F);
-   dst[GCOMP] = ((s >>  5) & 0x1f) * (1.0F / 31.0F);
-   dst[BCOMP] = ((s >>  0) & 0x1f) * (1.0F / 31.0F);
-   dst[ACOMP] = ((s >> 15) & 0x01) * 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >> 10) & 0x1f) * (1.0F / 31.0F);
+      dst[i][GCOMP] = ((s[i] >>  5) & 0x1f) * (1.0F / 31.0F);
+      dst[i][BCOMP] = ((s[i] >>  0) & 0x1f) * (1.0F / 31.0F);
+      dst[i][ACOMP] = ((s[i] >> 15) & 0x01) * 1.0F;
+   }
 }
 
 static void
-unpack_ARGB1555_REV(const void *src, GLfloat dst[4])
+unpack_ARGB1555_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( ((s >>  7) & 0xf8) | ((s >> 12) & 0x7) );
-   dst[GCOMP] = UBYTE_TO_FLOAT( ((s >>  2) & 0xf8) | ((s >>  7) & 0x7) );
-   dst[BCOMP] = UBYTE_TO_FLOAT( ((s <<  3) & 0xf8) | ((s >>  2) & 0x7) );
-   dst[ACOMP] = UBYTE_TO_FLOAT( ((s >> 15) & 0x01) * 255 );
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( ((s[i] >>  7) & 0xf8) | ((s[i] >> 12) & 0x7) );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( ((s[i] >>  2) & 0xf8) | ((s[i] >>  7) & 0x7) );
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( ((s[i] <<  3) & 0xf8) | ((s[i] >>  2) & 0x7) );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( ((s[i] >> 15) & 0x01) * 255 );
+   }
 }
 
 static void
-unpack_AL44(const void *src, GLfloat dst[4])
+unpack_AL44(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = (s & 0xf) * (1.0F / 15.0F);
-   dst[ACOMP] = ((s >> 4) & 0xf) * (1.0F / 15.0F);
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = (s[i] & 0xf) * (1.0F / 15.0F);
+      dst[i][ACOMP] = ((s[i] >> 4) & 0xf) * (1.0F / 15.0F);
+   }
 }
 
 static void
-unpack_AL88(const void *src, GLfloat dst[4])
+unpack_AL88(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = 
-   dst[GCOMP] = 
-   dst[BCOMP] = UBYTE_TO_FLOAT( s & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( s >> 8 );
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = 
+      dst[i][GCOMP] = 
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( s[i] & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( s[i] >> 8 );
+   }
 }
 
 static void
-unpack_AL88_REV(const void *src, GLfloat dst[4])
+unpack_AL88_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = 
-   dst[GCOMP] = 
-   dst[BCOMP] = UBYTE_TO_FLOAT( s >> 8 );
-   dst[ACOMP] = UBYTE_TO_FLOAT( s & 0xff );
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = 
+      dst[i][GCOMP] = 
+      dst[i][BCOMP] = UBYTE_TO_FLOAT( s[i] >> 8 );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( s[i] & 0xff );
+   }
 }
 
 static void
-unpack_AL1616(const void *src, GLfloat dst[4])
+unpack_AL1616(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = USHORT_TO_FLOAT( s & 0xffff );
-   dst[ACOMP] = USHORT_TO_FLOAT( s >> 16 );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = USHORT_TO_FLOAT( s[i] & 0xffff );
+      dst[i][ACOMP] = USHORT_TO_FLOAT( s[i] >> 16 );
+   }
 }
 
 static void
-unpack_AL1616_REV(const void *src, GLfloat dst[4])
+unpack_AL1616_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = USHORT_TO_FLOAT( s >> 16 );
-   dst[ACOMP] = USHORT_TO_FLOAT( s & 0xffff );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = USHORT_TO_FLOAT( s[i] >> 16 );
+      dst[i][ACOMP] = USHORT_TO_FLOAT( s[i] & 0xffff );
+   }
 }
 
 static void
-unpack_RGB332(const void *src, GLfloat dst[4])
+unpack_RGB332(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] = ((s >> 5) & 0x7) * (1.0F / 7.0F);
-   dst[GCOMP] = ((s >> 2) & 0x7) * (1.0F / 7.0F);
-   dst[BCOMP] = ((s     ) & 0x3) * (1.0F / 3.0F);
-   dst[ACOMP] = 1.0F;
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >> 5) & 0x7) * (1.0F / 7.0F);
+      dst[i][GCOMP] = ((s[i] >> 2) & 0x7) * (1.0F / 7.0F);
+      dst[i][BCOMP] = ((s[i]     ) & 0x3) * (1.0F / 3.0F);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 
 static void
-unpack_A8(const void *src, GLfloat dst[4])
+unpack_A8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = UBYTE_TO_FLOAT(s);
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = UBYTE_TO_FLOAT(s[i]);
+   }
 }
 
 static void
-unpack_A16(const void *src, GLfloat dst[4])
+unpack_A16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = USHORT_TO_FLOAT(s);
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = USHORT_TO_FLOAT(s[i]);
+   }
 }
 
 static void
-unpack_L8(const void *src, GLfloat dst[4])
+unpack_L8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = UBYTE_TO_FLOAT(s);
-   dst[ACOMP] = 1.0F;
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = UBYTE_TO_FLOAT(s[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_L16(const void *src, GLfloat dst[4])
+unpack_L16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = USHORT_TO_FLOAT(s);
-   dst[ACOMP] = 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = USHORT_TO_FLOAT(s[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_I8(const void *src, GLfloat dst[4])
+unpack_I8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = UBYTE_TO_FLOAT(s);
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = UBYTE_TO_FLOAT(s[i]);
+   }
 }
 
 static void
-unpack_I16(const void *src, GLfloat dst[4])
+unpack_I16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = USHORT_TO_FLOAT(s);
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = USHORT_TO_FLOAT(s[i]);
+   }
 }
 
 static void
-unpack_YCBCR(const void *src, GLfloat dst[4])
+unpack_YCBCR(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint i = 0;
-   const GLushort *src0 = (const GLushort *) src;
-   const GLushort *src1 = src0 + 1;                               /* odd */
-   const GLubyte y0 = (*src0 >> 8) & 0xff;  /* luminance */
-   const GLubyte cb = *src0 & 0xff;         /* chroma U */
-   const GLubyte y1 = (*src1 >> 8) & 0xff;  /* luminance */
-   const GLubyte cr = *src1 & 0xff;         /* chroma V */
-   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
-   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
-   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
-   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
-   r *= (1.0F / 255.0F);
-   g *= (1.0F / 255.0F);
-   b *= (1.0F / 255.0F);
-   dst[RCOMP] = CLAMP(r, 0.0F, 1.0F);
-   dst[GCOMP] = CLAMP(g, 0.0F, 1.0F);
-   dst[BCOMP] = CLAMP(b, 0.0F, 1.0F);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      const GLushort *src0 = ((const GLushort *) src) + i * 2; /* even */
+      const GLushort *src1 = src0 + 1;         /* odd */
+      const GLubyte y0 = (*src0 >> 8) & 0xff;  /* luminance */
+      const GLubyte cb = *src0 & 0xff;         /* chroma U */
+      const GLubyte y1 = (*src1 >> 8) & 0xff;  /* luminance */
+      const GLubyte cr = *src1 & 0xff;         /* chroma V */
+      const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
+      GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
+      GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
+      GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
+      r *= (1.0F / 255.0F);
+      g *= (1.0F / 255.0F);
+      b *= (1.0F / 255.0F);
+      dst[i][RCOMP] = CLAMP(r, 0.0F, 1.0F);
+      dst[i][GCOMP] = CLAMP(g, 0.0F, 1.0F);
+      dst[i][BCOMP] = CLAMP(b, 0.0F, 1.0F);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_YCBCR_REV(const void *src, GLfloat dst[4])
+unpack_YCBCR_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint i = 0;
-   const GLushort *src0 = (const GLushort *) src;
-   const GLushort *src1 = src0 + 1;                               /* odd */
-   const GLubyte y0 = *src0 & 0xff;         /* luminance */
-   const GLubyte cr = (*src0 >> 8) & 0xff;  /* chroma V */
-   const GLubyte y1 = *src1 & 0xff;         /* luminance */
-   const GLubyte cb = (*src1 >> 8) & 0xff;  /* chroma U */
-   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
-   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
-   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
-   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
-   r *= (1.0F / 255.0F);
-   g *= (1.0F / 255.0F);
-   b *= (1.0F / 255.0F);
-   dst[RCOMP] = CLAMP(r, 0.0F, 1.0F);
-   dst[GCOMP] = CLAMP(g, 0.0F, 1.0F);
-   dst[BCOMP] = CLAMP(b, 0.0F, 1.0F);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      const GLushort *src0 = ((const GLushort *) src) + i * 2; /* even */
+      const GLushort *src1 = src0 + 1;         /* odd */
+      const GLubyte y0 = *src0 & 0xff;         /* luminance */
+      const GLubyte cr = (*src0 >> 8) & 0xff;  /* chroma V */
+      const GLubyte y1 = *src1 & 0xff;         /* luminance */
+      const GLubyte cb = (*src1 >> 8) & 0xff;  /* chroma U */
+      const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
+      GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
+      GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
+      GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
+      r *= (1.0F / 255.0F);
+      g *= (1.0F / 255.0F);
+      b *= (1.0F / 255.0F);
+      dst[i][RCOMP] = CLAMP(r, 0.0F, 1.0F);
+      dst[i][GCOMP] = CLAMP(g, 0.0F, 1.0F);
+      dst[i][BCOMP] = CLAMP(b, 0.0F, 1.0F);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_R8(const void *src, GLfloat dst[4])
+unpack_R8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[0] = UBYTE_TO_FLOAT(s);
-   dst[1] = dst[2] = 0.0F;
-   dst[3] = 1.0F;
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] = UBYTE_TO_FLOAT(s[i]);
+      dst[i][1] =
+      dst[i][2] = 0.0F;
+      dst[i][3] = 1.0F;
+   }
 }
 
 static void
-unpack_RG88(const void *src, GLfloat dst[4])
+unpack_RG88(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );
-   dst[BCOMP] = 0.0;
-   dst[ACOMP] = 1.0;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( s[i] & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( s[i] >> 8 );
+      dst[i][BCOMP] = 0.0;
+      dst[i][ACOMP] = 1.0;
+   }
 }
 
 static void
-unpack_RG88_REV(const void *src, GLfloat dst[4])
+unpack_RG88_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );
-   dst[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );
-   dst[BCOMP] = 0.0;
-   dst[ACOMP] = 1.0;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = UBYTE_TO_FLOAT( s[i] & 0xff );
+      dst[i][GCOMP] = UBYTE_TO_FLOAT( s[i] >> 8 );
+      dst[i][BCOMP] = 0.0;
+      dst[i][ACOMP] = 1.0;
+   }
 }
 
 static void
-unpack_R16(const void *src, GLfloat dst[4])
+unpack_R16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = USHORT_TO_FLOAT(s);
-   dst[GCOMP] = 0.0;
-   dst[BCOMP] = 0.0;
-   dst[ACOMP] = 1.0;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = USHORT_TO_FLOAT(s[i]);
+      dst[i][GCOMP] = 0.0;
+      dst[i][BCOMP] = 0.0;
+      dst[i][ACOMP] = 1.0;
+   }
 }
 
 static void
-unpack_RG1616(const void *src, GLfloat dst[4])
+unpack_RG1616(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = USHORT_TO_FLOAT( s & 0xffff );
-   dst[GCOMP] = USHORT_TO_FLOAT( s >> 16 );
-   dst[BCOMP] = 0.0;
-   dst[ACOMP] = 1.0;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = USHORT_TO_FLOAT( s[i] & 0xffff );
+      dst[i][GCOMP] = USHORT_TO_FLOAT( s[i] >> 16 );
+      dst[i][BCOMP] = 0.0;
+      dst[i][ACOMP] = 1.0;
+   }
 }
 
 static void
-unpack_RG1616_REV(const void *src, GLfloat dst[4])
+unpack_RG1616_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = USHORT_TO_FLOAT( s >> 16 );
-   dst[GCOMP] = USHORT_TO_FLOAT( s & 0xffff );
-   dst[BCOMP] = 0.0;
-   dst[ACOMP] = 1.0;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = USHORT_TO_FLOAT( s[i] >> 16 );
+      dst[i][GCOMP] = USHORT_TO_FLOAT( s[i] & 0xffff );
+      dst[i][BCOMP] = 0.0;
+      dst[i][ACOMP] = 1.0;
+   }
 }
 
 static void
-unpack_ARGB2101010(const void *src, GLfloat dst[4])
+unpack_ARGB2101010(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = ((s >> 20) & 0x3ff) * (1.0F / 1023.0F);
-   dst[GCOMP] = ((s >> 10) & 0x3ff) * (1.0F / 1023.0F);
-   dst[BCOMP] = ((s >>  0) & 0x3ff) * (1.0F / 1023.0F);
-   dst[ACOMP] = ((s >> 30) & 0x03) * (1.0F / 3.0F);
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = ((s[i] >> 20) & 0x3ff) * (1.0F / 1023.0F);
+      dst[i][GCOMP] = ((s[i] >> 10) & 0x3ff) * (1.0F / 1023.0F);
+      dst[i][BCOMP] = ((s[i] >>  0) & 0x3ff) * (1.0F / 1023.0F);
+      dst[i][ACOMP] = ((s[i] >> 30) &  0x03) * (1.0F / 3.0F);
+   }
 }
 
 
 static void
-unpack_Z24_S8(const void *src, GLfloat dst[4])
+unpack_Z24_S8(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
+   const GLuint *s = ((const GLuint *) src);
    const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
-   dst[0] = dst[1] = dst[2] = (s >> 8) * scale;
-   dst[3] = 1.0F;
-   ASSERT(dst[0] >= 0.0F);
-   ASSERT(dst[0] <= 1.0F);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = (s[i] >> 8) * scale;
+      dst[i][3] = 1.0F;
+      ASSERT(dst[i][0] >= 0.0F);
+      ASSERT(dst[i][0] <= 1.0F);
+   }
 }
 
 static void
-unpack_S8_Z24(const void *src, GLfloat dst[4])
+unpack_S8_Z24(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
+   const GLuint *s = ((const GLuint *) src);
    const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
-   dst[0] = dst[1] = dst[2] = (s & 0x00ffffff) * scale;
-   dst[3] = 1.0F;
-   ASSERT(dst[0] >= 0.0F);
-   ASSERT(dst[0] <= 1.0F);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = (s[i] & 0x00ffffff) * scale;
+      dst[i][3] = 1.0F;
+      ASSERT(dst[i][0] >= 0.0F);
+      ASSERT(dst[i][0] <= 1.0F);
+   }
 }
 
 static void
-unpack_Z16(const void *src, GLfloat dst[4])
+unpack_Z16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[0] = dst[1] = dst[2] = s * (1.0F / 65535.0F);
-   dst[3] = 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = s[i] * (1.0F / 65535.0F);
+      dst[i][3] = 1.0F;
+   }
 }
 
 static void
-unpack_X8_Z24(const void *src, GLfloat dst[4])
+unpack_X8_Z24(const void *src, GLfloat dst[][4], GLuint n)
 {
-   unpack_S8_Z24(src, dst);
+   unpack_S8_Z24(src, dst, n);
 }
 
 static void
-unpack_Z24_X8(const void *src, GLfloat dst[4])
+unpack_Z24_X8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   unpack_Z24_S8(src, dst);
+   unpack_Z24_S8(src, dst, n);
 }
 
 static void
-unpack_Z32(const void *src, GLfloat dst[4])
+unpack_Z32(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[0] = dst[1] = dst[2] = s * (1.0F / 0xffffffff);
-   dst[3] = 1.0F;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = s[i] * (1.0F / 0xffffffff);
+      dst[i][3] = 1.0F;
+   }
 }
 
 static void
-unpack_Z32_FLOAT(const void *src, GLfloat dst[4])
+unpack_Z32_FLOAT(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLfloat s = *((const GLfloat *) src);
-   dst[0] = dst[1] = dst[2] = s;
-   dst[3] = 1.0F;
+   const GLfloat *s = ((const GLfloat *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = s[i];
+      dst[i][3] = 1.0F;
+   }
 }
 
 static void
-unpack_Z32_FLOAT_X24S8(const void *src, GLfloat dst[4])
+unpack_Z32_FLOAT_X24S8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLfloat s = *((const GLfloat *) src);
-   dst[0] = dst[1] = dst[2] = s;
-   dst[3] = 1.0F;
+   const GLfloat *s = ((const GLfloat *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = s[i];
+      dst[i][3] = 1.0F;
+   }
 }
 
 
 static void
-unpack_S8(const void *src, GLfloat dst[4])
+unpack_S8(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* should never be used */
-   dst[0] = dst[1] = dst[2] = 0.0F;
-   dst[3] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][0] =
+      dst[i][1] =
+      dst[i][2] = 0.0F;
+      dst[i][3] = 1.0F;
+   }
 }
 
 
 static void
-unpack_SRGB8(const void *src, GLfloat dst[4])
+unpack_SRGB8(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLubyte *s = (const GLubyte *) src;
-   dst[RCOMP] = nonlinear_to_linear(s[2]);
-   dst[GCOMP] = nonlinear_to_linear(s[1]);
-   dst[BCOMP] = nonlinear_to_linear(s[0]);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = nonlinear_to_linear(s[i*3+2]);
+      dst[i][GCOMP] = nonlinear_to_linear(s[i*3+1]);
+      dst[i][BCOMP] = nonlinear_to_linear(s[i*3+0]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SRGBA8(const void *src, GLfloat dst[4])
+unpack_SRGBA8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = nonlinear_to_linear( (s >> 24) );
-   dst[GCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );
-   dst[BCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff ); /* linear! */
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = nonlinear_to_linear( (s[i] >> 24) );
+      dst[i][GCOMP] = nonlinear_to_linear( (s[i] >> 16) & 0xff );
+      dst[i][BCOMP] = nonlinear_to_linear( (s[i] >>  8) & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( s[i] & 0xff ); /* linear! */
+   }
 }
 
 static void
-unpack_SARGB8(const void *src, GLfloat dst[4])
+unpack_SARGB8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );
-   dst[GCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );
-   dst[BCOMP] = nonlinear_to_linear( (s      ) & 0xff );
-   dst[ACOMP] = UBYTE_TO_FLOAT( (s >> 24) ); /* linear! */
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = nonlinear_to_linear( (s[i] >> 16) & 0xff );
+      dst[i][GCOMP] = nonlinear_to_linear( (s[i] >>  8) & 0xff );
+      dst[i][BCOMP] = nonlinear_to_linear( (s[i]      ) & 0xff );
+      dst[i][ACOMP] = UBYTE_TO_FLOAT( s[i] >> 24 ); /* linear! */
+   }
 }
 
 static void
-unpack_SL8(const void *src, GLfloat dst[4])
+unpack_SL8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLubyte s = *((const GLubyte *) src);
-   dst[RCOMP] = 
-   dst[GCOMP] = 
-   dst[BCOMP] = nonlinear_to_linear(s);
-   dst[ACOMP] = 1.0F;
+   const GLubyte *s = ((const GLubyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = 
+      dst[i][GCOMP] = 
+      dst[i][BCOMP] = nonlinear_to_linear(s[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SLA8(const void *src, GLfloat dst[4])
+unpack_SLA8(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLubyte *s = (const GLubyte *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = nonlinear_to_linear(s[0]);
-   dst[ACOMP] = UBYTE_TO_FLOAT(s[1]); /* linear */
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = nonlinear_to_linear(s[i*2+0]);
+      dst[i][ACOMP] = UBYTE_TO_FLOAT(s[i*2+1]); /* linear! */
+   }
 }
 
 static void
-unpack_SRGB_DXT1(const void *src, GLfloat dst[4])
+unpack_SRGB_DXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_SRGBA_DXT1(const void *src, GLfloat dst[4])
+unpack_SRGBA_DXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_SRGBA_DXT3(const void *src, GLfloat dst[4])
+unpack_SRGBA_DXT3(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_SRGBA_DXT5(const void *src, GLfloat dst[4])
+unpack_SRGBA_DXT5(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGB_FXT1(const void *src, GLfloat dst[4])
+unpack_RGB_FXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGBA_FXT1(const void *src, GLfloat dst[4])
+unpack_RGBA_FXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGB_DXT1(const void *src, GLfloat dst[4])
+unpack_RGB_DXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGBA_DXT1(const void *src, GLfloat dst[4])
+unpack_RGBA_DXT1(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGBA_DXT3(const void *src, GLfloat dst[4])
+unpack_RGBA_DXT3(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 static void
-unpack_RGBA_DXT5(const void *src, GLfloat dst[4])
+unpack_RGBA_DXT5(const void *src, GLfloat dst[][4], GLuint n)
 {
 }
 
 
 static void
-unpack_RGBA_FLOAT32(const void *src, GLfloat dst[4])
+unpack_RGBA_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] = s[0];
-   dst[GCOMP] = s[1];
-   dst[BCOMP] = s[2];
-   dst[ACOMP] = s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = s[i*4+0];
+      dst[i][GCOMP] = s[i*4+1];
+      dst[i][BCOMP] = s[i*4+2];
+      dst[i][ACOMP] = s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_FLOAT16(const void *src, GLfloat dst[4])
+unpack_RGBA_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] = _mesa_half_to_float(s[0]);
-   dst[GCOMP] = _mesa_half_to_float(s[1]);
-   dst[BCOMP] = _mesa_half_to_float(s[2]);
-   dst[ACOMP] = _mesa_half_to_float(s[3]);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = _mesa_half_to_float(s[i*4+0]);
+      dst[i][GCOMP] = _mesa_half_to_float(s[i*4+1]);
+      dst[i][BCOMP] = _mesa_half_to_float(s[i*4+2]);
+      dst[i][ACOMP] = _mesa_half_to_float(s[i*4+3]);
+   }
 }
 
 static void
-unpack_RGB_FLOAT32(const void *src, GLfloat dst[4])
+unpack_RGB_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] = s[0];
-   dst[GCOMP] = s[1];
-   dst[BCOMP] = s[2];
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = s[i*3+0];
+      dst[i][GCOMP] = s[i*3+1];
+      dst[i][BCOMP] = s[i*3+2];
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_RGB_FLOAT16(const void *src, GLfloat dst[4])
+unpack_RGB_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] = _mesa_half_to_float(s[0]);
-   dst[GCOMP] = _mesa_half_to_float(s[1]);
-   dst[BCOMP] = _mesa_half_to_float(s[2]);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = _mesa_half_to_float(s[i*3+0]);
+      dst[i][GCOMP] = _mesa_half_to_float(s[i*3+1]);
+      dst[i][BCOMP] = _mesa_half_to_float(s[i*3+2]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_ALPHA_FLOAT32(const void *src, GLfloat dst[4])
+unpack_ALPHA_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = s[0];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = s[i];
+   }
 }
 
 static void
-unpack_ALPHA_FLOAT16(const void *src, GLfloat dst[4])
+unpack_ALPHA_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = _mesa_half_to_float(s[0]);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = _mesa_half_to_float(s[i]);
+   }
 }
 
 static void
-unpack_LUMINANCE_FLOAT32(const void *src, GLfloat dst[4])
+unpack_LUMINANCE_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = s[0];
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = s[i];
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_LUMINANCE_FLOAT16(const void *src, GLfloat dst[4])
+unpack_LUMINANCE_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = _mesa_half_to_float(s[0]);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = _mesa_half_to_float(s[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_LUMINANCE_ALPHA_FLOAT32(const void *src, GLfloat dst[4])
+unpack_LUMINANCE_ALPHA_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = s[0];
-   dst[ACOMP] = s[1];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = s[i*2+0];
+      dst[i][ACOMP] = s[i*2+1];
+   }
 }
 
 static void
-unpack_LUMINANCE_ALPHA_FLOAT16(const void *src, GLfloat dst[4])
+unpack_LUMINANCE_ALPHA_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = _mesa_half_to_float(s[0]);
-   dst[ACOMP] = _mesa_half_to_float(s[1]);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = _mesa_half_to_float(s[i*2+0]);
+      dst[i][ACOMP] = _mesa_half_to_float(s[i*2+1]);
+   }
 }
 
 static void
-unpack_INTENSITY_FLOAT32(const void *src, GLfloat dst[4])
+unpack_INTENSITY_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = s[0];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = s[i];
+   }
 }
 
 static void
-unpack_INTENSITY_FLOAT16(const void *src, GLfloat dst[4])
+unpack_INTENSITY_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = s[0];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = s[i];
+   }
 }
 
 static void
-unpack_R_FLOAT32(const void *src, GLfloat dst[4])
+unpack_R_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] = s[0];
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = s[i];
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_R_FLOAT16(const void *src, GLfloat dst[4])
+unpack_R_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] = _mesa_half_to_float(s[0]);
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = _mesa_half_to_float(s[i]);
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_RG_FLOAT32(const void *src, GLfloat dst[4])
+unpack_RG_FLOAT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLfloat *s = (const GLfloat *) src;
-   dst[RCOMP] = s[0];
-   dst[GCOMP] = s[1];
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = s[i*2+0];
+      dst[i][GCOMP] = s[i*2+1];
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_RG_FLOAT16(const void *src, GLfloat dst[4])
+unpack_RG_FLOAT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLhalfARB *s = (const GLhalfARB *) src;
-   dst[RCOMP] = _mesa_half_to_float(s[0]);
-   dst[GCOMP] = _mesa_half_to_float(s[1]);
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = _mesa_half_to_float(s[i*2+0]);
+      dst[i][GCOMP] = _mesa_half_to_float(s[i*2+1]);
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 
 static void
-unpack_RGBA_INT8(const void *src, GLfloat dst[4])
+unpack_RGBA_INT8(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLbyte *s = (const GLbyte *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_INT16(const void *src, GLfloat dst[4])
+unpack_RGBA_INT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLshort *s = (const GLshort *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_INT32(const void *src, GLfloat dst[4])
+unpack_RGBA_INT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLint *s = (const GLint *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_UINT8(const void *src, GLfloat dst[4])
+unpack_RGBA_UINT8(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLubyte *s = (const GLubyte *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_UINT16(const void *src, GLfloat dst[4])
+unpack_RGBA_UINT16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLushort *s = (const GLushort *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_RGBA_UINT32(const void *src, GLfloat dst[4])
+unpack_RGBA_UINT32(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLuint *s = (const GLuint *) src;
-   dst[RCOMP] = (GLfloat) s[0];
-   dst[GCOMP] = (GLfloat) s[1];
-   dst[BCOMP] = (GLfloat) s[2];
-   dst[ACOMP] = (GLfloat) s[3];
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = (GLfloat) s[i*4+0];
+      dst[i][GCOMP] = (GLfloat) s[i*4+1];
+      dst[i][BCOMP] = (GLfloat) s[i*4+2];
+      dst[i][ACOMP] = (GLfloat) s[i*4+3];
+   }
 }
 
 static void
-unpack_DUDV8(const void *src, GLfloat dst[4])
+unpack_DUDV8(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLbyte *s = (const GLbyte *) src;
-   dst[RCOMP] = BYTE_TO_FLOAT(s[0]);
-   dst[GCOMP] = BYTE_TO_FLOAT(s[1]);
-   dst[BCOMP] = 0;
-   dst[ACOMP] = 0;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT(s[i*2+0]);
+      dst[i][GCOMP] = BYTE_TO_FLOAT(s[i*2+1]);
+      dst[i][BCOMP] = 0;
+      dst[i][ACOMP] = 0;
+   }
 }
 
 static void
-unpack_SIGNED_R8(const void *src, GLfloat dst[4])
+unpack_SIGNED_R8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLbyte s = *((const GLbyte *) src);
-   dst[RCOMP] = BYTE_TO_FLOAT_TEX( s );
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   const GLbyte *s = ((const GLbyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT_TEX( s[i] );
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_RG88_REV(const void *src, GLfloat dst[4])
+unpack_SIGNED_RG88_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLushort *) src);
-   dst[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );
-   dst[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] & 0xff) );
+      dst[i][GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 8) );
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_RGBX8888(const void *src, GLfloat dst[4])
+unpack_SIGNED_RGBX8888(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
-   dst[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
-   dst[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
-   dst[ACOMP] = 1.0f;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 24) );
+      dst[i][GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 16) );
+      dst[i][BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >>  8) );
+      dst[i][ACOMP] = 1.0f;
+   }
 }
 
 static void
-unpack_SIGNED_RGBA8888(const void *src, GLfloat dst[4])
+unpack_SIGNED_RGBA8888(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
-   dst[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
-   dst[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
-   dst[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 24) );
+      dst[i][GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 16) );
+      dst[i][BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >>  8) );
+      dst[i][ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i]      ) );
+   }
 }
 
 static void
-unpack_SIGNED_RGBA8888_REV(const void *src, GLfloat dst[4])
+unpack_SIGNED_RGBA8888_REV(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );
-   dst[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
-   dst[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
-   dst[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i]      ) );
+      dst[i][GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >>  8) );
+      dst[i][BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 16) );
+      dst[i][ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 24) );
+   }
 }
 
 static void
-unpack_SIGNED_R16(const void *src, GLfloat dst[4])
+unpack_SIGNED_R16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLshort s = *((const GLshort *) src);
-   dst[RCOMP] = SHORT_TO_FLOAT_TEX( s );
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   const GLshort *s = ((const GLshort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = SHORT_TO_FLOAT_TEX( s[i] );
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_GR1616(const void *src, GLfloat dst[4])
+unpack_SIGNED_GR1616(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLuint s = *((const GLuint *) src);
-   dst[RCOMP] = SHORT_TO_FLOAT_TEX( s & 0xffff );
-   dst[GCOMP] = SHORT_TO_FLOAT_TEX( s >> 16 );
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = 1.0F;
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = SHORT_TO_FLOAT_TEX( s[i] & 0xffff );
+      dst[i][GCOMP] = SHORT_TO_FLOAT_TEX( s[i] >> 16 );
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_RGB_16(const void *src, GLfloat dst[4])
+unpack_SIGNED_RGB_16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLshort *s = (const GLshort *) src;
-   dst[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
-   dst[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );
-   dst[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = SHORT_TO_FLOAT_TEX( s[i*3+0] );
+      dst[i][GCOMP] = SHORT_TO_FLOAT_TEX( s[i*3+1] );
+      dst[i][BCOMP] = SHORT_TO_FLOAT_TEX( s[i*3+2] );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_RGBA_16(const void *src, GLfloat dst[4])
+unpack_SIGNED_RGBA_16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLshort *s = (const GLshort *) src;
-   dst[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
-   dst[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );
-   dst[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );
-   dst[ACOMP] = SHORT_TO_FLOAT_TEX( s[3] );
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = SHORT_TO_FLOAT_TEX( s[i*4+0] );
+      dst[i][GCOMP] = SHORT_TO_FLOAT_TEX( s[i*4+1] );
+      dst[i][BCOMP] = SHORT_TO_FLOAT_TEX( s[i*4+2] );
+      dst[i][ACOMP] = SHORT_TO_FLOAT_TEX( s[i*4+3] );
+   }
 }
 
 static void
-unpack_RGBA_16(const void *src, GLfloat dst[4])
+unpack_RGBA_16(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLushort *s = (const GLushort *) src;
-   dst[RCOMP] = USHORT_TO_FLOAT( s[0] );
-   dst[GCOMP] = USHORT_TO_FLOAT( s[1] );
-   dst[BCOMP] = USHORT_TO_FLOAT( s[2] );
-   dst[ACOMP] = USHORT_TO_FLOAT( s[3] );
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = USHORT_TO_FLOAT( s[i*4+0] );
+      dst[i][GCOMP] = USHORT_TO_FLOAT( s[i*4+1] );
+      dst[i][BCOMP] = USHORT_TO_FLOAT( s[i*4+2] );
+      dst[i][ACOMP] = USHORT_TO_FLOAT( s[i*4+3] );
+   }
 }
 
 static void
-unpack_RED_RGTC1(const void *src, GLfloat dst[4])
+unpack_RED_RGTC1(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_SIGNED_RED_RGTC1(const void *src, GLfloat dst[4])
+unpack_SIGNED_RED_RGTC1(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_RG_RGTC2(const void *src, GLfloat dst[4])
+unpack_RG_RGTC2(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_SIGNED_RG_RGTC2(const void *src, GLfloat dst[4])
+unpack_SIGNED_RG_RGTC2(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_L_LATC1(const void *src, GLfloat dst[4])
+unpack_L_LATC1(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_SIGNED_L_LATC1(const void *src, GLfloat dst[4])
+unpack_SIGNED_L_LATC1(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_LA_LATC2(const void *src, GLfloat dst[4])
+unpack_LA_LATC2(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_SIGNED_LA_LATC2(const void *src, GLfloat dst[4])
+unpack_SIGNED_LA_LATC2(const void *src, GLfloat dst[][4], GLuint n)
 {
    /* XXX to do */
 }
 
 static void
-unpack_SIGNED_A8(const void *src, GLfloat dst[4])
+unpack_SIGNED_A8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLbyte s = *((const GLbyte *) src);
-   dst[RCOMP] = 0.0F;
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = BYTE_TO_FLOAT_TEX( s );
+   const GLbyte *s = ((const GLbyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = 0.0F;
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = BYTE_TO_FLOAT_TEX( s[i] );
+   }
 }
 
 static void
-unpack_SIGNED_L8(const void *src, GLfloat dst[4])
+unpack_SIGNED_L8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLbyte s = *((const GLbyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = BYTE_TO_FLOAT_TEX( s );
-   dst[ACOMP] = 1.0F;
+   const GLbyte *s = ((const GLbyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = BYTE_TO_FLOAT_TEX( s[i] );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_AL88(const void *src, GLfloat dst[4])
+unpack_SIGNED_AL88(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLushort s = *((const GLshort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );
-   dst[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );
+   const GLshort *s = ((const GLshort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] & 0xff) );
+      dst[i][ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s[i] >> 8) );
+   }
 }
 
 static void
-unpack_SIGNED_I8(const void *src, GLfloat dst[4])
+unpack_SIGNED_I8(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLbyte s = *((const GLbyte *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = BYTE_TO_FLOAT_TEX( s );
+   const GLbyte *s = ((const GLbyte *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = BYTE_TO_FLOAT_TEX( s[i] );
+   }
 }
 
 static void
-unpack_SIGNED_A16(const void *src, GLfloat dst[4])
+unpack_SIGNED_A16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLshort s = *((const GLshort *) src);
-   dst[RCOMP] = 0.0F;
-   dst[GCOMP] = 0.0F;
-   dst[BCOMP] = 0.0F;
-   dst[ACOMP] = SHORT_TO_FLOAT_TEX( s );
+   const GLshort *s = ((const GLshort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] = 0.0F;
+      dst[i][GCOMP] = 0.0F;
+      dst[i][BCOMP] = 0.0F;
+      dst[i][ACOMP] = SHORT_TO_FLOAT_TEX( s[i] );
+   }
 }
 
 static void
-unpack_SIGNED_L16(const void *src, GLfloat dst[4])
+unpack_SIGNED_L16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLshort s = *((const GLshort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = SHORT_TO_FLOAT_TEX( s );
-   dst[ACOMP] = 1.0F;
+   const GLshort *s = ((const GLshort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = SHORT_TO_FLOAT_TEX( s[i] );
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_SIGNED_AL1616(const void *src, GLfloat dst[4])
+unpack_SIGNED_AL1616(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLshort *s = (const GLshort *) src;
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
-   dst[ACOMP] = SHORT_TO_FLOAT_TEX( s[1] );
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] = SHORT_TO_FLOAT_TEX( s[i*2+0] );
+      dst[i][ACOMP] = SHORT_TO_FLOAT_TEX( s[i*2+1] );
+   }
 }
 
 static void
-unpack_SIGNED_I16(const void *src, GLfloat dst[4])
+unpack_SIGNED_I16(const void *src, GLfloat dst[][4], GLuint n)
 {
-   const GLshort s = *((const GLshort *) src);
-   dst[RCOMP] =
-   dst[GCOMP] =
-   dst[BCOMP] =
-   dst[ACOMP] = SHORT_TO_FLOAT_TEX( s );
+   const GLshort *s = ((const GLshort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i][RCOMP] =
+      dst[i][GCOMP] =
+      dst[i][BCOMP] =
+      dst[i][ACOMP] = SHORT_TO_FLOAT_TEX( s[i] );
+   }
 }
 
 static void
-unpack_RGB9_E5_FLOAT(const void *src, GLfloat dst[4])
+unpack_RGB9_E5_FLOAT(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLuint *s = (const GLuint *) src;
-   rgb9e5_to_float3(*s, dst);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      rgb9e5_to_float3(s[i], dst[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 static void
-unpack_R11_G11_B10_FLOAT(const void *src, GLfloat dst[4])
+unpack_R11_G11_B10_FLOAT(const void *src, GLfloat dst[][4], GLuint n)
 {
    const GLuint *s = (const GLuint *) src;
-   r11g11b10f_to_float3(*s, dst);
-   dst[ACOMP] = 1.0F;
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      r11g11b10f_to_float3(s[i], dst[i]);
+      dst[i][ACOMP] = 1.0F;
+   }
 }
 
 
@@ -1248,14 +1534,7 @@ _mesa_unpack_rgba_row(gl_format format, GLuint n,
                       const void *src, GLfloat dst[][4])
 {
    unpack_rgba_func unpack = get_unpack_rgba_function(format);
-   GLuint srcStride = _mesa_get_format_bytes(format);
-   const GLubyte *srcPtr = (GLubyte *) src;
-   GLuint i;
-
-   for (i = 0; i < n; i++) {
-      unpack(srcPtr, dst[i]);
-      srcPtr += srcStride;
-   }
+   unpack(src, dst, n);
 }
 
 static void
@@ -1401,10 +1680,9 @@ _mesa_unpack_rgba_block(gl_format format,
    unpack_rgba_func unpack = get_unpack_rgba_function(format);
    const GLuint srcPixStride = _mesa_get_format_bytes(format);
    const GLuint dstPixStride = 4 * sizeof(GLfloat);
-   const GLubyte *srcRow, *srcPix;
+   const GLubyte *srcRow;
    GLubyte *dstRow;
-   GLfloat *dstPix;
-   GLuint i, j;
+   GLuint i;
 
    /* XXX needs to be fixed for compressed formats */
 
@@ -1412,14 +1690,7 @@ _mesa_unpack_rgba_block(gl_format format,
    dstRow = ((GLubyte *) dst) + dstRowStride * y + dstPixStride * x;
 
    for (i = 0; i < height; i++) {
-      srcPix = srcRow;
-      dstPix = (GLfloat *) dstRow;
-
-      for (j = 0; j < width; j++) {
-         unpack(srcPix, dstPix);
-         srcPix += srcPixStride;
-         dstPix += dstPixStride;
-      }
+      unpack(srcRow, (GLfloat (*)[4]) dstRow, width);
 
       dstRow += dstRowStride;
       srcRow += srcRowStride;
@@ -1429,60 +1700,64 @@ _mesa_unpack_rgba_block(gl_format format,
 
 
 
-typedef void (*unpack_float_z_func)(const void *src, GLfloat *dst);
+typedef void (*unpack_float_z_func)(GLuint n, const void *src, GLfloat *dst);
 
 static void
-unpack_float_z_Z24_S8(const void *src, GLfloat *dst)
+unpack_float_z_Z24_X8(GLuint n, const void *src, GLfloat *dst)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
+   const GLuint *s = ((const GLuint *) src);
    const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
-   *dst = (s >> 8) * scale;
-   ASSERT(*dst >= 0.0F);
-   ASSERT(*dst <= 1.0F);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = (s[i] >> 8) * scale;
+      ASSERT(dst[i] >= 0.0F);
+      ASSERT(dst[i] <= 1.0F);
+   }
 }
 
 static void
-unpack_float_z_S8_Z24(const void *src, GLfloat *dst)
+unpack_float_z_X8_Z24(GLuint n, const void *src, GLfloat *dst)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
+   const GLuint *s = ((const GLuint *) src);
    const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
-   *dst = (s & 0x00ffffff) * scale;
-   ASSERT(*dst >= 0.0F);
-   ASSERT(*dst <= 1.0F);
-}
-
-static void
-unpack_float_z_Z16(const void *src, GLfloat *dst)
-{
-   const GLushort s = *((const GLushort *) src);
-   *dst = s * (1.0F / 65535.0F);
-}
-
-static void
-unpack_float_z_X8_Z24(const void *src, GLfloat *dst)
-{
-   unpack_float_z_S8_Z24(src, dst);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = (s[i] & 0x00ffffff) * scale;
+      ASSERT(dst[i] >= 0.0F);
+      ASSERT(dst[i] <= 1.0F);
+   }
 }
 
 static void
-unpack_float_z_Z24_X8(const void *src, GLfloat *dst)
+unpack_float_z_Z16(GLuint n, const void *src, GLfloat *dst)
 {
-   unpack_float_z_Z24_S8(src, dst);
+   const GLushort *s = ((const GLushort *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = s[i] * (1.0F / 65535.0F);
+   }
 }
 
 static void
-unpack_float_z_Z32(const void *src, GLfloat *dst)
+unpack_float_z_Z32(GLuint n, const void *src, GLfloat *dst)
 {
-   const GLuint s = *((const GLuint *) src);
-   *dst = s * (1.0F / 0xffffffff);
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = s[i] * (1.0F / 0xffffffff);
+   }
 }
 
 static void
-unpack_float_z_Z32X24S8(const void *src, GLfloat *dst)
+unpack_float_z_Z32X24S8(GLuint n, const void *src, GLfloat *dst)
 {
-   *dst = *((const GLfloat *) src);
+   const GLfloat *s = ((const GLfloat *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = s[i * 2];
+   }
 }
 
 
@@ -1492,25 +1767,18 @@ _mesa_unpack_float_z_row(gl_format format, GLuint n,
                          const void *src, GLfloat *dst)
 {
    unpack_float_z_func unpack;
-   GLuint srcStride = _mesa_get_format_bytes(format);
-   const GLubyte *srcPtr = (GLubyte *) src;
-   GLuint i;
 
    switch (format) {
    case MESA_FORMAT_Z24_S8:
-      unpack = unpack_float_z_Z24_S8;
+   case MESA_FORMAT_Z24_X8:
+      unpack = unpack_float_z_Z24_X8;
       break;
    case MESA_FORMAT_S8_Z24:
-      unpack = unpack_float_z_S8_Z24;
-      break;
-   case MESA_FORMAT_Z16:
-      unpack = unpack_float_z_Z16;
-      break;
    case MESA_FORMAT_X8_Z24:
       unpack = unpack_float_z_X8_Z24;
       break;
-   case MESA_FORMAT_Z24_X8:
-      unpack = unpack_float_z_Z24_X8;
+   case MESA_FORMAT_Z16:
+      unpack = unpack_float_z_Z16;
       break;
    case MESA_FORMAT_Z32:
       unpack = unpack_float_z_Z32;
@@ -1524,43 +1792,49 @@ _mesa_unpack_float_z_row(gl_format format, GLuint n,
       return;
    }
 
-   for (i = 0; i < n; i++) {
-      unpack(srcPtr, &dst[i]);
-      srcPtr += srcStride;
-   }
+   unpack(n, src, dst);
 }
 
 
 
-typedef void (*unpack_uint_z_func)(const void *src, GLuint *dst);
+typedef void (*unpack_uint_z_func)(const void *src, GLuint *dst, GLuint n);
 
 static void
-unpack_uint_z_Z24_X8(const void *src, GLuint *dst)
+unpack_uint_z_Z24_X8(const void *src, GLuint *dst, GLuint n)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
-   *dst = (s & 0xffffff00) | (s >> 24);
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = (s[i] & 0xffffff00) | (s[i] >> 24);
+   }
 }
 
 static void
-unpack_uint_z_X8_Z24(const void *src, GLuint *dst)
+unpack_uint_z_X8_Z24(const void *src, GLuint *dst, GLuint n)
 {
    /* only return Z, not stencil data */
-   const GLuint s = *((const GLuint *) src);
-   *dst = (s << 8) | ((s >> 16) & 0xff);
+   const GLuint *s = ((const GLuint *) src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = (s[i] << 8) | ((s[i] >> 16) & 0xff);
+   }
 }
 
 static void
-unpack_uint_z_Z16(const void *src, GLuint *dst)
+unpack_uint_z_Z16(const void *src, GLuint *dst, GLuint n)
 {
-   const GLushort s = *((const GLushort *)src);
-   *dst = (s << 16) | s;
+   const GLushort *s = ((const GLushort *)src);
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      dst[i] = (s[i] << 16) | s[i];
+   }
 }
 
 static void
-unpack_uint_z_Z32(const void *src, GLuint *dst)
+unpack_uint_z_Z32(const void *src, GLuint *dst, GLuint n)
 {
-   *dst = *((const GLuint *) src);
+   memcpy(dst, src, n * sizeof(GLuint));
 }
 
 
@@ -1569,9 +1843,7 @@ _mesa_unpack_uint_z_row(gl_format format, GLuint n,
                         const void *src, GLuint *dst)
 {
    unpack_uint_z_func unpack;
-   GLuint srcStride = _mesa_get_format_bytes(format);
    const GLubyte *srcPtr = (GLubyte *) src;
-   GLuint i;
 
    switch (format) {
    case MESA_FORMAT_Z24_S8:
@@ -1594,12 +1866,10 @@ _mesa_unpack_uint_z_row(gl_format format, GLuint n,
       return;
    }
 
-   for (i = 0; i < n; i++) {
-      unpack(srcPtr, &dst[i]);
-      srcPtr += srcStride;
-   }
+   unpack(srcPtr, dst, n);
 }
 
+
 static void
 unpack_ubyte_s_S8(const void *src, GLubyte *dst, GLuint n)
 {
diff --git a/mesalib/src/mesa/main/image.c b/mesalib/src/mesa/main/image.c
index 7d95dd6be..914a99923 100644
--- a/mesalib/src/mesa/main/image.c
+++ b/mesalib/src/mesa/main/image.c
@@ -39,25 +39,6 @@
 #include "mtypes.h"
 
 
-/**
- * NOTE:
- * Normally, BYTE_TO_FLOAT(0) returns 0.00392  That causes problems when
- * we later convert the float to a packed integer value (such as for
- * GL_RGB5_A1) because we'll wind up with a non-zero value.
- *
- * We redefine the macros here so zero is handled correctly.
- */
-#undef BYTE_TO_FLOAT
-#define BYTE_TO_FLOAT(B)    ((B) == 0 ? 0.0F : ((2.0F * (B) + 1.0F) * (1.0F/255.0F)))
-
-#undef SHORT_TO_FLOAT
-#define SHORT_TO_FLOAT(S)   ((S) == 0 ? 0.0F : ((2.0F * (S) + 1.0F) * (1.0F/65535.0F)))
-
-
-
-/** Compute ceiling of integer quotient of A divided by B. */
-#define CEILING( A, B )  ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )
-
 
 /**
  * \return GL_TRUE if type is packed pixel type, GL_FALSE otherwise.
@@ -195,38 +176,24 @@ _mesa_sizeof_packed_type( GLenum type )
       case GL_FLOAT:
 	 return sizeof(GLfloat);
       case GL_UNSIGNED_BYTE_3_3_2:
-         return sizeof(GLubyte);
       case GL_UNSIGNED_BYTE_2_3_3_REV:
-         return sizeof(GLubyte);
       case MESA_UNSIGNED_BYTE_4_4:
          return sizeof(GLubyte);
       case GL_UNSIGNED_SHORT_5_6_5:
-         return sizeof(GLushort);
       case GL_UNSIGNED_SHORT_5_6_5_REV:
-         return sizeof(GLushort);
       case GL_UNSIGNED_SHORT_4_4_4_4:
-         return sizeof(GLushort);
       case GL_UNSIGNED_SHORT_4_4_4_4_REV:
-         return sizeof(GLushort);
       case GL_UNSIGNED_SHORT_5_5_5_1:
-         return sizeof(GLushort);
       case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+      case GL_UNSIGNED_SHORT_8_8_MESA:
+      case GL_UNSIGNED_SHORT_8_8_REV_MESA:
          return sizeof(GLushort);
       case GL_UNSIGNED_INT_8_8_8_8:
-         return sizeof(GLuint);
       case GL_UNSIGNED_INT_8_8_8_8_REV:
-         return sizeof(GLuint);
       case GL_UNSIGNED_INT_10_10_10_2:
-         return sizeof(GLuint);
       case GL_UNSIGNED_INT_2_10_10_10_REV:
-         return sizeof(GLuint);
-      case GL_UNSIGNED_SHORT_8_8_MESA:
-      case GL_UNSIGNED_SHORT_8_8_REV_MESA:
-         return sizeof(GLushort);      
       case GL_UNSIGNED_INT_24_8_EXT:
-         return sizeof(GLuint);
       case GL_UNSIGNED_INT_5_9_9_9_REV:
-         return sizeof(GLuint);
       case GL_UNSIGNED_INT_10F_11F_11F_REV:
          return sizeof(GLuint);
       case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
@@ -263,29 +230,27 @@ _mesa_components_in_format( GLenum format )
       case GL_LUMINANCE_INTEGER_EXT:
       case GL_INTENSITY:
          return 1;
+
       case GL_LUMINANCE_ALPHA:
       case GL_LUMINANCE_ALPHA_INTEGER_EXT:
       case GL_RG:
+      case GL_YCBCR_MESA:
+      case GL_DEPTH_STENCIL_EXT:
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
 	 return 2;
+
       case GL_RGB:
+      case GL_BGR:
       case GL_RGB_INTEGER_EXT:
 	 return 3;
+
       case GL_RGBA:
-      case GL_RGBA_INTEGER_EXT:
-	 return 4;
-      case GL_BGR:
-	 return 3;
       case GL_BGRA:
-	 return 4;
       case GL_ABGR_EXT:
+      case GL_RGBA_INTEGER_EXT:
          return 4;
-      case GL_YCBCR_MESA:
-         return 2;
-      case GL_DEPTH_STENCIL_EXT:
-         return 2;
-      case GL_DUDV_ATI:
-      case GL_DU8DV8_ATI:
-         return 2;
+
       default:
          return -1;
    }
@@ -987,36 +952,48 @@ _mesa_is_integer_format(GLenum format)
    /* specific integer formats */
    case GL_RGBA32UI_EXT:
    case GL_RGB32UI_EXT:
+   case GL_RG32UI:
+   case GL_R32UI:
    case GL_ALPHA32UI_EXT:
    case GL_INTENSITY32UI_EXT:
    case GL_LUMINANCE32UI_EXT:
    case GL_LUMINANCE_ALPHA32UI_EXT:
    case GL_RGBA16UI_EXT:
    case GL_RGB16UI_EXT:
+   case GL_RG16UI:
+   case GL_R16UI:
    case GL_ALPHA16UI_EXT:
    case GL_INTENSITY16UI_EXT:
    case GL_LUMINANCE16UI_EXT:
    case GL_LUMINANCE_ALPHA16UI_EXT:
    case GL_RGBA8UI_EXT:
    case GL_RGB8UI_EXT:
+   case GL_RG8UI:
+   case GL_R8UI:
    case GL_ALPHA8UI_EXT:
    case GL_INTENSITY8UI_EXT:
    case GL_LUMINANCE8UI_EXT:
    case GL_LUMINANCE_ALPHA8UI_EXT:
    case GL_RGBA32I_EXT:
    case GL_RGB32I_EXT:
+   case GL_RG32I:
+   case GL_R32I:
    case GL_ALPHA32I_EXT:
    case GL_INTENSITY32I_EXT:
    case GL_LUMINANCE32I_EXT:
    case GL_LUMINANCE_ALPHA32I_EXT:
    case GL_RGBA16I_EXT:
    case GL_RGB16I_EXT:
+   case GL_RG16I:
+   case GL_R16I:
    case GL_ALPHA16I_EXT:
    case GL_INTENSITY16I_EXT:
    case GL_LUMINANCE16I_EXT:
    case GL_LUMINANCE_ALPHA16I_EXT:
    case GL_RGBA8I_EXT:
    case GL_RGB8I_EXT:
+   case GL_RG8I:
+   case GL_R8I:
    case GL_ALPHA8I_EXT:
    case GL_INTENSITY8I_EXT:
    case GL_LUMINANCE8I_EXT:
diff --git a/mesalib/src/mesa/main/light.c b/mesalib/src/mesa/main/light.c
index 888e5622e..60daa89a3 100644
--- a/mesalib/src/mesa/main/light.c
+++ b/mesalib/src/mesa/main/light.c
@@ -1,1430 +1,1433 @@
-/*

- * Mesa 3-D graphics library

- * Version:  7.5

- *

- * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.

- * Copyright (C) 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 "Software"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- */

-

-

-#include "glheader.h"

-#include "imports.h"

-#include "context.h"

-#include "enums.h"

-#include "light.h"

-#include "macros.h"

-#include "simple_list.h"

-#include "mtypes.h"

-#include "math/m_matrix.h"

-

-

-void GLAPIENTRY

-_mesa_ShadeModel( GLenum mode )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (MESA_VERBOSE & VERBOSE_API)

-      _mesa_debug(ctx, "glShadeModel %s\n", _mesa_lookup_enum_by_nr(mode));

-

-   if (mode != GL_FLAT && mode != GL_SMOOTH) {

-      _mesa_error(ctx, GL_INVALID_ENUM, "glShadeModel");

-      return;

-   }

-

-   if (ctx->Light.ShadeModel == mode)

-      return;

-

-   FLUSH_VERTICES(ctx, _NEW_LIGHT);

-   ctx->Light.ShadeModel = mode;

-   if (mode == GL_FLAT)

-      ctx->_TriangleCaps |= DD_FLATSHADE;

-   else

-      ctx->_TriangleCaps &= ~DD_FLATSHADE;

-

-   if (ctx->Driver.ShadeModel)

-      ctx->Driver.ShadeModel( ctx, mode );

-}

-

-

-/**

- * Set the provoking vertex (the vertex which specifies the prim's

- * color when flat shading) to either the first or last vertex of the

- * triangle or line.

- */

-void GLAPIENTRY

-_mesa_ProvokingVertexEXT(GLenum mode)

-{

-   GET_CURRENT_CONTEXT(ctx);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (MESA_VERBOSE&VERBOSE_API)

-      _mesa_debug(ctx, "glProvokingVertexEXT 0x%x\n", mode);

-

-   switch (mode) {

-   case GL_FIRST_VERTEX_CONVENTION_EXT:

-   case GL_LAST_VERTEX_CONVENTION_EXT:

-      break;

-   default:

-      _mesa_error(ctx, GL_INVALID_ENUM, "glProvokingVertexEXT(0x%x)", mode);

-      return;

-   }

-

-   if (ctx->Light.ProvokingVertex == mode)

-      return;

-

-   FLUSH_VERTICES(ctx, _NEW_LIGHT);

-   ctx->Light.ProvokingVertex = mode;

-}

-

-

-/**

- * Helper function called by _mesa_Lightfv and _mesa_PopAttrib to set

- * per-light state.

- * For GL_POSITION and GL_SPOT_DIRECTION the params position/direction

- * will have already been transformed by the modelview matrix!

- * Also, all error checking should have already been done.

- */

-void

-_mesa_light(struct gl_context *ctx, GLuint lnum, GLenum pname, const GLfloat *params)

-{

-   struct gl_light *light;

-

-   ASSERT(lnum < MAX_LIGHTS);

-   light = &ctx->Light.Light[lnum];

-

-   switch (pname) {

-   case GL_AMBIENT:

-      if (TEST_EQ_4V(light->Ambient, params))

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      COPY_4V( light->Ambient, params );

-      break;

-   case GL_DIFFUSE:

-      if (TEST_EQ_4V(light->Diffuse, params))

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      COPY_4V( light->Diffuse, params );

-      break;

-   case GL_SPECULAR:

-      if (TEST_EQ_4V(light->Specular, params))

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      COPY_4V( light->Specular, params );

-      break;

-   case GL_POSITION:

-      /* NOTE: position has already been transformed by ModelView! */

-      if (TEST_EQ_4V(light->EyePosition, params))

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      COPY_4V(light->EyePosition, params);

-      if (light->EyePosition[3] != 0.0F)

-	 light->_Flags |= LIGHT_POSITIONAL;

-      else

-	 light->_Flags &= ~LIGHT_POSITIONAL;

-      break;

-   case GL_SPOT_DIRECTION:

-      /* NOTE: Direction already transformed by inverse ModelView! */

-      if (TEST_EQ_3V(light->SpotDirection, params))

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      COPY_3V(light->SpotDirection, params);

-      break;

-   case GL_SPOT_EXPONENT:

-      ASSERT(params[0] >= 0.0);

-      ASSERT(params[0] <= ctx->Const.MaxSpotExponent);

-      if (light->SpotExponent == params[0])

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      light->SpotExponent = params[0];

-      _mesa_invalidate_spot_exp_table(light);

-      break;

-   case GL_SPOT_CUTOFF:

-      ASSERT(params[0] == 180.0 || (params[0] >= 0.0 && params[0] <= 90.0));

-      if (light->SpotCutoff == params[0])

-         return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      light->SpotCutoff = params[0];

-      light->_CosCutoffNeg = (GLfloat) (cos(light->SpotCutoff * DEG2RAD));

-      if (light->_CosCutoffNeg < 0)

-         light->_CosCutoff = 0;

-      else

-         light->_CosCutoff = light->_CosCutoffNeg;

-      if (light->SpotCutoff != 180.0F)

-         light->_Flags |= LIGHT_SPOT;

-      else

-         light->_Flags &= ~LIGHT_SPOT;

-      break;

-   case GL_CONSTANT_ATTENUATION:

-      ASSERT(params[0] >= 0.0);

-      if (light->ConstantAttenuation == params[0])

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      light->ConstantAttenuation = params[0];

-      break;

-   case GL_LINEAR_ATTENUATION:

-      ASSERT(params[0] >= 0.0);

-      if (light->LinearAttenuation == params[0])

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      light->LinearAttenuation = params[0];

-      break;

-   case GL_QUADRATIC_ATTENUATION:

-      ASSERT(params[0] >= 0.0);

-      if (light->QuadraticAttenuation == params[0])

-	 return;

-      FLUSH_VERTICES(ctx, _NEW_LIGHT);

-      light->QuadraticAttenuation = params[0];

-      break;

-   default:

-      _mesa_problem(ctx, "Unexpected pname in _mesa_light()");

-      return;

-   }

-

-   if (ctx->Driver.Lightfv)

-      ctx->Driver.Lightfv( ctx, GL_LIGHT0 + lnum, pname, params );

-}

-

-

-void GLAPIENTRY

-_mesa_Lightf( GLenum light, GLenum pname, GLfloat param )

-{

-   GLfloat fparam[4];

-   fparam[0] = param;

-   fparam[1] = fparam[2] = fparam[3] = 0.0F;

-   _mesa_Lightfv( light, pname, fparam );

-}

-

-

-void GLAPIENTRY

-_mesa_Lightfv( GLenum light, GLenum pname, const GLfloat *params )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLint i = (GLint) (light - GL_LIGHT0);

-   GLfloat temp[4];

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (i < 0 || i >= (GLint) ctx->Const.MaxLights) {

-      _mesa_error( ctx, GL_INVALID_ENUM, "glLight(light=0x%x)", light );

-      return;

-   }

-

-   /* do particular error checks, transformations */

-   switch (pname) {

-   case GL_AMBIENT:

-   case GL_DIFFUSE:

-   case GL_SPECULAR:

-      /* nothing */

-      break;

-   case GL_POSITION:

-      /* transform position by ModelView matrix */

-      TRANSFORM_POINT(temp, ctx->ModelviewMatrixStack.Top->m, params);

-      params = temp;

-      break;

-   case GL_SPOT_DIRECTION:

-      /* transform direction by inverse modelview */

-      if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) {

-	 _math_matrix_analyse(ctx->ModelviewMatrixStack.Top);

-      }

-      TRANSFORM_DIRECTION(temp, params, ctx->ModelviewMatrixStack.Top->m);

-      params = temp;

-      break;

-   case GL_SPOT_EXPONENT:

-      if (params[0] < 0.0 || params[0] > ctx->Const.MaxSpotExponent) {

-	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");

-	 return;

-      }

-      break;

-   case GL_SPOT_CUTOFF:

-      if ((params[0] < 0.0 || params[0] > 90.0) && params[0] != 180.0) {

-	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");

-	 return;

-      }

-      break;

-   case GL_CONSTANT_ATTENUATION:

-      if (params[0] < 0.0) {

-	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");

-	 return;

-      }

-      break;

-   case GL_LINEAR_ATTENUATION:

-      if (params[0] < 0.0) {

-	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");

-	 return;

-      }

-      break;

-   case GL_QUADRATIC_ATTENUATION:

-      if (params[0] < 0.0) {

-	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");

-	 return;

-      }

-      break;

-   default:

-      _mesa_error(ctx, GL_INVALID_ENUM, "glLight(pname=0x%x)", pname);

-      return;

-   }

-

-   _mesa_light(ctx, i, pname, params);

-}

-

-

-void GLAPIENTRY

-_mesa_Lighti( GLenum light, GLenum pname, GLint param )

-{

-   GLint iparam[4];

-   iparam[0] = param;

-   iparam[1] = iparam[2] = iparam[3] = 0;

-   _mesa_Lightiv( light, pname, iparam );

-}

-

-

-void GLAPIENTRY

-_mesa_Lightiv( GLenum light, GLenum pname, const GLint *params )

-{

-   GLfloat fparam[4];

-

-   switch (pname) {

-      case GL_AMBIENT:

-      case GL_DIFFUSE:

-      case GL_SPECULAR:

-         fparam[0] = INT_TO_FLOAT( params[0] );

-         fparam[1] = INT_TO_FLOAT( params[1] );

-         fparam[2] = INT_TO_FLOAT( params[2] );

-         fparam[3] = INT_TO_FLOAT( params[3] );

-         break;

-      case GL_POSITION:

-         fparam[0] = (GLfloat) params[0];

-         fparam[1] = (GLfloat) params[1];

-         fparam[2] = (GLfloat) params[2];

-         fparam[3] = (GLfloat) params[3];

-         break;

-      case GL_SPOT_DIRECTION:

-         fparam[0] = (GLfloat) params[0];

-         fparam[1] = (GLfloat) params[1];

-         fparam[2] = (GLfloat) params[2];

-         break;

-      case GL_SPOT_EXPONENT:

-      case GL_SPOT_CUTOFF:

-      case GL_CONSTANT_ATTENUATION:

-      case GL_LINEAR_ATTENUATION:

-      case GL_QUADRATIC_ATTENUATION:

-         fparam[0] = (GLfloat) params[0];

-         break;

-      default:

-         /* error will be caught later in gl_Lightfv */

-         ;

-   }

-

-   _mesa_Lightfv( light, pname, fparam );

-}

-

-

-

-void GLAPIENTRY

-_mesa_GetLightfv( GLenum light, GLenum pname, GLfloat *params )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLint l = (GLint) (light - GL_LIGHT0);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (l < 0 || l >= (GLint) ctx->Const.MaxLights) {

-      _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightfv" );

-      return;

-   }

-

-   switch (pname) {

-      case GL_AMBIENT:

-         COPY_4V( params, ctx->Light.Light[l].Ambient );

-         break;

-      case GL_DIFFUSE:

-         COPY_4V( params, ctx->Light.Light[l].Diffuse );

-         break;

-      case GL_SPECULAR:

-         COPY_4V( params, ctx->Light.Light[l].Specular );

-         break;

-      case GL_POSITION:

-         COPY_4V( params, ctx->Light.Light[l].EyePosition );

-         break;

-      case GL_SPOT_DIRECTION:

-         COPY_3V( params, ctx->Light.Light[l].SpotDirection );

-         break;

-      case GL_SPOT_EXPONENT:

-         params[0] = ctx->Light.Light[l].SpotExponent;

-         break;

-      case GL_SPOT_CUTOFF:

-         params[0] = ctx->Light.Light[l].SpotCutoff;

-         break;

-      case GL_CONSTANT_ATTENUATION:

-         params[0] = ctx->Light.Light[l].ConstantAttenuation;

-         break;

-      case GL_LINEAR_ATTENUATION:

-         params[0] = ctx->Light.Light[l].LinearAttenuation;

-         break;

-      case GL_QUADRATIC_ATTENUATION:

-         params[0] = ctx->Light.Light[l].QuadraticAttenuation;

-         break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightfv" );

-         break;

-   }

-}

-

-

-void GLAPIENTRY

-_mesa_GetLightiv( GLenum light, GLenum pname, GLint *params )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLint l = (GLint) (light - GL_LIGHT0);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (l < 0 || l >= (GLint) ctx->Const.MaxLights) {

-      _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightiv" );

-      return;

-   }

-

-   switch (pname) {

-      case GL_AMBIENT:

-         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[0]);

-         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[1]);

-         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[2]);

-         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[3]);

-         break;

-      case GL_DIFFUSE:

-         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[0]);

-         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[1]);

-         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[2]);

-         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[3]);

-         break;

-      case GL_SPECULAR:

-         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[0]);

-         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[1]);

-         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[2]);

-         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[3]);

-         break;

-      case GL_POSITION:

-         params[0] = (GLint) ctx->Light.Light[l].EyePosition[0];

-         params[1] = (GLint) ctx->Light.Light[l].EyePosition[1];

-         params[2] = (GLint) ctx->Light.Light[l].EyePosition[2];

-         params[3] = (GLint) ctx->Light.Light[l].EyePosition[3];

-         break;

-      case GL_SPOT_DIRECTION:

-         params[0] = (GLint) ctx->Light.Light[l].SpotDirection[0];

-         params[1] = (GLint) ctx->Light.Light[l].SpotDirection[1];

-         params[2] = (GLint) ctx->Light.Light[l].SpotDirection[2];

-         break;

-      case GL_SPOT_EXPONENT:

-         params[0] = (GLint) ctx->Light.Light[l].SpotExponent;

-         break;

-      case GL_SPOT_CUTOFF:

-         params[0] = (GLint) ctx->Light.Light[l].SpotCutoff;

-         break;

-      case GL_CONSTANT_ATTENUATION:

-         params[0] = (GLint) ctx->Light.Light[l].ConstantAttenuation;

-         break;

-      case GL_LINEAR_ATTENUATION:

-         params[0] = (GLint) ctx->Light.Light[l].LinearAttenuation;

-         break;

-      case GL_QUADRATIC_ATTENUATION:

-         params[0] = (GLint) ctx->Light.Light[l].QuadraticAttenuation;

-         break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightiv" );

-         break;

-   }

-}

-

-

-

-/**********************************************************************/

-/***                        Light Model                             ***/

-/**********************************************************************/

-

-

-void GLAPIENTRY

-_mesa_LightModelfv( GLenum pname, const GLfloat *params )

-{

-   GLenum newenum;

-   GLboolean newbool;

-   GET_CURRENT_CONTEXT(ctx);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   switch (pname) {

-      case GL_LIGHT_MODEL_AMBIENT:

-         if (TEST_EQ_4V( ctx->Light.Model.Ambient, params ))

-	    return;

-	 FLUSH_VERTICES(ctx, _NEW_LIGHT);

-         COPY_4V( ctx->Light.Model.Ambient, params );

-         break;

-      case GL_LIGHT_MODEL_LOCAL_VIEWER:

-         newbool = (params[0]!=0.0);

-	 if (ctx->Light.Model.LocalViewer == newbool)

-	    return;

-	 FLUSH_VERTICES(ctx, _NEW_LIGHT);

-	 ctx->Light.Model.LocalViewer = newbool;

-         break;

-      case GL_LIGHT_MODEL_TWO_SIDE:

-         newbool = (params[0]!=0.0);

-	 if (ctx->Light.Model.TwoSide == newbool)

-	    return;

-	 FLUSH_VERTICES(ctx, _NEW_LIGHT);

-	 ctx->Light.Model.TwoSide = newbool;

-         if (ctx->Light.Enabled && ctx->Light.Model.TwoSide)

-            ctx->_TriangleCaps |= DD_TRI_LIGHT_TWOSIDE;

-         else

-            ctx->_TriangleCaps &= ~DD_TRI_LIGHT_TWOSIDE;

-         break;

-      case GL_LIGHT_MODEL_COLOR_CONTROL:

-         if (params[0] == (GLfloat) GL_SINGLE_COLOR)

-	    newenum = GL_SINGLE_COLOR;

-         else if (params[0] == (GLfloat) GL_SEPARATE_SPECULAR_COLOR)

-	    newenum = GL_SEPARATE_SPECULAR_COLOR;

-	 else {

-            _mesa_error( ctx, GL_INVALID_ENUM, "glLightModel(param=0x0%x)",

-                         (GLint) params[0] );

-	    return;

-         }

-	 if (ctx->Light.Model.ColorControl == newenum)

-	    return;

-	 FLUSH_VERTICES(ctx, _NEW_LIGHT);

-	 ctx->Light.Model.ColorControl = newenum;

-         break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "glLightModel(pname=0x%x)", pname );

-         break;

-   }

-

-   if (ctx->Driver.LightModelfv)

-      ctx->Driver.LightModelfv( ctx, pname, params );

-}

-

-

-void GLAPIENTRY

-_mesa_LightModeliv( GLenum pname, const GLint *params )

-{

-   GLfloat fparam[4];

-

-   switch (pname) {

-      case GL_LIGHT_MODEL_AMBIENT:

-         fparam[0] = INT_TO_FLOAT( params[0] );

-         fparam[1] = INT_TO_FLOAT( params[1] );

-         fparam[2] = INT_TO_FLOAT( params[2] );

-         fparam[3] = INT_TO_FLOAT( params[3] );

-         break;

-      case GL_LIGHT_MODEL_LOCAL_VIEWER:

-      case GL_LIGHT_MODEL_TWO_SIDE:

-      case GL_LIGHT_MODEL_COLOR_CONTROL:

-         fparam[0] = (GLfloat) params[0];

-         break;

-      default:

-         /* Error will be caught later in gl_LightModelfv */

-         ASSIGN_4V(fparam, 0.0F, 0.0F, 0.0F, 0.0F);

-   }

-   _mesa_LightModelfv( pname, fparam );

-}

-

-

-void GLAPIENTRY

-_mesa_LightModeli( GLenum pname, GLint param )

-{

-   GLint iparam[4];

-   iparam[0] = param;

-   iparam[1] = iparam[2] = iparam[3] = 0;

-   _mesa_LightModeliv( pname, iparam );

-}

-

-

-void GLAPIENTRY

-_mesa_LightModelf( GLenum pname, GLfloat param )

-{

-   GLfloat fparam[4];

-   fparam[0] = param;

-   fparam[1] = fparam[2] = fparam[3] = 0.0F;

-   _mesa_LightModelfv( pname, fparam );

-}

-

-

-

-/********** MATERIAL **********/

-

-

-/*

- * Given a face and pname value (ala glColorMaterial), compute a bitmask

- * of the targeted material values.

- */

-GLuint

-_mesa_material_bitmask( struct gl_context *ctx, GLenum face, GLenum pname,

-                        GLuint legal, const char *where )

-{

-   GLuint bitmask = 0;

-

-   /* Make a bitmask indicating what material attribute(s) we're updating */

-   switch (pname) {

-      case GL_EMISSION:

-         bitmask |= MAT_BIT_FRONT_EMISSION | MAT_BIT_BACK_EMISSION;

-         break;

-      case GL_AMBIENT:

-         bitmask |= MAT_BIT_FRONT_AMBIENT | MAT_BIT_BACK_AMBIENT;

-         break;

-      case GL_DIFFUSE:

-         bitmask |= MAT_BIT_FRONT_DIFFUSE | MAT_BIT_BACK_DIFFUSE;

-         break;

-      case GL_SPECULAR:

-         bitmask |= MAT_BIT_FRONT_SPECULAR | MAT_BIT_BACK_SPECULAR;

-         break;

-      case GL_SHININESS:

-         bitmask |= MAT_BIT_FRONT_SHININESS | MAT_BIT_BACK_SHININESS;

-         break;

-      case GL_AMBIENT_AND_DIFFUSE:

-         bitmask |= MAT_BIT_FRONT_AMBIENT | MAT_BIT_BACK_AMBIENT;

-         bitmask |= MAT_BIT_FRONT_DIFFUSE | MAT_BIT_BACK_DIFFUSE;

-         break;

-      case GL_COLOR_INDEXES:

-         bitmask |= MAT_BIT_FRONT_INDEXES  | MAT_BIT_BACK_INDEXES;

-         break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );

-         return 0;

-   }

-

-   if (face==GL_FRONT) {

-      bitmask &= FRONT_MATERIAL_BITS;

-   }

-   else if (face==GL_BACK) {

-      bitmask &= BACK_MATERIAL_BITS;

-   }

-   else if (face != GL_FRONT_AND_BACK) {

-      _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );

-      return 0;

-   }

-

-   if (bitmask & ~legal) {

-      _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );

-      return 0;

-   }

-

-   return bitmask;

-}

-

-

-

-/* Perform a straight copy between materials.

- */

-void

-_mesa_copy_materials( struct gl_material *dst,

-                      const struct gl_material *src,

-                      GLuint bitmask )

-{

-   int i;

-

-   for (i = 0 ; i < MAT_ATTRIB_MAX ; i++) 

-      if (bitmask & (1<<i))

-	 COPY_4FV( dst->Attrib[i], src->Attrib[i] );

-}

-

-

-

-/* Update derived values following a change in ctx->Light.Material

- */

-void

-_mesa_update_material( struct gl_context *ctx, GLuint bitmask )

-{

-   struct gl_light *light, *list = &ctx->Light.EnabledList;

-   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;

-

-   if (MESA_VERBOSE & VERBOSE_MATERIAL) 

-      _mesa_debug(ctx, "_mesa_update_material, mask 0x%x\n", bitmask);

-

-   if (!bitmask)

-      return;

-

-   /* update material ambience */

-   if (bitmask & MAT_BIT_FRONT_AMBIENT) {

-      foreach (light, list) {

-         SCALE_3V( light->_MatAmbient[0], light->Ambient, 

-		   mat[MAT_ATTRIB_FRONT_AMBIENT]);

-      }

-   }

-

-   if (bitmask & MAT_BIT_BACK_AMBIENT) {

-      foreach (light, list) {

-         SCALE_3V( light->_MatAmbient[1], light->Ambient, 

-		   mat[MAT_ATTRIB_BACK_AMBIENT]);

-      }

-   }

-

-   /* update BaseColor = emission + scene's ambience * material's ambience */

-   if (bitmask & (MAT_BIT_FRONT_EMISSION | MAT_BIT_FRONT_AMBIENT)) {

-      COPY_3V( ctx->Light._BaseColor[0], mat[MAT_ATTRIB_FRONT_EMISSION] );

-      ACC_SCALE_3V( ctx->Light._BaseColor[0], mat[MAT_ATTRIB_FRONT_AMBIENT],

-		    ctx->Light.Model.Ambient );

-   }

-

-   if (bitmask & (MAT_BIT_BACK_EMISSION | MAT_BIT_BACK_AMBIENT)) {

-      COPY_3V( ctx->Light._BaseColor[1], mat[MAT_ATTRIB_BACK_EMISSION] );

-      ACC_SCALE_3V( ctx->Light._BaseColor[1], mat[MAT_ATTRIB_BACK_AMBIENT],

-		    ctx->Light.Model.Ambient );

-   }

-

-   /* update material diffuse values */

-   if (bitmask & MAT_BIT_FRONT_DIFFUSE) {

-      foreach (light, list) {

-	 SCALE_3V( light->_MatDiffuse[0], light->Diffuse, 

-		   mat[MAT_ATTRIB_FRONT_DIFFUSE] );

-      }

-   }

-

-   if (bitmask & MAT_BIT_BACK_DIFFUSE) {

-      foreach (light, list) {

-	 SCALE_3V( light->_MatDiffuse[1], light->Diffuse, 

-		   mat[MAT_ATTRIB_BACK_DIFFUSE] );

-      }

-   }

-

-   /* update material specular values */

-   if (bitmask & MAT_BIT_FRONT_SPECULAR) {

-      foreach (light, list) {

-	 SCALE_3V( light->_MatSpecular[0], light->Specular, 

-		   mat[MAT_ATTRIB_FRONT_SPECULAR]);

-      }

-   }

-

-   if (bitmask & MAT_BIT_BACK_SPECULAR) {

-      foreach (light, list) {

-	 SCALE_3V( light->_MatSpecular[1], light->Specular,

-		   mat[MAT_ATTRIB_BACK_SPECULAR]);

-      }

-   }

-

-   if (bitmask & MAT_BIT_FRONT_SHININESS) {

-      _mesa_invalidate_shine_table( ctx, 0 );

-   }

-

-   if (bitmask & MAT_BIT_BACK_SHININESS) {

-      _mesa_invalidate_shine_table( ctx, 1 );

-   }

-}

-

-

-/*

- * Update the current materials from the given rgba color

- * according to the bitmask in ColorMaterialBitmask, which is

- * set by glColorMaterial().

- */

-void

-_mesa_update_color_material( struct gl_context *ctx, const GLfloat color[4] )

-{

-   GLuint bitmask = ctx->Light.ColorMaterialBitmask;

-   struct gl_material *mat = &ctx->Light.Material;

-   int i;

-

-   for (i = 0 ; i < MAT_ATTRIB_MAX ; i++) 

-      if (bitmask & (1<<i))

-	 COPY_4FV( mat->Attrib[i], color );

-

-   _mesa_update_material( ctx, bitmask );

-}

-

-

-void GLAPIENTRY

-_mesa_ColorMaterial( GLenum face, GLenum mode )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLuint bitmask;

-   GLuint legal = (MAT_BIT_FRONT_EMISSION | MAT_BIT_BACK_EMISSION |

-		   MAT_BIT_FRONT_SPECULAR | MAT_BIT_BACK_SPECULAR |

-		   MAT_BIT_FRONT_DIFFUSE  | MAT_BIT_BACK_DIFFUSE  |

-		   MAT_BIT_FRONT_AMBIENT  | MAT_BIT_BACK_AMBIENT);

-   ASSERT_OUTSIDE_BEGIN_END(ctx);

-

-   if (MESA_VERBOSE&VERBOSE_API)

-      _mesa_debug(ctx, "glColorMaterial %s %s\n",

-                  _mesa_lookup_enum_by_nr(face),

-                  _mesa_lookup_enum_by_nr(mode));

-

-   bitmask = _mesa_material_bitmask(ctx, face, mode, legal, "glColorMaterial");

-

-   if (ctx->Light.ColorMaterialBitmask == bitmask &&

-       ctx->Light.ColorMaterialFace == face &&

-       ctx->Light.ColorMaterialMode == mode)

-      return;

-

-   FLUSH_VERTICES(ctx, _NEW_LIGHT);

-   ctx->Light.ColorMaterialBitmask = bitmask;

-   ctx->Light.ColorMaterialFace = face;

-   ctx->Light.ColorMaterialMode = mode;

-

-   if (ctx->Light.ColorMaterialEnabled) {

-      FLUSH_CURRENT( ctx, 0 );

-      _mesa_update_color_material(ctx,ctx->Current.Attrib[VERT_ATTRIB_COLOR0]);

-   }

-

-   if (ctx->Driver.ColorMaterial)

-      ctx->Driver.ColorMaterial( ctx, face, mode );

-}

-

-

-void GLAPIENTRY

-_mesa_GetMaterialfv( GLenum face, GLenum pname, GLfloat *params )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLuint f;

-   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;

-   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); /* update materials */

-

-   FLUSH_CURRENT(ctx, 0); /* update ctx->Light.Material from vertex buffer */

-

-   if (face==GL_FRONT) {

-      f = 0;

-   }

-   else if (face==GL_BACK) {

-      f = 1;

-   }

-   else {

-      _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(face)" );

-      return;

-   }

-

-   switch (pname) {

-      case GL_AMBIENT:

-         COPY_4FV( params, mat[MAT_ATTRIB_AMBIENT(f)] );

-         break;

-      case GL_DIFFUSE:

-         COPY_4FV( params, mat[MAT_ATTRIB_DIFFUSE(f)] );

-	 break;

-      case GL_SPECULAR:

-         COPY_4FV( params, mat[MAT_ATTRIB_SPECULAR(f)] );

-	 break;

-      case GL_EMISSION:

-	 COPY_4FV( params, mat[MAT_ATTRIB_EMISSION(f)] );

-	 break;

-      case GL_SHININESS:

-	 *params = mat[MAT_ATTRIB_SHININESS(f)][0];

-	 break;

-      case GL_COLOR_INDEXES:

-	 params[0] = mat[MAT_ATTRIB_INDEXES(f)][0];

-	 params[1] = mat[MAT_ATTRIB_INDEXES(f)][1];

-	 params[2] = mat[MAT_ATTRIB_INDEXES(f)][2];

-	 break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(pname)" );

-   }

-}

-

-

-void GLAPIENTRY

-_mesa_GetMaterialiv( GLenum face, GLenum pname, GLint *params )

-{

-   GET_CURRENT_CONTEXT(ctx);

-   GLuint f;

-   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;

-   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); /* update materials */

-

-   FLUSH_CURRENT(ctx, 0); /* update ctx->Light.Material from vertex buffer */

-

-   if (face==GL_FRONT) {

-      f = 0;

-   }

-   else if (face==GL_BACK) {

-      f = 1;

-   }

-   else {

-      _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialiv(face)" );

-      return;

-   }

-   switch (pname) {

-      case GL_AMBIENT:

-         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][0] );

-         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][1] );

-         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][2] );

-         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][3] );

-         break;

-      case GL_DIFFUSE:

-         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][0] );

-         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][1] );

-         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][2] );

-         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][3] );

-	 break;

-      case GL_SPECULAR:

-         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][0] );

-         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][1] );

-         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][2] );

-         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][3] );

-	 break;

-      case GL_EMISSION:

-         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][0] );

-         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][1] );

-         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][2] );

-         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][3] );

-	 break;

-      case GL_SHININESS:

-         *params = IROUND( mat[MAT_ATTRIB_SHININESS(f)][0] );

-	 break;

-      case GL_COLOR_INDEXES:

-	 params[0] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][0] );

-	 params[1] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][1] );

-	 params[2] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][2] );

-	 break;

-      default:

-         _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(pname)" );

-   }

-}

-

-

-

-/**********************************************************************/

-/*****                  Lighting computation                      *****/

-/**********************************************************************/

-

-

-/*

- * Notes:

- *   When two-sided lighting is enabled we compute the color (or index)

- *   for both the front and back side of the primitive.  Then, when the

- *   orientation of the facet is later learned, we can determine which

- *   color (or index) to use for rendering.

- *

- *   KW: We now know orientation in advance and only shade for

- *       the side or sides which are actually required.

- *

- * Variables:

- *   n = normal vector

- *   V = vertex position

- *   P = light source position

- *   Pe = (0,0,0,1)

- *

- * Precomputed:

- *   IF P[3]==0 THEN

- *       // light at infinity

- *       IF local_viewer THEN

- *           _VP_inf_norm = unit vector from V to P      // Precompute

- *       ELSE

- *           // eye at infinity

- *           _h_inf_norm = Normalize( VP + <0,0,1> )     // Precompute

- *       ENDIF

- *   ENDIF

- *

- * Functions:

- *   Normalize( v ) = normalized vector v

- *   Magnitude( v ) = length of vector v

- */

-

-

-

-/*

- * Whenever the spotlight exponent for a light changes we must call

- * this function to recompute the exponent lookup table.

- */

-void

-_mesa_invalidate_spot_exp_table( struct gl_light *l )

-{

-   l->_SpotExpTable[0][0] = -1;

-}

-

-

-static void

-validate_spot_exp_table( struct gl_light *l )

-{

-   GLint i;

-   GLdouble exponent = l->SpotExponent;

-   GLdouble tmp = 0;

-   GLint clamp = 0;

-

-   l->_SpotExpTable[0][0] = 0.0;

-

-   for (i = EXP_TABLE_SIZE - 1; i > 0 ;i--) {

-      if (clamp == 0) {

-	 tmp = pow(i / (GLdouble) (EXP_TABLE_SIZE - 1), exponent);

-	 if (tmp < FLT_MIN * 100.0) {

-	    tmp = 0.0;

-	    clamp = 1;

-	 }

-      }

-      l->_SpotExpTable[i][0] = (GLfloat) tmp;

-   }

-   for (i = 0; i < EXP_TABLE_SIZE - 1; i++) {

-      l->_SpotExpTable[i][1] = (l->_SpotExpTable[i+1][0] -

-				l->_SpotExpTable[i][0]);

-   }

-   l->_SpotExpTable[EXP_TABLE_SIZE-1][1] = 0.0;

-}

-

-

-

-/* Calculate a new shine table.  Doing this here saves a branch in

- * lighting, and the cost of doing it early may be partially offset

- * by keeping a MRU cache of shine tables for various shine values.

- */

-void

-_mesa_invalidate_shine_table( struct gl_context *ctx, GLuint side )

-{

-   ASSERT(side < 2);

-   if (ctx->_ShineTable[side])

-      ctx->_ShineTable[side]->refcount--;

-   ctx->_ShineTable[side] = NULL;

-}

-

-

-static void

-validate_shine_table( struct gl_context *ctx, GLuint side, GLfloat shininess )

-{

-   struct gl_shine_tab *list = ctx->_ShineTabList;

-   struct gl_shine_tab *s;

-

-   ASSERT(side < 2);

-

-   foreach(s, list)

-      if ( s->shininess == shininess )

-	 break;

-

-   if (s == list) {

-      GLint j;

-      GLfloat *m;

-

-      foreach(s, list)

-	 if (s->refcount == 0)

-	    break;

-

-      m = s->tab;

-      m[0] = 0.0;

-      if (shininess == 0.0) {

-	 for (j = 1 ; j <= SHINE_TABLE_SIZE ; j++)

-	    m[j] = 1.0;

-      }

-      else {

-	 for (j = 1 ; j < SHINE_TABLE_SIZE ; j++) {

-            GLdouble t, x = j / (GLfloat) (SHINE_TABLE_SIZE - 1);

-            if (x < 0.005) /* underflow check */

-               x = 0.005;

-            t = pow(x, shininess);

-	    if (t > 1e-20)

-	       m[j] = (GLfloat) t;

-	    else

-	       m[j] = 0.0;

-	 }

-	 m[SHINE_TABLE_SIZE] = 1.0;

-      }

-

-      s->shininess = shininess;

-   }

-

-   if (ctx->_ShineTable[side])

-      ctx->_ShineTable[side]->refcount--;

-

-   ctx->_ShineTable[side] = s;

-   move_to_tail( list, s );

-   s->refcount++;

-}

-

-

-void

-_mesa_validate_all_lighting_tables( struct gl_context *ctx )

-{

-   GLuint i;

-   GLfloat shininess;

-   

-   shininess = ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_SHININESS][0];

-   if (!ctx->_ShineTable[0] || ctx->_ShineTable[0]->shininess != shininess)

-      validate_shine_table( ctx, 0, shininess );

-

-   shininess = ctx->Light.Material.Attrib[MAT_ATTRIB_BACK_SHININESS][0];

-   if (!ctx->_ShineTable[1] || ctx->_ShineTable[1]->shininess != shininess)

-      validate_shine_table( ctx, 1, shininess );

-

-   for (i = 0; i < ctx->Const.MaxLights; i++)

-      if (ctx->Light.Light[i]._SpotExpTable[0][0] == -1)

-	 validate_spot_exp_table( &ctx->Light.Light[i] );

-}

-

-

-/**

- * Examine current lighting parameters to determine if the optimized lighting

- * function can be used.

- * Also, precompute some lighting values such as the products of light

- * source and material ambient, diffuse and specular coefficients.

- */

-void

-_mesa_update_lighting( struct gl_context *ctx )

-{

-   struct gl_light *light;

-   ctx->Light._NeedEyeCoords = GL_FALSE;

-   ctx->Light._Flags = 0;

-

-   if (!ctx->Light.Enabled)

-      return;

-

-   foreach(light, &ctx->Light.EnabledList) {

-      ctx->Light._Flags |= light->_Flags;

-   }

-

-   ctx->Light._NeedVertices =

-      ((ctx->Light._Flags & (LIGHT_POSITIONAL|LIGHT_SPOT)) ||

-       ctx->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR ||

-       ctx->Light.Model.LocalViewer);

-

-   ctx->Light._NeedEyeCoords = ((ctx->Light._Flags & LIGHT_POSITIONAL) ||

-				ctx->Light.Model.LocalViewer);

-

-   /* XXX: This test is overkill & needs to be fixed both for software and

-    * hardware t&l drivers.  The above should be sufficient & should

-    * be tested to verify this.

-    */

-   if (ctx->Light._NeedVertices)

-      ctx->Light._NeedEyeCoords = GL_TRUE;

-

-   /* Precompute some shading values.  Although we reference

-    * Light.Material here, we can get away without flushing

-    * FLUSH_UPDATE_CURRENT, as when any outstanding material changes

-    * are flushed, they will update the derived state at that time.

-    */

-   if (ctx->Light.Model.TwoSide)

-      _mesa_update_material(ctx,

-			    MAT_BIT_FRONT_EMISSION |

-			    MAT_BIT_FRONT_AMBIENT |

-			    MAT_BIT_FRONT_DIFFUSE |

-			    MAT_BIT_FRONT_SPECULAR |

-			    MAT_BIT_BACK_EMISSION |

-			    MAT_BIT_BACK_AMBIENT |

-			    MAT_BIT_BACK_DIFFUSE |

-			    MAT_BIT_BACK_SPECULAR);

-   else

-      _mesa_update_material(ctx,

-			    MAT_BIT_FRONT_EMISSION |

-			    MAT_BIT_FRONT_AMBIENT |

-			    MAT_BIT_FRONT_DIFFUSE |

-			    MAT_BIT_FRONT_SPECULAR);

-}

-

-

-/**

- * Update state derived from light position, spot direction.

- * Called upon:

- *   _NEW_MODELVIEW

- *   _NEW_LIGHT

- *   _TNL_NEW_NEED_EYE_COORDS

- *

- * Update on (_NEW_MODELVIEW | _NEW_LIGHT) when lighting is enabled.

- * Also update on lighting space changes.

- */

-static void

-compute_light_positions( struct gl_context *ctx )

-{

-   struct gl_light *light;

-   static const GLfloat eye_z[3] = { 0, 0, 1 };

-

-   if (!ctx->Light.Enabled)

-      return;

-

-   if (ctx->_NeedEyeCoords) {

-      COPY_3V( ctx->_EyeZDir, eye_z );

-   }

-   else {

-      TRANSFORM_NORMAL( ctx->_EyeZDir, eye_z, ctx->ModelviewMatrixStack.Top->m );

-   }

-

-   foreach (light, &ctx->Light.EnabledList) {

-

-      if (ctx->_NeedEyeCoords) {

-         /* _Position is in eye coordinate space */

-	 COPY_4FV( light->_Position, light->EyePosition );

-      }

-      else {

-         /* _Position is in object coordinate space */

-	 TRANSFORM_POINT( light->_Position, ctx->ModelviewMatrixStack.Top->inv,

-			  light->EyePosition );

-      }

-

-      if (!(light->_Flags & LIGHT_POSITIONAL)) {

-	 /* VP (VP) = Normalize( Position ) */

-	 COPY_3V( light->_VP_inf_norm, light->_Position );

-	 NORMALIZE_3FV( light->_VP_inf_norm );

-

-	 if (!ctx->Light.Model.LocalViewer) {

-	    /* _h_inf_norm = Normalize( V_to_P + <0,0,1> ) */

-	    ADD_3V( light->_h_inf_norm, light->_VP_inf_norm, ctx->_EyeZDir);

-	    NORMALIZE_3FV( light->_h_inf_norm );

-	 }

-	 light->_VP_inf_spot_attenuation = 1.0;

-      }

-      else {

-         /* positional light w/ homogeneous coordinate, divide by W */

-         GLfloat wInv = (GLfloat)1.0 / light->_Position[3];

-         light->_Position[0] *= wInv;

-         light->_Position[1] *= wInv;

-         light->_Position[2] *= wInv;

-      }

-

-      if (light->_Flags & LIGHT_SPOT) {

-         /* Note: we normalize the spot direction now */

-

-	 if (ctx->_NeedEyeCoords) {

-	    COPY_3V( light->_NormSpotDirection, light->SpotDirection );

-            NORMALIZE_3FV( light->_NormSpotDirection );

-	 }

-         else {

-            GLfloat spotDir[3];

-            COPY_3V(spotDir, light->SpotDirection);

-            NORMALIZE_3FV(spotDir);

-	    TRANSFORM_NORMAL( light->_NormSpotDirection,

-			      spotDir,

-			      ctx->ModelviewMatrixStack.Top->m);

-	 }

-

-	 NORMALIZE_3FV( light->_NormSpotDirection );

-

-	 if (!(light->_Flags & LIGHT_POSITIONAL)) {

-	    GLfloat PV_dot_dir = - DOT3(light->_VP_inf_norm,

-					light->_NormSpotDirection);

-

-	    if (PV_dot_dir > light->_CosCutoff) {

-	       double x = PV_dot_dir * (EXP_TABLE_SIZE-1);

-	       int k = (int) x;

-	       light->_VP_inf_spot_attenuation =

-		  (GLfloat) (light->_SpotExpTable[k][0] +

-		   (x-k)*light->_SpotExpTable[k][1]);

-	    }

-	    else {

-	       light->_VP_inf_spot_attenuation = 0;

-            }

-	 }

-      }

-   }

-}

-

-

-

-static void

-update_modelview_scale( struct gl_context *ctx )

-{

-   ctx->_ModelViewInvScale = 1.0F;

-   if (!_math_matrix_is_length_preserving(ctx->ModelviewMatrixStack.Top)) {

-      const GLfloat *m = ctx->ModelviewMatrixStack.Top->inv;

-      GLfloat f = m[2] * m[2] + m[6] * m[6] + m[10] * m[10];

-      if (f < 1e-12) f = 1.0;

-      if (ctx->_NeedEyeCoords)

-	 ctx->_ModelViewInvScale = (GLfloat) INV_SQRTF(f);

-      else

-	 ctx->_ModelViewInvScale = (GLfloat) SQRTF(f);

-   }

-}

-

-

-/**

- * Bring up to date any state that relies on _NeedEyeCoords.

- */

-void

-_mesa_update_tnl_spaces( struct gl_context *ctx, GLuint new_state )

-{

-   const GLuint oldneedeyecoords = ctx->_NeedEyeCoords;

-

-   (void) new_state;

-   ctx->_NeedEyeCoords = GL_FALSE;

-

-   if (ctx->_ForceEyeCoords ||

-       (ctx->Texture._GenFlags & TEXGEN_NEED_EYE_COORD) ||

-       ctx->Point._Attenuated ||

-       ctx->Light._NeedEyeCoords)

-      ctx->_NeedEyeCoords = GL_TRUE;

-

-   if (ctx->Light.Enabled &&

-       !_math_matrix_is_length_preserving(ctx->ModelviewMatrixStack.Top))

-      ctx->_NeedEyeCoords = GL_TRUE;

-

-   /* Check if the truth-value interpretations of the bitfields have

-    * changed:

-    */

-   if (oldneedeyecoords != ctx->_NeedEyeCoords) {

-      /* Recalculate all state that depends on _NeedEyeCoords.

-       */

-      update_modelview_scale(ctx);

-      compute_light_positions( ctx );

-

-      if (ctx->Driver.LightingSpaceChange)

-	 ctx->Driver.LightingSpaceChange( ctx );

-   }

-   else {

-      GLuint new_state2 = ctx->NewState;

-

-      /* Recalculate that same state only if it has been invalidated

-       * by other statechanges.

-       */

-      if (new_state2 & _NEW_MODELVIEW)

-	 update_modelview_scale(ctx);

-

-      if (new_state2 & (_NEW_LIGHT|_NEW_MODELVIEW))

-	 compute_light_positions( ctx );

-   }

-}

-

-

-/**

- * Drivers may need this if the hardware tnl unit doesn't support the

- * light-in-modelspace optimization.  It's also useful for debugging.

- */

-void

-_mesa_allow_light_in_model( struct gl_context *ctx, GLboolean flag )

-{

-   ctx->_ForceEyeCoords = !flag;

-   ctx->NewState |= _NEW_POINT;	/* one of the bits from

-				 * _MESA_NEW_NEED_EYE_COORDS.

-				 */

-}

-

-

-

-/**********************************************************************/

-/*****                      Initialization                        *****/

-/**********************************************************************/

-

-/**

- * Initialize the n-th light data structure.

- *

- * \param l pointer to the gl_light structure to be initialized.

- * \param n number of the light. 

- * \note The defaults for light 0 are different than the other lights.

- */

-static void

-init_light( struct gl_light *l, GLuint n )

-{

-   make_empty_list( l );

-

-   ASSIGN_4V( l->Ambient, 0.0, 0.0, 0.0, 1.0 );

-   if (n==0) {

-      ASSIGN_4V( l->Diffuse, 1.0, 1.0, 1.0, 1.0 );

-      ASSIGN_4V( l->Specular, 1.0, 1.0, 1.0, 1.0 );

-   }

-   else {

-      ASSIGN_4V( l->Diffuse, 0.0, 0.0, 0.0, 1.0 );

-      ASSIGN_4V( l->Specular, 0.0, 0.0, 0.0, 1.0 );

-   }

-   ASSIGN_4V( l->EyePosition, 0.0, 0.0, 1.0, 0.0 );

-   ASSIGN_3V( l->SpotDirection, 0.0, 0.0, -1.0 );

-   l->SpotExponent = 0.0;

-   _mesa_invalidate_spot_exp_table( l );

-   l->SpotCutoff = 180.0;

-   l->_CosCutoffNeg = -1.0f;

-   l->_CosCutoff = 0.0;		/* KW: -ve values not admitted */

-   l->ConstantAttenuation = 1.0;

-   l->LinearAttenuation = 0.0;

-   l->QuadraticAttenuation = 0.0;

-   l->Enabled = GL_FALSE;

-}

-

-

-/**

- * Initialize the light model data structure.

- *

- * \param lm pointer to the gl_lightmodel structure to be initialized.

- */

-static void

-init_lightmodel( struct gl_lightmodel *lm )

-{

-   ASSIGN_4V( lm->Ambient, 0.2F, 0.2F, 0.2F, 1.0F );

-   lm->LocalViewer = GL_FALSE;

-   lm->TwoSide = GL_FALSE;

-   lm->ColorControl = GL_SINGLE_COLOR;

-}

-

-

-/**

- * Initialize the material data structure.

- * 

- * \param m pointer to the gl_material structure to be initialized.

- */

-static void

-init_material( struct gl_material *m )

-{

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_AMBIENT],  0.2F, 0.2F, 0.2F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_DIFFUSE],  0.8F, 0.8F, 0.8F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_SPECULAR], 0.0F, 0.0F, 0.0F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_EMISSION], 0.0F, 0.0F, 0.0F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_SHININESS], 0.0F, 0.0F, 0.0F, 0.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_INDEXES], 0.0F, 1.0F, 1.0F, 0.0F );

- 

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_AMBIENT],  0.2F, 0.2F, 0.2F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_DIFFUSE],  0.8F, 0.8F, 0.8F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_SPECULAR], 0.0F, 0.0F, 0.0F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_EMISSION], 0.0F, 0.0F, 0.0F, 1.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_SHININESS], 0.0F, 0.0F, 0.0F, 0.0F );

-   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_INDEXES], 0.0F, 1.0F, 1.0F, 0.0F );

-}

-

-

-/**

- * Initialize all lighting state for the given context.

- */

-void

-_mesa_init_lighting( struct gl_context *ctx )

-{

-   GLuint i;

-

-   /* Lighting group */

-   for (i = 0; i < MAX_LIGHTS; i++) {

-      init_light( &ctx->Light.Light[i], i );

-   }

-   make_empty_list( &ctx->Light.EnabledList );

-

-   init_lightmodel( &ctx->Light.Model );

-   init_material( &ctx->Light.Material );

-   ctx->Light.ShadeModel = GL_SMOOTH;

-   ctx->Light.ProvokingVertex = GL_LAST_VERTEX_CONVENTION_EXT;

-   ctx->Light.Enabled = GL_FALSE;

-   ctx->Light.ColorMaterialFace = GL_FRONT_AND_BACK;

-   ctx->Light.ColorMaterialMode = GL_AMBIENT_AND_DIFFUSE;

-   ctx->Light.ColorMaterialBitmask = _mesa_material_bitmask( ctx,

-                                               GL_FRONT_AND_BACK,

-                                               GL_AMBIENT_AND_DIFFUSE, ~0,

-                                               NULL );

-

-   ctx->Light.ColorMaterialEnabled = GL_FALSE;

-   ctx->Light.ClampVertexColor = GL_TRUE;

-

-   /* Lighting miscellaneous */

-   ctx->_ShineTabList = MALLOC_STRUCT( gl_shine_tab );

-   make_empty_list( ctx->_ShineTabList );

-   /* Allocate 10 (arbitrary) shininess lookup tables */

-   for (i = 0 ; i < 10 ; i++) {

-      struct gl_shine_tab *s = MALLOC_STRUCT( gl_shine_tab );

-      s->shininess = -1;

-      s->refcount = 0;

-      insert_at_tail( ctx->_ShineTabList, s );

-   }

-

-   /* Miscellaneous */

-   ctx->Light._NeedEyeCoords = GL_FALSE;

-   ctx->_NeedEyeCoords = GL_FALSE;

-   ctx->_ForceEyeCoords = GL_FALSE;

-   ctx->_ModelViewInvScale = 1.0;

-}

-

-

-/**

- * Deallocate malloc'd lighting state attached to given context.

- */

-void

-_mesa_free_lighting_data( struct gl_context *ctx )

-{

-   struct gl_shine_tab *s, *tmps;

-

-   /* Free lighting shininess exponentiation table */

-   foreach_s( s, tmps, ctx->_ShineTabList ) {

-      free( s );

-   }

-   free( ctx->_ShineTabList );

-}

+/*
+ * Mesa 3-D graphics library
+ * Version:  7.5
+ *
+ * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
+ * Copyright (C) 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 "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+#include "glheader.h"
+#include "imports.h"
+#include "context.h"
+#include "enums.h"
+#include "light.h"
+#include "macros.h"
+#include "simple_list.h"
+#include "mtypes.h"
+#include "math/m_matrix.h"
+
+
+void GLAPIENTRY
+_mesa_ShadeModel( GLenum mode )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (MESA_VERBOSE & VERBOSE_API)
+      _mesa_debug(ctx, "glShadeModel %s\n", _mesa_lookup_enum_by_nr(mode));
+
+   if (mode != GL_FLAT && mode != GL_SMOOTH) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glShadeModel");
+      return;
+   }
+
+   if (ctx->Light.ShadeModel == mode)
+      return;
+
+   FLUSH_VERTICES(ctx, _NEW_LIGHT);
+   ctx->Light.ShadeModel = mode;
+   if (mode == GL_FLAT)
+      ctx->_TriangleCaps |= DD_FLATSHADE;
+   else
+      ctx->_TriangleCaps &= ~DD_FLATSHADE;
+
+   if (ctx->Driver.ShadeModel)
+      ctx->Driver.ShadeModel( ctx, mode );
+}
+
+
+/**
+ * Set the provoking vertex (the vertex which specifies the prim's
+ * color when flat shading) to either the first or last vertex of the
+ * triangle or line.
+ */
+void GLAPIENTRY
+_mesa_ProvokingVertexEXT(GLenum mode)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (MESA_VERBOSE&VERBOSE_API)
+      _mesa_debug(ctx, "glProvokingVertexEXT 0x%x\n", mode);
+
+   switch (mode) {
+   case GL_FIRST_VERTEX_CONVENTION_EXT:
+   case GL_LAST_VERTEX_CONVENTION_EXT:
+      break;
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "glProvokingVertexEXT(0x%x)", mode);
+      return;
+   }
+
+   if (ctx->Light.ProvokingVertex == mode)
+      return;
+
+   FLUSH_VERTICES(ctx, _NEW_LIGHT);
+   ctx->Light.ProvokingVertex = mode;
+}
+
+
+/**
+ * Helper function called by _mesa_Lightfv and _mesa_PopAttrib to set
+ * per-light state.
+ * For GL_POSITION and GL_SPOT_DIRECTION the params position/direction
+ * will have already been transformed by the modelview matrix!
+ * Also, all error checking should have already been done.
+ */
+void
+_mesa_light(struct gl_context *ctx, GLuint lnum, GLenum pname, const GLfloat *params)
+{
+   struct gl_light *light;
+
+   ASSERT(lnum < MAX_LIGHTS);
+   light = &ctx->Light.Light[lnum];
+
+   switch (pname) {
+   case GL_AMBIENT:
+      if (TEST_EQ_4V(light->Ambient, params))
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      COPY_4V( light->Ambient, params );
+      break;
+   case GL_DIFFUSE:
+      if (TEST_EQ_4V(light->Diffuse, params))
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      COPY_4V( light->Diffuse, params );
+      break;
+   case GL_SPECULAR:
+      if (TEST_EQ_4V(light->Specular, params))
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      COPY_4V( light->Specular, params );
+      break;
+   case GL_POSITION:
+      /* NOTE: position has already been transformed by ModelView! */
+      if (TEST_EQ_4V(light->EyePosition, params))
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      COPY_4V(light->EyePosition, params);
+      if (light->EyePosition[3] != 0.0F)
+	 light->_Flags |= LIGHT_POSITIONAL;
+      else
+	 light->_Flags &= ~LIGHT_POSITIONAL;
+      break;
+   case GL_SPOT_DIRECTION:
+      /* NOTE: Direction already transformed by inverse ModelView! */
+      if (TEST_EQ_3V(light->SpotDirection, params))
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      COPY_3V(light->SpotDirection, params);
+      break;
+   case GL_SPOT_EXPONENT:
+      ASSERT(params[0] >= 0.0);
+      ASSERT(params[0] <= ctx->Const.MaxSpotExponent);
+      if (light->SpotExponent == params[0])
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      light->SpotExponent = params[0];
+      _mesa_invalidate_spot_exp_table(light);
+      break;
+   case GL_SPOT_CUTOFF:
+      ASSERT(params[0] == 180.0 || (params[0] >= 0.0 && params[0] <= 90.0));
+      if (light->SpotCutoff == params[0])
+         return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      light->SpotCutoff = params[0];
+      light->_CosCutoffNeg = (GLfloat) (cos(light->SpotCutoff * DEG2RAD));
+      if (light->_CosCutoffNeg < 0)
+         light->_CosCutoff = 0;
+      else
+         light->_CosCutoff = light->_CosCutoffNeg;
+      if (light->SpotCutoff != 180.0F)
+         light->_Flags |= LIGHT_SPOT;
+      else
+         light->_Flags &= ~LIGHT_SPOT;
+      break;
+   case GL_CONSTANT_ATTENUATION:
+      ASSERT(params[0] >= 0.0);
+      if (light->ConstantAttenuation == params[0])
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      light->ConstantAttenuation = params[0];
+      break;
+   case GL_LINEAR_ATTENUATION:
+      ASSERT(params[0] >= 0.0);
+      if (light->LinearAttenuation == params[0])
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      light->LinearAttenuation = params[0];
+      break;
+   case GL_QUADRATIC_ATTENUATION:
+      ASSERT(params[0] >= 0.0);
+      if (light->QuadraticAttenuation == params[0])
+	 return;
+      FLUSH_VERTICES(ctx, _NEW_LIGHT);
+      light->QuadraticAttenuation = params[0];
+      break;
+   default:
+      _mesa_problem(ctx, "Unexpected pname in _mesa_light()");
+      return;
+   }
+
+   if (ctx->Driver.Lightfv)
+      ctx->Driver.Lightfv( ctx, GL_LIGHT0 + lnum, pname, params );
+}
+
+
+void GLAPIENTRY
+_mesa_Lightf( GLenum light, GLenum pname, GLfloat param )
+{
+   GLfloat fparam[4];
+   fparam[0] = param;
+   fparam[1] = fparam[2] = fparam[3] = 0.0F;
+   _mesa_Lightfv( light, pname, fparam );
+}
+
+
+void GLAPIENTRY
+_mesa_Lightfv( GLenum light, GLenum pname, const GLfloat *params )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint i = (GLint) (light - GL_LIGHT0);
+   GLfloat temp[4];
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (i < 0 || i >= (GLint) ctx->Const.MaxLights) {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glLight(light=0x%x)", light );
+      return;
+   }
+
+   /* do particular error checks, transformations */
+   switch (pname) {
+   case GL_AMBIENT:
+   case GL_DIFFUSE:
+   case GL_SPECULAR:
+      /* nothing */
+      break;
+   case GL_POSITION:
+      /* transform position by ModelView matrix */
+      TRANSFORM_POINT(temp, ctx->ModelviewMatrixStack.Top->m, params);
+      params = temp;
+      break;
+   case GL_SPOT_DIRECTION:
+      /* transform direction by inverse modelview */
+      if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) {
+	 _math_matrix_analyse(ctx->ModelviewMatrixStack.Top);
+      }
+      TRANSFORM_DIRECTION(temp, params, ctx->ModelviewMatrixStack.Top->m);
+      params = temp;
+      break;
+   case GL_SPOT_EXPONENT:
+      if (params[0] < 0.0 || params[0] > ctx->Const.MaxSpotExponent) {
+	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");
+	 return;
+      }
+      break;
+   case GL_SPOT_CUTOFF:
+      if ((params[0] < 0.0 || params[0] > 90.0) && params[0] != 180.0) {
+	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");
+	 return;
+      }
+      break;
+   case GL_CONSTANT_ATTENUATION:
+      if (params[0] < 0.0) {
+	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");
+	 return;
+      }
+      break;
+   case GL_LINEAR_ATTENUATION:
+      if (params[0] < 0.0) {
+	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");
+	 return;
+      }
+      break;
+   case GL_QUADRATIC_ATTENUATION:
+      if (params[0] < 0.0) {
+	 _mesa_error(ctx, GL_INVALID_VALUE, "glLight");
+	 return;
+      }
+      break;
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "glLight(pname=0x%x)", pname);
+      return;
+   }
+
+   _mesa_light(ctx, i, pname, params);
+}
+
+
+void GLAPIENTRY
+_mesa_Lighti( GLenum light, GLenum pname, GLint param )
+{
+   GLint iparam[4];
+   iparam[0] = param;
+   iparam[1] = iparam[2] = iparam[3] = 0;
+   _mesa_Lightiv( light, pname, iparam );
+}
+
+
+void GLAPIENTRY
+_mesa_Lightiv( GLenum light, GLenum pname, const GLint *params )
+{
+   GLfloat fparam[4];
+
+   switch (pname) {
+      case GL_AMBIENT:
+      case GL_DIFFUSE:
+      case GL_SPECULAR:
+         fparam[0] = INT_TO_FLOAT( params[0] );
+         fparam[1] = INT_TO_FLOAT( params[1] );
+         fparam[2] = INT_TO_FLOAT( params[2] );
+         fparam[3] = INT_TO_FLOAT( params[3] );
+         break;
+      case GL_POSITION:
+         fparam[0] = (GLfloat) params[0];
+         fparam[1] = (GLfloat) params[1];
+         fparam[2] = (GLfloat) params[2];
+         fparam[3] = (GLfloat) params[3];
+         break;
+      case GL_SPOT_DIRECTION:
+         fparam[0] = (GLfloat) params[0];
+         fparam[1] = (GLfloat) params[1];
+         fparam[2] = (GLfloat) params[2];
+         break;
+      case GL_SPOT_EXPONENT:
+      case GL_SPOT_CUTOFF:
+      case GL_CONSTANT_ATTENUATION:
+      case GL_LINEAR_ATTENUATION:
+      case GL_QUADRATIC_ATTENUATION:
+         fparam[0] = (GLfloat) params[0];
+         break;
+      default:
+         /* error will be caught later in gl_Lightfv */
+         ;
+   }
+
+   _mesa_Lightfv( light, pname, fparam );
+}
+
+
+
+void GLAPIENTRY
+_mesa_GetLightfv( GLenum light, GLenum pname, GLfloat *params )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint l = (GLint) (light - GL_LIGHT0);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (l < 0 || l >= (GLint) ctx->Const.MaxLights) {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightfv" );
+      return;
+   }
+
+   switch (pname) {
+      case GL_AMBIENT:
+         COPY_4V( params, ctx->Light.Light[l].Ambient );
+         break;
+      case GL_DIFFUSE:
+         COPY_4V( params, ctx->Light.Light[l].Diffuse );
+         break;
+      case GL_SPECULAR:
+         COPY_4V( params, ctx->Light.Light[l].Specular );
+         break;
+      case GL_POSITION:
+         COPY_4V( params, ctx->Light.Light[l].EyePosition );
+         break;
+      case GL_SPOT_DIRECTION:
+         COPY_3V( params, ctx->Light.Light[l].SpotDirection );
+         break;
+      case GL_SPOT_EXPONENT:
+         params[0] = ctx->Light.Light[l].SpotExponent;
+         break;
+      case GL_SPOT_CUTOFF:
+         params[0] = ctx->Light.Light[l].SpotCutoff;
+         break;
+      case GL_CONSTANT_ATTENUATION:
+         params[0] = ctx->Light.Light[l].ConstantAttenuation;
+         break;
+      case GL_LINEAR_ATTENUATION:
+         params[0] = ctx->Light.Light[l].LinearAttenuation;
+         break;
+      case GL_QUADRATIC_ATTENUATION:
+         params[0] = ctx->Light.Light[l].QuadraticAttenuation;
+         break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightfv" );
+         break;
+   }
+}
+
+
+void GLAPIENTRY
+_mesa_GetLightiv( GLenum light, GLenum pname, GLint *params )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint l = (GLint) (light - GL_LIGHT0);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (l < 0 || l >= (GLint) ctx->Const.MaxLights) {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightiv" );
+      return;
+   }
+
+   switch (pname) {
+      case GL_AMBIENT:
+         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[0]);
+         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[1]);
+         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[2]);
+         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Ambient[3]);
+         break;
+      case GL_DIFFUSE:
+         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[0]);
+         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[1]);
+         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[2]);
+         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Diffuse[3]);
+         break;
+      case GL_SPECULAR:
+         params[0] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[0]);
+         params[1] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[1]);
+         params[2] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[2]);
+         params[3] = FLOAT_TO_INT(ctx->Light.Light[l].Specular[3]);
+         break;
+      case GL_POSITION:
+         params[0] = (GLint) ctx->Light.Light[l].EyePosition[0];
+         params[1] = (GLint) ctx->Light.Light[l].EyePosition[1];
+         params[2] = (GLint) ctx->Light.Light[l].EyePosition[2];
+         params[3] = (GLint) ctx->Light.Light[l].EyePosition[3];
+         break;
+      case GL_SPOT_DIRECTION:
+         params[0] = (GLint) ctx->Light.Light[l].SpotDirection[0];
+         params[1] = (GLint) ctx->Light.Light[l].SpotDirection[1];
+         params[2] = (GLint) ctx->Light.Light[l].SpotDirection[2];
+         break;
+      case GL_SPOT_EXPONENT:
+         params[0] = (GLint) ctx->Light.Light[l].SpotExponent;
+         break;
+      case GL_SPOT_CUTOFF:
+         params[0] = (GLint) ctx->Light.Light[l].SpotCutoff;
+         break;
+      case GL_CONSTANT_ATTENUATION:
+         params[0] = (GLint) ctx->Light.Light[l].ConstantAttenuation;
+         break;
+      case GL_LINEAR_ATTENUATION:
+         params[0] = (GLint) ctx->Light.Light[l].LinearAttenuation;
+         break;
+      case GL_QUADRATIC_ATTENUATION:
+         params[0] = (GLint) ctx->Light.Light[l].QuadraticAttenuation;
+         break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "glGetLightiv" );
+         break;
+   }
+}
+
+
+
+/**********************************************************************/
+/***                        Light Model                             ***/
+/**********************************************************************/
+
+
+void GLAPIENTRY
+_mesa_LightModelfv( GLenum pname, const GLfloat *params )
+{
+   GLenum newenum;
+   GLboolean newbool;
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   switch (pname) {
+      case GL_LIGHT_MODEL_AMBIENT:
+         if (TEST_EQ_4V( ctx->Light.Model.Ambient, params ))
+	    return;
+	 FLUSH_VERTICES(ctx, _NEW_LIGHT);
+         COPY_4V( ctx->Light.Model.Ambient, params );
+         break;
+      case GL_LIGHT_MODEL_LOCAL_VIEWER:
+         newbool = (params[0]!=0.0);
+	 if (ctx->Light.Model.LocalViewer == newbool)
+	    return;
+	 FLUSH_VERTICES(ctx, _NEW_LIGHT);
+	 ctx->Light.Model.LocalViewer = newbool;
+         break;
+      case GL_LIGHT_MODEL_TWO_SIDE:
+         newbool = (params[0]!=0.0);
+	 if (ctx->Light.Model.TwoSide == newbool)
+	    return;
+	 FLUSH_VERTICES(ctx, _NEW_LIGHT);
+	 ctx->Light.Model.TwoSide = newbool;
+         if (ctx->Light.Enabled && ctx->Light.Model.TwoSide)
+            ctx->_TriangleCaps |= DD_TRI_LIGHT_TWOSIDE;
+         else
+            ctx->_TriangleCaps &= ~DD_TRI_LIGHT_TWOSIDE;
+         break;
+      case GL_LIGHT_MODEL_COLOR_CONTROL:
+         if (params[0] == (GLfloat) GL_SINGLE_COLOR)
+	    newenum = GL_SINGLE_COLOR;
+         else if (params[0] == (GLfloat) GL_SEPARATE_SPECULAR_COLOR)
+	    newenum = GL_SEPARATE_SPECULAR_COLOR;
+	 else {
+            _mesa_error( ctx, GL_INVALID_ENUM, "glLightModel(param=0x0%x)",
+                         (GLint) params[0] );
+	    return;
+         }
+	 if (ctx->Light.Model.ColorControl == newenum)
+	    return;
+	 FLUSH_VERTICES(ctx, _NEW_LIGHT);
+	 ctx->Light.Model.ColorControl = newenum;
+         break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "glLightModel(pname=0x%x)", pname );
+         break;
+   }
+
+   if (ctx->Driver.LightModelfv)
+      ctx->Driver.LightModelfv( ctx, pname, params );
+}
+
+
+void GLAPIENTRY
+_mesa_LightModeliv( GLenum pname, const GLint *params )
+{
+   GLfloat fparam[4];
+
+   switch (pname) {
+      case GL_LIGHT_MODEL_AMBIENT:
+         fparam[0] = INT_TO_FLOAT( params[0] );
+         fparam[1] = INT_TO_FLOAT( params[1] );
+         fparam[2] = INT_TO_FLOAT( params[2] );
+         fparam[3] = INT_TO_FLOAT( params[3] );
+         break;
+      case GL_LIGHT_MODEL_LOCAL_VIEWER:
+      case GL_LIGHT_MODEL_TWO_SIDE:
+      case GL_LIGHT_MODEL_COLOR_CONTROL:
+         fparam[0] = (GLfloat) params[0];
+         break;
+      default:
+         /* Error will be caught later in gl_LightModelfv */
+         ASSIGN_4V(fparam, 0.0F, 0.0F, 0.0F, 0.0F);
+   }
+   _mesa_LightModelfv( pname, fparam );
+}
+
+
+void GLAPIENTRY
+_mesa_LightModeli( GLenum pname, GLint param )
+{
+   GLint iparam[4];
+   iparam[0] = param;
+   iparam[1] = iparam[2] = iparam[3] = 0;
+   _mesa_LightModeliv( pname, iparam );
+}
+
+
+void GLAPIENTRY
+_mesa_LightModelf( GLenum pname, GLfloat param )
+{
+   GLfloat fparam[4];
+   fparam[0] = param;
+   fparam[1] = fparam[2] = fparam[3] = 0.0F;
+   _mesa_LightModelfv( pname, fparam );
+}
+
+
+
+/********** MATERIAL **********/
+
+
+/*
+ * Given a face and pname value (ala glColorMaterial), compute a bitmask
+ * of the targeted material values.
+ */
+GLuint
+_mesa_material_bitmask( struct gl_context *ctx, GLenum face, GLenum pname,
+                        GLuint legal, const char *where )
+{
+   GLuint bitmask = 0;
+
+   /* Make a bitmask indicating what material attribute(s) we're updating */
+   switch (pname) {
+      case GL_EMISSION:
+         bitmask |= MAT_BIT_FRONT_EMISSION | MAT_BIT_BACK_EMISSION;
+         break;
+      case GL_AMBIENT:
+         bitmask |= MAT_BIT_FRONT_AMBIENT | MAT_BIT_BACK_AMBIENT;
+         break;
+      case GL_DIFFUSE:
+         bitmask |= MAT_BIT_FRONT_DIFFUSE | MAT_BIT_BACK_DIFFUSE;
+         break;
+      case GL_SPECULAR:
+         bitmask |= MAT_BIT_FRONT_SPECULAR | MAT_BIT_BACK_SPECULAR;
+         break;
+      case GL_SHININESS:
+         bitmask |= MAT_BIT_FRONT_SHININESS | MAT_BIT_BACK_SHININESS;
+         break;
+      case GL_AMBIENT_AND_DIFFUSE:
+         bitmask |= MAT_BIT_FRONT_AMBIENT | MAT_BIT_BACK_AMBIENT;
+         bitmask |= MAT_BIT_FRONT_DIFFUSE | MAT_BIT_BACK_DIFFUSE;
+         break;
+      case GL_COLOR_INDEXES:
+         bitmask |= MAT_BIT_FRONT_INDEXES  | MAT_BIT_BACK_INDEXES;
+         break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );
+         return 0;
+   }
+
+   if (face==GL_FRONT) {
+      bitmask &= FRONT_MATERIAL_BITS;
+   }
+   else if (face==GL_BACK) {
+      bitmask &= BACK_MATERIAL_BITS;
+   }
+   else if (face != GL_FRONT_AND_BACK) {
+      _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );
+      return 0;
+   }
+
+   if (bitmask & ~legal) {
+      _mesa_error( ctx, GL_INVALID_ENUM, "%s", where );
+      return 0;
+   }
+
+   return bitmask;
+}
+
+
+
+/* Perform a straight copy between materials.
+ */
+void
+_mesa_copy_materials( struct gl_material *dst,
+                      const struct gl_material *src,
+                      GLuint bitmask )
+{
+   int i;
+
+   for (i = 0 ; i < MAT_ATTRIB_MAX ; i++) 
+      if (bitmask & (1<<i))
+	 COPY_4FV( dst->Attrib[i], src->Attrib[i] );
+}
+
+
+
+/* Update derived values following a change in ctx->Light.Material
+ */
+void
+_mesa_update_material( struct gl_context *ctx, GLuint bitmask )
+{
+   struct gl_light *light, *list = &ctx->Light.EnabledList;
+   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;
+
+   if (MESA_VERBOSE & VERBOSE_MATERIAL) 
+      _mesa_debug(ctx, "_mesa_update_material, mask 0x%x\n", bitmask);
+
+   if (!bitmask)
+      return;
+
+   /* update material ambience */
+   if (bitmask & MAT_BIT_FRONT_AMBIENT) {
+      foreach (light, list) {
+         SCALE_3V( light->_MatAmbient[0], light->Ambient, 
+		   mat[MAT_ATTRIB_FRONT_AMBIENT]);
+      }
+   }
+
+   if (bitmask & MAT_BIT_BACK_AMBIENT) {
+      foreach (light, list) {
+         SCALE_3V( light->_MatAmbient[1], light->Ambient, 
+		   mat[MAT_ATTRIB_BACK_AMBIENT]);
+      }
+   }
+
+   /* update BaseColor = emission + scene's ambience * material's ambience */
+   if (bitmask & (MAT_BIT_FRONT_EMISSION | MAT_BIT_FRONT_AMBIENT)) {
+      COPY_3V( ctx->Light._BaseColor[0], mat[MAT_ATTRIB_FRONT_EMISSION] );
+      ACC_SCALE_3V( ctx->Light._BaseColor[0], mat[MAT_ATTRIB_FRONT_AMBIENT],
+		    ctx->Light.Model.Ambient );
+   }
+
+   if (bitmask & (MAT_BIT_BACK_EMISSION | MAT_BIT_BACK_AMBIENT)) {
+      COPY_3V( ctx->Light._BaseColor[1], mat[MAT_ATTRIB_BACK_EMISSION] );
+      ACC_SCALE_3V( ctx->Light._BaseColor[1], mat[MAT_ATTRIB_BACK_AMBIENT],
+		    ctx->Light.Model.Ambient );
+   }
+
+   /* update material diffuse values */
+   if (bitmask & MAT_BIT_FRONT_DIFFUSE) {
+      foreach (light, list) {
+	 SCALE_3V( light->_MatDiffuse[0], light->Diffuse, 
+		   mat[MAT_ATTRIB_FRONT_DIFFUSE] );
+      }
+   }
+
+   if (bitmask & MAT_BIT_BACK_DIFFUSE) {
+      foreach (light, list) {
+	 SCALE_3V( light->_MatDiffuse[1], light->Diffuse, 
+		   mat[MAT_ATTRIB_BACK_DIFFUSE] );
+      }
+   }
+
+   /* update material specular values */
+   if (bitmask & MAT_BIT_FRONT_SPECULAR) {
+      foreach (light, list) {
+	 SCALE_3V( light->_MatSpecular[0], light->Specular, 
+		   mat[MAT_ATTRIB_FRONT_SPECULAR]);
+      }
+   }
+
+   if (bitmask & MAT_BIT_BACK_SPECULAR) {
+      foreach (light, list) {
+	 SCALE_3V( light->_MatSpecular[1], light->Specular,
+		   mat[MAT_ATTRIB_BACK_SPECULAR]);
+      }
+   }
+
+   if (bitmask & MAT_BIT_FRONT_SHININESS) {
+      _mesa_invalidate_shine_table( ctx, 0 );
+   }
+
+   if (bitmask & MAT_BIT_BACK_SHININESS) {
+      _mesa_invalidate_shine_table( ctx, 1 );
+   }
+}
+
+
+/*
+ * Update the current materials from the given rgba color
+ * according to the bitmask in ColorMaterialBitmask, which is
+ * set by glColorMaterial().
+ */
+void
+_mesa_update_color_material( struct gl_context *ctx, const GLfloat color[4] )
+{
+   GLuint bitmask = ctx->Light.ColorMaterialBitmask;
+   struct gl_material *mat = &ctx->Light.Material;
+   int i;
+
+   for (i = 0 ; i < MAT_ATTRIB_MAX ; i++) 
+      if (bitmask & (1<<i))
+	 COPY_4FV( mat->Attrib[i], color );
+
+   _mesa_update_material( ctx, bitmask );
+}
+
+
+void GLAPIENTRY
+_mesa_ColorMaterial( GLenum face, GLenum mode )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint bitmask;
+   GLuint legal = (MAT_BIT_FRONT_EMISSION | MAT_BIT_BACK_EMISSION |
+		   MAT_BIT_FRONT_SPECULAR | MAT_BIT_BACK_SPECULAR |
+		   MAT_BIT_FRONT_DIFFUSE  | MAT_BIT_BACK_DIFFUSE  |
+		   MAT_BIT_FRONT_AMBIENT  | MAT_BIT_BACK_AMBIENT);
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
+   if (MESA_VERBOSE&VERBOSE_API)
+      _mesa_debug(ctx, "glColorMaterial %s %s\n",
+                  _mesa_lookup_enum_by_nr(face),
+                  _mesa_lookup_enum_by_nr(mode));
+
+   bitmask = _mesa_material_bitmask(ctx, face, mode, legal, "glColorMaterial");
+
+   if (ctx->Light.ColorMaterialBitmask == bitmask &&
+       ctx->Light.ColorMaterialFace == face &&
+       ctx->Light.ColorMaterialMode == mode)
+      return;
+
+   FLUSH_VERTICES(ctx, _NEW_LIGHT);
+   ctx->Light.ColorMaterialBitmask = bitmask;
+   ctx->Light.ColorMaterialFace = face;
+   ctx->Light.ColorMaterialMode = mode;
+
+   if (ctx->Light.ColorMaterialEnabled) {
+      FLUSH_CURRENT( ctx, 0 );
+      _mesa_update_color_material(ctx,ctx->Current.Attrib[VERT_ATTRIB_COLOR0]);
+   }
+
+   if (ctx->Driver.ColorMaterial)
+      ctx->Driver.ColorMaterial( ctx, face, mode );
+}
+
+
+void GLAPIENTRY
+_mesa_GetMaterialfv( GLenum face, GLenum pname, GLfloat *params )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint f;
+   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;
+   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); /* update materials */
+
+   FLUSH_CURRENT(ctx, 0); /* update ctx->Light.Material from vertex buffer */
+
+   if (face==GL_FRONT) {
+      f = 0;
+   }
+   else if (face==GL_BACK) {
+      f = 1;
+   }
+   else {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(face)" );
+      return;
+   }
+
+   switch (pname) {
+      case GL_AMBIENT:
+         COPY_4FV( params, mat[MAT_ATTRIB_AMBIENT(f)] );
+         break;
+      case GL_DIFFUSE:
+         COPY_4FV( params, mat[MAT_ATTRIB_DIFFUSE(f)] );
+	 break;
+      case GL_SPECULAR:
+         COPY_4FV( params, mat[MAT_ATTRIB_SPECULAR(f)] );
+	 break;
+      case GL_EMISSION:
+	 COPY_4FV( params, mat[MAT_ATTRIB_EMISSION(f)] );
+	 break;
+      case GL_SHININESS:
+	 *params = mat[MAT_ATTRIB_SHININESS(f)][0];
+	 break;
+      case GL_COLOR_INDEXES:
+	 params[0] = mat[MAT_ATTRIB_INDEXES(f)][0];
+	 params[1] = mat[MAT_ATTRIB_INDEXES(f)][1];
+	 params[2] = mat[MAT_ATTRIB_INDEXES(f)][2];
+	 break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(pname)" );
+   }
+}
+
+
+void GLAPIENTRY
+_mesa_GetMaterialiv( GLenum face, GLenum pname, GLint *params )
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint f;
+   GLfloat (*mat)[4] = ctx->Light.Material.Attrib;
+   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); /* update materials */
+
+   FLUSH_CURRENT(ctx, 0); /* update ctx->Light.Material from vertex buffer */
+
+   if (face==GL_FRONT) {
+      f = 0;
+   }
+   else if (face==GL_BACK) {
+      f = 1;
+   }
+   else {
+      _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialiv(face)" );
+      return;
+   }
+   switch (pname) {
+      case GL_AMBIENT:
+         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][0] );
+         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][1] );
+         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][2] );
+         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_AMBIENT(f)][3] );
+         break;
+      case GL_DIFFUSE:
+         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][0] );
+         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][1] );
+         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][2] );
+         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_DIFFUSE(f)][3] );
+	 break;
+      case GL_SPECULAR:
+         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][0] );
+         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][1] );
+         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][2] );
+         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_SPECULAR(f)][3] );
+	 break;
+      case GL_EMISSION:
+         params[0] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][0] );
+         params[1] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][1] );
+         params[2] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][2] );
+         params[3] = FLOAT_TO_INT( mat[MAT_ATTRIB_EMISSION(f)][3] );
+	 break;
+      case GL_SHININESS:
+         *params = IROUND( mat[MAT_ATTRIB_SHININESS(f)][0] );
+	 break;
+      case GL_COLOR_INDEXES:
+	 params[0] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][0] );
+	 params[1] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][1] );
+	 params[2] = IROUND( mat[MAT_ATTRIB_INDEXES(f)][2] );
+	 break;
+      default:
+         _mesa_error( ctx, GL_INVALID_ENUM, "glGetMaterialfv(pname)" );
+   }
+}
+
+
+
+/**********************************************************************/
+/*****                  Lighting computation                      *****/
+/**********************************************************************/
+
+
+/*
+ * Notes:
+ *   When two-sided lighting is enabled we compute the color (or index)
+ *   for both the front and back side of the primitive.  Then, when the
+ *   orientation of the facet is later learned, we can determine which
+ *   color (or index) to use for rendering.
+ *
+ *   KW: We now know orientation in advance and only shade for
+ *       the side or sides which are actually required.
+ *
+ * Variables:
+ *   n = normal vector
+ *   V = vertex position
+ *   P = light source position
+ *   Pe = (0,0,0,1)
+ *
+ * Precomputed:
+ *   IF P[3]==0 THEN
+ *       // light at infinity
+ *       IF local_viewer THEN
+ *           _VP_inf_norm = unit vector from V to P      // Precompute
+ *       ELSE
+ *           // eye at infinity
+ *           _h_inf_norm = Normalize( VP + <0,0,1> )     // Precompute
+ *       ENDIF
+ *   ENDIF
+ *
+ * Functions:
+ *   Normalize( v ) = normalized vector v
+ *   Magnitude( v ) = length of vector v
+ */
+
+
+
+/*
+ * Whenever the spotlight exponent for a light changes we must call
+ * this function to recompute the exponent lookup table.
+ */
+void
+_mesa_invalidate_spot_exp_table( struct gl_light *l )
+{
+   l->_SpotExpTable[0][0] = -1;
+}
+
+
+static void
+validate_spot_exp_table( struct gl_light *l )
+{
+   GLint i;
+   GLdouble exponent = l->SpotExponent;
+   GLdouble tmp = 0;
+   GLint clamp = 0;
+
+   l->_SpotExpTable[0][0] = 0.0;
+
+   for (i = EXP_TABLE_SIZE - 1; i > 0 ;i--) {
+      if (clamp == 0) {
+	 tmp = pow(i / (GLdouble) (EXP_TABLE_SIZE - 1), exponent);
+	 if (tmp < FLT_MIN * 100.0) {
+	    tmp = 0.0;
+	    clamp = 1;
+	 }
+      }
+      l->_SpotExpTable[i][0] = (GLfloat) tmp;
+   }
+   for (i = 0; i < EXP_TABLE_SIZE - 1; i++) {
+      l->_SpotExpTable[i][1] = (l->_SpotExpTable[i+1][0] -
+				l->_SpotExpTable[i][0]);
+   }
+   l->_SpotExpTable[EXP_TABLE_SIZE-1][1] = 0.0;
+}
+
+
+
+/* Calculate a new shine table.  Doing this here saves a branch in
+ * lighting, and the cost of doing it early may be partially offset
+ * by keeping a MRU cache of shine tables for various shine values.
+ */
+void
+_mesa_invalidate_shine_table( struct gl_context *ctx, GLuint side )
+{
+   ASSERT(side < 2);
+   if (ctx->_ShineTable[side])
+      ctx->_ShineTable[side]->refcount--;
+   ctx->_ShineTable[side] = NULL;
+}
+
+
+static void
+validate_shine_table( struct gl_context *ctx, GLuint side, GLfloat shininess )
+{
+   struct gl_shine_tab *list = ctx->_ShineTabList;
+   struct gl_shine_tab *s;
+
+   ASSERT(side < 2);
+
+   foreach(s, list)
+      if ( s->shininess == shininess )
+	 break;
+
+   if (s == list) {
+      GLint j;
+      GLfloat *m;
+
+      foreach(s, list)
+	 if (s->refcount == 0)
+	    break;
+
+      m = s->tab;
+      m[0] = 0.0;
+      if (shininess == 0.0) {
+	 for (j = 1 ; j <= SHINE_TABLE_SIZE ; j++)
+	    m[j] = 1.0;
+      }
+      else {
+	 for (j = 1 ; j < SHINE_TABLE_SIZE ; j++) {
+            GLdouble t, x = j / (GLfloat) (SHINE_TABLE_SIZE - 1);
+            if (x < 0.005) /* underflow check */
+               x = 0.005;
+            t = pow(x, shininess);
+	    if (t > 1e-20)
+	       m[j] = (GLfloat) t;
+	    else
+	       m[j] = 0.0;
+	 }
+	 m[SHINE_TABLE_SIZE] = 1.0;
+      }
+
+      s->shininess = shininess;
+   }
+
+   if (ctx->_ShineTable[side])
+      ctx->_ShineTable[side]->refcount--;
+
+   ctx->_ShineTable[side] = s;
+   move_to_tail( list, s );
+   s->refcount++;
+}
+
+
+void
+_mesa_validate_all_lighting_tables( struct gl_context *ctx )
+{
+   GLuint i;
+   GLfloat shininess;
+   
+   shininess = ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_SHININESS][0];
+   if (!ctx->_ShineTable[0] || ctx->_ShineTable[0]->shininess != shininess)
+      validate_shine_table( ctx, 0, shininess );
+
+   shininess = ctx->Light.Material.Attrib[MAT_ATTRIB_BACK_SHININESS][0];
+   if (!ctx->_ShineTable[1] || ctx->_ShineTable[1]->shininess != shininess)
+      validate_shine_table( ctx, 1, shininess );
+
+   for (i = 0; i < ctx->Const.MaxLights; i++)
+      if (ctx->Light.Light[i]._SpotExpTable[0][0] == -1)
+	 validate_spot_exp_table( &ctx->Light.Light[i] );
+}
+
+
+/**
+ * Examine current lighting parameters to determine if the optimized lighting
+ * function can be used.
+ * Also, precompute some lighting values such as the products of light
+ * source and material ambient, diffuse and specular coefficients.
+ */
+void
+_mesa_update_lighting( struct gl_context *ctx )
+{
+   struct gl_light *light;
+   ctx->Light._NeedEyeCoords = GL_FALSE;
+   ctx->Light._Flags = 0;
+
+   if (!ctx->Light.Enabled)
+      return;
+
+   foreach(light, &ctx->Light.EnabledList) {
+      ctx->Light._Flags |= light->_Flags;
+   }
+
+   ctx->Light._NeedVertices =
+      ((ctx->Light._Flags & (LIGHT_POSITIONAL|LIGHT_SPOT)) ||
+       ctx->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR ||
+       ctx->Light.Model.LocalViewer);
+
+   ctx->Light._NeedEyeCoords = ((ctx->Light._Flags & LIGHT_POSITIONAL) ||
+				ctx->Light.Model.LocalViewer);
+
+   /* XXX: This test is overkill & needs to be fixed both for software and
+    * hardware t&l drivers.  The above should be sufficient & should
+    * be tested to verify this.
+    */
+   if (ctx->Light._NeedVertices)
+      ctx->Light._NeedEyeCoords = GL_TRUE;
+
+   /* Precompute some shading values.  Although we reference
+    * Light.Material here, we can get away without flushing
+    * FLUSH_UPDATE_CURRENT, as when any outstanding material changes
+    * are flushed, they will update the derived state at that time.
+    */
+   if (ctx->Light.Model.TwoSide)
+      _mesa_update_material(ctx,
+			    MAT_BIT_FRONT_EMISSION |
+			    MAT_BIT_FRONT_AMBIENT |
+			    MAT_BIT_FRONT_DIFFUSE |
+			    MAT_BIT_FRONT_SPECULAR |
+			    MAT_BIT_BACK_EMISSION |
+			    MAT_BIT_BACK_AMBIENT |
+			    MAT_BIT_BACK_DIFFUSE |
+			    MAT_BIT_BACK_SPECULAR);
+   else
+      _mesa_update_material(ctx,
+			    MAT_BIT_FRONT_EMISSION |
+			    MAT_BIT_FRONT_AMBIENT |
+			    MAT_BIT_FRONT_DIFFUSE |
+			    MAT_BIT_FRONT_SPECULAR);
+}
+
+
+/**
+ * Update state derived from light position, spot direction.
+ * Called upon:
+ *   _NEW_MODELVIEW
+ *   _NEW_LIGHT
+ *   _TNL_NEW_NEED_EYE_COORDS
+ *
+ * Update on (_NEW_MODELVIEW | _NEW_LIGHT) when lighting is enabled.
+ * Also update on lighting space changes.
+ */
+static void
+compute_light_positions( struct gl_context *ctx )
+{
+   struct gl_light *light;
+   static const GLfloat eye_z[3] = { 0, 0, 1 };
+
+   if (!ctx->Light.Enabled)
+      return;
+
+   if (ctx->_NeedEyeCoords) {
+      COPY_3V( ctx->_EyeZDir, eye_z );
+   }
+   else {
+      TRANSFORM_NORMAL( ctx->_EyeZDir, eye_z, ctx->ModelviewMatrixStack.Top->m );
+   }
+
+   /* Make sure all the light tables are updated before the computation */
+   _mesa_validate_all_lighting_tables(ctx);
+
+   foreach (light, &ctx->Light.EnabledList) {
+
+      if (ctx->_NeedEyeCoords) {
+         /* _Position is in eye coordinate space */
+	 COPY_4FV( light->_Position, light->EyePosition );
+      }
+      else {
+         /* _Position is in object coordinate space */
+	 TRANSFORM_POINT( light->_Position, ctx->ModelviewMatrixStack.Top->inv,
+			  light->EyePosition );
+      }
+
+      if (!(light->_Flags & LIGHT_POSITIONAL)) {
+	 /* VP (VP) = Normalize( Position ) */
+	 COPY_3V( light->_VP_inf_norm, light->_Position );
+	 NORMALIZE_3FV( light->_VP_inf_norm );
+
+	 if (!ctx->Light.Model.LocalViewer) {
+	    /* _h_inf_norm = Normalize( V_to_P + <0,0,1> ) */
+	    ADD_3V( light->_h_inf_norm, light->_VP_inf_norm, ctx->_EyeZDir);
+	    NORMALIZE_3FV( light->_h_inf_norm );
+	 }
+	 light->_VP_inf_spot_attenuation = 1.0;
+      }
+      else {
+         /* positional light w/ homogeneous coordinate, divide by W */
+         GLfloat wInv = (GLfloat)1.0 / light->_Position[3];
+         light->_Position[0] *= wInv;
+         light->_Position[1] *= wInv;
+         light->_Position[2] *= wInv;
+      }
+
+      if (light->_Flags & LIGHT_SPOT) {
+         /* Note: we normalize the spot direction now */
+
+	 if (ctx->_NeedEyeCoords) {
+	    COPY_3V( light->_NormSpotDirection, light->SpotDirection );
+            NORMALIZE_3FV( light->_NormSpotDirection );
+	 }
+         else {
+            GLfloat spotDir[3];
+            COPY_3V(spotDir, light->SpotDirection);
+            NORMALIZE_3FV(spotDir);
+	    TRANSFORM_NORMAL( light->_NormSpotDirection,
+			      spotDir,
+			      ctx->ModelviewMatrixStack.Top->m);
+	 }
+
+	 NORMALIZE_3FV( light->_NormSpotDirection );
+
+	 if (!(light->_Flags & LIGHT_POSITIONAL)) {
+	    GLfloat PV_dot_dir = - DOT3(light->_VP_inf_norm,
+					light->_NormSpotDirection);
+
+	    if (PV_dot_dir > light->_CosCutoff) {
+	       double x = PV_dot_dir * (EXP_TABLE_SIZE-1);
+	       int k = (int) x;
+	       light->_VP_inf_spot_attenuation =
+		  (GLfloat) (light->_SpotExpTable[k][0] +
+		   (x-k)*light->_SpotExpTable[k][1]);
+	    }
+	    else {
+	       light->_VP_inf_spot_attenuation = 0;
+            }
+	 }
+      }
+   }
+}
+
+
+
+static void
+update_modelview_scale( struct gl_context *ctx )
+{
+   ctx->_ModelViewInvScale = 1.0F;
+   if (!_math_matrix_is_length_preserving(ctx->ModelviewMatrixStack.Top)) {
+      const GLfloat *m = ctx->ModelviewMatrixStack.Top->inv;
+      GLfloat f = m[2] * m[2] + m[6] * m[6] + m[10] * m[10];
+      if (f < 1e-12) f = 1.0;
+      if (ctx->_NeedEyeCoords)
+	 ctx->_ModelViewInvScale = (GLfloat) INV_SQRTF(f);
+      else
+	 ctx->_ModelViewInvScale = (GLfloat) SQRTF(f);
+   }
+}
+
+
+/**
+ * Bring up to date any state that relies on _NeedEyeCoords.
+ */
+void
+_mesa_update_tnl_spaces( struct gl_context *ctx, GLuint new_state )
+{
+   const GLuint oldneedeyecoords = ctx->_NeedEyeCoords;
+
+   (void) new_state;
+   ctx->_NeedEyeCoords = GL_FALSE;
+
+   if (ctx->_ForceEyeCoords ||
+       (ctx->Texture._GenFlags & TEXGEN_NEED_EYE_COORD) ||
+       ctx->Point._Attenuated ||
+       ctx->Light._NeedEyeCoords)
+      ctx->_NeedEyeCoords = GL_TRUE;
+
+   if (ctx->Light.Enabled &&
+       !_math_matrix_is_length_preserving(ctx->ModelviewMatrixStack.Top))
+      ctx->_NeedEyeCoords = GL_TRUE;
+
+   /* Check if the truth-value interpretations of the bitfields have
+    * changed:
+    */
+   if (oldneedeyecoords != ctx->_NeedEyeCoords) {
+      /* Recalculate all state that depends on _NeedEyeCoords.
+       */
+      update_modelview_scale(ctx);
+      compute_light_positions( ctx );
+
+      if (ctx->Driver.LightingSpaceChange)
+	 ctx->Driver.LightingSpaceChange( ctx );
+   }
+   else {
+      GLuint new_state2 = ctx->NewState;
+
+      /* Recalculate that same state only if it has been invalidated
+       * by other statechanges.
+       */
+      if (new_state2 & _NEW_MODELVIEW)
+	 update_modelview_scale(ctx);
+
+      if (new_state2 & (_NEW_LIGHT|_NEW_MODELVIEW))
+	 compute_light_positions( ctx );
+   }
+}
+
+
+/**
+ * Drivers may need this if the hardware tnl unit doesn't support the
+ * light-in-modelspace optimization.  It's also useful for debugging.
+ */
+void
+_mesa_allow_light_in_model( struct gl_context *ctx, GLboolean flag )
+{
+   ctx->_ForceEyeCoords = !flag;
+   ctx->NewState |= _NEW_POINT;	/* one of the bits from
+				 * _MESA_NEW_NEED_EYE_COORDS.
+				 */
+}
+
+
+
+/**********************************************************************/
+/*****                      Initialization                        *****/
+/**********************************************************************/
+
+/**
+ * Initialize the n-th light data structure.
+ *
+ * \param l pointer to the gl_light structure to be initialized.
+ * \param n number of the light. 
+ * \note The defaults for light 0 are different than the other lights.
+ */
+static void
+init_light( struct gl_light *l, GLuint n )
+{
+   make_empty_list( l );
+
+   ASSIGN_4V( l->Ambient, 0.0, 0.0, 0.0, 1.0 );
+   if (n==0) {
+      ASSIGN_4V( l->Diffuse, 1.0, 1.0, 1.0, 1.0 );
+      ASSIGN_4V( l->Specular, 1.0, 1.0, 1.0, 1.0 );
+   }
+   else {
+      ASSIGN_4V( l->Diffuse, 0.0, 0.0, 0.0, 1.0 );
+      ASSIGN_4V( l->Specular, 0.0, 0.0, 0.0, 1.0 );
+   }
+   ASSIGN_4V( l->EyePosition, 0.0, 0.0, 1.0, 0.0 );
+   ASSIGN_3V( l->SpotDirection, 0.0, 0.0, -1.0 );
+   l->SpotExponent = 0.0;
+   _mesa_invalidate_spot_exp_table( l );
+   l->SpotCutoff = 180.0;
+   l->_CosCutoffNeg = -1.0f;
+   l->_CosCutoff = 0.0;		/* KW: -ve values not admitted */
+   l->ConstantAttenuation = 1.0;
+   l->LinearAttenuation = 0.0;
+   l->QuadraticAttenuation = 0.0;
+   l->Enabled = GL_FALSE;
+}
+
+
+/**
+ * Initialize the light model data structure.
+ *
+ * \param lm pointer to the gl_lightmodel structure to be initialized.
+ */
+static void
+init_lightmodel( struct gl_lightmodel *lm )
+{
+   ASSIGN_4V( lm->Ambient, 0.2F, 0.2F, 0.2F, 1.0F );
+   lm->LocalViewer = GL_FALSE;
+   lm->TwoSide = GL_FALSE;
+   lm->ColorControl = GL_SINGLE_COLOR;
+}
+
+
+/**
+ * Initialize the material data structure.
+ * 
+ * \param m pointer to the gl_material structure to be initialized.
+ */
+static void
+init_material( struct gl_material *m )
+{
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_AMBIENT],  0.2F, 0.2F, 0.2F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_DIFFUSE],  0.8F, 0.8F, 0.8F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_SPECULAR], 0.0F, 0.0F, 0.0F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_EMISSION], 0.0F, 0.0F, 0.0F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_SHININESS], 0.0F, 0.0F, 0.0F, 0.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_FRONT_INDEXES], 0.0F, 1.0F, 1.0F, 0.0F );
+ 
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_AMBIENT],  0.2F, 0.2F, 0.2F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_DIFFUSE],  0.8F, 0.8F, 0.8F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_SPECULAR], 0.0F, 0.0F, 0.0F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_EMISSION], 0.0F, 0.0F, 0.0F, 1.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_SHININESS], 0.0F, 0.0F, 0.0F, 0.0F );
+   ASSIGN_4V( m->Attrib[MAT_ATTRIB_BACK_INDEXES], 0.0F, 1.0F, 1.0F, 0.0F );
+}
+
+
+/**
+ * Initialize all lighting state for the given context.
+ */
+void
+_mesa_init_lighting( struct gl_context *ctx )
+{
+   GLuint i;
+
+   /* Lighting group */
+   for (i = 0; i < MAX_LIGHTS; i++) {
+      init_light( &ctx->Light.Light[i], i );
+   }
+   make_empty_list( &ctx->Light.EnabledList );
+
+   init_lightmodel( &ctx->Light.Model );
+   init_material( &ctx->Light.Material );
+   ctx->Light.ShadeModel = GL_SMOOTH;
+   ctx->Light.ProvokingVertex = GL_LAST_VERTEX_CONVENTION_EXT;
+   ctx->Light.Enabled = GL_FALSE;
+   ctx->Light.ColorMaterialFace = GL_FRONT_AND_BACK;
+   ctx->Light.ColorMaterialMode = GL_AMBIENT_AND_DIFFUSE;
+   ctx->Light.ColorMaterialBitmask = _mesa_material_bitmask( ctx,
+                                               GL_FRONT_AND_BACK,
+                                               GL_AMBIENT_AND_DIFFUSE, ~0,
+                                               NULL );
+
+   ctx->Light.ColorMaterialEnabled = GL_FALSE;
+   ctx->Light.ClampVertexColor = GL_TRUE;
+
+   /* Lighting miscellaneous */
+   ctx->_ShineTabList = MALLOC_STRUCT( gl_shine_tab );
+   make_empty_list( ctx->_ShineTabList );
+   /* Allocate 10 (arbitrary) shininess lookup tables */
+   for (i = 0 ; i < 10 ; i++) {
+      struct gl_shine_tab *s = MALLOC_STRUCT( gl_shine_tab );
+      s->shininess = -1;
+      s->refcount = 0;
+      insert_at_tail( ctx->_ShineTabList, s );
+   }
+
+   /* Miscellaneous */
+   ctx->Light._NeedEyeCoords = GL_FALSE;
+   ctx->_NeedEyeCoords = GL_FALSE;
+   ctx->_ForceEyeCoords = GL_FALSE;
+   ctx->_ModelViewInvScale = 1.0;
+}
+
+
+/**
+ * Deallocate malloc'd lighting state attached to given context.
+ */
+void
+_mesa_free_lighting_data( struct gl_context *ctx )
+{
+   struct gl_shine_tab *s, *tmps;
+
+   /* Free lighting shininess exponentiation table */
+   foreach_s( s, tmps, ctx->_ShineTabList ) {
+      free( s );
+   }
+   free( ctx->_ShineTabList );
+}
diff --git a/mesalib/src/mesa/main/macros.h b/mesalib/src/mesa/main/macros.h
index 2a849e36a..dbe5b867c 100644
--- a/mesalib/src/mesa/main/macros.h
+++ b/mesalib/src/mesa/main/macros.h
@@ -54,6 +54,10 @@ extern GLfloat _mesa_ubyte_to_float_color_tab[256];
 #define FLOAT_TO_BYTE(X)    ( (((GLint) (255.0F * (X))) - 1) / 2 )
 
 
+/** Convert GLbyte to GLfloat while preserving zero */
+#define BYTE_TO_FLOATZ(B)   ((B) == 0 ? 0.0F : BYTE_TO_FLOAT(B))
+
+
 /** Convert GLbyte in [-128,127] to GLfloat in [-1.0,1.0], texture/fb data */
 #define BYTE_TO_FLOAT_TEX(B)    ((B) == -128 ? -1.0F : (B) * (1.0F/127.0F))
 
@@ -73,6 +77,9 @@ extern GLfloat _mesa_ubyte_to_float_color_tab[256];
 /** Convert GLfloat in [-1.0,1.0] to GLshort in [-32768,32767] */
 #define FLOAT_TO_SHORT(X)   ( (((GLint) (65535.0F * (X))) - 1) / 2 )
 
+/** Convert GLshort to GLfloat while preserving zero */
+#define SHORT_TO_FLOATZ(S)   ((S) == 0 ? 0.0F : SHORT_TO_FLOAT(S))
+
 
 /** Convert GLshort in [-32768,32767] to GLfloat in [-1.0,1.0], texture/fb data */
 #define SHORT_TO_FLOAT_TEX(S)    ((S) == -32768 ? -1.0F : (S) * (1.0F/32767.0F))
@@ -588,19 +595,6 @@ do {				\
  */
 #define LINTERP(T, OUT, IN) ((OUT) + (T) * ((IN) - (OUT)))
 
-/* Can do better with integer math
- */
-#define INTERP_UB( t, dstub, outub, inub )  \
-do {                        \
-   GLfloat inf = UBYTE_TO_FLOAT( inub );    \
-   GLfloat outf = UBYTE_TO_FLOAT( outub );  \
-   GLfloat dstf = LINTERP( t, outf, inf );  \
-   UNCLAMPED_FLOAT_TO_UBYTE( dstub, dstf ); \
-} while (0)
-
-#define INTERP_UI( t, dstui, outui, inui )  \
-   dstui = (GLuint) (GLint) LINTERP( (t), (GLfloat) (outui), (GLfloat) (inui) )
-
 #define INTERP_F( t, dstf, outf, inf )      \
    dstf = LINTERP( t, outf, inf )
 
@@ -619,16 +613,6 @@ do {                        \
    dst[2] = LINTERP( (t), (out)[2], (in)[2] );  \
 } while (0)
 
-#define INTERP_SZ( t, vec, to, out, in, sz )                \
-do {                                    \
-   switch (sz) {                            \
-   case 4: vec[to][3] = LINTERP( (t), (vec)[out][3], (vec)[in][3] );    \
-   case 3: vec[to][2] = LINTERP( (t), (vec)[out][2], (vec)[in][2] );    \
-   case 2: vec[to][1] = LINTERP( (t), (vec)[out][1], (vec)[in][1] );    \
-   case 1: vec[to][0] = LINTERP( (t), (vec)[out][0], (vec)[in][0] );    \
-   }                                    \
-} while(0)
-
 /*@}*/
 
 
@@ -656,9 +640,6 @@ do {                                    \
 #define DOT4( a, b )  ( (a)[0]*(b)[0] + (a)[1]*(b)[1] + \
             (a)[2]*(b)[2] + (a)[3]*(b)[3] )
 
-/** Dot product of two 4-element vectors */
-#define DOT4V(v,a,b,c,d) (v[0]*(a) + v[1]*(b) + v[2]*(c) + v[3]*(d))
-
 
 /** Cross product of two 3-element vectors */
 #define CROSS3(n, u, v)             \
@@ -688,6 +669,10 @@ do {                        \
 #define LEN_SQUARED_2FV( V ) ((V)[0]*(V)[0]+(V)[1]*(V)[1])
 
 
+/** Compute ceiling of integer quotient of A divided by B. */
+#define CEILING( A, B )  ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )
+
+
 /** casts to silence warnings with some compilers */
 #define ENUM_TO_INT(E)     ((GLint)(E))
 #define ENUM_TO_FLOAT(E)   ((GLfloat)(GLint)(E))
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index deab97d3e..285ec0783 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -49,18 +49,6 @@ extern "C" {
 
 
 /**
- * Stencil buffer data type.
- */
-#if STENCIL_BITS==8
-   typedef GLubyte GLstencil;
-#elif STENCIL_BITS==16
-   typedef GLushort GLstencil;
-#else
-#  error "illegal number of stencil bits"
-#endif
-
-
-/**
  * \name 64-bit extension of GLbitfield.
  */
 /*@{*/
@@ -2158,6 +2146,16 @@ struct gl_shader
 
    unsigned Version;       /**< GLSL version used for linking */
 
+   unsigned num_samplers;	/**< Number of samplers used by this shader.
+				 * This field is only set post-linking.
+				 */
+   /**
+    * Number of uniform components used by this shader.
+    *
+    * This field is only set post-linking.
+    */
+   unsigned num_uniform_components;
+
    struct exec_list *ir;
    struct glsl_symbol_table *symbols;
 
diff --git a/mesalib/src/mesa/main/pack.c b/mesalib/src/mesa/main/pack.c
index 539a06c9a..6f48a2e7c 100644
--- a/mesalib/src/mesa/main/pack.c
+++ b/mesalib/src/mesa/main/pack.c
@@ -34,6 +34,7 @@
 #include "enums.h"
 #include "image.h"
 #include "imports.h"
+#include "macros.h"
 #include "mtypes.h"
 #include "pack.h"
 #include "pixeltransfer.h"
@@ -43,26 +44,6 @@
 
 
 /**
- * NOTE:
- * Normally, BYTE_TO_FLOAT(0) returns 0.00392  That causes problems when
- * we later convert the float to a packed integer value (such as for
- * GL_RGB5_A1) because we'll wind up with a non-zero value.
- *
- * We redefine the macros here so zero is handled correctly.
- */
-#undef BYTE_TO_FLOAT
-#define BYTE_TO_FLOAT(B)    ((B) == 0 ? 0.0F : ((2.0F * (B) + 1.0F) * (1.0F/255.0F)))
-
-#undef SHORT_TO_FLOAT
-#define SHORT_TO_FLOAT(S)   ((S) == 0 ? 0.0F : ((2.0F * (S) + 1.0F) * (1.0F/65535.0F)))
-
-
-
-/** Compute ceiling of integer quotient of A divided by B. */
-#define CEILING( A, B )  ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )
-
-
-/**
  * Flip the 8 bits in each byte of the given array.
  *
  * \param p array.
@@ -2507,10 +2488,10 @@ extract_float_rgba(GLuint n, GLfloat rgba[][4],
          PROCESS(aSrc, ACOMP, 1.0F, 255, GLubyte, UBYTE_TO_FLOAT);
          break;
       case GL_BYTE:
-         PROCESS(rSrc, RCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOAT);
-         PROCESS(gSrc, GCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOAT);
-         PROCESS(bSrc, BCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOAT);
-         PROCESS(aSrc, ACOMP, 1.0F, 127, GLbyte, BYTE_TO_FLOAT);
+         PROCESS(rSrc, RCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOATZ);
+         PROCESS(gSrc, GCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOATZ);
+         PROCESS(bSrc, BCOMP, 0.0F,   0, GLbyte, BYTE_TO_FLOATZ);
+         PROCESS(aSrc, ACOMP, 1.0F, 127, GLbyte, BYTE_TO_FLOATZ);
          break;
       case GL_UNSIGNED_SHORT:
          PROCESS(rSrc, RCOMP, 0.0F,      0, GLushort, USHORT_TO_FLOAT);
@@ -2519,10 +2500,10 @@ extract_float_rgba(GLuint n, GLfloat rgba[][4],
          PROCESS(aSrc, ACOMP, 1.0F, 0xffff, GLushort, USHORT_TO_FLOAT);
          break;
       case GL_SHORT:
-         PROCESS(rSrc, RCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOAT);
-         PROCESS(gSrc, GCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOAT);
-         PROCESS(bSrc, BCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOAT);
-         PROCESS(aSrc, ACOMP, 1.0F, 32767, GLshort, SHORT_TO_FLOAT);
+         PROCESS(rSrc, RCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOATZ);
+         PROCESS(gSrc, GCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOATZ);
+         PROCESS(bSrc, BCOMP, 0.0F,     0, GLshort, SHORT_TO_FLOATZ);
+         PROCESS(aSrc, ACOMP, 1.0F, 32767, GLshort, SHORT_TO_FLOATZ);
          break;
       case GL_UNSIGNED_INT:
          PROCESS(rSrc, RCOMP, 0.0F,          0, GLuint, UINT_TO_FLOAT);
@@ -4586,10 +4567,10 @@ _mesa_unpack_stencil_span( struct gl_context *ctx, GLuint n,
 
 void
 _mesa_pack_stencil_span( struct gl_context *ctx, GLuint n,
-                         GLenum dstType, GLvoid *dest, const GLstencil *source,
+                         GLenum dstType, GLvoid *dest, const GLubyte *source,
                          const struct gl_pixelstore_attrib *dstPacking )
 {
-   GLstencil *stencil = (GLstencil *) malloc(n * sizeof(GLstencil));
+   GLubyte *stencil = (GLubyte *) malloc(n * sizeof(GLubyte));
 
    if (!stencil) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "stencil packing");
@@ -4599,23 +4580,14 @@ _mesa_pack_stencil_span( struct gl_context *ctx, GLuint n,
    if (ctx->Pixel.IndexShift || ctx->Pixel.IndexOffset ||
        ctx->Pixel.MapStencilFlag) {
       /* make a copy of input */
-      memcpy(stencil, source, n * sizeof(GLstencil));
+      memcpy(stencil, source, n * sizeof(GLubyte));
       _mesa_apply_stencil_transfer_ops(ctx, n, stencil);
       source = stencil;
    }
 
    switch (dstType) {
    case GL_UNSIGNED_BYTE:
-      if (sizeof(GLstencil) == 1) {
-         memcpy( dest, source, n );
-      }
-      else {
-         GLubyte *dst = (GLubyte *) dest;
-         GLuint i;
-         for (i=0;i<n;i++) {
-            dst[i] = (GLubyte) source[i];
-         }
-      }
+      memcpy(dest, source, n);
       break;
    case GL_BYTE:
       {
@@ -4834,14 +4806,14 @@ _mesa_unpack_depth_span( struct gl_context *ctx, GLuint n,
     */
    switch (srcType) {
       case GL_BYTE:
-         DEPTH_VALUES(GLbyte, BYTE_TO_FLOAT);
+         DEPTH_VALUES(GLbyte, BYTE_TO_FLOATZ);
          needClamp = GL_TRUE;
          break;
       case GL_UNSIGNED_BYTE:
          DEPTH_VALUES(GLubyte, UBYTE_TO_FLOAT);
          break;
       case GL_SHORT:
-         DEPTH_VALUES(GLshort, SHORT_TO_FLOAT);
+         DEPTH_VALUES(GLshort, SHORT_TO_FLOATZ);
          needClamp = GL_TRUE;
          break;
       case GL_UNSIGNED_SHORT:
@@ -5120,11 +5092,11 @@ void
 _mesa_pack_depth_stencil_span(struct gl_context *ctx,GLuint n,
                               GLenum dstType, GLuint *dest,
                               const GLfloat *depthVals,
-                              const GLstencil *stencilVals,
+                              const GLubyte *stencilVals,
                               const struct gl_pixelstore_attrib *dstPacking)
 {
    GLfloat *depthCopy = (GLfloat *) malloc(n * sizeof(GLfloat));
-   GLstencil *stencilCopy = (GLstencil *) malloc(n * sizeof(GLstencil));
+   GLubyte *stencilCopy = (GLubyte *) malloc(n * sizeof(GLubyte));
    GLuint i;
 
    if (!depthCopy || !stencilCopy) {
@@ -5143,7 +5115,7 @@ _mesa_pack_depth_stencil_span(struct gl_context *ctx,GLuint n,
    if (ctx->Pixel.IndexShift ||
        ctx->Pixel.IndexOffset ||
        ctx->Pixel.MapStencilFlag) {
-      memcpy(stencilCopy, stencilVals, n * sizeof(GLstencil));
+      memcpy(stencilCopy, stencilVals, n * sizeof(GLubyte));
       _mesa_apply_stencil_transfer_ops(ctx, n, stencilCopy);
       stencilVals = stencilCopy;
    }
diff --git a/mesalib/src/mesa/main/pack.h b/mesalib/src/mesa/main/pack.h
index 7a0089c2f..b1853cd59 100644
--- a/mesalib/src/mesa/main/pack.h
+++ b/mesalib/src/mesa/main/pack.h
@@ -113,7 +113,7 @@ _mesa_unpack_stencil_span(struct gl_context *ctx, GLuint n,
 
 extern void
 _mesa_pack_stencil_span(struct gl_context *ctx, GLuint n,
-                        GLenum dstType, GLvoid *dest, const GLstencil *source,
+                        GLenum dstType, GLvoid *dest, const GLubyte *source,
                         const struct gl_pixelstore_attrib *dstPacking);
 
 
@@ -133,7 +133,7 @@ extern void
 _mesa_pack_depth_stencil_span(struct gl_context *ctx,GLuint n,
                               GLenum dstType, GLuint *dest,
                               const GLfloat *depthVals,
-                              const GLstencil *stencilVals,
+                              const GLubyte *stencilVals,
                               const struct gl_pixelstore_attrib *dstPacking);
 
 
diff --git a/mesalib/src/mesa/main/pixeltransfer.c b/mesalib/src/mesa/main/pixeltransfer.c
index 5e881436a..5c167e0a9 100644
--- a/mesalib/src/mesa/main/pixeltransfer.c
+++ b/mesalib/src/mesa/main/pixeltransfer.c
@@ -273,7 +273,7 @@ _mesa_apply_ci_transfer_ops(const struct gl_context *ctx,
  */
 void
 _mesa_apply_stencil_transfer_ops(const struct gl_context *ctx, GLuint n,
-                                 GLstencil stencil[])
+                                 GLubyte stencil[])
 {
    if (ctx->Pixel.IndexShift != 0 || ctx->Pixel.IndexOffset != 0) {
       const GLint offset = ctx->Pixel.IndexOffset;
@@ -300,7 +300,7 @@ _mesa_apply_stencil_transfer_ops(const struct gl_context *ctx, GLuint n,
       GLuint mask = ctx->PixelMaps.StoS.Size - 1;
       GLuint i;
       for (i = 0; i < n; i++) {
-         stencil[i] = (GLstencil)ctx->PixelMaps.StoS.Map[ stencil[i] & mask ];
+         stencil[i] = (GLubyte) ctx->PixelMaps.StoS.Map[ stencil[i] & mask ];
       }
    }
 }
diff --git a/mesalib/src/mesa/main/pixeltransfer.h b/mesalib/src/mesa/main/pixeltransfer.h
index 8af2e9ee2..a8c14757f 100644
--- a/mesalib/src/mesa/main/pixeltransfer.h
+++ b/mesalib/src/mesa/main/pixeltransfer.h
@@ -75,7 +75,7 @@ _mesa_apply_ci_transfer_ops(const struct gl_context *ctx,
 
 extern void
 _mesa_apply_stencil_transfer_ops(const struct gl_context *ctx, GLuint n,
-                                 GLstencil stencil[]);
+                                 GLubyte stencil[]);
 
 
 #endif
diff --git a/mesalib/src/mesa/main/readpix.c b/mesalib/src/mesa/main/readpix.c
index 84c5b2228..86b87534d 100644
--- a/mesalib/src/mesa/main/readpix.c
+++ b/mesalib/src/mesa/main/readpix.c
@@ -30,13 +30,509 @@
 #include "readpix.h"
 #include "framebuffer.h"
 #include "formats.h"
+#include "format_unpack.h"
 #include "image.h"
 #include "mtypes.h"
+#include "pack.h"
 #include "pbo.h"
 #include "state.h"
 
 
 /**
+ * Tries to implement glReadPixels() of GL_DEPTH_COMPONENT using memcpy of the
+ * mapping.
+ */
+static GLboolean
+fast_read_depth_pixels( struct gl_context *ctx,
+			GLint x, GLint y,
+			GLsizei width, GLsizei height,
+			GLenum type, GLvoid *pixels,
+			const struct gl_pixelstore_attrib *packing )
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   GLubyte *map, *dst;
+   int stride, dstStride, j;
+
+   if (ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0)
+      return GL_FALSE;
+
+   if (packing->SwapBytes)
+      return GL_FALSE;
+
+   if (_mesa_get_format_datatype(rb->Format) != GL_UNSIGNED_INT)
+      return GL_FALSE;
+
+   if (!((type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16) ||
+	 type == GL_UNSIGNED_INT))
+      return GL_FALSE;
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   dstStride = _mesa_image_row_stride(packing, width, GL_DEPTH_COMPONENT, type);
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+					   GL_DEPTH_COMPONENT, type, 0, 0);
+
+   for (j = 0; j < height; j++) {
+      if (type == GL_UNSIGNED_INT) {
+	 _mesa_unpack_uint_z_row(rb->Format, width, map, (GLuint *)dst);
+      } else {
+	 ASSERT(type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16);
+	 memcpy(dst, map, width * 2);
+      }
+
+      map += stride;
+      dst += dstStride;
+   }
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+
+   return GL_TRUE;
+}
+
+/**
+ * Read pixels for format=GL_DEPTH_COMPONENT.
+ */
+static void
+read_depth_pixels( struct gl_context *ctx,
+                   GLint x, GLint y,
+                   GLsizei width, GLsizei height,
+                   GLenum type, GLvoid *pixels,
+                   const struct gl_pixelstore_attrib *packing )
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   GLint j;
+   GLubyte *dst, *map;
+   int dstStride, stride;
+
+   if (!rb)
+      return;
+
+   /* clipping should have been done already */
+   ASSERT(x >= 0);
+   ASSERT(y >= 0);
+   ASSERT(x + width <= (GLint) rb->Width);
+   ASSERT(y + height <= (GLint) rb->Height);
+   /* width should never be > MAX_WIDTH since we did clipping earlier */
+   ASSERT(width <= MAX_WIDTH);
+
+   if (fast_read_depth_pixels(ctx, x, y, width, height, type, pixels, packing))
+      return;
+
+   dstStride = _mesa_image_row_stride(packing, width, GL_DEPTH_COMPONENT, type);
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+					   GL_DEPTH_COMPONENT, type, 0, 0);
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   /* General case (slower) */
+   for (j = 0; j < height; j++, y++) {
+      GLfloat depthValues[MAX_WIDTH];
+      _mesa_unpack_float_z_row(rb->Format, width, map, depthValues);
+      _mesa_pack_depth_span(ctx, width, dst, type, depthValues, packing);
+
+      dst += dstStride;
+      map += stride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+}
+
+
+/**
+ * Read pixels for format=GL_STENCIL_INDEX.
+ */
+static void
+read_stencil_pixels( struct gl_context *ctx,
+                     GLint x, GLint y,
+                     GLsizei width, GLsizei height,
+                     GLenum type, GLvoid *pixels,
+                     const struct gl_pixelstore_attrib *packing )
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
+   GLint j;
+   GLubyte *map;
+   GLint stride;
+
+   if (!rb)
+      return;
+
+   /* width should never be > MAX_WIDTH since we did clipping earlier */
+   ASSERT(width <= MAX_WIDTH);
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   /* process image row by row */
+   for (j = 0; j < height; j++) {
+      GLvoid *dest;
+      GLubyte stencil[MAX_WIDTH];
+
+      _mesa_unpack_ubyte_stencil_row(rb->Format, width, map, stencil);
+      dest = _mesa_image_address2d(packing, pixels, width, height,
+                                   GL_STENCIL_INDEX, type, j, 0);
+
+      _mesa_pack_stencil_span(ctx, width, type, dest, stencil, packing);
+
+      map += stride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+}
+
+static GLboolean
+fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
+			      GLint x, GLint y,
+			      GLsizei width, GLsizei height,
+			      GLenum format, GLenum type,
+			      GLvoid *pixels,
+			      const struct gl_pixelstore_attrib *packing,
+			      GLbitfield transferOps )
+{
+   struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
+   GLubyte *dst, *map;
+   int dstStride, stride, j, texelBytes;
+
+   if (!_mesa_format_matches_format_and_type(rb->Format, format, type))
+      return GL_FALSE;
+
+   /* check for things we can't handle here */
+   if (packing->SwapBytes ||
+       packing->LsbFirst) {
+      return GL_FALSE;
+   }
+
+   dstStride = _mesa_image_row_stride(packing, width, format, type);
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+					   format, type, 0, 0);
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   texelBytes = _mesa_get_format_bytes(rb->Format);
+   for (j = 0; j < height; j++) {
+      memcpy(dst, map, width * texelBytes);
+      dst += dstStride;
+      map += stride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+
+   return GL_TRUE;
+}
+
+static GLboolean
+slow_read_rgba_pixels( struct gl_context *ctx,
+		       GLint x, GLint y,
+		       GLsizei width, GLsizei height,
+		       GLenum format, GLenum type,
+		       GLvoid *pixels,
+		       const struct gl_pixelstore_attrib *packing,
+		       GLbitfield transferOps )
+{
+   struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
+   const gl_format rbFormat = _mesa_get_srgb_format_linear(rb->Format);
+   union {
+      float f[MAX_WIDTH][4];
+      unsigned int i[MAX_WIDTH][4];
+   } rgba;
+   GLubyte *dst, *map;
+   int dstStride, stride, j;
+
+   dstStride = _mesa_image_row_stride(packing, width, format, type);
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+					   format, type, 0, 0);
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   for (j = 0; j < height; j++) {
+      if (_mesa_is_integer_format(format)) {
+	 _mesa_unpack_int_rgba_row(rbFormat, width, map, rgba.i);
+	 _mesa_pack_rgba_span_int(ctx, width, rgba.i, format, type, dst);
+      } else {
+	 _mesa_unpack_rgba_row(rbFormat, width, map, rgba.f);
+	 _mesa_pack_rgba_span_float(ctx, width, rgba.f, format, type, dst,
+				    packing, transferOps);
+      }
+      dst += dstStride;
+      map += stride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+
+   return GL_TRUE;
+}
+
+/*
+ * Read R, G, B, A, RGB, L, or LA pixels.
+ */
+static void
+read_rgba_pixels( struct gl_context *ctx,
+                  GLint x, GLint y,
+                  GLsizei width, GLsizei height,
+                  GLenum format, GLenum type, GLvoid *pixels,
+                  const struct gl_pixelstore_attrib *packing )
+{
+   GLbitfield transferOps = ctx->_ImageTransferState;
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->_ColorReadBuffer;
+
+   if (!rb)
+      return;
+
+   if ((ctx->Color._ClampReadColor == GL_TRUE || type != GL_FLOAT) &&
+       !_mesa_is_integer_format(format)) {
+      transferOps |= IMAGE_CLAMP_BIT;
+   }
+
+   if (!transferOps) {
+      /* Try the optimized paths first. */
+      if (fast_read_rgba_pixels_memcpy(ctx, x, y, width, height,
+				       format, type, pixels, packing,
+				       transferOps)) {
+	 return;
+      }
+   }
+
+   slow_read_rgba_pixels(ctx, x, y, width, height,
+			 format, type, pixels, packing, transferOps);
+}
+
+/**
+ * For a packed depth/stencil buffer being read as depth/stencil, just memcpy the
+ * data (possibly swapping 8/24 vs 24/8 as we go).
+ */
+static GLboolean
+fast_read_depth_stencil_pixels(struct gl_context *ctx,
+			       GLint x, GLint y,
+			       GLsizei width, GLsizei height,
+			       GLubyte *dst, int dstStride)
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
+   GLubyte *map;
+   int stride, i;
+
+   if (rb != stencilRb)
+      return GL_FALSE;
+
+   if (rb->Format != MESA_FORMAT_Z24_S8 &&
+       rb->Format != MESA_FORMAT_S8_Z24)
+      return GL_FALSE;
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+			       &map, &stride);
+
+   for (i = 0; i < height; i++) {
+      _mesa_unpack_uint_24_8_depth_stencil_row(rb->Format, width,
+					       map, (GLuint *)dst);
+      map += stride;
+      dst += dstStride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
+
+   return GL_TRUE;
+}
+
+
+/**
+ * For non-float-depth and stencil buffers being read as 24/8 depth/stencil,
+ * copy the integer data directly instead of converting depth to float and
+ * re-packing.
+ */
+static GLboolean
+fast_read_depth_stencil_pixels_separate(struct gl_context *ctx,
+					GLint x, GLint y,
+					GLsizei width, GLsizei height,
+					uint32_t *dst, int dstStride)
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *depthRb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
+   GLubyte *depthMap, *stencilMap;
+   int depthStride, stencilStride, i, j;
+
+   if (_mesa_get_format_datatype(depthRb->Format) != GL_UNSIGNED_INT)
+      return GL_FALSE;
+
+   ctx->Driver.MapRenderbuffer(ctx, depthRb, x, y, width, height,
+			       GL_MAP_READ_BIT, &depthMap, &depthStride);
+   ctx->Driver.MapRenderbuffer(ctx, stencilRb, x, y, width, height,
+			       GL_MAP_READ_BIT, &stencilMap, &stencilStride);
+
+   for (j = 0; j < height; j++) {
+      GLubyte stencilVals[MAX_WIDTH];
+
+      _mesa_unpack_uint_z_row(depthRb->Format, width, depthMap, dst);
+      _mesa_unpack_ubyte_stencil_row(stencilRb->Format, width,
+				     stencilMap, stencilVals);
+
+      for (i = 0; i < width; i++) {
+	 dst[i] = (dst[i] & 0xffffff00) | stencilVals[i];
+      }
+
+      depthMap += depthStride;
+      stencilMap += stencilStride;
+      dst += dstStride / 4;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, depthRb);
+   ctx->Driver.UnmapRenderbuffer(ctx, stencilRb);
+
+   return GL_TRUE;
+}
+
+static void
+slow_read_depth_stencil_pixels_separate(struct gl_context *ctx,
+					GLint x, GLint y,
+					GLsizei width, GLsizei height,
+					GLenum type,
+					const struct gl_pixelstore_attrib *packing,
+					GLubyte *dst, int dstStride)
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *depthRb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
+   GLubyte *depthMap, *stencilMap;
+   int depthStride, stencilStride, j;
+
+   /* The depth and stencil buffers might be separate, or a single buffer.
+    * If one buffer, only map it once.
+    */
+   ctx->Driver.MapRenderbuffer(ctx, depthRb, x, y, width, height,
+			       GL_MAP_READ_BIT, &depthMap, &depthStride);
+   if (stencilRb != depthRb) {
+      ctx->Driver.MapRenderbuffer(ctx, stencilRb, x, y, width, height,
+                                  GL_MAP_READ_BIT, &stencilMap,
+                                  &stencilStride);
+   }
+   else {
+      stencilMap = depthMap;
+      stencilStride = depthStride;
+   }
+
+   for (j = 0; j < height; j++) {
+      GLubyte stencilVals[MAX_WIDTH];
+      GLfloat depthVals[MAX_WIDTH];
+
+      _mesa_unpack_float_z_row(depthRb->Format, width, depthMap, depthVals);
+      _mesa_unpack_ubyte_stencil_row(stencilRb->Format, width,
+				     stencilMap, stencilVals);
+
+      _mesa_pack_depth_stencil_span(ctx, width, type, (GLuint *)dst,
+				    depthVals, stencilVals, packing);
+
+      depthMap += depthStride;
+      stencilMap += stencilStride;
+      dst += dstStride;
+   }
+
+   ctx->Driver.UnmapRenderbuffer(ctx, depthRb);
+   if (stencilRb != depthRb) {
+      ctx->Driver.UnmapRenderbuffer(ctx, stencilRb);
+   }
+}
+
+
+/**
+ * Read combined depth/stencil values.
+ * We'll have already done error checking to be sure the expected
+ * depth and stencil buffers really exist.
+ */
+static void
+read_depth_stencil_pixels(struct gl_context *ctx,
+                          GLint x, GLint y,
+                          GLsizei width, GLsizei height,
+                          GLenum type, GLvoid *pixels,
+                          const struct gl_pixelstore_attrib *packing )
+{
+   const GLboolean scaleOrBias
+      = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
+   const GLboolean stencilTransfer = ctx->Pixel.IndexShift
+      || ctx->Pixel.IndexOffset || ctx->Pixel.MapStencilFlag;
+   GLubyte *dst;
+   int dstStride;
+
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels,
+					   width, height,
+					   GL_DEPTH_STENCIL_EXT,
+					   type, 0, 0);
+   dstStride = _mesa_image_row_stride(packing, width,
+				      GL_DEPTH_STENCIL_EXT, type);
+
+   /* Fast 24/8 reads. */
+   if (type == GL_UNSIGNED_INT_24_8 &&
+       !scaleOrBias && !stencilTransfer && !packing->SwapBytes) {
+      if (fast_read_depth_stencil_pixels(ctx, x, y, width, height,
+					 dst, dstStride))
+	 return;
+
+      if (fast_read_depth_stencil_pixels_separate(ctx, x, y, width, height,
+						  (uint32_t *)dst, dstStride))
+	 return;
+   }
+
+   slow_read_depth_stencil_pixels_separate(ctx, x, y, width, height,
+					   type, packing,
+					   dst, dstStride);
+}
+
+
+
+/**
+ * Software fallback routine for ctx->Driver.ReadPixels().
+ * By time we get here, all error checking will have been done.
+ */
+void
+_mesa_readpixels(struct gl_context *ctx,
+                 GLint x, GLint y, GLsizei width, GLsizei height,
+                 GLenum format, GLenum type,
+                 const struct gl_pixelstore_attrib *packing,
+                 GLvoid *pixels)
+{
+   struct gl_pixelstore_attrib clippedPacking = *packing;
+
+   if (ctx->NewState)
+      _mesa_update_state(ctx);
+
+   /* Do all needed clipping here, so that we can forget about it later */
+   if (_mesa_clip_readpixels(ctx, &x, &y, &width, &height, &clippedPacking)) {
+
+      pixels = _mesa_map_pbo_dest(ctx, &clippedPacking, pixels);
+
+      if (pixels) {
+         switch (format) {
+         case GL_STENCIL_INDEX:
+            read_stencil_pixels(ctx, x, y, width, height, type, pixels,
+                                &clippedPacking);
+            break;
+         case GL_DEPTH_COMPONENT:
+            read_depth_pixels(ctx, x, y, width, height, type, pixels,
+                              &clippedPacking);
+            break;
+         case GL_DEPTH_STENCIL_EXT:
+            read_depth_stencil_pixels(ctx, x, y, width, height, type, pixels,
+                                      &clippedPacking);
+            break;
+         default:
+            /* all other formats should be color formats */
+            read_rgba_pixels(ctx, x, y, width, height, format, type, pixels,
+                             &clippedPacking);
+         }
+
+         _mesa_unmap_pbo_dest(ctx, &clippedPacking);
+      }
+   }
+}
+
+
+/**
  * Do error checking of the format/type parameters to glReadPixels and
  * glDrawPixels.
  * \param drawing if GL_TRUE do checking for DrawPixels, else do checking
diff --git a/mesalib/src/mesa/main/readpix.h b/mesalib/src/mesa/main/readpix.h
index f6bb3d6e2..6caaf3adc 100644
--- a/mesalib/src/mesa/main/readpix.h
+++ b/mesalib/src/mesa/main/readpix.h
@@ -30,11 +30,20 @@
 #include "glheader.h"
 
 struct gl_context;
+struct gl_pixelstore_attrib;
+
 
 extern GLboolean
 _mesa_error_check_format_type(struct gl_context *ctx, GLenum format, GLenum type,
                               GLboolean drawing);
 
+extern void
+_mesa_readpixels(struct gl_context *ctx,
+                 GLint x, GLint y, GLsizei width, GLsizei height,
+                 GLenum format, GLenum type,
+                 const struct gl_pixelstore_attrib *packing,
+                 GLvoid *pixels);
+
 extern void GLAPIENTRY
 _mesa_ReadPixels( GLint x, GLint y, GLsizei width, GLsizei height,
                   GLenum format, GLenum type, GLvoid *pixels );
diff --git a/mesalib/src/mesa/main/renderbuffer.c b/mesalib/src/mesa/main/renderbuffer.c
index 4415dbd4f..c9a16a989 100644
--- a/mesalib/src/mesa/main/renderbuffer.c
+++ b/mesalib/src/mesa/main/renderbuffer.c
@@ -28,11 +28,6 @@
  * Also, routines for reading/writing software-based renderbuffer data as
  * ubytes, ushorts, uints, etc.
  *
- * The 'alpha8' renderbuffer is interesting.  It's used to add a software-based
- * alpha channel to RGB renderbuffers.  This is done by wrapping the RGB
- * renderbuffer with the alpha renderbuffer.  We can do this because of the
- * OO-nature of renderbuffers.
- *
  * Down the road we'll use this for run-time support of 8, 16 and 32-bit
  * color channels.  For example, Mesa may use 32-bit/float color channels
  * internally (swrast) and use wrapper renderbuffers to convert 32-bit
@@ -1470,18 +1465,6 @@ _mesa_set_renderbuffer_accessors(struct gl_renderbuffer *rb)
       rb->PutMonoValues = put_mono_values_ushort4;
       break;
 
-#if 0
-   case MESA_FORMAT_A8:
-      rb->DataType = GL_UNSIGNED_BYTE;
-      rb->GetValues = get_values_alpha8;
-      rb->PutRow = put_row_alpha8;
-      rb->PutRowRGB = NULL;
-      rb->PutMonoRow = put_mono_row_alpha8;
-      rb->PutValues = put_values_alpha8;
-      rb->PutMonoValues = put_mono_values_alpha8;
-      break;
-#endif
-
    case MESA_FORMAT_S8:
       rb->DataType = GL_UNSIGNED_BYTE;
       rb->GetValues = get_values_ubyte;
@@ -1639,11 +1622,6 @@ _mesa_soft_renderbuffer_storage(struct gl_context *ctx, struct gl_renderbuffer *
       /* for accum buffer */
       rb->Format = MESA_FORMAT_SIGNED_RGBA_16;
       break;
-#if 0
-   case GL_ALPHA8:
-      rb->Format = MESA_FORMAT_A8;
-      break;
-#endif
    case GL_STENCIL_INDEX:
    case GL_STENCIL_INDEX1_EXT:
    case GL_STENCIL_INDEX4_EXT:
@@ -1754,236 +1732,6 @@ _mesa_unmap_soft_renderbuffer(struct gl_context *ctx,
 }
 
 
-/**********************************************************************/
-/**********************************************************************/
-/**********************************************************************/
-
-
-/**
- * Here we utilize the gl_renderbuffer->Wrapper field to put an alpha
- * buffer wrapper around an existing RGB renderbuffer (hw or sw).
- *
- * When PutRow is called (for example), we store the alpha values in
- * this buffer, then pass on the PutRow call to the wrapped RGB
- * buffer.
- */
-
-
-static GLboolean
-alloc_storage_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb,
-                     GLenum internalFormat, GLuint width, GLuint height)
-{
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->Format == MESA_FORMAT_A8);
-
-   /* first, pass the call to the wrapped RGB buffer */
-   if (!arb->Wrapped->AllocStorage(ctx, arb->Wrapped, internalFormat,
-                                  width, height)) {
-      return GL_FALSE;
-   }
-
-   /* next, resize my alpha buffer */
-   if (arb->Data) {
-      free(arb->Data);
-   }
-
-   arb->Data = malloc(width * height * sizeof(GLubyte));
-   if (arb->Data == NULL) {
-      arb->Width = 0;
-      arb->Height = 0;
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "software alpha buffer allocation");
-      return GL_FALSE;
-   }
-
-   arb->Width = width;
-   arb->Height = height;
-   arb->RowStride = width;
-
-   return GL_TRUE;
-}
-
-
-/**
- * Delete an alpha_renderbuffer object, as well as the wrapped RGB buffer.
- */
-static void
-delete_renderbuffer_alpha8(struct gl_renderbuffer *arb)
-{
-   if (arb->Data) {
-      free(arb->Data);
-   }
-   ASSERT(arb->Wrapped);
-   ASSERT(arb != arb->Wrapped);
-   arb->Wrapped->Delete(arb->Wrapped);
-   arb->Wrapped = NULL;
-   free(arb);
-}
-
-
-static void *
-get_pointer_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb,
-                   GLint x, GLint y)
-{
-   return NULL;   /* don't allow direct access! */
-}
-
-
-static void
-get_row_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-               GLint x, GLint y, void *values)
-{
-   /* NOTE: 'values' is RGBA format! */
-   const GLubyte *src = (const GLubyte *) arb->Data + y * arb->RowStride + x;
-   GLubyte *dst = (GLubyte *) values;
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->GetRow(ctx, arb->Wrapped, count, x, y, values);
-   /* second, fill in alpha values from this buffer! */
-   for (i = 0; i < count; i++) {
-      dst[i * 4 + 3] = src[i];
-   }
-}
-
-
-static void
-get_values_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-                  const GLint x[], const GLint y[], void *values)
-{
-   GLubyte *dst = (GLubyte *) values;
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->GetValues(ctx, arb->Wrapped, count, x, y, values);
-   /* second, fill in alpha values from this buffer! */
-   for (i = 0; i < count; i++) {
-      const GLubyte *src = (GLubyte *) arb->Data + y[i] * arb->RowStride + x[i];
-      dst[i * 4 + 3] = *src;
-   }
-}
-
-
-static void
-put_row_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-               GLint x, GLint y, const void *values, const GLubyte *mask)
-{
-   const GLubyte *src = (const GLubyte *) values;
-   GLubyte *dst = (GLubyte *) arb->Data + y * arb->RowStride + x;
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->PutRow(ctx, arb->Wrapped, count, x, y, values, mask);
-   /* second, store alpha in our buffer */
-   for (i = 0; i < count; i++) {
-      if (!mask || mask[i]) {
-         dst[i] = src[i * 4 + 3];
-      }
-   }
-}
-
-
-static void
-put_row_rgb_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-                   GLint x, GLint y, const void *values, const GLubyte *mask)
-{
-   const GLubyte *src = (const GLubyte *) values;
-   GLubyte *dst = (GLubyte *) arb->Data + y * arb->RowStride + x;
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->PutRowRGB(ctx, arb->Wrapped, count, x, y, values, mask);
-   /* second, store alpha in our buffer */
-   for (i = 0; i < count; i++) {
-      if (!mask || mask[i]) {
-         dst[i] = src[i * 4 + 3];
-      }
-   }
-}
-
-
-static void
-put_mono_row_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-                    GLint x, GLint y, const void *value, const GLubyte *mask)
-{
-   const GLubyte val = ((const GLubyte *) value)[3];
-   GLubyte *dst = (GLubyte *) arb->Data + y * arb->RowStride + x;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->PutMonoRow(ctx, arb->Wrapped, count, x, y, value, mask);
-   /* second, store alpha in our buffer */
-   if (mask) {
-      GLuint i;
-      for (i = 0; i < count; i++) {
-         if (mask[i]) {
-            dst[i] = val;
-         }
-      }
-   }
-   else {
-      memset(dst, val, count);
-   }
-}
-
-
-static void
-put_values_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb, GLuint count,
-                  const GLint x[], const GLint y[],
-                  const void *values, const GLubyte *mask)
-{
-   const GLubyte *src = (const GLubyte *) values;
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->PutValues(ctx, arb->Wrapped, count, x, y, values, mask);
-   /* second, store alpha in our buffer */
-   for (i = 0; i < count; i++) {
-      if (!mask || mask[i]) {
-         GLubyte *dst = (GLubyte *) arb->Data + y[i] * arb->RowStride + x[i];
-         *dst = src[i * 4 + 3];
-      }
-   }
-}
-
-
-static void
-put_mono_values_alpha8(struct gl_context *ctx, struct gl_renderbuffer *arb,
-                       GLuint count, const GLint x[], const GLint y[],
-                       const void *value, const GLubyte *mask)
-{
-   const GLubyte val = ((const GLubyte *) value)[3];
-   GLuint i;
-   ASSERT(arb != arb->Wrapped);
-   ASSERT(arb->DataType == GL_UNSIGNED_BYTE);
-   /* first, pass the call to the wrapped RGB buffer */
-   arb->Wrapped->PutValues(ctx, arb->Wrapped, count, x, y, value, mask);
-   /* second, store alpha in our buffer */
-   for (i = 0; i < count; i++) {
-      if (!mask || mask[i]) {
-         GLubyte *dst = (GLubyte *) arb->Data + y[i] * arb->RowStride + x[i];
-         *dst = val;
-      }
-   }
-}
-
-
-static void
-copy_buffer_alpha8(struct gl_renderbuffer* dst, struct gl_renderbuffer* src)
-{
-   ASSERT(dst->Format == MESA_FORMAT_A8);
-   ASSERT(src->Format == MESA_FORMAT_A8);
-   ASSERT(dst->Width == src->Width);
-   ASSERT(dst->Height == src->Height);
-   ASSERT(dst->RowStride == src->RowStride);
-
-   memcpy(dst->Data, src->Data, dst->RowStride * dst->Height * sizeof(GLubyte));
-}
-
 
 /**********************************************************************/
 /**********************************************************************/
@@ -2148,114 +1896,6 @@ _mesa_add_color_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
 
 
 /**
- * Add software-based alpha renderbuffers to the given framebuffer.
- * This is a helper routine for device drivers when creating a
- * window system framebuffer (not a user-created render/framebuffer).
- * Once this function is called, you can basically forget about this
- * renderbuffer; core Mesa will handle all the buffer management and
- * rendering!
- */
-GLboolean
-_mesa_add_alpha_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
-                              GLuint alphaBits,
-                              GLboolean frontLeft, GLboolean backLeft,
-                              GLboolean frontRight, GLboolean backRight)
-{
-   gl_buffer_index b;
-
-   /* for window system framebuffers only! */
-   assert(fb->Name == 0);
-
-   if (alphaBits > 8) {
-      _mesa_problem(ctx,
-                    "Unsupported bit depth in _mesa_add_alpha_renderbuffers");
-      return GL_FALSE;
-   }
-
-   assert(MAX_COLOR_ATTACHMENTS >= 4);
-
-   /* Wrap each of the RGB color buffers with an alpha renderbuffer.
-    */
-   for (b = BUFFER_FRONT_LEFT; b <= BUFFER_BACK_RIGHT; b++) {
-      struct gl_renderbuffer *arb;
-
-      if (b == BUFFER_FRONT_LEFT && !frontLeft)
-         continue;
-      else if (b == BUFFER_BACK_LEFT && !backLeft)
-         continue;
-      else if (b == BUFFER_FRONT_RIGHT && !frontRight)
-         continue;
-      else if (b == BUFFER_BACK_RIGHT && !backRight)
-         continue;
-
-      /* the RGB buffer to wrap must already exist!! */
-      assert(fb->Attachment[b].Renderbuffer);
-
-      /* only GLubyte supported for now */
-      assert(fb->Attachment[b].Renderbuffer->DataType == GL_UNSIGNED_BYTE);
-
-      /* allocate alpha renderbuffer */
-      arb = _mesa_new_renderbuffer(ctx, 0);
-      if (!arb) {
-         _mesa_error(ctx, GL_OUT_OF_MEMORY, "Allocating alpha buffer");
-         return GL_FALSE;
-      }
-
-      /* wrap the alpha renderbuffer around the RGB renderbuffer */
-      arb->Wrapped = fb->Attachment[b].Renderbuffer;
-
-      /* Set up my alphabuffer fields and plug in my functions.
-       * The functions will put/get the alpha values from/to RGBA arrays
-       * and then call the wrapped buffer's functions to handle the RGB
-       * values.
-       */
-      arb->InternalFormat = arb->Wrapped->InternalFormat;
-      arb->Format         = MESA_FORMAT_A8;
-      arb->DataType       = arb->Wrapped->DataType;
-      arb->AllocStorage   = alloc_storage_alpha8;
-      arb->Delete         = delete_renderbuffer_alpha8;
-      arb->GetPointer     = get_pointer_alpha8;
-      arb->GetRow         = get_row_alpha8;
-      arb->GetValues      = get_values_alpha8;
-      arb->PutRow         = put_row_alpha8;
-      arb->PutRowRGB      = put_row_rgb_alpha8;
-      arb->PutMonoRow     = put_mono_row_alpha8;
-      arb->PutValues      = put_values_alpha8;
-      arb->PutMonoValues  = put_mono_values_alpha8;
-
-      /* clear the pointer to avoid assertion/sanity check failure later */
-      fb->Attachment[b].Renderbuffer = NULL;
-
-      /* plug the alpha renderbuffer into the colorbuffer attachment */
-      _mesa_add_renderbuffer(fb, b, arb);
-   }
-
-   return GL_TRUE;
-}
-
-
-/**
- * For framebuffers that use a software alpha channel wrapper
- * created by _mesa_add_alpha_renderbuffer or _mesa_add_soft_renderbuffers,
- * copy the back buffer alpha channel into the front buffer alpha channel.
- */
-void
-_mesa_copy_soft_alpha_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb)
-{
-   if (fb->Attachment[BUFFER_FRONT_LEFT].Renderbuffer &&
-       fb->Attachment[BUFFER_BACK_LEFT].Renderbuffer)
-      copy_buffer_alpha8(fb->Attachment[BUFFER_FRONT_LEFT].Renderbuffer,
-                         fb->Attachment[BUFFER_BACK_LEFT].Renderbuffer);
-
-
-   if (fb->Attachment[BUFFER_FRONT_RIGHT].Renderbuffer &&
-       fb->Attachment[BUFFER_BACK_RIGHT].Renderbuffer)
-      copy_buffer_alpha8(fb->Attachment[BUFFER_FRONT_RIGHT].Renderbuffer,
-                         fb->Attachment[BUFFER_BACK_RIGHT].Renderbuffer);
-}
-
-
-/**
  * Add a software-based depth renderbuffer to the given framebuffer.
  * This is a helper routine for device drivers when creating a
  * window system framebuffer (not a user-created render/framebuffer).
@@ -2476,13 +2116,6 @@ _mesa_add_soft_renderbuffers(struct gl_framebuffer *fb,
                                   fb->Visual.numAuxBuffers);
    }
 
-   if (alpha) {
-      assert(fb->Visual.alphaBits > 0);
-      _mesa_add_alpha_renderbuffers(NULL, fb, fb->Visual.alphaBits,
-                                    frontLeft, backLeft,
-                                    frontRight, backRight);
-   }
-
 #if 0
    if (multisample) {
       /* maybe someday */
diff --git a/mesalib/src/mesa/main/renderbuffer.h b/mesalib/src/mesa/main/renderbuffer.h
index cb0d712eb..3194fc3fe 100644
--- a/mesalib/src/mesa/main/renderbuffer.h
+++ b/mesalib/src/mesa/main/renderbuffer.h
@@ -73,15 +73,6 @@ _mesa_add_color_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
                               GLboolean frontRight, GLboolean backRight);
 
 extern GLboolean
-_mesa_add_alpha_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
-                              GLuint alphaBits,
-                              GLboolean frontLeft, GLboolean backLeft,
-                              GLboolean frontRight, GLboolean backRight);
-
-extern void
-_mesa_copy_soft_alpha_renderbuffers(struct gl_context *ctx, struct gl_framebuffer *fb);
-
-extern GLboolean
 _mesa_add_depth_renderbuffer(struct gl_context *ctx, struct gl_framebuffer *fb,
                              GLuint depthBits);
 
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index 5371d6a17..33ba53c2e 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -698,11 +698,27 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
 
 	 prog = shProg->_LinkedShaders[i]->Program;
 
+	 /* If the shader stage doesn't use any samplers, don't bother
+	  * checking if any samplers have changed.
+	  */
+	 if (prog->SamplersUsed == 0)
+	    continue;
+
 	 assert(sizeof(prog->SamplerUnits) == sizeof(shProg->SamplerUnits));
 
-	 if (memcmp(prog->SamplerUnits,
-		    shProg->SamplerUnits,
-		    sizeof(shProg->SamplerUnits)) != 0) {
+	 /* Determine if any of the samplers used by this shader stage have
+	  * been modified.
+	  */
+	 bool changed = false;
+	 for (unsigned j = 0; j < Elements(prog->SamplerUnits); j++) {
+	    if ((prog->SamplersUsed & (1U << j)) != 0
+		&& (prog->SamplerUnits[j] != shProg->SamplerUnits[j])) {
+	       changed = true;
+	       break;
+	    }
+	 }
+
+	 if (changed) {
 	    if (!flushed) {
 	       FLUSH_VERTICES(ctx, _NEW_TEXTURE | _NEW_PROGRAM);
 	       flushed = true;
diff --git a/mesalib/src/mesa/program/ir_to_mesa.cpp b/mesalib/src/mesa/program/ir_to_mesa.cpp
index 1b8b48e53..5cee83778 100644
--- a/mesalib/src/mesa/program/ir_to_mesa.cpp
+++ b/mesalib/src/mesa/program/ir_to_mesa.cpp
@@ -2541,58 +2541,6 @@ count_resources(struct gl_program *prog)
    _mesa_update_shader_textures_used(prog);
 }
 
-
-/**
- * Check if the given vertex/fragment/shader program is within the
- * resource limits of the context (number of texture units, etc).
- * If any of those checks fail, record a linker error.
- *
- * XXX more checks are needed...
- */
-static bool
-check_resources(const struct gl_context *ctx,
-                struct gl_shader_program *shader_program,
-                struct gl_program *prog)
-{
-   switch (prog->Target) {
-   case GL_VERTEX_PROGRAM_ARB:
-      if (_mesa_bitcount(prog->SamplersUsed) >
-          ctx->Const.MaxVertexTextureImageUnits) {
-         linker_error(shader_program,
-		      "Too many vertex shader texture samplers");
-      }
-      if (prog->Parameters->NumParameters > MAX_UNIFORMS) {
-         linker_error(shader_program, "Too many vertex shader constants");
-      }
-      break;
-   case MESA_GEOMETRY_PROGRAM:
-      if (_mesa_bitcount(prog->SamplersUsed) >
-          ctx->Const.MaxGeometryTextureImageUnits) {
-         linker_error(shader_program,
-		      "Too many geometry shader texture samplers");
-      }
-      if (prog->Parameters->NumParameters >
-          MAX_GEOMETRY_UNIFORM_COMPONENTS / 4) {
-         linker_error(shader_program, "Too many geometry shader constants");
-      }
-      break;
-   case GL_FRAGMENT_PROGRAM_ARB:
-      if (_mesa_bitcount(prog->SamplersUsed) >
-          ctx->Const.MaxTextureImageUnits) {
-         linker_error(shader_program,
-		      "Too many fragment shader texture samplers");
-      }
-      if (prog->Parameters->NumParameters > MAX_UNIFORMS) {
-         linker_error(shader_program, "Too many fragment shader constants");
-      }
-      break;
-   default:
-      _mesa_problem(ctx, "unexpected program type in check_resources()");
-   }
-
-   return shader_program->LinkStatus;
-}
-
 class add_uniform_to_shader : public uniform_field_visitor {
 public:
    add_uniform_to_shader(struct gl_shader_program *shader_program,
@@ -3274,9 +3222,6 @@ get_mesa_program(struct gl_context *ctx,
    do_set_program_inouts(shader->ir, prog, shader->Type == GL_FRAGMENT_SHADER);
    count_resources(prog);
 
-   if (!check_resources(ctx, shader_program, prog))
-      goto fail_exit;
-
    _mesa_reference_program(ctx, &shader->Program, prog);
 
    if ((ctx->Shader.Flags & GLSL_NO_OPT) == 0) {
@@ -3513,7 +3458,9 @@ _mesa_glsl_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
       }
    }
 
-   set_uniform_initializers(ctx, prog);
+   if (prog->LinkStatus) {
+      set_uniform_initializers(ctx, prog);
+   }
 
    if (ctx->Shader.Flags & GLSL_DUMP) {
       if (!prog->LinkStatus) {
diff --git a/mesalib/src/mesa/sources.mak b/mesalib/src/mesa/sources.mak
index 102562ace..e72a1ce1e 100644
--- a/mesalib/src/mesa/sources.mak
+++ b/mesalib/src/mesa/sources.mak
@@ -141,7 +141,6 @@ SWRAST_SOURCES = \
 	swrast/s_logic.c \
 	swrast/s_masking.c \
 	swrast/s_points.c \
-	swrast/s_readpix.c \
 	swrast/s_span.c \
 	swrast/s_stencil.c \
 	swrast/s_texcombine.c \
diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
index 5714d343d..95805fd3c 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -38,6 +38,7 @@
 #include "main/mtypes.h"
 #include "main/pack.h"
 #include "main/pbo.h"
+#include "main/readpix.h"
 #include "main/texformat.h"
 #include "main/teximage.h"
 #include "main/texstore.h"
@@ -1189,9 +1190,9 @@ copy_stencil_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,
       rbDraw = st_renderbuffer(rbDraw->Base.Wrapped);
 
    /* this will do stencil pixel transfer ops */
-   st_read_stencil_pixels(ctx, srcx, srcy, width, height,
-                          GL_STENCIL_INDEX, GL_UNSIGNED_BYTE,
-                          &ctx->DefaultPacking, buffer);
+   _mesa_readpixels(ctx, srcx, srcy, width, height,
+                    GL_STENCIL_INDEX, GL_UNSIGNED_BYTE,
+                    &ctx->DefaultPacking, buffer);
 
    if (0) {
       /* debug code: dump stencil values */
@@ -1296,6 +1297,20 @@ copy_stencil_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,
 }
 
 
+/**
+ * Return renderbuffer to use for reading color pixels for glCopyPixels
+ */
+static struct st_renderbuffer *
+st_get_color_read_renderbuffer(struct gl_context *ctx)
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct st_renderbuffer *strb =
+      st_renderbuffer(fb->_ColorReadBuffer);
+
+   return strb;
+}
+
+
 /** Do the src/dest regions overlap? */
 static GLboolean
 regions_overlap(GLint srcX, GLint srcY, GLint dstX, GLint dstY,
diff --git a/mesalib/src/mesa/state_tracker/st_cb_fbo.c b/mesalib/src/mesa/state_tracker/st_cb_fbo.c
index e04cb4406..a351d1197 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_fbo.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_fbo.c
@@ -355,8 +355,12 @@ st_render_texture(struct gl_context *ctx,
    /* get pointer to texture image we're rendeing to */
    texImage = _mesa_get_attachment_teximage(att);
 
-   /* create new renderbuffer which wraps the texture image */
-   rb = st_new_renderbuffer(ctx, 0);
+   /* create new renderbuffer which wraps the texture image.
+    * Use the texture's name as the renderbuffer's name so that we have
+    * something that's non-zero (to determine vertical orientation) and
+    * possibly helpful for debugging.
+    */
+   rb = st_new_renderbuffer(ctx, att->Texture->Name);
    if (!rb) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "glFramebufferTexture()");
       return;
@@ -644,7 +648,9 @@ st_MapRenderbuffer(struct gl_context *ctx,
    struct st_context *st = st_context(ctx);
    struct st_renderbuffer *strb = st_renderbuffer(rb);
    struct pipe_context *pipe = st->pipe;
+   const GLboolean invert = rb->Name == 0;
    unsigned usage;
+   GLuint y2;
 
    usage = 0x0;
    if (mode & GL_MAP_READ_BIT)
@@ -652,14 +658,30 @@ st_MapRenderbuffer(struct gl_context *ctx,
    if (mode & GL_MAP_WRITE_BIT)
       usage |= PIPE_TRANSFER_WRITE;
 
+   /* Note: y=0=bottom of buffer while y2=0=top of buffer.
+    * 'invert' will be true for window-system buffers and false for
+    * user-allocated renderbuffers and textures.
+    */
+   if (invert)
+      y2 = strb->Base.Height - y - h;
+   else
+      y2 = y;
+
    strb->transfer = pipe_get_transfer(pipe,
                                       strb->texture,
                                       strb->rtt_level,
                                       strb->rtt_face + strb->rtt_slice,
-                                      usage, x, y, w, h);
+                                      usage, x, y2, w, h);
    if (strb->transfer) {
-      *mapOut = pipe_transfer_map(pipe, strb->transfer);
-      *rowStrideOut = strb->transfer->stride;
+      GLubyte *map = pipe_transfer_map(pipe, strb->transfer);
+      if (invert) {
+         *rowStrideOut = -strb->transfer->stride;
+         map += (h - 1) * strb->transfer->stride;
+      }
+      else {
+         *rowStrideOut = strb->transfer->stride;
+      }
+      *mapOut = map;
    }
    else {
       *mapOut = NULL;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_readpixels.c b/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
index bd73f3bf0..6b824b161 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
@@ -26,649 +26,32 @@
  **************************************************************************/
 
 
-/**
- * glReadPixels interface to pipe
- *
- * \author Brian Paul
- */
-
-
 #include "main/imports.h"
-#include "main/bufferobj.h"
-#include "main/context.h"
-#include "main/image.h"
-#include "main/pack.h"
-#include "main/pbo.h"
-
-#include "pipe/p_context.h"
-#include "pipe/p_defines.h"
-#include "util/u_format.h"
-#include "util/u_inlines.h"
-#include "util/u_tile.h"
+#include "main/readpix.h"
 
-#include "st_debug.h"
-#include "st_context.h"
 #include "st_atom.h"
+#include "st_context.h"
 #include "st_cb_bitmap.h"
 #include "st_cb_readpixels.h"
-#include "st_cb_fbo.h"
-
-/**
- * Special case for reading stencil buffer.
- * For color/depth we use get_tile().  For stencil, map the stencil buffer.
- */
-void
-st_read_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
-                       GLsizei width, GLsizei height,
-                       GLenum format, GLenum type,
-                       const struct gl_pixelstore_attrib *packing,
-                       GLvoid *pixels)
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct pipe_context *pipe = st_context(ctx)->pipe;
-   struct st_renderbuffer *strb = st_renderbuffer(fb->_StencilBuffer);
-   struct pipe_transfer *pt;
-   ubyte *stmap;
-   GLint j;
-
-   if (strb->Base.Wrapped) {
-      strb = st_renderbuffer(strb->Base.Wrapped);
-   }
-
-   if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) {
-      y = ctx->DrawBuffer->Height - y - height;
-   }
-
-   /* Create a read transfer from the renderbuffer's texture */
-
-   pt = pipe_get_transfer(pipe, strb->texture,
-                          strb->rtt_level,
-                          strb->rtt_face + strb->rtt_slice,
-                          PIPE_TRANSFER_READ,
-                          x, y, width, height);
-
-   /* map the stencil buffer */
-   stmap = pipe_transfer_map(pipe, pt);
-
-   /* width should never be > MAX_WIDTH since we did clipping earlier */
-   ASSERT(width <= MAX_WIDTH);
-
-   /* process image row by row */
-   for (j = 0; j < height; j++) {
-      GLvoid *dest;
-      GLstencil sValues[MAX_WIDTH];
-      GLfloat zValues[MAX_WIDTH];
-      GLint srcY;
-
-      if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) {
-         srcY = height - j - 1;
-      }
-      else {
-         srcY = j;
-      }
-
-      /* get stencil (and Z) values */
-      switch (pt->resource->format) {
-      case PIPE_FORMAT_S8_UINT:
-         {
-            const ubyte *src = stmap + srcY * pt->stride;
-            memcpy(sValues, src, width);
-         }
-         break;
-      case PIPE_FORMAT_Z24_UNORM_S8_UINT:
-         if (format == GL_DEPTH_STENCIL) {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            const GLfloat scale = 1.0f / (0xffffff);
-            GLint k;
-            for (k = 0; k < width; k++) {
-               sValues[k] = src[k] >> 24;
-               zValues[k] = (src[k] & 0xffffff) * scale;
-            }
-         }
-         else {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            GLint k;
-            for (k = 0; k < width; k++) {
-               sValues[k] = src[k] >> 24;
-            }
-         }
-         break;
-      case PIPE_FORMAT_S8_UINT_Z24_UNORM:
-         if (format == GL_DEPTH_STENCIL) {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            const GLfloat scale = 1.0f / (0xffffff);
-            GLint k;
-            for (k = 0; k < width; k++) {
-               sValues[k] = src[k] & 0xff;
-               zValues[k] = (src[k] >> 8) * scale;
-            }
-         }
-         else {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            GLint k;
-            for (k = 0; k < width; k++) {
-               sValues[k] = src[k] & 0xff;
-            }
-         }
-         break;
-      case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
-         if (format == GL_DEPTH_STENCIL) {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            const GLfloat *srcf = (const GLfloat*)src;
-            GLint k;
-            for (k = 0; k < width; k++) {
-               zValues[k] = srcf[k*2];
-               sValues[k] = src[k*2+1] & 0xff;
-            }
-         }
-         else {
-            const uint *src = (uint *) (stmap + srcY * pt->stride);
-            GLint k;
-            for (k = 0; k < width; k++) {
-               sValues[k] = src[k*2+1] & 0xff;
-            }
-         }
-         break;
-      default:
-         assert(0);
-      }
-
-      /* store */
-      dest = _mesa_image_address2d(packing, pixels, width, height,
-                                   format, type, j, 0);
-      if (format == GL_DEPTH_STENCIL) {
-         _mesa_pack_depth_stencil_span(ctx, width, type, dest,
-                                       zValues, sValues, packing);
-      }
-      else {
-         _mesa_pack_stencil_span(ctx, width, type, dest, sValues, packing);
-      }
-   }
-
-   /* unmap the stencil buffer */
-   pipe_transfer_unmap(pipe, pt);
-   pipe->transfer_destroy(pipe, pt);
-}
-
-
-/**
- * Return renderbuffer to use for reading color pixels for glRead/CopyPixel
- * commands.
- */
-struct st_renderbuffer *
-st_get_color_read_renderbuffer(struct gl_context *ctx)
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct st_renderbuffer *strb =
-      st_renderbuffer(fb->_ColorReadBuffer);
-
-   return strb;
-}
-
-
-/**
- * Try to do glReadPixels in a fast manner for common cases.
- * \return GL_TRUE for success, GL_FALSE for failure
- */
-static GLboolean
-st_fast_readpixels(struct gl_context *ctx, struct st_renderbuffer *strb,
-                   GLint x, GLint y, GLsizei width, GLsizei height,
-                   GLenum format, GLenum type,
-                   const struct gl_pixelstore_attrib *pack,
-                   GLvoid *dest)
-{
-   GLubyte alphaORoperand;
-   enum combination {
-      A8R8G8B8_UNORM_TO_RGBA_UBYTE,
-      A8R8G8B8_UNORM_TO_RGB_UBYTE,
-      A8R8G8B8_UNORM_TO_BGRA_UINT,
-      A8R8G8B8_UNORM_TO_RGBA_UINT
-   } combo;
-
-   if (ctx->_ImageTransferState)
-      return GL_FALSE;
-
-   if (strb->format == PIPE_FORMAT_B8G8R8A8_UNORM) {
-      alphaORoperand = 0;
-   }
-   else if (strb->format == PIPE_FORMAT_B8G8R8X8_UNORM ) {
-      alphaORoperand = 0xff;
-   }
-   else {
-      return GL_FALSE;
-   }
-
-   if (format == GL_RGBA && type == GL_UNSIGNED_BYTE) {
-      combo = A8R8G8B8_UNORM_TO_RGBA_UBYTE;
-   }
-   else if (format == GL_RGB && type == GL_UNSIGNED_BYTE) {
-      combo = A8R8G8B8_UNORM_TO_RGB_UBYTE;
-   }
-   else if (format == GL_BGRA && type == GL_UNSIGNED_INT_8_8_8_8_REV) {
-      combo = A8R8G8B8_UNORM_TO_BGRA_UINT;
-   }
-   else if (format == GL_RGBA && type == GL_UNSIGNED_INT_8_8_8_8) {
-      combo = A8R8G8B8_UNORM_TO_RGBA_UINT;
-   }
-   else {
-      return GL_FALSE;
-   }
-
-   /*printf("st_fast_readpixels combo %d\n", (GLint) combo);*/
-
-   {
-      struct pipe_context *pipe = st_context(ctx)->pipe;
-      struct pipe_transfer *trans;
-      const GLubyte *map;
-      GLubyte *dst;
-      GLint row, col, dy, dstStride;
-
-      if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
-         /* convert GL Y to Gallium Y */
-         y = strb->texture->height0 - y - height;
-      }
-
-      trans = pipe_get_transfer(pipe, strb->texture,
-                                strb->rtt_level,
-                                strb->rtt_face + strb->rtt_slice,
-                                PIPE_TRANSFER_READ,
-                                x, y, width, height);
-      if (!trans) {
-         return GL_FALSE;
-      }
-
-      map = pipe_transfer_map(pipe, trans);
-      if (!map) {
-         pipe->transfer_destroy(pipe, trans);
-         return GL_FALSE;
-      }
-
-      /* We always write to the user/dest buffer from low addr to high addr
-       * but the read order depends on renderbuffer orientation
-       */
-      if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
-         /* read source rows from bottom to top */
-         y = height - 1;
-         dy = -1;
-      }
-      else {
-         /* read source rows from top to bottom */
-         y = 0;
-         dy = 1;
-      }
-
-      dst = _mesa_image_address2d(pack, dest, width, height,
-                                  format, type, 0, 0);
-      dstStride = _mesa_image_row_stride(pack, width, format, type);
-
-      switch (combo) {
-      case A8R8G8B8_UNORM_TO_RGBA_UBYTE:
-         for (row = 0; row < height; row++) {
-            const GLubyte *src = map + y * trans->stride;
-            for (col = 0; col < width; col++) {
-               GLuint pixel = ((GLuint *) src)[col];
-               dst[col*4+0] = (pixel >> 16) & 0xff;
-               dst[col*4+1] = (pixel >>  8) & 0xff;
-               dst[col*4+2] = (pixel >>  0) & 0xff;
-               dst[col*4+3] = ((pixel >> 24) & 0xff) | alphaORoperand;
-            }
-            dst += dstStride;
-            y += dy;
-         }
-         break;
-      case A8R8G8B8_UNORM_TO_RGB_UBYTE:
-         for (row = 0; row < height; row++) {
-            const GLubyte *src = map + y * trans->stride;
-            for (col = 0; col < width; col++) {
-               GLuint pixel = ((GLuint *) src)[col];
-               dst[col*3+0] = (pixel >> 16) & 0xff;
-               dst[col*3+1] = (pixel >>  8) & 0xff;
-               dst[col*3+2] = (pixel >>  0) & 0xff;
-            }
-            dst += dstStride;
-            y += dy;
-         }
-         break;
-      case A8R8G8B8_UNORM_TO_BGRA_UINT:
-         for (row = 0; row < height; row++) {
-            const GLubyte *src = map + y * trans->stride;
-            memcpy(dst, src, 4 * width);
-            if (alphaORoperand) {
-               assert(alphaORoperand == 0xff);
-               for (col = 0; col < width; col++) {
-                  dst[col*4+3] = 0xff;
-               }
-            }
-            dst += dstStride;
-            y += dy;
-         }
-         break;
-      case A8R8G8B8_UNORM_TO_RGBA_UINT:
-         for (row = 0; row < height; row++) {
-            const GLubyte *src = map + y * trans->stride;
-            for (col = 0; col < width; col++) {
-               GLuint pixel = ((GLuint *) src)[col];
-               dst[col*4+0] = ((pixel >> 24) & 0xff) | alphaORoperand;
-               dst[col*4+1] = (pixel >> 0) & 0xff;
-               dst[col*4+2] = (pixel >> 8) & 0xff;
-               dst[col*4+3] = (pixel >> 16) & 0xff;
-            }
-            dst += dstStride;
-            y += dy;
-         }
-         break;
-      default:
-         ; /* nothing */
-      }
-
-      pipe_transfer_unmap(pipe, trans);
-      pipe->transfer_destroy(pipe, trans);
-   }
-
-   return GL_TRUE;
-}
 
 
 /**
- * Do glReadPixels by getting rows from the framebuffer transfer with
- * get_tile().  Convert to requested format/type with Mesa image routines.
- * Image transfer ops are done in software too.
+ * The only special thing we need to do for the state tracker's
+ * glReadPixels is to validate state (to be sure we have up-to-date
+ * framebuffer surfaces) and flush the bitmap cache prior to reading.
  */
 static void
-st_readpixels(struct gl_context *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
+st_readpixels(struct gl_context *ctx, GLint x, GLint y,
+              GLsizei width, GLsizei height,
               GLenum format, GLenum type,
               const struct gl_pixelstore_attrib *pack,
               GLvoid *dest)
 {
    struct st_context *st = st_context(ctx);
-   struct pipe_context *pipe = st->pipe;
-   GLfloat (*temp)[4];
-   GLbitfield transferOps = ctx->_ImageTransferState;
-   GLsizei i, j;
-   GLint yStep, dfStride;
-   GLfloat *df;
-   GLuint *dui;
-   GLint *di;
-   struct st_renderbuffer *strb;
-   struct gl_pixelstore_attrib clippedPacking = *pack;
-   struct pipe_transfer *trans;
-   enum pipe_format pformat;
-
-   assert(ctx->ReadBuffer->Width > 0);
 
    st_validate_state(st);
-
-   /* Do all needed clipping here, so that we can forget about it later */
-   if (!_mesa_clip_readpixels(ctx, &x, &y, &width, &height, &clippedPacking)) {
-      /* The ReadPixels transfer is totally outside the window bounds */
-      return;
-   }
-
    st_flush_bitmap_cache(st);
-
-   dest = _mesa_map_pbo_dest(ctx, &clippedPacking, dest);
-   if (!dest)
-      return;
-
-   if (format == GL_STENCIL_INDEX ||
-       format == GL_DEPTH_STENCIL) {
-      st_read_stencil_pixels(ctx, x, y, width, height,
-                             format, type, pack, dest);
-      return;
-   }
-   else if (format == GL_DEPTH_COMPONENT) {
-      strb = st_renderbuffer(ctx->ReadBuffer->_DepthBuffer);
-      if (strb->Base.Wrapped) {
-         strb = st_renderbuffer(strb->Base.Wrapped);
-      }
-   }
-   else {
-      /* Read color buffer */
-      strb = st_get_color_read_renderbuffer(ctx);
-   }
-
-   if (!strb)
-      return;
-
-   /* try a fast-path readpixels before anything else */
-   if (st_fast_readpixels(ctx, strb, x, y, width, height,
-                          format, type, pack, dest)) {
-      /* success! */
-      _mesa_unmap_pbo_dest(ctx, &clippedPacking);
-      return;
-   }
-
-   /* allocate temp pixel row buffer */
-   temp = (GLfloat (*)[4]) malloc(4 * width * sizeof(GLfloat));
-   if (!temp) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glReadPixels");
-      return;
-   }
-
-   if(ctx->Color._ClampReadColor)
-      transferOps |= IMAGE_CLAMP_BIT;
-
-   if (format == GL_RGBA && type == GL_FLOAT && !transferOps) {
-      /* write tile(row) directly into user's buffer */
-      df = (GLfloat *) _mesa_image_address2d(&clippedPacking, dest, width,
-                                             height, format, type, 0, 0);
-      dfStride = width * 4;
-   }
-   else {
-      /* write tile(row) into temp row buffer */
-      df = (GLfloat *)temp;
-      dfStride = 0;
-   }
-
-   dui = (GLuint *)df;
-   di = (GLint *)df;
-
-   if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
-      /* convert GL Y to Gallium Y */
-      y = strb->Base.Height - y - height;
-   }
-
-   /* Create a read transfer from the renderbuffer's texture */
-   trans = pipe_get_transfer(pipe, strb->texture,
-                             strb->rtt_level, /* level */
-                             strb->rtt_face + strb->rtt_slice, /* layer */
-                             PIPE_TRANSFER_READ,
-                             x, y, width, height);
-
-   /* determine bottom-to-top vs. top-to-bottom order */
-   if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
-      y = height - 1;
-      yStep = -1;
-   }
-   else {
-      y = 0;
-      yStep = 1;
-   }
-
-   /* possibly convert sRGB format to linear RGB format */
-   pformat = util_format_linear(trans->resource->format);
-
-   if (ST_DEBUG & DEBUG_FALLBACK)
-      debug_printf("%s: fallback processing\n", __FUNCTION__);
-
-   /*
-    * Copy pixels from pipe_transfer to user memory
-    */
-   {
-      /* dest of first pixel in client memory */
-      GLubyte *dst = _mesa_image_address2d(&clippedPacking, dest, width,
-                                           height, format, type, 0, 0);
-      /* dest row stride */
-      const GLint dstStride = _mesa_image_row_stride(&clippedPacking, width,
-                                                     format, type);
-
-      if (pformat == PIPE_FORMAT_Z24_UNORM_S8_UINT ||
-          pformat == PIPE_FORMAT_Z24X8_UNORM) {
-         if (format == GL_DEPTH_COMPONENT) {
-            for (i = 0; i < height; i++) {
-               GLuint ztemp[MAX_WIDTH];
-               GLfloat zfloat[MAX_WIDTH];
-               const double scale = 1.0 / ((1 << 24) - 1);
-               pipe_get_tile_raw(pipe, trans, 0, y, width, 1, ztemp, 0);
-               y += yStep;
-               for (j = 0; j < width; j++) {
-                  zfloat[j] = (float) (scale * (ztemp[j] & 0xffffff));
-               }
-               _mesa_pack_depth_span(ctx, width, dst, type,
-                                     zfloat, &clippedPacking);
-               dst += dstStride;
-            }
-         }
-         else {
-            /* XXX: unreachable code -- should be before st_read_stencil_pixels */
-            assert(format == GL_DEPTH_STENCIL_EXT);
-            for (i = 0; i < height; i++) {
-               GLuint *zshort = (GLuint *)dst;
-               pipe_get_tile_raw(pipe, trans, 0, y, width, 1, dst, 0);
-               y += yStep;
-               /* Reverse into 24/8 */
-               for (j = 0; j < width; j++) {
-                  zshort[j] = (zshort[j] << 8) | (zshort[j] >> 24);
-               }
-               dst += dstStride;
-            }
-         }
-      }
-      else if (pformat == PIPE_FORMAT_S8_UINT_Z24_UNORM ||
-               pformat == PIPE_FORMAT_X8Z24_UNORM) {
-         if (format == GL_DEPTH_COMPONENT) {
-            for (i = 0; i < height; i++) {
-               GLuint ztemp[MAX_WIDTH];
-               GLfloat zfloat[MAX_WIDTH];
-               const double scale = 1.0 / ((1 << 24) - 1);
-               pipe_get_tile_raw(pipe, trans, 0, y, width, 1, ztemp, 0);
-               y += yStep;
-               for (j = 0; j < width; j++) {
-                  zfloat[j] = (float) (scale * ((ztemp[j] >> 8) & 0xffffff));
-               }
-               _mesa_pack_depth_span(ctx, width, dst, type,
-                                     zfloat, &clippedPacking);
-               dst += dstStride;
-            }
-         }
-         else {
-            /* XXX: unreachable code -- should be before st_read_stencil_pixels */
-            assert(format == GL_DEPTH_STENCIL_EXT);
-            for (i = 0; i < height; i++) {
-               pipe_get_tile_raw(pipe, trans, 0, y, width, 1, dst, 0);
-               y += yStep;
-               dst += dstStride;
-            }
-         }
-      }
-      else if (pformat == PIPE_FORMAT_Z16_UNORM) {
-         for (i = 0; i < height; i++) {
-            GLushort ztemp[MAX_WIDTH];
-            GLfloat zfloat[MAX_WIDTH];
-            const double scale = 1.0 / 0xffff;
-            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, ztemp, 0);
-            y += yStep;
-            for (j = 0; j < width; j++) {
-               zfloat[j] = (float) (scale * ztemp[j]);
-            }
-            _mesa_pack_depth_span(ctx, width, dst, type,
-                                  zfloat, &clippedPacking);
-            dst += dstStride;
-         }
-      }
-      else if (pformat == PIPE_FORMAT_Z32_UNORM) {
-         for (i = 0; i < height; i++) {
-            GLuint ztemp[MAX_WIDTH];
-            GLfloat zfloat[MAX_WIDTH];
-            const double scale = 1.0 / 0xffffffff;
-            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, ztemp, 0);
-            y += yStep;
-            for (j = 0; j < width; j++) {
-               zfloat[j] = (float) (scale * ztemp[j]);
-            }
-            _mesa_pack_depth_span(ctx, width, dst, type,
-                                  zfloat, &clippedPacking);
-            dst += dstStride;
-         }
-      }
-      else if (pformat == PIPE_FORMAT_Z32_FLOAT) {
-         for (i = 0; i < height; i++) {
-            GLfloat zfloat[MAX_WIDTH];
-            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, zfloat, 0);
-            y += yStep;
-            _mesa_pack_depth_span(ctx, width, dst, type,
-                                  zfloat, &clippedPacking);
-            dst += dstStride;
-         }
-      }
-      else if (pformat == PIPE_FORMAT_Z32_FLOAT_S8X24_UINT) {
-         assert(format == GL_DEPTH_COMPONENT);
-         for (i = 0; i < height; i++) {
-            GLfloat zfloat[MAX_WIDTH];    /* Z32 */
-            GLfloat zfloat2[MAX_WIDTH*2]; /* Z32X32 */
-            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, zfloat2, 0);
-            y += yStep;
-            for (j = 0; j < width; j++) {
-               zfloat[j] = zfloat2[j*2];
-            }
-            _mesa_pack_depth_span(ctx, width, dst, type,
-                                  zfloat, &clippedPacking);
-            dst += dstStride;
-         }
-      }
-      else if (util_format_is_pure_sint(pformat)) {
-         for (i = 0; i < height; i++) {
-            if (type == GL_UNSIGNED_INT)
-               pipe_get_tile_ui_format(pipe, trans, 0, y, width, 1,
-                                       pformat, dui);
-            else
-               pipe_get_tile_i_format(pipe, trans, 0, y, width, 1,
-                                      pformat, di);
-            y += yStep;
-            if (!dfStride) {
-               _mesa_pack_rgba_span_int(ctx, width, (GLuint (*)[4])temp,
-                                        format, type, dst);
-               dst += dstStride;
-            }
-        }
-      } else if (util_format_is_pure_uint(pformat)) {
-         for (i = 0; i < height; i++) {
-            if (type == GL_UNSIGNED_INT)
-               pipe_get_tile_ui_format(pipe, trans, 0, y, width, 1,
-                                       pformat, dui);
-            else
-               pipe_get_tile_i_format(pipe, trans, 0, y, width, 1,
-                                      pformat, di);
-            y += yStep;
-            df += dfStride;
-            if (!dfStride) {
-               _mesa_pack_rgba_span_int(ctx, width, (GLuint (*)[4])temp,
-                                        format, type, dst);
-               dst += dstStride;
-            }
-         }
-      } else {
-         /* RGBA format */
-         /* Do a row at a time to flip image data vertically */
-         for (i = 0; i < height; i++) {
-            pipe_get_tile_rgba_format(pipe, trans, 0, y, width, 1,
-                                      pformat, df);
-            y += yStep;
-            df += dfStride;
-            if (!dfStride) {
-               _mesa_pack_rgba_span_float(ctx, width, temp, format, type, dst,
-                                          &clippedPacking, transferOps);
-               dst += dstStride;
-            }
-         }
-      }
-   }
-
-   free(temp);
-
-   pipe->transfer_destroy(pipe, trans);
-
-   _mesa_unmap_pbo_dest(ctx, &clippedPacking);
+   _mesa_readpixels(ctx, x, y, width, height, format, type, pack, dest);
 }
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_readpixels.h b/mesalib/src/mesa/state_tracker/st_cb_readpixels.h
index fba1a6540..fdc854782 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_readpixels.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_readpixels.h
@@ -1,52 +1,40 @@
-/**************************************************************************

- * 

- * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.

- * 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"), to deal in the Software without restriction, including

- * without limitation the rights to use, copy, modify, merge, publish,

- * distribute, sub license, and/or sell copies of the Software, and to

- * permit persons to whom the Software is furnished to do so, subject to

- * the following conditions:

- * 

- * The above copyright notice and this permission notice (including the

- * next paragraph) shall be included in all copies or substantial portions

- * of the Software.

- * 

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF

- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.

- * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR

- * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,

- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE

- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- * 

- **************************************************************************/

-

-

-#ifndef ST_CB_READPIXELS_H

-#define ST_CB_READPIXELS_H

-

-#include "main/glheader.h"

-

-struct dd_function_table;

-struct gl_context;

-struct gl_pixelstore_attrib;

-

-extern struct st_renderbuffer *

-st_get_color_read_renderbuffer(struct gl_context *ctx);

-

-extern void

-st_read_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,

-                       GLsizei width, GLsizei height,

-                       GLenum format, GLenum type,

-                       const struct gl_pixelstore_attrib *packing,

-                       GLvoid *pixels);

-

-extern void

-st_init_readpixels_functions(struct dd_function_table *functions);

-

-

-#endif /* ST_CB_READPIXELS_H */

+/**************************************************************************
+ * 
+ * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * 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"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+
+#ifndef ST_CB_READPIXELS_H
+#define ST_CB_READPIXELS_H
+
+#include "main/glheader.h"
+
+struct dd_function_table;
+
+extern void
+st_init_readpixels_functions(struct dd_function_table *functions);
+
+
+#endif /* ST_CB_READPIXELS_H */
diff --git a/mesalib/src/mesa/state_tracker/st_draw.c b/mesalib/src/mesa/state_tracker/st_draw.c
index f2fbf4836..cb518e1af 100644
--- a/mesalib/src/mesa/state_tracker/st_draw.c
+++ b/mesalib/src/mesa/state_tracker/st_draw.c
@@ -436,8 +436,8 @@ setup_interleaved_attribs(struct gl_context *ctx,
       /* all interleaved arrays in a VBO */
       struct st_buffer_object *stobj = st_buffer_object(bufobj);
 
-      if (!stobj) {
-         /* probably out of memory */
+      if (!stobj || !stobj->buffer) {
+         /* probably out of memory (or zero-sized buffer) */
          return GL_FALSE;
       }
 
@@ -505,7 +505,7 @@ setup_non_interleaved_attribs(struct gl_context *ctx,
          struct st_buffer_object *stobj = st_buffer_object(bufobj);
 
          if (!stobj || !stobj->buffer) {
-            /* probably ran out of memory */
+            /* probably out of memory (or zero-sized buffer) */
             return GL_FALSE;
          }
 
diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index b133164c3..0bf6766f7 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -2873,55 +2873,6 @@ count_resources(glsl_to_tgsi_visitor *v, gl_program *prog)
    _mesa_update_shader_textures_used(prog);
 }
 
-
-/**
- * Check if the given vertex/fragment/shader program is within the
- * resource limits of the context (number of texture units, etc).
- * If any of those checks fail, record a linker error.
- *
- * XXX more checks are needed...
- */
-static void
-check_resources(const struct gl_context *ctx,
-                struct gl_shader_program *shader_program,
-                glsl_to_tgsi_visitor *prog,
-                struct gl_program *proginfo)
-{
-   switch (proginfo->Target) {
-   case GL_VERTEX_PROGRAM_ARB:
-      if (_mesa_bitcount(prog->samplers_used) >
-          ctx->Const.MaxVertexTextureImageUnits) {
-         fail_link(shader_program, "Too many vertex shader texture samplers");
-      }
-      if (proginfo->Parameters->NumParameters > MAX_UNIFORMS) {
-         fail_link(shader_program, "Too many vertex shader constants");
-      }
-      break;
-   case MESA_GEOMETRY_PROGRAM:
-      if (_mesa_bitcount(prog->samplers_used) >
-          ctx->Const.MaxGeometryTextureImageUnits) {
-         fail_link(shader_program, "Too many geometry shader texture samplers");
-      }
-      if (proginfo->Parameters->NumParameters >
-          MAX_GEOMETRY_UNIFORM_COMPONENTS / 4) {
-         fail_link(shader_program, "Too many geometry shader constants");
-      }
-      break;
-   case GL_FRAGMENT_PROGRAM_ARB:
-      if (_mesa_bitcount(prog->samplers_used) >
-          ctx->Const.MaxTextureImageUnits) {
-         fail_link(shader_program, "Too many fragment shader texture samplers");
-      }
-      if (proginfo->Parameters->NumParameters > MAX_UNIFORMS) {
-         fail_link(shader_program, "Too many fragment shader constants");
-      }
-      break;
-   default:
-      _mesa_problem(ctx, "unexpected program type in check_resources()");
-   }
-}
-
-
 static void
 set_uniform_initializer(struct gl_context *ctx, void *mem_ctx,
         		struct gl_shader_program *shader_program,
@@ -4015,6 +3966,7 @@ struct st_translate {
 /** Map Mesa's SYSTEM_VALUE_x to TGSI_SEMANTIC_x */
 static unsigned mesa_sysval_to_semantic[SYSTEM_VALUE_MAX] = {
    TGSI_SEMANTIC_FACE,
+   TGSI_SEMANTIC_VERTEXID,
    TGSI_SEMANTIC_INSTANCEID
 };
 
@@ -4991,6 +4943,7 @@ get_mesa_program(struct gl_context *ctx,
       _mesa_print_ir(shader->ir, NULL);
       printf("\n");
       printf("\n");
+      fflush(stdout);
    }
 
    prog->Instructions = NULL;
@@ -4999,8 +4952,6 @@ get_mesa_program(struct gl_context *ctx,
    do_set_program_inouts(shader->ir, prog, shader->Type == GL_FRAGMENT_SHADER);
    count_resources(v, prog);
 
-   check_resources(ctx, shader_program, v, prog);
-
    _mesa_reference_program(ctx, &shader->Program, prog);
    
    /* This has to be done last.  Any operation the can cause
diff --git a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
index 27ebb1ad7..ac615343a 100644
--- a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
+++ b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
@@ -108,6 +108,7 @@ struct st_translate {
 /** Map Mesa's SYSTEM_VALUE_x to TGSI_SEMANTIC_x */
 static unsigned mesa_sysval_to_semantic[SYSTEM_VALUE_MAX] = {
    TGSI_SEMANTIC_FACE,
+   TGSI_SEMANTIC_VERTEXID,
    TGSI_SEMANTIC_INSTANCEID
 };
 
diff --git a/mesalib/src/mesa/swrast/s_copypix.c b/mesalib/src/mesa/swrast/s_copypix.c
index e249e4ad5..3ba31f22c 100644
--- a/mesalib/src/mesa/swrast/s_copypix.c
+++ b/mesalib/src/mesa/swrast/s_copypix.c
@@ -1,720 +1,720 @@
-/*

- * Mesa 3-D graphics library

- * Version:  7.1

- *

- * 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"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- */

-

-

-#include "main/glheader.h"

-#include "main/context.h"

-#include "main/colormac.h"

-#include "main/condrender.h"

-#include "main/macros.h"

-#include "main/pixeltransfer.h"

-#include "main/imports.h"

-

-#include "s_context.h"

-#include "s_depth.h"

-#include "s_span.h"

-#include "s_stencil.h"

-#include "s_zoom.h"

-

-

-

-/**

- * Determine if there's overlap in an image copy.

- * This test also compensates for the fact that copies are done from

- * bottom to top and overlaps can sometimes be handled correctly

- * without making a temporary image copy.

- * \return GL_TRUE if the regions overlap, GL_FALSE otherwise.

- */

-static GLboolean

-regions_overlap(GLint srcx, GLint srcy,

-                GLint dstx, GLint dsty,

-                GLint width, GLint height,

-                GLfloat zoomX, GLfloat zoomY)

-{

-   if (zoomX == 1.0 && zoomY == 1.0) {

-      /* no zoom */

-      if (srcx >= dstx + width || (srcx + width <= dstx)) {

-         return GL_FALSE;

-      }

-      else if (srcy < dsty) { /* this is OK */

-         return GL_FALSE;

-      }

-      else if (srcy > dsty + height) {

-         return GL_FALSE;

-      }

-      else {

-         return GL_TRUE;

-      }

-   }

-   else {

-      /* add one pixel of slop when zooming, just to be safe */

-      if (srcx > (dstx + ((zoomX > 0.0F) ? (width * zoomX + 1.0F) : 0.0F))) {

-         /* src is completely right of dest */

-         return GL_FALSE;

-      }

-      else if (srcx + width + 1.0F < dstx + ((zoomX > 0.0F) ? 0.0F : (width * zoomX))) {

-         /* src is completely left of dest */

-         return GL_FALSE;

-      }

-      else if ((srcy < dsty) && (srcy + height < dsty + (height * zoomY))) {

-         /* src is completely below dest */

-         return GL_FALSE;

-      }

-      else if ((srcy > dsty) && (srcy + height > dsty + (height * zoomY))) {

-         /* src is completely above dest */

-         return GL_FALSE;

-      }

-      else {

-         return GL_TRUE;

-      }

-   }

-}

-

-

-/**

- * RGBA copypixels

- */

-static void

-copy_rgba_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,

-                 GLint width, GLint height, GLint destx, GLint desty)

-{

-   GLfloat *tmpImage, *p;

-   GLint sy, dy, stepy, row;

-   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;

-   GLint overlapping;

-   GLuint transferOps = ctx->_ImageTransferState;

-   SWspan span;

-

-   if (!ctx->ReadBuffer->_ColorReadBuffer) {

-      /* no readbuffer - OK */

-      return;

-   }

-

-   if (ctx->DrawBuffer == ctx->ReadBuffer) {

-      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,

-                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);

-   }

-   else {

-      overlapping = GL_FALSE;

-   }

-

-   /* Determine if copy should be done bottom-to-top or top-to-bottom */

-   if (!overlapping && srcy < desty) {

-      /* top-down  max-to-min */

-      sy = srcy + height - 1;

-      dy = desty + height - 1;

-      stepy = -1;

-   }

-   else {

-      /* bottom-up  min-to-max */

-      sy = srcy;

-      dy = desty;

-      stepy = 1;

-   }

-

-   INIT_SPAN(span, GL_BITMAP);

-   _swrast_span_default_attribs(ctx, &span);

-   span.arrayMask = SPAN_RGBA;

-   span.arrayAttribs = FRAG_BIT_COL0; /* we'll fill in COL0 attrib values */

-

-   if (overlapping) {

-      tmpImage = (GLfloat *) malloc(width * height * sizeof(GLfloat) * 4);

-      if (!tmpImage) {

-         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );

-         return;

-      }

-      /* read the source image as RGBA/float */

-      p = tmpImage;

-      for (row = 0; row < height; row++) {

-         _swrast_read_rgba_span( ctx, ctx->ReadBuffer->_ColorReadBuffer,

-                                 width, srcx, sy + row, GL_FLOAT, p );

-         p += width * 4;

-      }

-      p = tmpImage;

-   }

-   else {

-      tmpImage = NULL;  /* silence compiler warnings */

-      p = NULL;

-   }

-

-   ASSERT(width < MAX_WIDTH);

-

-   for (row = 0; row < height; row++, sy += stepy, dy += stepy) {

-      GLvoid *rgba = span.array->attribs[FRAG_ATTRIB_COL0];

-

-      /* Get row/span of source pixels */

-      if (overlapping) {

-         /* get from buffered image */

-         memcpy(rgba, p, width * sizeof(GLfloat) * 4);

-         p += width * 4;

-      }

-      else {

-         /* get from framebuffer */

-         _swrast_read_rgba_span( ctx, ctx->ReadBuffer->_ColorReadBuffer,

-                                 width, srcx, sy, GL_FLOAT, rgba );

-      }

-

-      if (transferOps) {

-         _mesa_apply_rgba_transfer_ops(ctx, transferOps, width,

-                                       (GLfloat (*)[4]) rgba);

-      }

-

-      /* Write color span */

-      span.x = destx;

-      span.y = dy;

-      span.end = width;

-      span.array->ChanType = GL_FLOAT;

-      if (zoom) {

-         _swrast_write_zoomed_rgba_span(ctx, destx, desty, &span, rgba);

-      }

-      else {

-         _swrast_write_rgba_span(ctx, &span);

-      }

-   }

-

-   span.array->ChanType = CHAN_TYPE; /* restore */

-

-   if (overlapping)

-      free(tmpImage);

-}

-

-

-/**

- * Convert floating point Z values to integer Z values with pixel transfer's

- * Z scale and bias.

- */

-static void

-scale_and_bias_z(struct gl_context *ctx, GLuint width,

-                 const GLfloat depth[], GLuint z[])

-{

-   const GLuint depthMax = ctx->DrawBuffer->_DepthMax;

-   GLuint i;

-

-   if (depthMax <= 0xffffff &&

-       ctx->Pixel.DepthScale == 1.0 &&

-       ctx->Pixel.DepthBias == 0.0) {

-      /* no scale or bias and no clamping and no worry of overflow */

-      const GLfloat depthMaxF = ctx->DrawBuffer->_DepthMaxF;

-      for (i = 0; i < width; i++) {

-         z[i] = (GLuint) (depth[i] * depthMaxF);

-      }

-   }

-   else {

-      /* need to be careful with overflow */

-      const GLdouble depthMaxF = ctx->DrawBuffer->_DepthMaxF;

-      for (i = 0; i < width; i++) {

-         GLdouble d = depth[i] * ctx->Pixel.DepthScale + ctx->Pixel.DepthBias;

-         d = CLAMP(d, 0.0, 1.0) * depthMaxF;

-         if (d >= depthMaxF)

-            z[i] = depthMax;

-         else

-            z[i] = (GLuint) d;

-      }

-   }

-}

-

-

-

-/*

- * TODO: Optimize!!!!

- */

-static void

-copy_depth_pixels( struct gl_context *ctx, GLint srcx, GLint srcy,

-                   GLint width, GLint height,

-                   GLint destx, GLint desty )

-{

-   struct gl_framebuffer *fb = ctx->ReadBuffer;

-   struct gl_renderbuffer *readRb = fb->_DepthBuffer;

-   GLfloat *p, *tmpImage;

-   GLint sy, dy, stepy;

-   GLint j;

-   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;

-   GLint overlapping;

-   SWspan span;

-

-   if (!readRb) {

-      /* no readbuffer - OK */

-      return;

-   }

-

-   INIT_SPAN(span, GL_BITMAP);

-   _swrast_span_default_attribs(ctx, &span);

-   span.arrayMask = SPAN_Z;

-

-   if (ctx->DrawBuffer == ctx->ReadBuffer) {

-      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,

-                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);

-   }

-   else {

-      overlapping = GL_FALSE;

-   }

-

-   /* Determine if copy should be bottom-to-top or top-to-bottom */

-   if (!overlapping && srcy < desty) {

-      /* top-down  max-to-min */

-      sy = srcy + height - 1;

-      dy = desty + height - 1;

-      stepy = -1;

-   }

-   else {

-      /* bottom-up  min-to-max */

-      sy = srcy;

-      dy = desty;

-      stepy = 1;

-   }

-

-   if (overlapping) {

-      GLint ssy = sy;

-      tmpImage = (GLfloat *) malloc(width * height * sizeof(GLfloat));

-      if (!tmpImage) {

-         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );

-         return;

-      }

-      p = tmpImage;

-      for (j = 0; j < height; j++, ssy += stepy) {

-         _swrast_read_depth_span_float(ctx, readRb, width, srcx, ssy, p);

-         p += width;

-      }

-      p = tmpImage;

-   }

-   else {

-      tmpImage = NULL;  /* silence compiler warning */

-      p = NULL;

-   }

-

-   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {

-      GLfloat depth[MAX_WIDTH];

-      /* get depth values */

-      if (overlapping) {

-         memcpy(depth, p, width * sizeof(GLfloat));

-         p += width;

-      }

-      else {

-         _swrast_read_depth_span_float(ctx, readRb, width, srcx, sy, depth);

-      }

-

-      /* apply scale and bias */

-      scale_and_bias_z(ctx, width, depth, span.array->z);

-

-      /* write depth values */

-      span.x = destx;

-      span.y = dy;

-      span.end = width;

-      if (zoom)

-         _swrast_write_zoomed_depth_span(ctx, destx, desty, &span);

-      else

-         _swrast_write_rgba_span(ctx, &span);

-   }

-

-   if (overlapping)

-      free(tmpImage);

-}

-

-

-

-static void

-copy_stencil_pixels( struct gl_context *ctx, GLint srcx, GLint srcy,

-                     GLint width, GLint height,

-                     GLint destx, GLint desty )

-{

-   struct gl_framebuffer *fb = ctx->ReadBuffer;

-   struct gl_renderbuffer *rb = fb->_StencilBuffer;

-   GLint sy, dy, stepy;

-   GLint j;

-   GLstencil *p, *tmpImage;

-   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;

-   GLint overlapping;

-

-   if (!rb) {

-      /* no readbuffer - OK */

-      return;

-   }

-

-   if (ctx->DrawBuffer == ctx->ReadBuffer) {

-      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,

-                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);

-   }

-   else {

-      overlapping = GL_FALSE;

-   }

-

-   /* Determine if copy should be bottom-to-top or top-to-bottom */

-   if (!overlapping && srcy < desty) {

-      /* top-down  max-to-min */

-      sy = srcy + height - 1;

-      dy = desty + height - 1;

-      stepy = -1;

-   }

-   else {

-      /* bottom-up  min-to-max */

-      sy = srcy;

-      dy = desty;

-      stepy = 1;

-   }

-

-   if (overlapping) {

-      GLint ssy = sy;

-      tmpImage = (GLstencil *) malloc(width * height * sizeof(GLstencil));

-      if (!tmpImage) {

-         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );

-         return;

-      }

-      p = tmpImage;

-      for (j = 0; j < height; j++, ssy += stepy) {

-         _swrast_read_stencil_span( ctx, rb, width, srcx, ssy, p );

-         p += width;

-      }

-      p = tmpImage;

-   }

-   else {

-      tmpImage = NULL;  /* silence compiler warning */

-      p = NULL;

-   }

-

-   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {

-      GLstencil stencil[MAX_WIDTH];

-

-      /* Get stencil values */

-      if (overlapping) {

-         memcpy(stencil, p, width * sizeof(GLstencil));

-         p += width;

-      }

-      else {

-         _swrast_read_stencil_span( ctx, rb, width, srcx, sy, stencil );

-      }

-

-      _mesa_apply_stencil_transfer_ops(ctx, width, stencil);

-

-      /* Write stencil values */

-      if (zoom) {

-         _swrast_write_zoomed_stencil_span(ctx, destx, desty, width,

-                                           destx, dy, stencil);

-      }

-      else {

-         _swrast_write_stencil_span( ctx, width, destx, dy, stencil );

-      }

-   }

-

-   if (overlapping)

-      free(tmpImage);

-}

-

-

-/**

- * This isn't terribly efficient.  If a driver really has combined

- * depth/stencil buffers the driver should implement an optimized

- * CopyPixels function.

- */

-static void

-copy_depth_stencil_pixels(struct gl_context *ctx,

-                          const GLint srcX, const GLint srcY,

-                          const GLint width, const GLint height,

-                          const GLint destX, const GLint destY)

-{

-   struct gl_renderbuffer *stencilReadRb, *depthReadRb, *depthDrawRb;

-   GLint sy, dy, stepy;

-   GLint j;

-   GLstencil *tempStencilImage = NULL, *stencilPtr = NULL;

-   GLfloat *tempDepthImage = NULL, *depthPtr = NULL;

-   const GLfloat depthScale = ctx->DrawBuffer->_DepthMaxF;

-   const GLuint stencilMask = ctx->Stencil.WriteMask[0];

-   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;

-   const GLboolean scaleOrBias

-      = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;

-   GLint overlapping;

-

-   depthDrawRb = ctx->DrawBuffer->_DepthBuffer;

-   depthReadRb = ctx->ReadBuffer->_DepthBuffer;

-   stencilReadRb = ctx->ReadBuffer->_StencilBuffer;

-

-   ASSERT(depthDrawRb);

-   ASSERT(depthReadRb);

-   ASSERT(stencilReadRb);

-

-   if (ctx->DrawBuffer == ctx->ReadBuffer) {

-      overlapping = regions_overlap(srcX, srcY, destX, destY, width, height,

-                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);

-   }

-   else {

-      overlapping = GL_FALSE;

-   }

-

-   /* Determine if copy should be bottom-to-top or top-to-bottom */

-   if (!overlapping && srcY < destY) {

-      /* top-down  max-to-min */

-      sy = srcY + height - 1;

-      dy = destY + height - 1;

-      stepy = -1;

-   }

-   else {

-      /* bottom-up  min-to-max */

-      sy = srcY;

-      dy = destY;

-      stepy = 1;

-   }

-

-   if (overlapping) {

-      GLint ssy = sy;

-

-      if (stencilMask != 0x0) {

-         tempStencilImage

-            = (GLstencil *) malloc(width * height * sizeof(GLstencil));

-         if (!tempStencilImage) {

-            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCopyPixels");

-            return;

-         }

-

-         /* get copy of stencil pixels */

-         stencilPtr = tempStencilImage;

-         for (j = 0; j < height; j++, ssy += stepy) {

-            _swrast_read_stencil_span(ctx, stencilReadRb,

-                                      width, srcX, ssy, stencilPtr);

-            stencilPtr += width;

-         }

-         stencilPtr = tempStencilImage;

-      }

-

-      if (ctx->Depth.Mask) {

-         tempDepthImage

-            = (GLfloat *) malloc(width * height * sizeof(GLfloat));

-         if (!tempDepthImage) {

-            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCopyPixels");

-            free(tempStencilImage);

-            return;

-         }

-

-         /* get copy of depth pixels */

-         depthPtr = tempDepthImage;

-         for (j = 0; j < height; j++, ssy += stepy) {

-            _swrast_read_depth_span_float(ctx, depthReadRb,

-                                          width, srcX, ssy, depthPtr);

-            depthPtr += width;

-         }

-         depthPtr = tempDepthImage;

-      }

-   }

-

-   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {

-      if (stencilMask != 0x0) {

-         GLstencil stencil[MAX_WIDTH];

-

-         /* Get stencil values */

-         if (overlapping) {

-            memcpy(stencil, stencilPtr, width * sizeof(GLstencil));

-            stencilPtr += width;

-         }

-         else {

-            _swrast_read_stencil_span(ctx, stencilReadRb,

-                                      width, srcX, sy, stencil);

-         }

-

-         _mesa_apply_stencil_transfer_ops(ctx, width, stencil);

-

-         /* Write values */

-         if (zoom) {

-            _swrast_write_zoomed_stencil_span(ctx, destX, destY, width,

-                                              destX, dy, stencil);

-         }

-         else {

-            _swrast_write_stencil_span( ctx, width, destX, dy, stencil );

-         }

-      }

-

-      if (ctx->Depth.Mask) {

-         GLfloat depth[MAX_WIDTH];

-         GLuint zVals32[MAX_WIDTH];

-         GLushort zVals16[MAX_WIDTH];

-         GLvoid *zVals;

-         GLuint zBytes;

-

-         /* get depth values */

-         if (overlapping) {

-            memcpy(depth, depthPtr, width * sizeof(GLfloat));

-            depthPtr += width;

-         }

-         else {

-            _swrast_read_depth_span_float(ctx, depthReadRb,

-                                          width, srcX, sy, depth);

-         }

-

-         /* scale & bias */

-         if (scaleOrBias) {

-            _mesa_scale_and_bias_depth(ctx, width, depth);

-         }

-         /* convert to integer Z values */

-         if (depthDrawRb->DataType == GL_UNSIGNED_SHORT) {

-            GLint k;

-            for (k = 0; k < width; k++)

-               zVals16[k] = (GLushort) (depth[k] * depthScale);

-            zVals = zVals16;

-            zBytes = 2;

-         }

-         else {

-            GLint k;

-            for (k = 0; k < width; k++)

-               zVals32[k] = (GLuint) (depth[k] * depthScale);

-            zVals = zVals32;

-            zBytes = 4;

-         }

-

-         /* Write values */

-         if (zoom) {

-            _swrast_write_zoomed_z_span(ctx, destX, destY, width,

-                                        destX, dy, zVals);

-         }

-         else {

-            _swrast_put_row(ctx, depthDrawRb, width, destX, dy, zVals, zBytes);

-         }

-      }

-   }

-

-   if (tempStencilImage)

-      free(tempStencilImage);

-

-   if (tempDepthImage)

-      free(tempDepthImage);

-}

-

-

-

-/**

- * Try to do a fast copy pixels.

- */

-static GLboolean

-fast_copy_pixels(struct gl_context *ctx,

-                 GLint srcX, GLint srcY, GLsizei width, GLsizei height,

-                 GLint dstX, GLint dstY, GLenum type)

-{

-   struct gl_framebuffer *srcFb = ctx->ReadBuffer;

-   struct gl_framebuffer *dstFb = ctx->DrawBuffer;

-   struct gl_renderbuffer *srcRb, *dstRb;

-   GLint row, yStep;

-

-   if (SWRAST_CONTEXT(ctx)->_RasterMask != 0x0 ||

-       ctx->Pixel.ZoomX != 1.0F ||

-       ctx->Pixel.ZoomY != 1.0F ||

-       ctx->_ImageTransferState) {

-      /* can't handle these */

-      return GL_FALSE;

-   }

-

-   if (type == GL_COLOR) {

-      if (dstFb->_NumColorDrawBuffers != 1)

-         return GL_FALSE;

-      srcRb = srcFb->_ColorReadBuffer;

-      dstRb = dstFb->_ColorDrawBuffers[0];

-   }

-   else if (type == GL_STENCIL) {

-      srcRb = srcFb->_StencilBuffer;

-      dstRb = dstFb->_StencilBuffer;

-   }

-   else if (type == GL_DEPTH) {

-      srcRb = srcFb->_DepthBuffer;

-      dstRb = dstFb->_DepthBuffer;

-   }

-   else {

-      ASSERT(type == GL_DEPTH_STENCIL_EXT);

-      /* XXX correct? */

-      srcRb = srcFb->Attachment[BUFFER_DEPTH].Renderbuffer;

-      dstRb = dstFb->Attachment[BUFFER_DEPTH].Renderbuffer;

-   }

-

-   /* src and dst renderbuffers must be same format and type */

-   if (!srcRb || !dstRb ||

-       srcRb->DataType != dstRb->DataType ||

-       srcRb->_BaseFormat != dstRb->_BaseFormat) {

-      return GL_FALSE;

-   }

-

-   /* clipping not supported */

-   if (srcX < 0 || srcX + width > (GLint) srcFb->Width ||

-       srcY < 0 || srcY + height > (GLint) srcFb->Height ||

-       dstX < dstFb->_Xmin || dstX + width > dstFb->_Xmax ||

-       dstY < dstFb->_Ymin || dstY + height > dstFb->_Ymax) {

-      return GL_FALSE;

-   }

-

-   /* overlapping src/dst doesn't matter, just determine Y direction */

-   if (srcY < dstY) {

-      /* top-down  max-to-min */

-      srcY = srcY + height - 1;

-      dstY = dstY + height - 1;

-      yStep = -1;

-   }

-   else {

-      /* bottom-up  min-to-max */

-      yStep = 1;

-   }

-

-   for (row = 0; row < height; row++) {

-      GLuint temp[MAX_WIDTH][4];

-      srcRb->GetRow(ctx, srcRb, width, srcX, srcY, temp);

-      dstRb->PutRow(ctx, dstRb, width, dstX, dstY, temp, NULL);

-      srcY += yStep;

-      dstY += yStep;

-   }

-

-   return GL_TRUE;

-}

-

-

-/**

- * Do software-based glCopyPixels.

- * By time we get here, all parameters will have been error-checked.

- */

-void

-_swrast_CopyPixels( struct gl_context *ctx,

-		    GLint srcx, GLint srcy, GLsizei width, GLsizei height,

-		    GLint destx, GLint desty, GLenum type )

-{

-   SWcontext *swrast = SWRAST_CONTEXT(ctx);

-   swrast_render_start(ctx);

-      

-   if (!_mesa_check_conditional_render(ctx))

-      return; /* don't copy */

-

-   if (swrast->NewState)

-      _swrast_validate_derived( ctx );

-

-   if (!fast_copy_pixels(ctx, srcx, srcy, width, height, destx, desty, type)) {

-      switch (type) {

-      case GL_COLOR:

-         copy_rgba_pixels( ctx, srcx, srcy, width, height, destx, desty );

-         break;

-      case GL_DEPTH:

-         copy_depth_pixels( ctx, srcx, srcy, width, height, destx, desty );

-         break;

-      case GL_STENCIL:

-         copy_stencil_pixels( ctx, srcx, srcy, width, height, destx, desty );

-         break;

-      case GL_DEPTH_STENCIL_EXT:

-         copy_depth_stencil_pixels(ctx, srcx, srcy, width, height, destx, desty);

-         break;

-      default:

-         _mesa_problem(ctx, "unexpected type in _swrast_CopyPixels");

-      }

-   }

-

-   swrast_render_finish(ctx);

-}

+/*
+ * Mesa 3-D graphics library
+ * Version:  7.1
+ *
+ * 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"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/colormac.h"
+#include "main/condrender.h"
+#include "main/macros.h"
+#include "main/pixeltransfer.h"
+#include "main/imports.h"
+
+#include "s_context.h"
+#include "s_depth.h"
+#include "s_span.h"
+#include "s_stencil.h"
+#include "s_zoom.h"
+
+
+
+/**
+ * Determine if there's overlap in an image copy.
+ * This test also compensates for the fact that copies are done from
+ * bottom to top and overlaps can sometimes be handled correctly
+ * without making a temporary image copy.
+ * \return GL_TRUE if the regions overlap, GL_FALSE otherwise.
+ */
+static GLboolean
+regions_overlap(GLint srcx, GLint srcy,
+                GLint dstx, GLint dsty,
+                GLint width, GLint height,
+                GLfloat zoomX, GLfloat zoomY)
+{
+   if (zoomX == 1.0 && zoomY == 1.0) {
+      /* no zoom */
+      if (srcx >= dstx + width || (srcx + width <= dstx)) {
+         return GL_FALSE;
+      }
+      else if (srcy < dsty) { /* this is OK */
+         return GL_FALSE;
+      }
+      else if (srcy > dsty + height) {
+         return GL_FALSE;
+      }
+      else {
+         return GL_TRUE;
+      }
+   }
+   else {
+      /* add one pixel of slop when zooming, just to be safe */
+      if (srcx > (dstx + ((zoomX > 0.0F) ? (width * zoomX + 1.0F) : 0.0F))) {
+         /* src is completely right of dest */
+         return GL_FALSE;
+      }
+      else if (srcx + width + 1.0F < dstx + ((zoomX > 0.0F) ? 0.0F : (width * zoomX))) {
+         /* src is completely left of dest */
+         return GL_FALSE;
+      }
+      else if ((srcy < dsty) && (srcy + height < dsty + (height * zoomY))) {
+         /* src is completely below dest */
+         return GL_FALSE;
+      }
+      else if ((srcy > dsty) && (srcy + height > dsty + (height * zoomY))) {
+         /* src is completely above dest */
+         return GL_FALSE;
+      }
+      else {
+         return GL_TRUE;
+      }
+   }
+}
+
+
+/**
+ * RGBA copypixels
+ */
+static void
+copy_rgba_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,
+                 GLint width, GLint height, GLint destx, GLint desty)
+{
+   GLfloat *tmpImage, *p;
+   GLint sy, dy, stepy, row;
+   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;
+   GLint overlapping;
+   GLuint transferOps = ctx->_ImageTransferState;
+   SWspan span;
+
+   if (!ctx->ReadBuffer->_ColorReadBuffer) {
+      /* no readbuffer - OK */
+      return;
+   }
+
+   if (ctx->DrawBuffer == ctx->ReadBuffer) {
+      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,
+                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);
+   }
+   else {
+      overlapping = GL_FALSE;
+   }
+
+   /* Determine if copy should be done bottom-to-top or top-to-bottom */
+   if (!overlapping && srcy < desty) {
+      /* top-down  max-to-min */
+      sy = srcy + height - 1;
+      dy = desty + height - 1;
+      stepy = -1;
+   }
+   else {
+      /* bottom-up  min-to-max */
+      sy = srcy;
+      dy = desty;
+      stepy = 1;
+   }
+
+   INIT_SPAN(span, GL_BITMAP);
+   _swrast_span_default_attribs(ctx, &span);
+   span.arrayMask = SPAN_RGBA;
+   span.arrayAttribs = FRAG_BIT_COL0; /* we'll fill in COL0 attrib values */
+
+   if (overlapping) {
+      tmpImage = (GLfloat *) malloc(width * height * sizeof(GLfloat) * 4);
+      if (!tmpImage) {
+         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );
+         return;
+      }
+      /* read the source image as RGBA/float */
+      p = tmpImage;
+      for (row = 0; row < height; row++) {
+         _swrast_read_rgba_span( ctx, ctx->ReadBuffer->_ColorReadBuffer,
+                                 width, srcx, sy + row, GL_FLOAT, p );
+         p += width * 4;
+      }
+      p = tmpImage;
+   }
+   else {
+      tmpImage = NULL;  /* silence compiler warnings */
+      p = NULL;
+   }
+
+   ASSERT(width < MAX_WIDTH);
+
+   for (row = 0; row < height; row++, sy += stepy, dy += stepy) {
+      GLvoid *rgba = span.array->attribs[FRAG_ATTRIB_COL0];
+
+      /* Get row/span of source pixels */
+      if (overlapping) {
+         /* get from buffered image */
+         memcpy(rgba, p, width * sizeof(GLfloat) * 4);
+         p += width * 4;
+      }
+      else {
+         /* get from framebuffer */
+         _swrast_read_rgba_span( ctx, ctx->ReadBuffer->_ColorReadBuffer,
+                                 width, srcx, sy, GL_FLOAT, rgba );
+      }
+
+      if (transferOps) {
+         _mesa_apply_rgba_transfer_ops(ctx, transferOps, width,
+                                       (GLfloat (*)[4]) rgba);
+      }
+
+      /* Write color span */
+      span.x = destx;
+      span.y = dy;
+      span.end = width;
+      span.array->ChanType = GL_FLOAT;
+      if (zoom) {
+         _swrast_write_zoomed_rgba_span(ctx, destx, desty, &span, rgba);
+      }
+      else {
+         _swrast_write_rgba_span(ctx, &span);
+      }
+   }
+
+   span.array->ChanType = CHAN_TYPE; /* restore */
+
+   if (overlapping)
+      free(tmpImage);
+}
+
+
+/**
+ * Convert floating point Z values to integer Z values with pixel transfer's
+ * Z scale and bias.
+ */
+static void
+scale_and_bias_z(struct gl_context *ctx, GLuint width,
+                 const GLfloat depth[], GLuint z[])
+{
+   const GLuint depthMax = ctx->DrawBuffer->_DepthMax;
+   GLuint i;
+
+   if (depthMax <= 0xffffff &&
+       ctx->Pixel.DepthScale == 1.0 &&
+       ctx->Pixel.DepthBias == 0.0) {
+      /* no scale or bias and no clamping and no worry of overflow */
+      const GLfloat depthMaxF = ctx->DrawBuffer->_DepthMaxF;
+      for (i = 0; i < width; i++) {
+         z[i] = (GLuint) (depth[i] * depthMaxF);
+      }
+   }
+   else {
+      /* need to be careful with overflow */
+      const GLdouble depthMaxF = ctx->DrawBuffer->_DepthMaxF;
+      for (i = 0; i < width; i++) {
+         GLdouble d = depth[i] * ctx->Pixel.DepthScale + ctx->Pixel.DepthBias;
+         d = CLAMP(d, 0.0, 1.0) * depthMaxF;
+         if (d >= depthMaxF)
+            z[i] = depthMax;
+         else
+            z[i] = (GLuint) d;
+      }
+   }
+}
+
+
+
+/*
+ * TODO: Optimize!!!!
+ */
+static void
+copy_depth_pixels( struct gl_context *ctx, GLint srcx, GLint srcy,
+                   GLint width, GLint height,
+                   GLint destx, GLint desty )
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *readRb = fb->_DepthBuffer;
+   GLfloat *p, *tmpImage;
+   GLint sy, dy, stepy;
+   GLint j;
+   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;
+   GLint overlapping;
+   SWspan span;
+
+   if (!readRb) {
+      /* no readbuffer - OK */
+      return;
+   }
+
+   INIT_SPAN(span, GL_BITMAP);
+   _swrast_span_default_attribs(ctx, &span);
+   span.arrayMask = SPAN_Z;
+
+   if (ctx->DrawBuffer == ctx->ReadBuffer) {
+      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,
+                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);
+   }
+   else {
+      overlapping = GL_FALSE;
+   }
+
+   /* Determine if copy should be bottom-to-top or top-to-bottom */
+   if (!overlapping && srcy < desty) {
+      /* top-down  max-to-min */
+      sy = srcy + height - 1;
+      dy = desty + height - 1;
+      stepy = -1;
+   }
+   else {
+      /* bottom-up  min-to-max */
+      sy = srcy;
+      dy = desty;
+      stepy = 1;
+   }
+
+   if (overlapping) {
+      GLint ssy = sy;
+      tmpImage = (GLfloat *) malloc(width * height * sizeof(GLfloat));
+      if (!tmpImage) {
+         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );
+         return;
+      }
+      p = tmpImage;
+      for (j = 0; j < height; j++, ssy += stepy) {
+         _swrast_read_depth_span_float(ctx, readRb, width, srcx, ssy, p);
+         p += width;
+      }
+      p = tmpImage;
+   }
+   else {
+      tmpImage = NULL;  /* silence compiler warning */
+      p = NULL;
+   }
+
+   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {
+      GLfloat depth[MAX_WIDTH];
+      /* get depth values */
+      if (overlapping) {
+         memcpy(depth, p, width * sizeof(GLfloat));
+         p += width;
+      }
+      else {
+         _swrast_read_depth_span_float(ctx, readRb, width, srcx, sy, depth);
+      }
+
+      /* apply scale and bias */
+      scale_and_bias_z(ctx, width, depth, span.array->z);
+
+      /* write depth values */
+      span.x = destx;
+      span.y = dy;
+      span.end = width;
+      if (zoom)
+         _swrast_write_zoomed_depth_span(ctx, destx, desty, &span);
+      else
+         _swrast_write_rgba_span(ctx, &span);
+   }
+
+   if (overlapping)
+      free(tmpImage);
+}
+
+
+
+static void
+copy_stencil_pixels( struct gl_context *ctx, GLint srcx, GLint srcy,
+                     GLint width, GLint height,
+                     GLint destx, GLint desty )
+{
+   struct gl_framebuffer *fb = ctx->ReadBuffer;
+   struct gl_renderbuffer *rb = fb->_StencilBuffer;
+   GLint sy, dy, stepy;
+   GLint j;
+   GLubyte *p, *tmpImage;
+   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;
+   GLint overlapping;
+
+   if (!rb) {
+      /* no readbuffer - OK */
+      return;
+   }
+
+   if (ctx->DrawBuffer == ctx->ReadBuffer) {
+      overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,
+                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);
+   }
+   else {
+      overlapping = GL_FALSE;
+   }
+
+   /* Determine if copy should be bottom-to-top or top-to-bottom */
+   if (!overlapping && srcy < desty) {
+      /* top-down  max-to-min */
+      sy = srcy + height - 1;
+      dy = desty + height - 1;
+      stepy = -1;
+   }
+   else {
+      /* bottom-up  min-to-max */
+      sy = srcy;
+      dy = desty;
+      stepy = 1;
+   }
+
+   if (overlapping) {
+      GLint ssy = sy;
+      tmpImage = (GLubyte *) malloc(width * height * sizeof(GLubyte));
+      if (!tmpImage) {
+         _mesa_error( ctx, GL_OUT_OF_MEMORY, "glCopyPixels" );
+         return;
+      }
+      p = tmpImage;
+      for (j = 0; j < height; j++, ssy += stepy) {
+         _swrast_read_stencil_span( ctx, rb, width, srcx, ssy, p );
+         p += width;
+      }
+      p = tmpImage;
+   }
+   else {
+      tmpImage = NULL;  /* silence compiler warning */
+      p = NULL;
+   }
+
+   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {
+      GLubyte stencil[MAX_WIDTH];
+
+      /* Get stencil values */
+      if (overlapping) {
+         memcpy(stencil, p, width * sizeof(GLubyte));
+         p += width;
+      }
+      else {
+         _swrast_read_stencil_span( ctx, rb, width, srcx, sy, stencil );
+      }
+
+      _mesa_apply_stencil_transfer_ops(ctx, width, stencil);
+
+      /* Write stencil values */
+      if (zoom) {
+         _swrast_write_zoomed_stencil_span(ctx, destx, desty, width,
+                                           destx, dy, stencil);
+      }
+      else {
+         _swrast_write_stencil_span( ctx, width, destx, dy, stencil );
+      }
+   }
+
+   if (overlapping)
+      free(tmpImage);
+}
+
+
+/**
+ * This isn't terribly efficient.  If a driver really has combined
+ * depth/stencil buffers the driver should implement an optimized
+ * CopyPixels function.
+ */
+static void
+copy_depth_stencil_pixels(struct gl_context *ctx,
+                          const GLint srcX, const GLint srcY,
+                          const GLint width, const GLint height,
+                          const GLint destX, const GLint destY)
+{
+   struct gl_renderbuffer *stencilReadRb, *depthReadRb, *depthDrawRb;
+   GLint sy, dy, stepy;
+   GLint j;
+   GLubyte *tempStencilImage = NULL, *stencilPtr = NULL;
+   GLfloat *tempDepthImage = NULL, *depthPtr = NULL;
+   const GLfloat depthScale = ctx->DrawBuffer->_DepthMaxF;
+   const GLuint stencilMask = ctx->Stencil.WriteMask[0];
+   const GLboolean zoom = ctx->Pixel.ZoomX != 1.0F || ctx->Pixel.ZoomY != 1.0F;
+   const GLboolean scaleOrBias
+      = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
+   GLint overlapping;
+
+   depthDrawRb = ctx->DrawBuffer->_DepthBuffer;
+   depthReadRb = ctx->ReadBuffer->_DepthBuffer;
+   stencilReadRb = ctx->ReadBuffer->_StencilBuffer;
+
+   ASSERT(depthDrawRb);
+   ASSERT(depthReadRb);
+   ASSERT(stencilReadRb);
+
+   if (ctx->DrawBuffer == ctx->ReadBuffer) {
+      overlapping = regions_overlap(srcX, srcY, destX, destY, width, height,
+                                    ctx->Pixel.ZoomX, ctx->Pixel.ZoomY);
+   }
+   else {
+      overlapping = GL_FALSE;
+   }
+
+   /* Determine if copy should be bottom-to-top or top-to-bottom */
+   if (!overlapping && srcY < destY) {
+      /* top-down  max-to-min */
+      sy = srcY + height - 1;
+      dy = destY + height - 1;
+      stepy = -1;
+   }
+   else {
+      /* bottom-up  min-to-max */
+      sy = srcY;
+      dy = destY;
+      stepy = 1;
+   }
+
+   if (overlapping) {
+      GLint ssy = sy;
+
+      if (stencilMask != 0x0) {
+         tempStencilImage
+            = (GLubyte *) malloc(width * height * sizeof(GLubyte));
+         if (!tempStencilImage) {
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCopyPixels");
+            return;
+         }
+
+         /* get copy of stencil pixels */
+         stencilPtr = tempStencilImage;
+         for (j = 0; j < height; j++, ssy += stepy) {
+            _swrast_read_stencil_span(ctx, stencilReadRb,
+                                      width, srcX, ssy, stencilPtr);
+            stencilPtr += width;
+         }
+         stencilPtr = tempStencilImage;
+      }
+
+      if (ctx->Depth.Mask) {
+         tempDepthImage
+            = (GLfloat *) malloc(width * height * sizeof(GLfloat));
+         if (!tempDepthImage) {
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCopyPixels");
+            free(tempStencilImage);
+            return;
+         }
+
+         /* get copy of depth pixels */
+         depthPtr = tempDepthImage;
+         for (j = 0; j < height; j++, ssy += stepy) {
+            _swrast_read_depth_span_float(ctx, depthReadRb,
+                                          width, srcX, ssy, depthPtr);
+            depthPtr += width;
+         }
+         depthPtr = tempDepthImage;
+      }
+   }
+
+   for (j = 0; j < height; j++, sy += stepy, dy += stepy) {
+      if (stencilMask != 0x0) {
+         GLubyte stencil[MAX_WIDTH];
+
+         /* Get stencil values */
+         if (overlapping) {
+            memcpy(stencil, stencilPtr, width * sizeof(GLubyte));
+            stencilPtr += width;
+         }
+         else {
+            _swrast_read_stencil_span(ctx, stencilReadRb,
+                                      width, srcX, sy, stencil);
+         }
+
+         _mesa_apply_stencil_transfer_ops(ctx, width, stencil);
+
+         /* Write values */
+         if (zoom) {
+            _swrast_write_zoomed_stencil_span(ctx, destX, destY, width,
+                                              destX, dy, stencil);
+         }
+         else {
+            _swrast_write_stencil_span( ctx, width, destX, dy, stencil );
+         }
+      }
+
+      if (ctx->Depth.Mask) {
+         GLfloat depth[MAX_WIDTH];
+         GLuint zVals32[MAX_WIDTH];
+         GLushort zVals16[MAX_WIDTH];
+         GLvoid *zVals;
+         GLuint zBytes;
+
+         /* get depth values */
+         if (overlapping) {
+            memcpy(depth, depthPtr, width * sizeof(GLfloat));
+            depthPtr += width;
+         }
+         else {
+            _swrast_read_depth_span_float(ctx, depthReadRb,
+                                          width, srcX, sy, depth);
+         }
+
+         /* scale & bias */
+         if (scaleOrBias) {
+            _mesa_scale_and_bias_depth(ctx, width, depth);
+         }
+         /* convert to integer Z values */
+         if (depthDrawRb->DataType == GL_UNSIGNED_SHORT) {
+            GLint k;
+            for (k = 0; k < width; k++)
+               zVals16[k] = (GLushort) (depth[k] * depthScale);
+            zVals = zVals16;
+            zBytes = 2;
+         }
+         else {
+            GLint k;
+            for (k = 0; k < width; k++)
+               zVals32[k] = (GLuint) (depth[k] * depthScale);
+            zVals = zVals32;
+            zBytes = 4;
+         }
+
+         /* Write values */
+         if (zoom) {
+            _swrast_write_zoomed_z_span(ctx, destX, destY, width,
+                                        destX, dy, zVals);
+         }
+         else {
+            _swrast_put_row(ctx, depthDrawRb, width, destX, dy, zVals, zBytes);
+         }
+      }
+   }
+
+   if (tempStencilImage)
+      free(tempStencilImage);
+
+   if (tempDepthImage)
+      free(tempDepthImage);
+}
+
+
+
+/**
+ * Try to do a fast copy pixels.
+ */
+static GLboolean
+fast_copy_pixels(struct gl_context *ctx,
+                 GLint srcX, GLint srcY, GLsizei width, GLsizei height,
+                 GLint dstX, GLint dstY, GLenum type)
+{
+   struct gl_framebuffer *srcFb = ctx->ReadBuffer;
+   struct gl_framebuffer *dstFb = ctx->DrawBuffer;
+   struct gl_renderbuffer *srcRb, *dstRb;
+   GLint row, yStep;
+
+   if (SWRAST_CONTEXT(ctx)->_RasterMask != 0x0 ||
+       ctx->Pixel.ZoomX != 1.0F ||
+       ctx->Pixel.ZoomY != 1.0F ||
+       ctx->_ImageTransferState) {
+      /* can't handle these */
+      return GL_FALSE;
+   }
+
+   if (type == GL_COLOR) {
+      if (dstFb->_NumColorDrawBuffers != 1)
+         return GL_FALSE;
+      srcRb = srcFb->_ColorReadBuffer;
+      dstRb = dstFb->_ColorDrawBuffers[0];
+   }
+   else if (type == GL_STENCIL) {
+      srcRb = srcFb->_StencilBuffer;
+      dstRb = dstFb->_StencilBuffer;
+   }
+   else if (type == GL_DEPTH) {
+      srcRb = srcFb->_DepthBuffer;
+      dstRb = dstFb->_DepthBuffer;
+   }
+   else {
+      ASSERT(type == GL_DEPTH_STENCIL_EXT);
+      /* XXX correct? */
+      srcRb = srcFb->Attachment[BUFFER_DEPTH].Renderbuffer;
+      dstRb = dstFb->Attachment[BUFFER_DEPTH].Renderbuffer;
+   }
+
+   /* src and dst renderbuffers must be same format and type */
+   if (!srcRb || !dstRb ||
+       srcRb->DataType != dstRb->DataType ||
+       srcRb->_BaseFormat != dstRb->_BaseFormat) {
+      return GL_FALSE;
+   }
+
+   /* clipping not supported */
+   if (srcX < 0 || srcX + width > (GLint) srcFb->Width ||
+       srcY < 0 || srcY + height > (GLint) srcFb->Height ||
+       dstX < dstFb->_Xmin || dstX + width > dstFb->_Xmax ||
+       dstY < dstFb->_Ymin || dstY + height > dstFb->_Ymax) {
+      return GL_FALSE;
+   }
+
+   /* overlapping src/dst doesn't matter, just determine Y direction */
+   if (srcY < dstY) {
+      /* top-down  max-to-min */
+      srcY = srcY + height - 1;
+      dstY = dstY + height - 1;
+      yStep = -1;
+   }
+   else {
+      /* bottom-up  min-to-max */
+      yStep = 1;
+   }
+
+   for (row = 0; row < height; row++) {
+      GLuint temp[MAX_WIDTH][4];
+      srcRb->GetRow(ctx, srcRb, width, srcX, srcY, temp);
+      dstRb->PutRow(ctx, dstRb, width, dstX, dstY, temp, NULL);
+      srcY += yStep;
+      dstY += yStep;
+   }
+
+   return GL_TRUE;
+}
+
+
+/**
+ * Do software-based glCopyPixels.
+ * By time we get here, all parameters will have been error-checked.
+ */
+void
+_swrast_CopyPixels( struct gl_context *ctx,
+		    GLint srcx, GLint srcy, GLsizei width, GLsizei height,
+		    GLint destx, GLint desty, GLenum type )
+{
+   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   swrast_render_start(ctx);
+      
+   if (!_mesa_check_conditional_render(ctx))
+      return; /* don't copy */
+
+   if (swrast->NewState)
+      _swrast_validate_derived( ctx );
+
+   if (!fast_copy_pixels(ctx, srcx, srcy, width, height, destx, desty, type)) {
+      switch (type) {
+      case GL_COLOR:
+         copy_rgba_pixels( ctx, srcx, srcy, width, height, destx, desty );
+         break;
+      case GL_DEPTH:
+         copy_depth_pixels( ctx, srcx, srcy, width, height, destx, desty );
+         break;
+      case GL_STENCIL:
+         copy_stencil_pixels( ctx, srcx, srcy, width, height, destx, desty );
+         break;
+      case GL_DEPTH_STENCIL_EXT:
+         copy_depth_stencil_pixels(ctx, srcx, srcy, width, height, destx, desty);
+         break;
+      default:
+         _mesa_problem(ctx, "unexpected type in _swrast_CopyPixels");
+      }
+   }
+
+   swrast_render_finish(ctx);
+}
diff --git a/mesalib/src/mesa/swrast/s_depth.c b/mesalib/src/mesa/swrast/s_depth.c
index 6ffa0a310..8d3ad105b 100644
--- a/mesalib/src/mesa/swrast/s_depth.c
+++ b/mesalib/src/mesa/swrast/s_depth.c
@@ -1248,8 +1248,7 @@ _swrast_depth_bounds_test( struct gl_context *ctx, SWspan *span )
  * Read a span of depth values from the given depth renderbuffer, returning
  * the values as GLfloats.
  * This function does clipping to prevent reading outside the depth buffer's
- * bounds.  Though the clipping is redundant when we're called from
- * _swrast_ReadPixels.
+ * bounds.
  */
 void
 _swrast_read_depth_span_float( struct gl_context *ctx, struct gl_renderbuffer *rb,
diff --git a/mesalib/src/mesa/swrast/s_drawpix.c b/mesalib/src/mesa/swrast/s_drawpix.c
index 20bf4d626..b6c433753 100644
--- a/mesalib/src/mesa/swrast/s_drawpix.c
+++ b/mesalib/src/mesa/swrast/s_drawpix.c
@@ -320,6 +320,7 @@ draw_stencil_pixels( struct gl_context *ctx, GLint x, GLint y,
                      const GLvoid *pixels )
 {
    const GLboolean zoom = ctx->Pixel.ZoomX != 1.0 || ctx->Pixel.ZoomY != 1.0;
+   const GLenum destType = GL_UNSIGNED_BYTE;
    GLint skipPixels;
 
    /* if width > MAX_WIDTH, have to process image in chunks */
@@ -330,9 +331,7 @@ draw_stencil_pixels( struct gl_context *ctx, GLint x, GLint y,
       GLint row;
       for (row = 0; row < height; row++) {
          const GLint spanY = y + row;
-         GLstencil values[MAX_WIDTH];
-         GLenum destType = (sizeof(GLstencil) == sizeof(GLubyte))
-                         ? GL_UNSIGNED_BYTE : GL_UNSIGNED_SHORT;
+         GLubyte values[MAX_WIDTH];
          const GLvoid *source = _mesa_image_address2d(unpack, pixels,
                                                       width, height,
                                                       GL_STENCIL_INDEX, type,
@@ -570,8 +569,7 @@ draw_depth_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
       = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
    const GLuint depthMax = ctx->DrawBuffer->_DepthMax;
    const GLuint stencilMask = ctx->Stencil.WriteMask[0];
-   const GLuint stencilType = (STENCIL_BITS == 8) ? 
-      GL_UNSIGNED_BYTE : GL_UNSIGNED_SHORT;
+   const GLenum stencilType = GL_UNSIGNED_BYTE;
    const GLboolean zoom = ctx->Pixel.ZoomX != 1.0 || ctx->Pixel.ZoomY != 1.0;
    struct gl_renderbuffer *depthRb, *stencilRb;
    struct gl_pixelstore_attrib clippedUnpack = *unpack;
@@ -672,7 +670,7 @@ draw_depth_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
          }
 
          if (stencilMask != 0x0) {
-            GLstencil stencilValues[MAX_WIDTH];
+            GLubyte stencilValues[MAX_WIDTH];
             /* get stencil values, with shift/offset/mapping */
             _mesa_unpack_stencil_span(ctx, width, stencilType, stencilValues,
                                       type, depthStencilSrc, &clippedUnpack,
diff --git a/mesalib/src/mesa/swrast/s_readpix.c b/mesalib/src/mesa/swrast/s_readpix.c
deleted file mode 100644
index 500686798..000000000
--- a/mesalib/src/mesa/swrast/s_readpix.c
+++ /dev/null
@@ -1,522 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  7.0.3
- *
- * 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"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-#include "main/glheader.h"
-#include "main/colormac.h"
-#include "main/feedback.h"
-#include "main/formats.h"
-#include "main/format_unpack.h"
-#include "main/image.h"
-#include "main/imports.h"
-#include "main/macros.h"
-#include "main/pack.h"
-#include "main/pbo.h"
-#include "main/state.h"
-
-#include "s_context.h"
-#include "s_depth.h"
-#include "s_span.h"
-#include "s_stencil.h"
-
-/**
- * Tries to implement glReadPixels() of GL_DEPTH_COMPONENT using memcpy of the
- * mapping.
- */
-static GLboolean
-fast_read_depth_pixels( struct gl_context *ctx,
-			GLint x, GLint y,
-			GLsizei width, GLsizei height,
-			GLenum type, GLvoid *pixels,
-			const struct gl_pixelstore_attrib *packing )
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   GLubyte *map, *dst;
-   int stride, dstStride, j;
-
-   if (ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0)
-      return GL_FALSE;
-
-   if (packing->SwapBytes)
-      return GL_FALSE;
-
-   if (_mesa_get_format_datatype(rb->Format) != GL_UNSIGNED_INT)
-      return GL_FALSE;
-
-   if (!((type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16) ||
-	 type == GL_UNSIGNED_INT))
-      return GL_FALSE;
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   dstStride = _mesa_image_row_stride(packing, width, GL_DEPTH_COMPONENT, type);
-   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
-					   GL_DEPTH_COMPONENT, type, 0, 0);
-
-   for (j = 0; j < height; j++) {
-      if (type == GL_UNSIGNED_INT) {
-	 _mesa_unpack_uint_z_row(rb->Format, width, map, (GLuint *)dst);
-      } else {
-	 ASSERT(type == GL_UNSIGNED_SHORT && rb->Format == MESA_FORMAT_Z16);
-	 memcpy(dst, map, width * 2);
-      }
-
-      map += stride;
-      dst += dstStride;
-   }
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-
-   return GL_TRUE;
-}
-
-/**
- * Read pixels for format=GL_DEPTH_COMPONENT.
- */
-static void
-read_depth_pixels( struct gl_context *ctx,
-                   GLint x, GLint y,
-                   GLsizei width, GLsizei height,
-                   GLenum type, GLvoid *pixels,
-                   const struct gl_pixelstore_attrib *packing )
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   GLint j;
-   GLubyte *dst, *map;
-   int dstStride, stride;
-
-   if (!rb)
-      return;
-
-   /* clipping should have been done already */
-   ASSERT(x >= 0);
-   ASSERT(y >= 0);
-   ASSERT(x + width <= (GLint) rb->Width);
-   ASSERT(y + height <= (GLint) rb->Height);
-   /* width should never be > MAX_WIDTH since we did clipping earlier */
-   ASSERT(width <= MAX_WIDTH);
-
-   if (fast_read_depth_pixels(ctx, x, y, width, height, type, pixels, packing))
-      return;
-
-   dstStride = _mesa_image_row_stride(packing, width, GL_DEPTH_COMPONENT, type);
-   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
-					   GL_DEPTH_COMPONENT, type, 0, 0);
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   /* General case (slower) */
-   for (j = 0; j < height; j++, y++) {
-      GLfloat depthValues[MAX_WIDTH];
-      _mesa_unpack_float_z_row(rb->Format, width, map, depthValues);
-      _mesa_pack_depth_span(ctx, width, dst, type, depthValues, packing);
-
-      dst += dstStride;
-      map += stride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-}
-
-
-/**
- * Read pixels for format=GL_STENCIL_INDEX.
- */
-static void
-read_stencil_pixels( struct gl_context *ctx,
-                     GLint x, GLint y,
-                     GLsizei width, GLsizei height,
-                     GLenum type, GLvoid *pixels,
-                     const struct gl_pixelstore_attrib *packing )
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
-   GLint j;
-   GLubyte *map;
-   GLint stride;
-
-   if (!rb)
-      return;
-
-   /* width should never be > MAX_WIDTH since we did clipping earlier */
-   ASSERT(width <= MAX_WIDTH);
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   /* process image row by row */
-   for (j = 0; j < height; j++) {
-      GLvoid *dest;
-      GLstencil stencil[MAX_WIDTH];
-
-      _mesa_unpack_ubyte_stencil_row(rb->Format, width, map, stencil);
-      dest = _mesa_image_address2d(packing, pixels, width, height,
-                                   GL_STENCIL_INDEX, type, j, 0);
-
-      _mesa_pack_stencil_span(ctx, width, type, dest, stencil, packing);
-
-      map += stride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-}
-
-static GLboolean
-fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
-			      GLint x, GLint y,
-			      GLsizei width, GLsizei height,
-			      GLenum format, GLenum type,
-			      GLvoid *pixels,
-			      const struct gl_pixelstore_attrib *packing,
-			      GLbitfield transferOps )
-{
-   struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
-   GLubyte *dst, *map;
-   int dstStride, stride, j, texelBytes;
-
-   if (!_mesa_format_matches_format_and_type(rb->Format, format, type))
-      return GL_FALSE;
-
-   /* check for things we can't handle here */
-   if (packing->SwapBytes ||
-       packing->LsbFirst) {
-      return GL_FALSE;
-   }
-
-   dstStride = _mesa_image_row_stride(packing, width, format, type);
-   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
-					   format, type, 0, 0);
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   texelBytes = _mesa_get_format_bytes(rb->Format);
-   for (j = 0; j < height; j++) {
-      memcpy(dst, map, width * texelBytes);
-      dst += dstStride;
-      map += stride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-
-   return GL_TRUE;
-}
-
-static GLboolean
-slow_read_rgba_pixels( struct gl_context *ctx,
-		       GLint x, GLint y,
-		       GLsizei width, GLsizei height,
-		       GLenum format, GLenum type,
-		       GLvoid *pixels,
-		       const struct gl_pixelstore_attrib *packing,
-		       GLbitfield transferOps )
-{
-   struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
-   const gl_format rbFormat = _mesa_get_srgb_format_linear(rb->Format);
-   union {
-      float f[MAX_WIDTH][4];
-      unsigned int i[MAX_WIDTH][4];
-   } rgba;
-   GLubyte *dst, *map;
-   int dstStride, stride, j;
-
-   dstStride = _mesa_image_row_stride(packing, width, format, type);
-   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
-					   format, type, 0, 0);
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   for (j = 0; j < height; j++) {
-      if (_mesa_is_integer_format(format)) {
-	 _mesa_unpack_int_rgba_row(rbFormat, width, map, rgba.i);
-	 _mesa_pack_rgba_span_int(ctx, width, rgba.i, format, type, dst);
-      } else {
-	 _mesa_unpack_rgba_row(rbFormat, width, map, rgba.f);
-	 _mesa_pack_rgba_span_float(ctx, width, rgba.f, format, type, dst,
-				    packing, transferOps);
-      }
-      dst += dstStride;
-      map += stride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-
-   return GL_TRUE;
-}
-
-/*
- * Read R, G, B, A, RGB, L, or LA pixels.
- */
-static void
-read_rgba_pixels( struct gl_context *ctx,
-                  GLint x, GLint y,
-                  GLsizei width, GLsizei height,
-                  GLenum format, GLenum type, GLvoid *pixels,
-                  const struct gl_pixelstore_attrib *packing )
-{
-   GLbitfield transferOps = ctx->_ImageTransferState;
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *rb = fb->_ColorReadBuffer;
-
-   if (!rb)
-      return;
-
-   if ((ctx->Color._ClampReadColor == GL_TRUE || type != GL_FLOAT) &&
-       !_mesa_is_integer_format(format)) {
-      transferOps |= IMAGE_CLAMP_BIT;
-   }
-
-   if (!transferOps) {
-      /* Try the optimized paths first. */
-      if (fast_read_rgba_pixels_memcpy(ctx, x, y, width, height,
-				       format, type, pixels, packing,
-				       transferOps)) {
-	 return;
-      }
-   }
-
-   slow_read_rgba_pixels(ctx, x, y, width, height,
-			 format, type, pixels, packing, transferOps);
-}
-
-/**
- * For a packed depth/stencil buffer being read as depth/stencil, just memcpy the
- * data (possibly swapping 8/24 vs 24/8 as we go).
- */
-static GLboolean
-fast_read_depth_stencil_pixels(struct gl_context *ctx,
-			       GLint x, GLint y,
-			       GLsizei width, GLsizei height,
-			       GLubyte *dst, int dstStride)
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *rb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
-   GLubyte *map;
-   int stride, i;
-
-   if (rb != stencilRb)
-      return GL_FALSE;
-
-   if (rb->Format != MESA_FORMAT_Z24_S8 &&
-       rb->Format != MESA_FORMAT_S8_Z24)
-      return GL_FALSE;
-
-   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
-			       &map, &stride);
-
-   for (i = 0; i < height; i++) {
-      _mesa_unpack_uint_24_8_depth_stencil_row(rb->Format, width,
-					       map, (GLuint *)dst);
-      map += stride;
-      dst = (char*)dst + dstStride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, rb);
-
-   return GL_TRUE;
-}
-
-
-/**
- * For non-float-depth and stencil buffers being read as 24/8 depth/stencil,
- * copy the integer data directly instead of converting depth to float and
- * re-packing.
- */
-static GLboolean
-fast_read_depth_stencil_pixels_separate(struct gl_context *ctx,
-					GLint x, GLint y,
-					GLsizei width, GLsizei height,
-					uint32_t *dst, int dstStride)
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *depthRb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
-   GLubyte *depthMap, *stencilMap;
-   int depthStride, stencilStride, i, j;
-
-   if (_mesa_get_format_datatype(depthRb->Format) != GL_UNSIGNED_INT)
-      return GL_FALSE;
-
-   ctx->Driver.MapRenderbuffer(ctx, depthRb, x, y, width, height,
-			       GL_MAP_READ_BIT, &depthMap, &depthStride);
-   ctx->Driver.MapRenderbuffer(ctx, stencilRb, x, y, width, height,
-			       GL_MAP_READ_BIT, &stencilMap, &stencilStride);
-
-   for (j = 0; j < height; j++) {
-      GLstencil stencilVals[MAX_WIDTH];
-
-      _mesa_unpack_uint_z_row(depthRb->Format, width, depthMap, dst);
-      _mesa_unpack_ubyte_stencil_row(stencilRb->Format, width,
-				     stencilMap, stencilVals);
-
-      for (i = 0; i < width; i++) {
-	 dst[i] = (dst[i] & 0xffffff00) | stencilVals[i];
-      }
-
-      depthMap += depthStride;
-      stencilMap += stencilStride;
-      dst += dstStride / 4;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, depthRb);
-   ctx->Driver.UnmapRenderbuffer(ctx, stencilRb);
-
-   return GL_TRUE;
-}
-
-static void
-slow_read_depth_stencil_pixels_separate(struct gl_context *ctx,
-					GLint x, GLint y,
-					GLsizei width, GLsizei height,
-					GLenum type,
-					const struct gl_pixelstore_attrib *packing,
-					GLubyte *dst, int dstStride)
-{
-   struct gl_framebuffer *fb = ctx->ReadBuffer;
-   struct gl_renderbuffer *depthRb = fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   struct gl_renderbuffer *stencilRb = fb->Attachment[BUFFER_STENCIL].Renderbuffer;
-   GLubyte *depthMap, *stencilMap;
-   int depthStride, stencilStride, j;
-
-   ctx->Driver.MapRenderbuffer(ctx, depthRb, x, y, width, height,
-			       GL_MAP_READ_BIT, &depthMap, &depthStride);
-   ctx->Driver.MapRenderbuffer(ctx, stencilRb, x, y, width, height,
-			       GL_MAP_READ_BIT, &stencilMap, &stencilStride);
-
-   for (j = 0; j < height; j++) {
-      GLstencil stencilVals[MAX_WIDTH];
-      GLfloat depthVals[MAX_WIDTH];
-
-      _mesa_unpack_float_z_row(depthRb->Format, width, depthMap, depthVals);
-      _mesa_unpack_ubyte_stencil_row(stencilRb->Format, width,
-				     stencilMap, stencilVals);
-
-      _mesa_pack_depth_stencil_span(ctx, width, type, (GLuint *)dst,
-				    depthVals, stencilVals, packing);
-
-      depthMap += depthStride;
-      stencilMap += stencilStride;
-      dst += dstStride;
-   }
-
-   ctx->Driver.UnmapRenderbuffer(ctx, depthRb);
-   ctx->Driver.UnmapRenderbuffer(ctx, stencilRb);
-}
-
-
-/**
- * Read combined depth/stencil values.
- * We'll have already done error checking to be sure the expected
- * depth and stencil buffers really exist.
- */
-static void
-read_depth_stencil_pixels(struct gl_context *ctx,
-                          GLint x, GLint y,
-                          GLsizei width, GLsizei height,
-                          GLenum type, GLvoid *pixels,
-                          const struct gl_pixelstore_attrib *packing )
-{
-   const GLboolean scaleOrBias
-      = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
-   const GLboolean stencilTransfer = ctx->Pixel.IndexShift
-      || ctx->Pixel.IndexOffset || ctx->Pixel.MapStencilFlag;
-   GLubyte *dst;
-   int dstStride;
-
-   dst = (GLubyte *) _mesa_image_address2d(packing, pixels,
-					   width, height,
-					   GL_DEPTH_STENCIL_EXT,
-					   type, 0, 0);
-   dstStride = _mesa_image_row_stride(packing, width,
-				      GL_DEPTH_STENCIL_EXT, type);
-
-   /* Fast 24/8 reads. */
-   if (type == GL_UNSIGNED_INT_24_8 &&
-       !scaleOrBias && !stencilTransfer && !packing->SwapBytes) {
-      if (fast_read_depth_stencil_pixels(ctx, x, y, width, height,
-					 dst, dstStride))
-	 return;
-
-      if (fast_read_depth_stencil_pixels_separate(ctx, x, y, width, height,
-						  (uint32_t *)dst, dstStride))
-	 return;
-   }
-
-   slow_read_depth_stencil_pixels_separate(ctx, x, y, width, height,
-					   type, packing,
-					   dst, dstStride);
-}
-
-
-
-/**
- * Software fallback routine for ctx->Driver.ReadPixels().
- * By time we get here, all error checking will have been done.
- */
-void
-_swrast_ReadPixels( struct gl_context *ctx,
-		    GLint x, GLint y, GLsizei width, GLsizei height,
-		    GLenum format, GLenum type,
-		    const struct gl_pixelstore_attrib *packing,
-		    GLvoid *pixels )
-{
-   struct gl_pixelstore_attrib clippedPacking = *packing;
-
-   if (ctx->NewState)
-      _mesa_update_state(ctx);
-
-   /* Do all needed clipping here, so that we can forget about it later */
-   if (_mesa_clip_readpixels(ctx, &x, &y, &width, &height, &clippedPacking)) {
-
-      pixels = _mesa_map_pbo_dest(ctx, &clippedPacking, pixels);
-
-      if (pixels) {
-         switch (format) {
-         case GL_STENCIL_INDEX:
-            read_stencil_pixels(ctx, x, y, width, height, type, pixels,
-                                &clippedPacking);
-            break;
-         case GL_DEPTH_COMPONENT:
-            read_depth_pixels(ctx, x, y, width, height, type, pixels,
-                              &clippedPacking);
-            break;
-         case GL_DEPTH_STENCIL_EXT:
-            read_depth_stencil_pixels(ctx, x, y, width, height, type, pixels,
-                                      &clippedPacking);
-            break;
-         default:
-            /* all other formats should be color formats */
-            read_rgba_pixels(ctx, x, y, width, height, format, type, pixels,
-                             &clippedPacking);
-         }
-
-         _mesa_unmap_pbo_dest(ctx, &clippedPacking);
-      }
-   }
-}
diff --git a/mesalib/src/mesa/swrast/s_stencil.c b/mesalib/src/mesa/swrast/s_stencil.c
index e713e2393..101ee5056 100644
--- a/mesalib/src/mesa/swrast/s_stencil.c
+++ b/mesalib/src/mesa/swrast/s_stencil.c
@@ -62,12 +62,12 @@ ENDIF
  */
 static void
 apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
-                  GLuint n, GLstencil stencil[], const GLubyte mask[] )
+                  GLuint n, GLubyte stencil[], const GLubyte mask[] )
 {
-   const GLstencil ref = ctx->Stencil.Ref[face];
-   const GLstencil wrtmask = ctx->Stencil.WriteMask[face];
-   const GLstencil invmask = (GLstencil) (~wrtmask);
-   const GLstencil stencilMax = (1 << ctx->DrawBuffer->Visual.stencilBits) - 1;
+   const GLubyte ref = ctx->Stencil.Ref[face];
+   const GLubyte wrtmask = ctx->Stencil.WriteMask[face];
+   const GLubyte invmask = (GLubyte) (~wrtmask);
+   const GLubyte stencilMax = (1 << ctx->DrawBuffer->Visual.stencilBits) - 1;
    GLuint i;
 
    switch (oper) {
@@ -85,7 +85,7 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  stencil[i] = (GLstencil) (stencil[i] & invmask);
+		  stencil[i] = (GLubyte) (stencil[i] & invmask);
 	       }
 	    }
 	 }
@@ -101,8 +101,8 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  GLstencil s = stencil[i];
-		  stencil[i] = (GLstencil) ((invmask & s ) | (wrtmask & ref));
+		  GLubyte s = stencil[i];
+		  stencil[i] = (GLubyte) ((invmask & s ) | (wrtmask & ref));
 	       }
 	    }
 	 }
@@ -111,9 +111,9 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  GLstencil s = stencil[i];
+		  GLubyte s = stencil[i];
 		  if (s < stencilMax) {
-		     stencil[i] = (GLstencil) (s+1);
+		     stencil[i] = (GLubyte) (s+1);
 		  }
 	       }
 	    }
@@ -122,9 +122,9 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
 		  /* VERIFY logic of adding 1 to a write-masked value */
-		  GLstencil s = stencil[i];
+		  GLubyte s = stencil[i];
 		  if (s < stencilMax) {
-		     stencil[i] = (GLstencil) ((invmask & s) | (wrtmask & (s+1)));
+		     stencil[i] = (GLubyte) ((invmask & s) | (wrtmask & (s+1)));
 		  }
 	       }
 	    }
@@ -134,9 +134,9 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  GLstencil s = stencil[i];
+		  GLubyte s = stencil[i];
 		  if (s>0) {
-		     stencil[i] = (GLstencil) (s-1);
+		     stencil[i] = (GLubyte) (s-1);
 		  }
 	       }
 	    }
@@ -145,9 +145,9 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
 		  /* VERIFY logic of subtracting 1 to a write-masked value */
-		  GLstencil s = stencil[i];
+		  GLubyte s = stencil[i];
 		  if (s>0) {
-		     stencil[i] = (GLstencil) ((invmask & s) | (wrtmask & (s-1)));
+		     stencil[i] = (GLubyte) ((invmask & s) | (wrtmask & (s-1)));
 		  }
 	       }
 	    }
@@ -164,8 +164,8 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil s = stencil[i];
-                  stencil[i] = (GLstencil) ((invmask & s) | (wrtmask & (s+1)));
+                  GLubyte s = stencil[i];
+                  stencil[i] = (GLubyte) ((invmask & s) | (wrtmask & (s+1)));
 	       }
 	    }
 	 }
@@ -181,8 +181,8 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil s = stencil[i];
-                  stencil[i] = (GLstencil) ((invmask & s) | (wrtmask & (s-1)));
+                  GLubyte s = stencil[i];
+                  stencil[i] = (GLubyte) ((invmask & s) | (wrtmask & (s-1)));
 	       }
 	    }
 	 }
@@ -191,16 +191,16 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  GLstencil s = stencil[i];
-		  stencil[i] = (GLstencil) ~s;
+		  GLubyte s = stencil[i];
+		  stencil[i] = (GLubyte) ~s;
 	       }
 	    }
 	 }
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-		  GLstencil s = stencil[i];
-		  stencil[i] = (GLstencil) ((invmask & s) | (wrtmask & ~s));
+		  GLubyte s = stencil[i];
+		  stencil[i] = (GLubyte) ((invmask & s) | (wrtmask & ~s));
 	       }
 	    }
 	 }
@@ -225,15 +225,15 @@ apply_stencil_op( const struct gl_context *ctx, GLenum oper, GLuint face,
  * Return:  GL_FALSE = all pixels failed, GL_TRUE = zero or more pixels passed.
  */
 static GLboolean
-do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stencil[],
+do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLubyte stencil[],
                  GLubyte mask[] )
 {
    GLubyte fail[MAX_WIDTH];
    GLboolean allfail = GL_FALSE;
    GLuint i;
    const GLuint valueMask = ctx->Stencil.ValueMask[face];
-   const GLstencil r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
-   GLstencil s;
+   const GLubyte r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
+   GLubyte s;
 
    ASSERT(n <= MAX_WIDTH);
 
@@ -263,7 +263,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_LESS:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r < s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -281,7 +281,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_LEQUAL:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r <= s) {
 		  /* pass */
 		  fail[i] = 0;
@@ -299,7 +299,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_GREATER:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r > s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -317,7 +317,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_GEQUAL:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r >= s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -335,7 +335,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_EQUAL:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r == s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -353,7 +353,7 @@ do_stencil_test( struct gl_context *ctx, GLuint face, GLuint n, GLstencil stenci
       case GL_NOTEQUAL:
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-	       s = (GLstencil) (stencil[i] & valueMask);
+	       s = (GLubyte) (stencil[i] & valueMask);
 	       if (r != s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -422,8 +422,8 @@ stencil_and_ztest_span(struct gl_context *ctx, SWspan *span, GLuint face)
 {
    struct gl_framebuffer *fb = ctx->DrawBuffer;
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
-   GLstencil stencilRow[MAX_WIDTH];
-   GLstencil *stencil;
+   GLubyte stencilRow[MAX_WIDTH];
+   GLubyte *stencil;
    const GLuint n = span->end;
    const GLint x = span->x;
    const GLint y = span->y;
@@ -438,7 +438,7 @@ stencil_and_ztest_span(struct gl_context *ctx, SWspan *span, GLuint face)
    }
 #endif
 
-   stencil = (GLstencil *) rb->GetPointer(ctx, rb, x, y);
+   stencil = (GLubyte *) rb->GetPointer(ctx, rb, x, y);
    if (!stencil) {
       rb->GetRow(ctx, rb, n, x, y, stencilRow);
       stencil = stencilRow;
@@ -531,16 +531,15 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 {
    struct gl_framebuffer *fb = ctx->DrawBuffer;
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
-   const GLstencil stencilMax = (1 << fb->Visual.stencilBits) - 1;
-   const GLstencil ref = ctx->Stencil.Ref[face];
-   const GLstencil wrtmask = ctx->Stencil.WriteMask[face];
-   const GLstencil invmask = (GLstencil) (~wrtmask);
+   const GLubyte stencilMax = (1 << fb->Visual.stencilBits) - 1;
+   const GLubyte ref = ctx->Stencil.Ref[face];
+   const GLubyte wrtmask = ctx->Stencil.WriteMask[face];
+   const GLubyte invmask = (GLubyte) (~wrtmask);
    GLuint i;
-   GLstencil *stencilStart = (GLubyte *) rb->Data;
+   GLubyte *stencilStart = (GLubyte *) rb->Data;
    const GLuint stride = rb->Width;
 
    ASSERT(rb->GetPointer(ctx, rb, 0, 0));
-   ASSERT(sizeof(GLstencil) == 1);
 
    switch (oper) {
       case GL_KEEP:
@@ -550,7 +549,7 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
                   *sptr = 0;
 	       }
 	    }
@@ -558,8 +557,8 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-		  *sptr = (GLstencil) (invmask & *sptr);
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+		  *sptr = (GLubyte) (invmask & *sptr);
 	       }
 	    }
 	 }
@@ -568,7 +567,7 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
                   *sptr = ref;
 	       }
 	    }
@@ -576,8 +575,8 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-		  *sptr = (GLstencil) ((invmask & *sptr ) | (wrtmask & ref));
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+		  *sptr = (GLubyte) ((invmask & *sptr ) | (wrtmask & ref));
 	       }
 	    }
 	 }
@@ -586,9 +585,9 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
 		  if (*sptr < stencilMax) {
-		     *sptr = (GLstencil) (*sptr + 1);
+		     *sptr = (GLubyte) (*sptr + 1);
 		  }
 	       }
 	    }
@@ -596,9 +595,9 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
 		  if (*sptr < stencilMax) {
-		     *sptr = (GLstencil) ((invmask & *sptr) | (wrtmask & (*sptr+1)));
+		     *sptr = (GLubyte) ((invmask & *sptr) | (wrtmask & (*sptr+1)));
 		  }
 	       }
 	    }
@@ -608,9 +607,9 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
 		  if (*sptr>0) {
-		     *sptr = (GLstencil) (*sptr - 1);
+		     *sptr = (GLubyte) (*sptr - 1);
 		  }
 	       }
 	    }
@@ -618,9 +617,9 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
 		  if (*sptr>0) {
-		     *sptr = (GLstencil) ((invmask & *sptr) | (wrtmask & (*sptr-1)));
+		     *sptr = (GLubyte) ((invmask & *sptr) | (wrtmask & (*sptr-1)));
 		  }
 	       }
 	    }
@@ -630,16 +629,16 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) (*sptr + 1);
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) (*sptr + 1);
 	       }
 	    }
 	 }
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) ((invmask & *sptr) | (wrtmask & (*sptr+1)));
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) ((invmask & *sptr) | (wrtmask & (*sptr+1)));
 	       }
 	    }
 	 }
@@ -648,16 +647,16 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) (*sptr - 1);
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) (*sptr - 1);
 	       }
 	    }
 	 }
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) ((invmask & *sptr) | (wrtmask & (*sptr-1)));
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) ((invmask & *sptr) | (wrtmask & (*sptr-1)));
 	       }
 	    }
 	 }
@@ -666,16 +665,16 @@ apply_stencil_op_to_pixels( struct gl_context *ctx,
 	 if (invmask==0) {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) (~*sptr);
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) (~*sptr);
 	       }
 	    }
 	 }
 	 else {
 	    for (i=0;i<n;i++) {
 	       if (mask[i]) {
-                  GLstencil *sptr = STENCIL_ADDRESS( x[i], y[i] );
-                  *sptr = (GLstencil) ((invmask & *sptr) | (wrtmask & ~*sptr));
+                  GLubyte *sptr = STENCIL_ADDRESS( x[i], y[i] );
+                  *sptr = (GLubyte) ((invmask & *sptr) | (wrtmask & ~*sptr));
 	       }
 	    }
 	 }
@@ -705,15 +704,14 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
    const struct gl_framebuffer *fb = ctx->DrawBuffer;
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
    GLubyte fail[MAX_WIDTH];
-   GLstencil r, s;
+   GLubyte r, s;
    GLuint i;
    GLboolean allfail = GL_FALSE;
    const GLuint valueMask = ctx->Stencil.ValueMask[face];
-   const GLstencil *stencilStart = (GLstencil *) rb->Data;
+   const GLubyte *stencilStart = (GLubyte *) rb->Data;
    const GLuint stride = rb->Width;
 
    ASSERT(rb->GetPointer(ctx, rb, 0, 0));
-   ASSERT(sizeof(GLstencil) == 1);
 
    /*
     * Perform stencil test.  The results of this operation are stored
@@ -740,11 +738,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 allfail = GL_TRUE;
 	 break;
       case GL_LESS:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r < s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -760,11 +758,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 }
 	 break;
       case GL_LEQUAL:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r <= s) {
 		  /* pass */
 		  fail[i] = 0;
@@ -780,11 +778,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 }
 	 break;
       case GL_GREATER:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r > s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -800,11 +798,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 }
 	 break;
       case GL_GEQUAL:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r >= s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -820,11 +818,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 }
 	 break;
       case GL_EQUAL:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r == s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -840,11 +838,11 @@ stencil_test_pixels( struct gl_context *ctx, GLuint face, GLuint n,
 	 }
 	 break;
       case GL_NOTEQUAL:
-	 r = (GLstencil) (ctx->Stencil.Ref[face] & valueMask);
+	 r = (GLubyte) (ctx->Stencil.Ref[face] & valueMask);
 	 for (i=0;i<n;i++) {
 	    if (mask[i]) {
-               const GLstencil *sptr = STENCIL_ADDRESS(x[i],y[i]);
-	       s = (GLstencil) (*sptr & valueMask);
+               const GLubyte *sptr = STENCIL_ADDRESS(x[i],y[i]);
+	       s = (GLubyte) (*sptr & valueMask);
 	       if (r != s) {
 		  /* passed */
 		  fail[i] = 0;
@@ -914,7 +912,7 @@ stencil_and_ztest_pixels( struct gl_context *ctx, SWspan *span, GLuint face )
 
    if (!rb->GetPointer(ctx, rb, 0, 0)) {
       /* No direct access */
-      GLstencil stencil[MAX_WIDTH];
+      GLubyte stencil[MAX_WIDTH];
 
       ASSERT(rb->DataType == GL_UNSIGNED_BYTE);
       _swrast_get_values(ctx, rb, n, x, y, stencil, sizeof(GLubyte));
@@ -1044,7 +1042,7 @@ clip_span(GLuint bufferWidth, GLuint bufferHeight,
  */
 void
 _swrast_read_stencil_span(struct gl_context *ctx, struct gl_renderbuffer *rb,
-                          GLint n, GLint x, GLint y, GLstencil stencil[])
+                          GLint n, GLint x, GLint y, GLubyte stencil[])
 {
    if (y < 0 || y >= (GLint) rb->Height ||
        x + n <= 0 || x >= (GLint) rb->Width) {
@@ -1081,7 +1079,7 @@ _swrast_read_stencil_span(struct gl_context *ctx, struct gl_renderbuffer *rb,
  */
 void
 _swrast_write_stencil_span(struct gl_context *ctx, GLint n, GLint x, GLint y,
-                           const GLstencil stencil[] )
+                           const GLubyte stencil[] )
 {
    struct gl_framebuffer *fb = ctx->DrawBuffer;
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
@@ -1109,7 +1107,7 @@ _swrast_write_stencil_span(struct gl_context *ctx, GLint n, GLint x, GLint y,
 
    if ((stencilMask & stencilMax) != stencilMax) {
       /* need to apply writemask */
-      GLstencil destVals[MAX_WIDTH], newVals[MAX_WIDTH];
+      GLubyte destVals[MAX_WIDTH], newVals[MAX_WIDTH];
       GLint i;
       rb->GetRow(ctx, rb, n, x, y, destVals);
       for (i = 0; i < n; i++) {
diff --git a/mesalib/src/mesa/swrast/s_stencil.h b/mesalib/src/mesa/swrast/s_stencil.h
index 0bcfb799f..37f3c8da1 100644
--- a/mesalib/src/mesa/swrast/s_stencil.h
+++ b/mesalib/src/mesa/swrast/s_stencil.h
@@ -1,53 +1,53 @@
-/*

- * Mesa 3-D graphics library

- * Version:  6.3

- *

- * Copyright (C) 1999-2005  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"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- */

-

-

-#ifndef S_STENCIL_H

-#define S_STENCIL_H

-

-

-#include "main/mtypes.h"

-#include "s_span.h"

-

-

-

-extern GLboolean

-_swrast_stencil_and_ztest_span(struct gl_context *ctx, SWspan *span);

-

-

-extern void

-_swrast_read_stencil_span(struct gl_context *ctx, struct gl_renderbuffer *rb,

-                          GLint n, GLint x, GLint y, GLstencil stencil[]);

-

-

-extern void

-_swrast_write_stencil_span( struct gl_context *ctx, GLint n, GLint x, GLint y,

-                          const GLstencil stencil[] );

-

-

-extern void

-_swrast_clear_stencil_buffer( struct gl_context *ctx, struct gl_renderbuffer *rb );

-

-

-#endif

+/*
+ * Mesa 3-D graphics library
+ * Version:  6.3
+ *
+ * Copyright (C) 1999-2005  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"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+#ifndef S_STENCIL_H
+#define S_STENCIL_H
+
+
+#include "main/mtypes.h"
+#include "s_span.h"
+
+
+
+extern GLboolean
+_swrast_stencil_and_ztest_span(struct gl_context *ctx, SWspan *span);
+
+
+extern void
+_swrast_read_stencil_span(struct gl_context *ctx, struct gl_renderbuffer *rb,
+                          GLint n, GLint x, GLint y, GLubyte stencil[]);
+
+
+extern void
+_swrast_write_stencil_span( struct gl_context *ctx, GLint n, GLint x, GLint y,
+                          const GLubyte stencil[] );
+
+
+extern void
+_swrast_clear_stencil_buffer( struct gl_context *ctx, struct gl_renderbuffer *rb );
+
+
+#endif
diff --git a/mesalib/src/mesa/swrast/s_texrender.c b/mesalib/src/mesa/swrast/s_texrender.c
index 83e7a6a03..3734ce6d5 100644
--- a/mesalib/src/mesa/swrast/s_texrender.c
+++ b/mesalib/src/mesa/swrast/s_texrender.c
@@ -567,6 +567,19 @@ update_wrapper(struct gl_context *ctx, struct gl_renderbuffer_attachment *att)
    trb->Base.InternalFormat = trb->TexImage->Base.InternalFormat;
    trb->Base.Format = trb->TexImage->Base.TexFormat;
 
+   /* Set the gl_renderbuffer::Data field so that mapping the buffer
+    * in renderbuffer.c succeeds.
+    */
+   if (att->Texture->Target == GL_TEXTURE_3D ||
+       att->Texture->Target == GL_TEXTURE_2D_ARRAY_EXT) {
+      trb->Base.Data = trb->TexImage->Buffer +
+         trb->TexImage->ImageOffsets[trb->Zoffset] *
+         _mesa_get_format_bytes(trb->TexImage->Base.TexFormat);
+   }
+   else {
+      trb->Base.Data = trb->TexImage->Buffer;
+   }
+
    /* XXX may need more special cases here */
    switch (trb->TexImage->Base.TexFormat) {
    case MESA_FORMAT_Z24_S8:
diff --git a/mesalib/src/mesa/swrast/s_zoom.c b/mesalib/src/mesa/swrast/s_zoom.c
index 3fb784847..16bb997f3 100644
--- a/mesalib/src/mesa/swrast/s_zoom.c
+++ b/mesalib/src/mesa/swrast/s_zoom.c
@@ -351,9 +351,9 @@ _swrast_write_zoomed_depth_span(struct gl_context *ctx, GLint imgX, GLint imgY,
 void
 _swrast_write_zoomed_stencil_span(struct gl_context *ctx, GLint imgX, GLint imgY,
                                   GLint width, GLint spanX, GLint spanY,
-                                  const GLstencil stencil[])
+                                  const GLubyte stencil[])
 {
-   GLstencil zoomedVals[MAX_WIDTH];
+   GLubyte zoomedVals[MAX_WIDTH];
    GLint x0, x1, y0, y1, y;
    GLint i, zoomedWidth;
 
diff --git a/mesalib/src/mesa/swrast/s_zoom.h b/mesalib/src/mesa/swrast/s_zoom.h
index b8d17b5a4..0b82bb824 100644
--- a/mesalib/src/mesa/swrast/s_zoom.h
+++ b/mesalib/src/mesa/swrast/s_zoom.h
@@ -1,56 +1,56 @@
-/*

- * Mesa 3-D graphics library

- * Version:  6.5

- *

- * Copyright (C) 1999-2005  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"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- */

-

-#ifndef S_ZOOM_H

-#define S_ZOOM_H

-

-#include "main/mtypes.h"

-#include "s_span.h"

-

-

-extern void

-_swrast_write_zoomed_rgba_span(struct gl_context *ctx, GLint imgX, GLint imgY,

-                               const SWspan *span, const GLvoid *rgba);

-

-extern void

-_swrast_write_zoomed_rgb_span(struct gl_context *ctx, GLint imgX, GLint imgY,

-                              const SWspan *span, const GLvoid *rgb);

-

-extern void

-_swrast_write_zoomed_depth_span(struct gl_context *ctx, GLint imgX, GLint imgY,

-                                const SWspan *span);

-

-

-extern void

-_swrast_write_zoomed_stencil_span(struct gl_context *ctx, GLint imgX, GLint imgY,

-                                  GLint width, GLint spanX, GLint spanY,

-                                  const GLstencil stencil[]);

-

-extern void

-_swrast_write_zoomed_z_span(struct gl_context *ctx, GLint imgX, GLint imgY,

-                            GLint width, GLint spanX, GLint spanY,

-                            const GLvoid *z);

-

-

-#endif

+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 1999-2005  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"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef S_ZOOM_H
+#define S_ZOOM_H
+
+#include "main/mtypes.h"
+#include "s_span.h"
+
+
+extern void
+_swrast_write_zoomed_rgba_span(struct gl_context *ctx, GLint imgX, GLint imgY,
+                               const SWspan *span, const GLvoid *rgba);
+
+extern void
+_swrast_write_zoomed_rgb_span(struct gl_context *ctx, GLint imgX, GLint imgY,
+                              const SWspan *span, const GLvoid *rgb);
+
+extern void
+_swrast_write_zoomed_depth_span(struct gl_context *ctx, GLint imgX, GLint imgY,
+                                const SWspan *span);
+
+
+extern void
+_swrast_write_zoomed_stencil_span(struct gl_context *ctx, GLint imgX, GLint imgY,
+                                  GLint width, GLint spanX, GLint spanY,
+                                  const GLubyte stencil[]);
+
+extern void
+_swrast_write_zoomed_z_span(struct gl_context *ctx, GLint imgX, GLint imgY,
+                            GLint width, GLint spanX, GLint spanY,
+                            const GLvoid *z);
+
+
+#endif
diff --git a/mesalib/src/mesa/swrast/swrast.h b/mesalib/src/mesa/swrast/swrast.h
index 08d565ba0..17b66c664 100644
--- a/mesalib/src/mesa/swrast/swrast.h
+++ b/mesalib/src/mesa/swrast/swrast.h
@@ -119,13 +119,6 @@ _swrast_DrawPixels( struct gl_context *ctx,
 		    const GLvoid *pixels );
 
 extern void
-_swrast_ReadPixels( struct gl_context *ctx,
-		    GLint x, GLint y, GLsizei width, GLsizei height,
-		    GLenum format, GLenum type,
-		    const struct gl_pixelstore_attrib *unpack,
-		    GLvoid *pixels );
-
-extern void
 _swrast_BlitFramebuffer(struct gl_context *ctx,
                         GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
                         GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,