Merge branch 'master' of git+ssh://joukj@git.freedesktop.org/git/mesa/mesa

author: Jouk <joukj@tarantella.(none)> 2007-07-25 10:16:11 +0200
committer: Jouk <joukj@tarantella.(none)> 2007-07-25 10:16:11 +0200
commit: eb9a5b6d5127858b01ec12672c999e7d25cd7aed (patch)
tree: 91c0302c4d007ddcd0c42fbb8d8de71bf23d3914 /src
parent: 55f8b7053065ce88296608071feed6f9540f6c0d (diff)
parent: 03ec41ddc51e539c989a546f33d22daa2af69095 (diff)
204 files changed, 4836 insertions, 4922 deletions
diff --git a/src/glu/sgi/Makefile.mgw b/src/glu/sgi/Makefile.mgw
new file mode 100644
index 0000000000..43b421e737
--- /dev/null
+++ b/src/glu/sgi/Makefile.mgw
@@ -0,0 +1,229 @@
+# Mesa 3-D graphics library
+# Version:  5.1
+# 
+# Copyright (C) 1999-2003  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.
+
+# MinGW core makefile v1.4 for Mesa
+#
+#  Copyright (C) 2002 - Daniel Borca
+#  Email : dborca@users.sourceforge.net
+#  Web   : http://www.geocities.com/dborca
+
+# MinGW core-glu makefile updated for Mesa 7.0
+#
+#  Updated : by Heromyth, on 2007-7-21
+#  Email   : zxpmyth@yahoo.com.cn
+#  Bugs    : 1) All the default settings work fine. But the setting X86=1 can't work. 
+#            The others havn't been tested yet.
+#            2) The generated DLLs are *not* compatible with the ones built
+#            with the other compilers like VC8, especially for GLUT. 
+#            3) Although more tests are needed, it can be used individually!
+
+#
+#  Available options:
+#
+#    Environment variables:
+#	CFLAGS
+#
+#	GLIDE		path to Glide3 SDK; used with FX.
+#			default = $(TOP)/glide3
+#	FX=1		build for 3dfx Glide3. Note that this disables
+#			compilation of most WMesa code and requires fxMesa.
+#			As a consequence, you'll need the Win32 Glide3
+#			library to build any application.
+#			default = no
+#	ICD=1		build the installable client driver interface
+#			(windows opengl driver interface)
+#			default = no
+#	X86=1		optimize for x86 (if possible, use MMX, SSE, 3DNow).
+#			default = no
+#
+#    Targets:
+#	all:		build GL
+#	clean:		remove object files
+#
+
+
+
+.PHONY: all clean
+.INTERMEDIATE: x86/gen_matypes.exe
+.SUFFIXES: .rc .res
+
+# Set this to the prefix of your build tools, i.e. mingw32-
+TOOLS_PREFIX = mingw32-
+
+TOP = ../../..
+
+LIBDIR = $(TOP)/lib
+
+GLU_DLL = glu32.dll
+GLU_IMP = libglu32.a
+GLU_DEF = glu.def
+
+include $(TOP)/configs/config.mgw
+GL_USING_STDCALL ?= 1
+
+LDLIBS = -L$(LIBDIR) -lopengl32
+LDFLAGS = -Wl,--out-implib=$(LIBDIR)/$(GLU_IMP) -Wl,--output-def=$(LIBDIR)/$(GLU_DEF)
+
+CFLAGS += -DBUILD_GLU32 -D_DLL
+
+ifeq ($(GL_USING_STDCALL),1)
+  LDFLAGS += -Wl,--add-stdcall-alias
+else
+  CFLAGS += -DGL_NO_STDCALL
+endif
+
+CC = gcc
+CFLAGS += -DNDEBUG -DLIBRARYBUILD -I$(TOP)/include -Iinclude
+CXX = g++
+CXXFLAGS = $(CFLAGS) -Ilibnurbs/internals -Ilibnurbs/interface -Ilibnurbs/nurbtess
+
+AR = ar
+ARFLAGS = crus
+
+UNLINK = del $(subst /,\,$(1))
+ifneq ($(wildcard $(addsuffix /rm.exe,$(subst ;, ,$(PATH)))),)
+UNLINK = $(RM) $(1)
+endif
+ifneq ($(wildcard $(addsuffix /rm,$(subst :, ,$(PATH)))),)
+UNLINK = $(RM) $(1)
+endif
+
+C_SOURCES = \
+	libutil/error.c \
+	libutil/glue.c \
+	libutil/mipmap.c \
+	libutil/project.c \
+	libutil/quad.c \
+	libutil/registry.c \
+	libtess/dict.c \
+	libtess/geom.c \
+	libtess/memalloc.c \
+	libtess/mesh.c \
+	libtess/normal.c \
+	libtess/priorityq.c \
+	libtess/render.c \
+	libtess/sweep.c \
+	libtess/tess.c \
+	libtess/tessmono.c
+
+CC_SOURCES = \
+	libnurbs/interface/bezierEval.cc \
+	libnurbs/interface/bezierPatch.cc \
+	libnurbs/interface/bezierPatchMesh.cc \
+	libnurbs/interface/glcurveval.cc \
+	libnurbs/interface/glinterface.cc \
+	libnurbs/interface/glrenderer.cc \
+	libnurbs/interface/glsurfeval.cc \
+	libnurbs/interface/incurveeval.cc \
+	libnurbs/interface/insurfeval.cc \
+	libnurbs/internals/arc.cc \
+	libnurbs/internals/arcsorter.cc \
+	libnurbs/internals/arctess.cc \
+	libnurbs/internals/backend.cc \
+	libnurbs/internals/basiccrveval.cc \
+	libnurbs/internals/basicsurfeval.cc \
+	libnurbs/internals/bin.cc \
+	libnurbs/internals/bufpool.cc \
+	libnurbs/internals/cachingeval.cc \
+	libnurbs/internals/ccw.cc \
+	libnurbs/internals/coveandtiler.cc \
+	libnurbs/internals/curve.cc \
+	libnurbs/internals/curvelist.cc \
+	libnurbs/internals/curvesub.cc \
+	libnurbs/internals/dataTransform.cc \
+	libnurbs/internals/displaylist.cc \
+	libnurbs/internals/flist.cc \
+	libnurbs/internals/flistsorter.cc \
+	libnurbs/internals/hull.cc \
+	libnurbs/internals/intersect.cc \
+	libnurbs/internals/knotvector.cc \
+	libnurbs/internals/mapdesc.cc \
+	libnurbs/internals/mapdescv.cc \
+	libnurbs/internals/maplist.cc \
+	libnurbs/internals/mesher.cc \
+	libnurbs/internals/monoTriangulationBackend.cc \
+	libnurbs/internals/monotonizer.cc \
+	libnurbs/internals/mycode.cc \
+	libnurbs/internals/nurbsinterfac.cc \
+	libnurbs/internals/nurbstess.cc \
+	libnurbs/internals/patch.cc \
+	libnurbs/internals/patchlist.cc \
+	libnurbs/internals/quilt.cc \
+	libnurbs/internals/reader.cc \
+	libnurbs/internals/renderhints.cc \
+	libnurbs/internals/slicer.cc \
+	libnurbs/internals/sorter.cc \
+	libnurbs/internals/splitarcs.cc \
+	libnurbs/internals/subdivider.cc \
+	libnurbs/internals/tobezier.cc \
+	libnurbs/internals/trimline.cc \
+	libnurbs/internals/trimregion.cc \
+	libnurbs/internals/trimvertpool.cc \
+	libnurbs/internals/uarray.cc \
+	libnurbs/internals/varray.cc \
+	libnurbs/nurbtess/directedLine.cc \
+	libnurbs/nurbtess/gridWrap.cc \
+	libnurbs/nurbtess/monoChain.cc \
+	libnurbs/nurbtess/monoPolyPart.cc \
+	libnurbs/nurbtess/monoTriangulation.cc \
+	libnurbs/nurbtess/partitionX.cc \
+	libnurbs/nurbtess/partitionY.cc \
+	libnurbs/nurbtess/polyDBG.cc \
+	libnurbs/nurbtess/polyUtil.cc \
+	libnurbs/nurbtess/primitiveStream.cc \
+	libnurbs/nurbtess/quicksort.cc \
+	libnurbs/nurbtess/rectBlock.cc \
+	libnurbs/nurbtess/sampleComp.cc \
+	libnurbs/nurbtess/sampleCompBot.cc \
+	libnurbs/nurbtess/sampleCompRight.cc \
+	libnurbs/nurbtess/sampleCompTop.cc \
+	libnurbs/nurbtess/sampleMonoPoly.cc \
+	libnurbs/nurbtess/sampledLine.cc \
+	libnurbs/nurbtess/searchTree.cc
+
+SOURCES = $(C_SOURCES) $(CC_SOURCES)
+
+OBJECTS = $(addsuffix .o,$(basename $(SOURCES)))
+
+.c.o:
+	$(CC) -o $@ $(CFLAGS) -c $<
+.cc.o:
+	$(CXX) -o $@ $(CXXFLAGS) -c $<
+
+
+all: $(LIBDIR) $(LIBDIR)/$(GLU_DLL) $(LIBDIR)/$(GLU_IMP)
+
+$(LIBDIR):
+	mkdir -p $(LIBDIR)
+
+$(LIBDIR)/$(GLU_DLL) $(LIBDIR)/$(GLU_IMP): $(OBJECTS)
+	g++ -shared -fPIC -o $(LIBDIR)/$(GLU_DLL) $(LDFLAGS) \
+    $^ $(LDLIBS)
+
+
+
+clean:
+	-$(call UNLINK,libutil/*.o)
+	-$(call UNLINK,libtess/*.o)
+	-$(call UNLINK,libnurbs/interface/*.o)
+	-$(call UNLINK,libnurbs/internals/*.o)
+	-$(call UNLINK,libnurbs/nurbtess/*.o)
diff --git a/src/glu/sgi/libnurbs/interface/glcurveval.h b/src/glu/sgi/libnurbs/interface/glcurveval.h
index 4b44f6e847..a09a74d04c 100644
--- a/src/glu/sgi/libnurbs/interface/glcurveval.h
+++ b/src/glu/sgi/libnurbs/interface/glcurveval.h
@@ -93,7 +93,7 @@ public:
 	output_triangles = flag;
       }
 #ifdef _WIN32
-    void               putCallBack(GLenum which, void (APIENTRY *fn)() );
+    void               putCallBack(GLenum which, void (GLAPIENTRY *fn)() );
 #else
     void               putCallBack(GLenum which, _GLUfuncptr fn );
 #endif
diff --git a/src/glu/sgi/libnurbs/interface/glsurfeval.cc b/src/glu/sgi/libnurbs/interface/glsurfeval.cc
index a36b304508..b5bfab1e28 100644
--- a/src/glu/sgi/libnurbs/interface/glsurfeval.cc
+++ b/src/glu/sgi/libnurbs/interface/glsurfeval.cc
@@ -1184,8 +1184,11 @@ return;
 
 }
 
-void
-OpenGLSurfaceEvaluator::putCallBack(GLenum which, _GLUfuncptr fn )
+#ifdef _WIN32
+void OpenGLSurfaceEvaluator::putCallBack(GLenum which, void (GLAPIENTRY *fn)() )
+#else
+void OpenGLSurfaceEvaluator::putCallBack(GLenum which, _GLUfuncptr fn )
+#endif
 {
   switch(which)
     {
diff --git a/src/glu/sgi/libnurbs/interface/glsurfeval.h b/src/glu/sgi/libnurbs/interface/glsurfeval.h
index c34a58cb9b..b7a88069f5 100644
--- a/src/glu/sgi/libnurbs/interface/glsurfeval.h
+++ b/src/glu/sgi/libnurbs/interface/glsurfeval.h
@@ -145,7 +145,7 @@ public:
     void		newtmeshvert( long, long );
 
 #ifdef _WIN32
-    void 	        putCallBack(GLenum which, void (APIENTRY *fn)() );
+    void 	        putCallBack(GLenum which, void (GLAPIENTRY *fn)() );
 #else
     void 	        putCallBack(GLenum which, _GLUfuncptr fn );
 #endif
diff --git a/src/glu/sgi/libnurbs/nurbtess/monoTriangulation.h b/src/glu/sgi/libnurbs/nurbtess/monoTriangulation.h
index 7ff7739434..002549ecbd 100644
--- a/src/glu/sgi/libnurbs/nurbtess/monoTriangulation.h
+++ b/src/glu/sgi/libnurbs/nurbtess/monoTriangulation.h
@@ -43,10 +43,9 @@
 #include "definitions.h"
 #include "primitiveStream.h"
 #include "directedLine.h"
+#include "arc.h"
 
 class Backend;
-class Arc;
-typedef Arc *Arc_ptr;
 
 class reflexChain{
   Real2 *queue; 
diff --git a/src/glut/directfb/callback.c b/src/glut/directfb/callback.c
index 38cfccbd4f..4f23441167 100644
--- a/src/glut/directfb/callback.c
+++ b/src/glut/directfb/callback.c
@@ -28,14 +28,13 @@
 typedef void (GLUTCALLBACK *__GlutTimerCallback) ( int value );
 
 typedef struct __GlutTimer_s {
-     unsigned int          interval;
+     struct timeval        interval;
      struct timeval        expire;
      
      __GlutTimerCallback   func;
      int                   value;
      
      struct __GlutTimer_s *next;
-     struct __GlutTimer_s *prev;
 } __GlutTimer;
 
 /*****************************************************************************/
@@ -207,8 +206,7 @@ glutIdleFunc( void (GLUTCALLBACK *func) (void) )
 void GLUTAPIENTRY
 glutTimerFunc( unsigned int msec, void (GLUTCALLBACK *func) (int value), int value )
 {
-     __GlutTimer    *timer;
-     struct timeval  now;     
+     __GlutTimer *timer;
      
      if (!func)
           return;
@@ -217,24 +215,19 @@ glutTimerFunc( unsigned int msec, void (GLUTCALLBACK *func) (int value), int val
      if (!timer)
           __glutFatalError( "out of memory" );
      
-     gettimeofday( &now, NULL );
-     
-     timer->interval = msec;
-     timer->expire.tv_sec  = now.tv_sec + (now.tv_usec/1000 + msec) / 1000;
-     timer->expire.tv_usec = (now.tv_usec + msec*1000) % 1000000;
+     timer->interval.tv_sec  = msec / 1000;
+     timer->interval.tv_usec = (msec % 1000) * 1000;
+
+     gettimeofday( &timer->expire, NULL );
+     timer->expire.tv_usec += timer->interval.tv_usec;
+     timer->expire.tv_sec  += timer->interval.tv_sec + timer->expire.tv_usec/1000000;
+     timer->expire.tv_usec %= 1000000;
      
      timer->func  = func;
      timer->value = value;
      
-     if (g_timers) {
-          timer->prev = g_timers->prev;
-          g_timers->prev->next = timer;
-          g_timers->prev = timer;
-     }
-     else {
-          g_timers = timer;
-          g_timers->prev = timer;
-     }
+     timer->next = g_timers;
+     g_timers    = timer;
 }
 
 
@@ -254,12 +247,41 @@ __glutHandleTimers( void )
                g_idle = GL_FALSE;
                
                cur->func( cur->value );
-               
-               cur->expire.tv_sec += (cur->expire.tv_usec/1000 + cur->interval) / 1000;
-               cur->expire.tv_usec = (cur->expire.tv_usec + cur->interval*1000) % 1000000;
+              
+               cur->expire.tv_usec += cur->interval.tv_usec;
+               cur->expire.tv_sec  += cur->interval.tv_sec + cur->expire.tv_usec/1000000;
+               cur->expire.tv_usec %= 1000000;
           }
      }
-}    
+}
+
+
+GLboolean
+__glutGetTimeout( int *ret_msec )
+{
+     __GlutTimer    *cur;
+     struct timeval *time = NULL;
+     struct timeval  now;
+     
+     for (cur = g_timers; cur; cur = cur->next) {
+          if (time == NULL ||
+              time->tv_sec > cur->expire.tv_sec ||
+             (time->tv_sec == cur->expire.tv_sec && 
+              time->tv_usec > cur->expire.tv_usec)) {
+               time = &cur->expire;
+          }
+     }
+     
+     if (time == NULL)
+          return GL_FALSE;
+          
+     gettimeofday( &now, NULL );
+     
+     *ret_msec = (time->tv_sec  - now.tv_sec) * 1000 +
+                 (time->tv_usec - now.tv_usec + 500) / 1000;
+                 
+     return GL_TRUE;
+}
      
           
 void 
@@ -275,4 +297,4 @@ __glutFreeTimers( void )
      
      g_timers = NULL;
 }
-    
+
diff --git a/src/glut/directfb/events.c b/src/glut/directfb/events.c
index 4c474710ab..6ebdd166ac 100644
--- a/src/glut/directfb/events.c
+++ b/src/glut/directfb/events.c
@@ -435,13 +435,14 @@ glutMainLoop( void )
 {
      __glutAssert( events != NULL );
      
+     __glutHandleWindows();
+     
      while (GL_TRUE) {
           DFBEvent evt, prev;
           
           g_idle = GL_TRUE;
           
           __glutHandleTimers();
-          __glutHandleWindows();
 
           prev.clazz = DFEC_NONE;
           
@@ -471,13 +472,19 @@ glutMainLoop( void )
                __glutHandleTimers();
           }
           
+          __glutHandleWindows();
+          
           if (g_idle) {
                if (idle_func) {
                     idle_func();
                }
                else {
+                    int msec;
                     __glutSetWindow( NULL );
-                    usleep( 500 );
+                    if (__glutGetTimeout( &msec ))
+                         events->WaitForEventWithTimeout( events, msec/1000, msec%1000 );
+                    else
+                         events->WaitForEvent( events );
                }
           }
      }
diff --git a/src/glut/directfb/internal.h b/src/glut/directfb/internal.h
index bc3e20e93e..47311c9342 100644
--- a/src/glut/directfb/internal.h
+++ b/src/glut/directfb/internal.h
@@ -124,6 +124,7 @@ extern void          __glutDestroyWindow( __GlutWindow *window );
 extern void          __glutDestroyWindows( void );
 /* callback.c */
 extern void          __glutHandleTimers( void );
+extern GLboolean     __glutGetTimeout( int *ret_msec );
 extern void          __glutFreeTimers( void );
 
 
diff --git a/src/glut/glx/Makefile.mgw b/src/glut/glx/Makefile.mgw
new file mode 100644
index 0000000000..ae4eb6addc
--- /dev/null
+++ b/src/glut/glx/Makefile.mgw
@@ -0,0 +1,198 @@
+# Mesa 3-D graphics library
+# Version:  5.1
+# 
+# Copyright (C) 1999-2003  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.
+
+# MinGW core makefile v1.4 for Mesa
+#
+#  Copyright (C) 2002 - Daniel Borca
+#  Email : dborca@users.sourceforge.net
+#  Web   : http://www.geocities.com/dborca
+
+# MinGW core-glut makefile updated for Mesa 7.0
+#
+#  Updated : by Heromyth, on 2007-7-21
+#  Email   : zxpmyth@yahoo.com.cn
+#  Bugs    : 1) All the default settings work fine. But the setting X86=1 can't work. 
+#            The others havn't been tested yet.
+#            2) The generated DLLs are *not* compatible with the ones built
+#            with the other compilers like VC8, especially for GLUT. 
+#            3) Although more tests are needed, it can be used individually!
+
+
+#
+#  Available options:
+#
+#    Environment variables:
+#	CFLAGS
+#
+#	GLIDE		path to Glide3 SDK; used with FX.
+#			default = $(TOP)/glide3
+#	FX=1		build for 3dfx Glide3. Note that this disables
+#			compilation of most WMesa code and requires fxMesa.
+#			As a consequence, you'll need the Win32 Glide3
+#			library to build any application.
+#			default = no
+#	ICD=1		build the installable client driver interface
+#			(windows opengl driver interface)
+#			default = no
+#	X86=1		optimize for x86 (if possible, use MMX, SSE, 3DNow).
+#			default = no
+#
+#    Targets:
+#	all:		build GL
+#	clean:		remove object files
+#
+
+
+
+.PHONY: all clean
+.INTERMEDIATE: x86/gen_matypes.exe
+.SUFFIXES: .rc .res
+
+# Set this to the prefix of your build tools, i.e. mingw32-
+TOOLS_PREFIX = mingw32-
+
+TOP = ../../..
+
+LIBDIR = $(TOP)/lib
+
+GLUT_DLL = glut32.dll
+GLUT_IMP = libglut32.a
+GLUT_DEF = glut.def
+
+include $(TOP)/configs/config.mgw
+GLUT_USING_STDCALL ?= 1
+
+
+
+LDLIBS = -L$(LIBDIR) -lwinmm -lgdi32 -luser32 -lopengl32 -lglu32 
+LDFLAGS = -Wl,--out-implib=$(LIBDIR)/$(GLUT_IMP) -Wl,--output-def=$(LIBDIR)/$(GLUT_DEF) 
+
+CFLAGS += -DBUILD_GLUT32 -DGLUT_BUILDING_LIB -DMESA -D_DLL
+
+ifeq ($(GL_USING_STDCALL),0)
+  CFLAGS += -DGL_NO_STDCALL
+endif
+
+ifeq ($(GLUT_USING_STDCALL),1)
+  CFLAGS += -D_STDCALL_SUPPORTED
+  LDFLAGS += -Wl,--add-stdcall-alias
+else
+  CFLAGS += -DGLUT_NO_STDCALL
+endif
+
+CFLAGS += -DNDEBUG -DLIBRARYBUILD -I$(TOP)/include 
+
+CC = gcc
+CXX = g++
+CXXFLAGS = $(CFLAGS)
+
+AR = ar
+ARFLAGS = crus
+
+UNLINK = del $(subst /,\,$(1))
+ifneq ($(wildcard $(addsuffix /rm.exe,$(subst ;, ,$(PATH)))),)
+UNLINK = $(RM) $(1)
+endif
+ifneq ($(wildcard $(addsuffix /rm,$(subst :, ,$(PATH)))),)
+UNLINK = $(RM) $(1)
+endif
+
+HDRS = glutint.h glutstroke.h glutbitmap.h glutwin32.h stroke.h win32_glx.h win32_x11.h
+
+SRCS = \
+	glut_bitmap.c \
+	glut_bwidth.c \
+	glut_cindex.c \
+	glut_cmap.c \
+	glut_cursor.c \
+	glut_dials.c \
+	glut_dstr.c \
+	glut_event.c \
+	glut_ext.c \
+	glut_fbc.c \
+	glut_fullscrn.c \
+	glut_gamemode.c \
+	glut_get.c \
+	glut_init.c \
+	glut_input.c \
+	glut_joy.c \
+	glut_key.c \
+	glut_keyctrl.c \
+	glut_keyup.c \
+	glut_mesa.c \
+	glut_modifier.c \
+	glut_overlay.c \
+	glut_shapes.c \
+	glut_space.c \
+	glut_stroke.c \
+	glut_swap.c \
+	glut_swidth.c \
+	glut_tablet.c \
+	glut_teapot.c \
+	glut_util.c \
+	glut_vidresize.c \
+	glut_warp.c \
+	glut_win.c \
+	glut_winmisc.c \
+	win32_glx.c \
+	win32_menu.c \
+	win32_util.c \
+	win32_winproc.c \
+	win32_x11.c
+	
+
+SRCSSEMIGENS = \
+	glut_8x13.c \
+	glut_9x15.c \
+	glut_hel10.c \
+	glut_hel12.c \
+	glut_hel18.c \
+	glut_mroman.c \
+	glut_roman.c \
+	glut_tr10.c \
+	glut_tr24.c
+
+
+
+SOURCES = $(SRCS) $(SRCSSEMIGENS)
+
+OBJECTS = $(addsuffix .o,$(basename $(SOURCES)))
+
+.c.o:
+	$(CC) -o $@ $(CFLAGS) -c $<
+.cc.o:
+	$(CXX) -o $@ $(CXXFLAGS) -c $<
+
+
+all: $(LIBDIR) $(LIBDIR)/$(GLUT_DLL) $(LIBDIR)/$(GLUT_IMP)
+
+$(LIBDIR):
+	mkdir -p $(LIBDIR)
+
+$(LIBDIR)/$(GLUT_DLL) $(LIBDIR)/$(GLUT_IMP): $(OBJECTS)
+	$(CXX) -shared -fPIC -o $(LIBDIR)/$(GLUT_DLL) $(LDFLAGS) \
+	$^ $(LDLIBS)
+
+
+
+clean:
+	-$(call UNLINK,*.o)
+\ No newline at end of file
diff --git a/src/glut/glx/glut_fbc.c b/src/glut/glx/glut_fbc.c
index deb46c3d8d..e93188b862 100644
--- a/src/glut/glx/glut_fbc.c
+++ b/src/glut/glx/glut_fbc.c
@@ -18,7 +18,7 @@
 
 /* Set a Fortran callback function. */
 
-void GLUTAPIENTRY
+void APIENTRY
 __glutSetFCB(int which, void *func)
 {
 #ifdef SUPPORT_FORTRAN
@@ -100,7 +100,7 @@ __glutSetFCB(int which, void *func)
 
 /* Get a Fortran callback function. */
 
-void* GLUTAPIENTRY
+void* APIENTRY
 __glutGetFCB(int which)
 {
 #ifdef SUPPORT_FORTRAN
diff --git a/src/glut/glx/glut_input.c b/src/glut/glx/glut_input.c
index add3df7c3f..a76ff9a435 100644
--- a/src/glut/glx/glut_input.c
+++ b/src/glut/glx/glut_input.c
@@ -23,6 +23,9 @@
 #endif
 #include <X11/Xutil.h>
 #else
+#ifdef __MINGW32__
+#include <GL/gl.h>
+#endif
 #include <windows.h>
 #ifndef __CYGWIN32__
 #include <mmsystem.h>  /* Win32 Multimedia API header. */
diff --git a/src/glut/glx/glut_joy.c b/src/glut/glx/glut_joy.c
index a4528ae1ce..5025607922 100644
--- a/src/glut/glx/glut_joy.c
+++ b/src/glut/glx/glut_joy.c
@@ -6,6 +6,9 @@
    implied. This program is -not- in the public domain. */
 
 #ifdef _WIN32
+#ifdef __MINGW32__
+#include <GL/gl.h>
+#endif
 #include <windows.h>
 #ifndef __CYGWIN32__
 #include <mmsystem.h>  /* Win32 Multimedia API header. */
diff --git a/src/glut/glx/glutint.h b/src/glut/glx/glutint.h
index 6fe09ffe7e..a962c78023 100644
--- a/src/glut/glx/glutint.h
+++ b/src/glut/glx/glutint.h
@@ -26,7 +26,10 @@
 #include <GL/glx.h>
 #endif
 
+#ifndef GLUT_BUILDING_LIB
 #define GLUT_BUILDING_LIB  /* Building the GLUT library itself. */
+#endif
+
 #include <GL/glut.h>
 
 #if defined(MESA) && defined(_WIN32) && !defined(__CYGWIN32__)
diff --git a/src/glut/glx/win32_util.c b/src/glut/glx/win32_util.c
index becd823a40..25af48a112 100644
--- a/src/glut/glx/win32_util.c
+++ b/src/glut/glx/win32_util.c
@@ -15,6 +15,7 @@
 
 /* The following added by Paul Garceau <pgarceau@teleport.com> */
 #if defined(__MINGW32__)
+#include <GL/gl.h>
 #include <time.h>
 #include <windows.h>
 struct timeval;
diff --git a/src/glut/glx/win32_winproc.c b/src/glut/glx/win32_winproc.c
index a54bac75fa..e1fc785ebb 100644
--- a/src/glut/glx/win32_winproc.c
+++ b/src/glut/glx/win32_winproc.c
@@ -9,6 +9,9 @@
 
 #include "glutint.h"
 #include <sys/timeb.h>
+#ifdef __MINGW32__
+#include <ctype.h>
+#endif
 
 #if defined(_WIN32) && !defined(__CYGWIN32__)
 #include <mmsystem.h>  /* Win32 Multimedia API header. */
diff --git a/src/glut/glx/win32_x11.h b/src/glut/glx/win32_x11.h
index 1d8d048b2e..6f5c3a9aea 100644
--- a/src/glut/glx/win32_x11.h
+++ b/src/glut/glx/win32_x11.h
@@ -6,16 +6,14 @@
 /* This program is freely distributable without licensing fees 
    and is provided without guarantee or warrantee expressed or 
    implied. This program is -not- in the public domain. */
-
+#ifdef __MINGW32__
+#include <GL/gl.h>
+#endif
 #include <stdlib.h>
 #include <windows.h>
 
 /* These definitions are missing from windows.h */
-WINGDIAPI int WINAPI wglChoosePixelFormat(HDC, PIXELFORMATDESCRIPTOR *);
-WINGDIAPI int WINAPI wglDescribePixelFormat(HDC, int, UINT, LPPIXELFORMATDESCRIPTOR);
-WINGDIAPI int WINAPI wglGetPixelFormat(HDC);
-WINGDIAPI BOOL WINAPI wglSetPixelFormat(HDC, int, PIXELFORMATDESCRIPTOR *);
-WINGDIAPI BOOL WINAPI wglSwapBuffers(HDC);
+
 
 /* Type definitions (conversions) */
 typedef int Visual;			/* Win32 equivalent of X11 type */
diff --git a/src/mesa/Makefile b/src/mesa/Makefile
index c03c5f80ff..6943219036 100644
--- a/src/mesa/Makefile
+++ b/src/mesa/Makefile
@@ -176,9 +176,9 @@ clean:
 	-rm -f */*/*.o
 	-rm -f depend depend.bak libmesa.a
 	-rm -f drivers/*/*.o
-	(cd drivers/dri ; $(MAKE) clean)
-	(cd x86 ; $(MAKE) clean)
-	(cd x86-64 ; $(MAKE) clean)
+	(cd drivers/dri && $(MAKE) clean)
+	(cd x86 && $(MAKE) clean)
+	(cd x86-64 && $(MAKE) clean)
 
 
 include depend
diff --git a/src/mesa/Makefile.mgw b/src/mesa/Makefile.mgw
index ebec5c055a..3b52834bd1 100644
--- a/src/mesa/Makefile.mgw
+++ b/src/mesa/Makefile.mgw
@@ -1,5 +1,5 @@
 # Mesa 3-D graphics library
-# Version:  5.1
+# Version:  7.0
 # 
 # Copyright (C) 1999-2003  Brian Paul   All Rights Reserved.
 # 
@@ -26,6 +26,16 @@
 #  Email : dborca@users.sourceforge.net
 #  Web   : http://www.geocities.com/dborca
 
+# MinGW core-gl makefile updated for Mesa 7.0
+#
+#  updated : by Heromyth, on 2007-7-21
+#  Email   : zxpmyth@yahoo.com.cn
+#  Bugs    : 1) All the default settings work fine. But the setting X86=1 can't work. 
+#            The others havn't been tested yet.
+#            2) The generated DLLs are *not* compatible with the ones built
+#            with the other compilers like VC8, especially for GLUT. 
+#            3) Although more tests are needed, it can be used individually!
+
 
 #
 #  Available options:
@@ -52,7 +62,6 @@
 #
 
 
-
 .PHONY: all clean
 .INTERMEDIATE: x86/gen_matypes.exe
 .SUFFIXES: .rc .res
@@ -60,6 +69,8 @@
 # Set this to the prefix of your build tools, i.e. mingw32-
 TOOLS_PREFIX = mingw32-
 
+
+
 TOP = ../..
 GLIDE ?= $(TOP)/glide3
 LIBDIR = $(TOP)/lib
@@ -71,11 +82,25 @@ else
   GL_IMP = libopengl32.a
 endif
 
-LDLIBS = -lgdi32
+GL_DEF = gl.def
+
+include $(TOP)/configs/config.mgw
+GL_USING_STDCALL ?= 1
+
+MESA_LIB = libmesa.a
+
+LDLIBS = -lgdi32 -luser32 -liberty
+LDFLAGS = -Wl,--out-implib=$(LIBDIR)/$(GL_IMP) -Wl,--output-def=$(LIBDIR)/gl.def
 
 CC = $(TOOLS_PREFIX)gcc
-CFLAGS += -DBUILD_GL32 -D_OPENGL32_
-CFLAGS += $(INCLUDE_DIRS)
+CFLAGS += -DBUILD_GL32 -D_OPENGL32_ -D_DLL -DMESA_MINWARN -DNDEBUG -D_USRDLL -DGDI_EXPORTS
+
+ifeq ($(GL_USING_STDCALL),1)
+  LDFLAGS += -Wl,--add-stdcall-alias
+else
+  CFLAGS += -DGL_NO_STDCALL
+endif
+
 CFLAGS += -DUSE_EXTERNAL_DXTN_LIB=1
 ifeq ($(FX),1)
   CFLAGS += -I$(GLIDE)/include -DFX
@@ -104,6 +129,8 @@ endif
 
 include sources
 
+CFLAGS += $(INCLUDE_DIRS) 
+
 ifeq ($(X86),1)
 CFLAGS += -DUSE_X86_ASM
 CFLAGS += -DUSE_MMX_ASM
@@ -140,10 +167,9 @@ RESOURCE = $(GL_RES:.rc=.res)
 
 .c.o:
 	$(CC) -o $@ $(CFLAGS) -c $<
-.S.o:
-	$(CC) -o $@ $(CFLAGS) -c $<
 .s.o:
 	$(CC) -o $@ $(CFLAGS) -x assembler-with-cpp -c $<
+
 .rc.res:
 	windres -o $@ -Irc -Ocoff $<
 
@@ -153,9 +179,8 @@ $(LIBDIR):
 	mkdir -p $(LIBDIR)
 
 $(LIBDIR)/$(GL_DLL) $(LIBDIR)/$(GL_IMP): $(OBJECTS) $(RESOURCE)
-	$(TOOLS_PREFIX)dllwrap -o $(LIBDIR)/$(GL_DLL) --output-lib $(LIBDIR)/$(GL_IMP) \
-	                       --target i386-mingw32 --def $(GL_DEF) -Wl,-enable-stdcall-fixup \
-	                       $^ $(LDLIBS)
+	$(CC) -shared -fPIC -o $(LIBDIR)/$(GL_DLL) $(LDFLAGS) \
+	$^ $(LDLIBS)
 
 $(X86_OBJECTS): x86/matypes.h
 
@@ -187,17 +212,21 @@ tnl/t_vtx_x86_gcc.o: tnl/t_vtx_x86_gcc.S
 	$(CC) -o $@ $(CFLAGS) -DSTDCALL_API -c $<
 
 clean:
-	-$(call UNLINK,array_cache/*.o)
 	-$(call UNLINK,glapi/*.o)
 	-$(call UNLINK,main/*.o)
 	-$(call UNLINK,math/*.o)
+	-$(call UNLINK,vbo/*.o)
 	-$(call UNLINK,shader/*.o)
+	-$(call UNLINK,shader/slang/*.o)
+	-$(call UNLINK,shader/grammar/*.o)
 	-$(call UNLINK,sparc/*.o)
 	-$(call UNLINK,ppc/*.o)
 	-$(call UNLINK,swrast/*.o)
 	-$(call UNLINK,swrast_setup/*.o)
 	-$(call UNLINK,tnl/*.o)
 	-$(call UNLINK,x86/*.o)
+	-$(call UNLINK,x86/rtasm/*.o)
+	-$(call UNLINK,x86-64/*.o)
 	-$(call UNLINK,drivers/common/*.o)
 	-$(call UNLINK,drivers/glide/*.o)
 	-$(call UNLINK,drivers/windows/fx/*.o)
diff --git a/src/mesa/drivers/directfb/idirectfbgl_mesa.c b/src/mesa/drivers/directfb/idirectfbgl_mesa.c
index 88ac4bb31d..694eeb054d 100644
--- a/src/mesa/drivers/directfb/idirectfbgl_mesa.c
+++ b/src/mesa/drivers/directfb/idirectfbgl_mesa.c
@@ -134,7 +134,6 @@ static bool  directfbgl_init_visual    ( GLvisual              *visual,
 static bool  directfbgl_create_context ( GLcontext             *context,
                                          GLframebuffer         *framebuffer,
                                          GLvisual              *visual,
-                                         DFBSurfacePixelFormat  format,
                                          IDirectFBGL_data      *data );
 static void  directfbgl_destroy_context( GLcontext             *context,
                                          GLframebuffer         *framebuffer );
@@ -340,8 +339,9 @@ Construct( IDirectFBGL *thiz, IDirectFBSurface *surface )
      }
      
      /* Create context. */
-     if (!directfbgl_create_context( &data->context, &data->framebuffer,
-                                     &data->visual, data->format, data )) {
+     if (!directfbgl_create_context( &data->context,
+                                     &data->framebuffer,
+                                     &data->visual, data )) {
           D_ERROR( "DirectFBGL/Mesa: failed to create context.\n" );
           IDirectFBGL_Mesa_Destruct( thiz );
           return DFB_UNSUPPORTED;
@@ -762,11 +762,10 @@ directfbgl_init_visual( GLvisual              *visual,
 }
 
 static bool
-directfbgl_create_context( GLcontext             *context,
-                           GLframebuffer         *framebuffer,
-                           GLvisual              *visual,
-                           DFBSurfacePixelFormat  format,
-                           IDirectFBGL_data      *data )
+directfbgl_create_context( GLcontext        *context,
+                           GLframebuffer    *framebuffer,
+                           GLvisual         *visual,
+                           IDirectFBGL_data *data )
 {
      struct dd_function_table functions;
      
@@ -800,7 +799,7 @@ directfbgl_create_context( GLcontext             *context,
      data->render.Delete         = dfbDeleteRenderbuffer;
      data->render.AllocStorage   = dfbRenderbufferStorage;
      
-     switch (format) {
+     switch (data->format) {
           case DSPF_RGB332:
                data->render.GetRow        = get_row_RGB332;
                data->render.GetValues     = get_values_RGB332;
@@ -887,6 +886,9 @@ directfbgl_create_context( GLcontext             *context,
                return false;
      }
 
+     data->render.Width = data->width;
+     data->render.Height = data->height;
+
      _mesa_add_renderbuffer( framebuffer, BUFFER_FRONT_LEFT, &data->render );
      
      _mesa_add_soft_renderbuffers( framebuffer,
@@ -908,8 +910,11 @@ static void
 directfbgl_destroy_context( GLcontext     *context,
                             GLframebuffer *framebuffer )
 {
-     _mesa_free_framebuffer_data( framebuffer );
-     _mesa_notifyDestroy( context );
+     _swsetup_DestroyContext( context );
+     _swrast_DestroyContext( context );
+     _tnl_DestroyContext( context );
+     _vbo_DestroyContext( context );
+     //_mesa_free_framebuffer_data( framebuffer );
      _mesa_free_context_data( context );
 }    
  
diff --git a/src/mesa/drivers/dri/common/dri_util.c b/src/mesa/drivers/dri/common/dri_util.c
index dd52f7e915..c30e66f172 100644
--- a/src/mesa/drivers/dri/common/dri_util.c
+++ b/src/mesa/drivers/dri/common/dri_util.c
@@ -428,10 +428,12 @@ __driUtilUpdateDrawableInfo(__DRIdrawablePrivate *pdp)
 
     if (pdp->pClipRects) {
 	_mesa_free(pdp->pClipRects); 
+	pdp->pClipRects = NULL;
     }
 
     if (pdp->pBackClipRects) {
 	_mesa_free(pdp->pBackClipRects); 
+	pdp->pBackClipRects = NULL;
     }
 
     DRM_SPINUNLOCK(&psp->pSAREA->drawable_lock, psp->drawLockID);
diff --git a/src/mesa/drivers/dri/common/extension_helper.h b/src/mesa/drivers/dri/common/extension_helper.h
index 10f75edaaa..bf103a3931 100644
--- a/src/mesa/drivers/dri/common/extension_helper.h
+++ b/src/mesa/drivers/dri/common/extension_helper.h
@@ -1478,9 +1478,10 @@ static const char ImageTransformParameterfvHP_names[] =
     "";
 #endif
 
-#if defined(need_GL_ARB_vertex_program)
+#if defined(need_GL_VERSION_2_0) || defined(need_GL_ARB_vertex_program)
 static const char VertexAttrib4ivARB_names[] = 
     "ip\0" /* Parameter signature */
+    "glVertexAttrib4iv\0"
     "glVertexAttrib4ivARB\0"
     "";
 #endif
@@ -1586,9 +1587,10 @@ static const char PixelTransformParameterfvEXT_names[] =
     "";
 #endif
 
-#if defined(need_GL_ARB_vertex_program)
+#if defined(need_GL_VERSION_2_0) || defined(need_GL_ARB_vertex_program)
 static const char VertexAttrib4bvARB_names[] = 
     "ip\0" /* Parameter signature */
+    "glVertexAttrib4bv\0"
     "glVertexAttrib4bvARB\0"
     "";
 #endif
@@ -2391,9 +2393,10 @@ static const char GetAttribLocationARB_names[] =
     "";
 #endif
 
-#if defined(need_GL_ARB_vertex_program)
+#if defined(need_GL_VERSION_2_0) || defined(need_GL_ARB_vertex_program)
 static const char VertexAttrib4ubvARB_names[] = 
     "ip\0" /* Parameter signature */
+    "glVertexAttrib4ubv\0"
     "glVertexAttrib4ubvARB\0"
     "";
 #endif
@@ -2910,9 +2913,10 @@ static const char ReplacementCodeuiColor4ubVertex3fSUN_names[] =
     "";
 #endif
 
-#if defined(need_GL_ARB_vertex_program)
+#if defined(need_GL_VERSION_2_0) || defined(need_GL_ARB_vertex_program)
 static const char VertexAttrib4usvARB_names[] = 
     "ip\0" /* Parameter signature */
+    "glVertexAttrib4usv\0"
     "glVertexAttrib4usvARB\0"
     "";
 #endif
@@ -4393,9 +4397,10 @@ static const char WindowPos4iMESA_names[] =
     "";
 #endif
 
-#if defined(need_GL_ARB_vertex_program)
+#if defined(need_GL_VERSION_2_0) || defined(need_GL_ARB_vertex_program)
 static const char VertexAttrib4uivARB_names[] = 
     "ip\0" /* Parameter signature */
+    "glVertexAttrib4uiv\0"
     "glVertexAttrib4uivARB\0"
     "";
 #endif
@@ -6257,6 +6262,8 @@ static const struct dri_extension_function GL_VERSION_2_0_functions[] = {
     { GetVertexAttribivARB_names, GetVertexAttribivARB_remap_index, -1 },
     { CreateProgram_names, CreateProgram_remap_index, -1 },
     { StencilFuncSeparate_names, StencilFuncSeparate_remap_index, -1 },
+    { VertexAttrib4ivARB_names, VertexAttrib4ivARB_remap_index, -1 },
+    { VertexAttrib4bvARB_names, VertexAttrib4bvARB_remap_index, -1 },
     { VertexAttrib3dARB_names, VertexAttrib3dARB_remap_index, -1 },
     { VertexAttrib4fARB_names, VertexAttrib4fARB_remap_index, -1 },
     { VertexAttrib4fvARB_names, VertexAttrib4fvARB_remap_index, -1 },
@@ -6270,6 +6277,7 @@ static const struct dri_extension_function GL_VERSION_2_0_functions[] = {
     { VertexAttrib1dvARB_names, VertexAttrib1dvARB_remap_index, -1 },
     { GetVertexAttribfvARB_names, GetVertexAttribfvARB_remap_index, -1 },
     { GetAttribLocationARB_names, GetAttribLocationARB_remap_index, -1 },
+    { VertexAttrib4ubvARB_names, VertexAttrib4ubvARB_remap_index, -1 },
     { Uniform3ivARB_names, Uniform3ivARB_remap_index, -1 },
     { VertexAttrib4sARB_names, VertexAttrib4sARB_remap_index, -1 },
     { VertexAttrib2dvARB_names, VertexAttrib2dvARB_remap_index, -1 },
@@ -6282,6 +6290,7 @@ static const struct dri_extension_function GL_VERSION_2_0_functions[] = {
     { VertexAttrib4NuivARB_names, VertexAttrib4NuivARB_remap_index, -1 },
     { Uniform4fARB_names, Uniform4fARB_remap_index, -1 },
     { VertexAttrib1dARB_names, VertexAttrib1dARB_remap_index, -1 },
+    { VertexAttrib4usvARB_names, VertexAttrib4usvARB_remap_index, -1 },
     { LinkProgramARB_names, LinkProgramARB_remap_index, -1 },
     { ShaderSourceARB_names, ShaderSourceARB_remap_index, -1 },
     { VertexAttrib3svARB_names, VertexAttrib3svARB_remap_index, -1 },
@@ -6311,6 +6320,7 @@ static const struct dri_extension_function GL_VERSION_2_0_functions[] = {
     { DrawBuffersARB_names, DrawBuffersARB_remap_index, -1 },
     { Uniform1fvARB_names, Uniform1fvARB_remap_index, -1 },
     { EnableVertexAttribArrayARB_names, EnableVertexAttribArrayARB_remap_index, -1 },
+    { VertexAttrib4uivARB_names, VertexAttrib4uivARB_remap_index, -1 },
     { VertexAttrib4svARB_names, VertexAttrib4svARB_remap_index, -1 },
     { GetShaderiv_names, GetShaderiv_remap_index, -1 },
     { VertexAttrib2svARB_names, VertexAttrib2svARB_remap_index, -1 },
diff --git a/src/mesa/drivers/dri/i810/i810state.c b/src/mesa/drivers/dri/i810/i810state.c
index 3ad25282d9..e0d5b2b487 100644
--- a/src/mesa/drivers/dri/i810/i810state.c
+++ b/src/mesa/drivers/dri/i810/i810state.c
@@ -380,7 +380,10 @@ static void i810CullFaceFrontFace(GLcontext *ctx, GLenum unused)
 static void i810LineWidth( GLcontext *ctx, GLfloat widthf )
 {
    i810ContextPtr imesa = I810_CONTEXT( ctx );
-   int width = (int)ctx->Line._Width;
+   /* AA, non-AA limits are same */
+   const int width = (int) CLAMP(ctx->Line.Width,
+                                 ctx->Const.MinLineWidth,
+                                 ctx->Const.MaxLineWidth);
 
    imesa->LcsLineWidth = 0;
    if (width & 1) imesa->LcsLineWidth |= LCS_LINEWIDTH_1_0;
@@ -396,7 +399,10 @@ static void i810LineWidth( GLcontext *ctx, GLfloat widthf )
 static void i810PointSize( GLcontext *ctx, GLfloat sz )
 {
    i810ContextPtr imesa = I810_CONTEXT( ctx );
-   int size = (int)ctx->Point._Size;
+   /* AA, non-AA limits are same */
+   const int size = (int) CLAMP(ctx->Point.Size,
+                                ctx->Const.MinPointSize,
+                                ctx->Const.MaxPointSize);
 
    imesa->LcsPointSize = 0;
    if (size & 1) imesa->LcsPointSize |= LCS_LINEWIDTH_1_0;
diff --git a/src/mesa/drivers/dri/i810/i810tris.c b/src/mesa/drivers/dri/i810/i810tris.c
index 3e09427bb9..2c4ee06633 100644
--- a/src/mesa/drivers/dri/i810/i810tris.c
+++ b/src/mesa/drivers/dri/i810/i810tris.c
@@ -112,7 +112,9 @@ static __inline__ void i810_draw_quad( i810ContextPtr imesa,
 static __inline__ void i810_draw_point( i810ContextPtr imesa,
 					i810VertexPtr tmp )
 {
-   GLfloat sz = imesa->glCtx->Point._Size * .5;
+   GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size,
+                            imesa->glCtx->Const.MinPointSize,
+                            imesa->glCtx->Const.MaxPointSize);
    int vertsize = imesa->vertex_size;
    GLuint *vb = i810AllocDmaLow( imesa, 2 * 4 * vertsize );
    int j;
diff --git a/src/mesa/drivers/dri/i915/i915_vtbl.c b/src/mesa/drivers/dri/i915/i915_vtbl.c
index b0e5f87fc7..cc8a605e50 100644
--- a/src/mesa/drivers/dri/i915/i915_vtbl.c
+++ b/src/mesa/drivers/dri/i915/i915_vtbl.c
@@ -64,6 +64,7 @@ static void i915_reduced_primitive_state( intelContextPtr intel,
     st1 &= ~ST1_ENABLE;
 
     switch (rprim) {
+    case GL_QUADS: /* from RASTERIZE(GL_QUADS) in t_dd_tritemp.h */
     case GL_TRIANGLES:
        if (intel->ctx.Polygon.StippleFlag &&
 	   intel->hw_stipple)
diff --git a/src/mesa/drivers/dri/i915/intel_pixel.c b/src/mesa/drivers/dri/i915/intel_pixel.c
index 535cbfcb26..c3030d42b0 100644
--- a/src/mesa/drivers/dri/i915/intel_pixel.c
+++ b/src/mesa/drivers/dri/i915/intel_pixel.c
@@ -439,10 +439,32 @@ intelDrawPixels( GLcontext *ctx,
    if (INTEL_DEBUG & DEBUG_PIXEL)
       fprintf(stderr, "%s\n", __FUNCTION__);
 
-   if (!intelTryDrawPixels( ctx, x, y, width, height, format, type,
-                            unpack, pixels ))
+   if (intelTryDrawPixels( ctx, x, y, width, height, format, type,
+                           unpack, pixels ))
+      return;
+
+   if (ctx->FragmentProgram._Current == ctx->FragmentProgram._TexEnvProgram) {
+      /*
+       * We don't want the i915 texenv program to be applied to DrawPixels.
+       * This is really just a performance optimization (mesa will other-
+       * wise happily run the fragment program on each pixel in the image).
+       */
+      struct gl_fragment_program *fpSave = ctx->FragmentProgram._Current;
+   /* can't just set current frag prog to 0 here as on buffer resize
+      we'll get new state checks which will segfault. Remains a hack. */
+      ctx->FragmentProgram._Current = NULL;
+      ctx->FragmentProgram._UseTexEnvProgram = GL_FALSE;
+      ctx->FragmentProgram._Active = GL_FALSE;
+      _swrast_DrawPixels( ctx, x, y, width, height, format, type,
+                          unpack, pixels );
+      ctx->FragmentProgram._Current = fpSave;
+      ctx->FragmentProgram._UseTexEnvProgram = GL_TRUE;
+      ctx->FragmentProgram._Active = GL_TRUE;
+   }
+   else {
       _swrast_DrawPixels( ctx, x, y, width, height, format, type,
-			  unpack, pixels );
+                          unpack, pixels );
+   }
 }
 
 
diff --git a/src/mesa/drivers/dri/i915tex/i830_vtbl.c b/src/mesa/drivers/dri/i915tex/i830_vtbl.c
index dd0670dec3..e432648ada 100644
--- a/src/mesa/drivers/dri/i915tex/i830_vtbl.c
+++ b/src/mesa/drivers/dri/i915tex/i830_vtbl.c
@@ -490,11 +490,13 @@ i830_emit_state(struct intel_context *intel)
                       DRM_BO_MASK_MEM | DRM_BO_FLAG_READ,
                       state->tex_offset[i] | TM0S0_USE_FENCE);
          }
-         else {
-            assert(i == 0);
-            assert(state == &i830->meta);
-            OUT_BATCH(0);
-         }
+	 else if (state == &i830->meta) {
+	    assert(i == 0);
+	    OUT_BATCH(0);
+	 }
+	 else {
+	    OUT_BATCH(state->tex_offset[i]);
+	 }
 
          OUT_BATCH(state->Tex[i][I830_TEXREG_TM0S1]);
          OUT_BATCH(state->Tex[i][I830_TEXREG_TM0S2]);
diff --git a/src/mesa/drivers/dri/i915tex/i915_vtbl.c b/src/mesa/drivers/dri/i915tex/i915_vtbl.c
index f80e8d6327..ad333b490b 100644
--- a/src/mesa/drivers/dri/i915tex/i915_vtbl.c
+++ b/src/mesa/drivers/dri/i915tex/i915_vtbl.c
@@ -61,6 +61,7 @@ i915_reduced_primitive_state(struct intel_context *intel, GLenum rprim)
    st1 &= ~ST1_ENABLE;
 
    switch (rprim) {
+   case GL_QUADS: /* from RASTERIZE(GL_QUADS) in t_dd_tritemp.h */
    case GL_TRIANGLES:
       if (intel->ctx.Polygon.StippleFlag && intel->hw_stipple)
          st1 |= ST1_ENABLE;
diff --git a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
index 59261f7274..212f130101 100644
--- a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
+++ b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
@@ -9,7 +9,7 @@ struct intel_context;
 #define BATCH_SZ 16384
 #define BATCH_RESERVED 16
 
-#define MAX_RELOCS 400
+#define MAX_RELOCS 4096
 
 #define INTEL_BATCH_NO_CLIPRECTS 0x1
 #define INTEL_BATCH_CLIPRECTS    0x2
diff --git a/src/mesa/drivers/dri/i915tex/intel_buffers.c b/src/mesa/drivers/dri/i915tex/intel_buffers.c
index 62ff54b007..15d02f8e2c 100644
--- a/src/mesa/drivers/dri/i915tex/intel_buffers.c
+++ b/src/mesa/drivers/dri/i915tex/intel_buffers.c
@@ -316,7 +316,8 @@ intelWindowMoved(struct intel_context *intel)
 	 flags = intel_fb->vblank_flags & ~VBLANK_FLAG_SECONDARY;
       }
 
-      if (flags != intel_fb->vblank_flags) {
+      if (flags != intel_fb->vblank_flags && intel_fb->vblank_flags &&
+	  !(intel_fb->vblank_flags & VBLANK_FLAG_NO_IRQ)) {
 	 drmVBlank vbl;
 	 int i;
 
@@ -327,7 +328,9 @@ intelWindowMoved(struct intel_context *intel)
 	 }
 
 	 for (i = 0; i < intel_fb->pf_num_pages; i++) {
-	    if (!intel_fb->color_rb[i])
+	    if (!intel_fb->color_rb[i] ||
+		(intel_fb->vbl_waited - intel_fb->color_rb[i]->vbl_pending) <=
+		(1<<23))
 	       continue;
 
 	    vbl.request.sequence = intel_fb->color_rb[i]->vbl_pending;
@@ -828,7 +831,8 @@ intelScheduleSwap(const __DRIdrawablePrivate * dPriv, GLboolean *missed_target)
    drm_i915_vblank_swap_t swap;
    GLboolean ret;
 
-   if ((intel_fb->vblank_flags & VBLANK_FLAG_NO_IRQ) ||
+   if (!intel_fb->vblank_flags ||
+       (intel_fb->vblank_flags & VBLANK_FLAG_NO_IRQ) ||
        intelScreen->current_rotation != 0 ||
        intelScreen->drmMinor < (intel_fb->pf_active ? 9 : 6))
       return GL_FALSE;
diff --git a/src/mesa/drivers/dri/i915tex/intel_context.c b/src/mesa/drivers/dri/i915tex/intel_context.c
index c927dca8e5..40ea756412 100644
--- a/src/mesa/drivers/dri/i915tex/intel_context.c
+++ b/src/mesa/drivers/dri/i915tex/intel_context.c
@@ -619,12 +619,23 @@ intelMakeCurrent(__DRIcontextPrivate * driContextPriv,
       if (intel->ctx.DrawBuffer == &intel_fb->Base) {
 
 	 if (intel->driDrawable != driDrawPriv) {
-	    intel_fb->vblank_flags = (intel->intelScreen->irq_active != 0)
-	       ? driGetDefaultVBlankFlags(&intel->optionCache)
-	       : VBLANK_FLAG_NO_IRQ;
-	    (*dri_interface->getUST) (&intel_fb->swap_ust);
-	    driDrawableInitVBlank(driDrawPriv, intel_fb->vblank_flags,
-				  &intel_fb->vbl_seq);
+	    if (driDrawPriv->pdraw->swap_interval == (unsigned)-1) {
+	       int i;
+
+	       intel_fb->vblank_flags = (intel->intelScreen->irq_active != 0)
+		  ? driGetDefaultVBlankFlags(&intel->optionCache)
+		 : VBLANK_FLAG_NO_IRQ;
+
+	       (*dri_interface->getUST) (&intel_fb->swap_ust);
+	       driDrawableInitVBlank(driDrawPriv, intel_fb->vblank_flags,
+				     &intel_fb->vbl_seq);
+	       intel_fb->vbl_waited = intel_fb->vbl_seq;
+
+	       for (i = 0; i < (intel->intelScreen->third.handle ? 3 : 2); i++) {
+		  if (intel_fb->color_rb[i])
+		     intel_fb->color_rb[i]->vbl_pending = intel_fb->vbl_seq;
+	       }
+	    }
 	    intel->driDrawable = driDrawPriv;
 	    intelWindowMoved(intel);
 	 }
@@ -670,37 +681,27 @@ intelContendedLock(struct intel_context *intel, GLuint flags)
    if (sarea->width != intel->width ||
        sarea->height != intel->height ||
        sarea->rotation != intel->current_rotation) {
-      
-      void *batchMap = intel->batch->map;
-      
+      int numClipRects = intel->numClipRects;
+
       /*
        * FIXME: Really only need to do this when drawing to a
        * common back- or front buffer.
        */
 
       /*
-       * This will drop the outstanding batchbuffer on the floor
+       * This will essentially drop the outstanding batchbuffer on the floor.
        */
+      intel->numClipRects = 0;
 
-      if (batchMap != NULL) {
-	 driBOUnmap(intel->batch->buffer);
-	 intel->batch->map = NULL;
-      }
-
-      intel_batchbuffer_reset(intel->batch);
+      if (intel->Fallback)
+	 _swrast_flush(&intel->ctx);
 
-      if (batchMap == NULL) {
-	 driBOUnmap(intel->batch->buffer);
-	 intel->batch->map = NULL;
-      }
+      INTEL_FIREVERTICES(intel);
 
-      /* lose all primitives */
-      intel->prim.primitive = ~0;
-      intel->prim.start_ptr = 0;
-      intel->prim.flush = 0;
+      if (intel->batch->map != intel->batch->ptr)
+	 intel_batchbuffer_flush(intel->batch);
 
-      /* re-emit all state */
-      intel->vtbl.lost_hardware(intel);
+      intel->numClipRects = numClipRects;
 
       /* force window update */
       intel->lastStamp = 0;
@@ -741,7 +742,9 @@ void LOCK_HARDWARE( struct intel_context *intel )
 				    BUFFER_BACK_LEFT);
     }
 
-    if (intel_rb && (intel_fb->vbl_waited - intel_rb->vbl_pending) > (1<<23)) {
+    if (intel_rb && intel_fb->vblank_flags &&
+	!(intel_fb->vblank_flags & VBLANK_FLAG_NO_IRQ) &&
+	(intel_fb->vbl_waited - intel_rb->vbl_pending) > (1<<23)) {
 	drmVBlank vbl;
 
 	vbl.request.type = DRM_VBLANK_ABSOLUTE;
diff --git a/src/mesa/drivers/dri/i915tex/intel_fbo.c b/src/mesa/drivers/dri/i915tex/intel_fbo.c
index 349912ffec..6f99f401c7 100644
--- a/src/mesa/drivers/dri/i915tex/intel_fbo.c
+++ b/src/mesa/drivers/dri/i915tex/intel_fbo.c
@@ -488,7 +488,7 @@ intel_new_renderbuffer(GLcontext * ctx, GLuint name)
  */
 static void
 intel_bind_framebuffer(GLcontext * ctx, GLenum target,
-                       struct gl_framebuffer *fb)
+                       struct gl_framebuffer *fb, struct gl_framebuffer *fbread)
 {
    if (target == GL_FRAMEBUFFER_EXT || target == GL_DRAW_FRAMEBUFFER_EXT) {
       intel_draw_buffer(ctx, fb);
diff --git a/src/mesa/drivers/dri/i915tex/intel_pixel_draw.c b/src/mesa/drivers/dri/i915tex/intel_pixel_draw.c
index 10a079896a..e4e57cb3a7 100644
--- a/src/mesa/drivers/dri/i915tex/intel_pixel_draw.c
+++ b/src/mesa/drivers/dri/i915tex/intel_pixel_draw.c
@@ -363,5 +363,26 @@ intelDrawPixels(GLcontext * ctx,
    if (INTEL_DEBUG & DEBUG_PIXEL)
       _mesa_printf("%s: fallback to swrast\n", __FUNCTION__);
 
-   _swrast_DrawPixels(ctx, x, y, width, height, format, type, unpack, pixels);
+   if (ctx->FragmentProgram._Current == ctx->FragmentProgram._TexEnvProgram) {
+      /*
+       * We don't want the i915 texenv program to be applied to DrawPixels.
+       * This is really just a performance optimization (mesa will other-
+       * wise happily run the fragment program on each pixel in the image).
+       */
+      struct gl_fragment_program *fpSave = ctx->FragmentProgram._Current;
+   /* can't just set current frag prog to 0 here as on buffer resize
+      we'll get new state checks which will segfault. Remains a hack. */
+      ctx->FragmentProgram._Current = NULL;
+      ctx->FragmentProgram._UseTexEnvProgram = GL_FALSE;
+      ctx->FragmentProgram._Active = GL_FALSE;
+      _swrast_DrawPixels( ctx, x, y, width, height, format, type,
+                          unpack, pixels );
+      ctx->FragmentProgram._Current = fpSave;
+      ctx->FragmentProgram._UseTexEnvProgram = GL_TRUE;
+      ctx->FragmentProgram._Active = GL_TRUE;
+   }
+   else {
+      _swrast_DrawPixels( ctx, x, y, width, height, format, type,
+                          unpack, pixels );
+   }
 }
diff --git a/src/mesa/drivers/dri/i915tex/intel_tex_validate.c b/src/mesa/drivers/dri/i915tex/intel_tex_validate.c
index 0ae4fee1ba..af18c26d55 100644
--- a/src/mesa/drivers/dri/i915tex/intel_tex_validate.c
+++ b/src/mesa/drivers/dri/i915tex/intel_tex_validate.c
@@ -116,7 +116,7 @@ intel_finalize_mipmap_tree(struct intel_context *intel, GLuint unit)
 
    /* We know/require this is true by now: 
     */
-   assert(intelObj->base.Complete);
+   assert(intelObj->base._Complete);
 
    /* What levels must the tree include at a minimum?
     */
diff --git a/src/mesa/drivers/dri/i965/brw_sf_state.c b/src/mesa/drivers/dri/i965/brw_sf_state.c
index bfac52d765..9a6e5f5f19 100644
--- a/src/mesa/drivers/dri/i965/brw_sf_state.c
+++ b/src/mesa/drivers/dri/i965/brw_sf_state.c
@@ -173,7 +173,8 @@ static void upload_sf_unit( struct brw_context *brw )
       
 
    /* _NEW_LINE */
-   sf.sf6.line_width = brw->attribs.Line->_Width * (1<<1);
+   /* XXX use ctx->Const.Min/MaxLineWidth here */
+   sf.sf6.line_width = CLAMP(brw->attribs.Line->Width, 1.0, 5.0) * (1<<1);
 
    sf.sf6.line_endcap_aa_region_width = 1;
    if (brw->attribs.Line->SmoothFlag)
@@ -183,7 +184,8 @@ static void upload_sf_unit( struct brw_context *brw )
 
    /* _NEW_POINT */
    sf.sf6.point_rast_rule = 1;	/* opengl conventions */
-   sf.sf7.point_size = brw->attribs.Point->_Size * (1<<3);
+   /* XXX clamp max depends on AA vs. non-AA */
+   sf.sf7.point_size = CLAMP(brw->attribs.Point->Size, 1.0, 3.0) * (1<<3);
    sf.sf7.use_point_size_state = !brw->attribs.Point->_Attenuated;
       
    /* might be BRW_NEW_PRIMITIVE if we have to adjust pv for polygons:
diff --git a/src/mesa/drivers/dri/i965/intel_blit.c b/src/mesa/drivers/dri/i965/intel_blit.c
index 173d1d5b6c..f88cbb2328 100644
--- a/src/mesa/drivers/dri/i965/intel_blit.c
+++ b/src/mesa/drivers/dri/i965/intel_blit.c
@@ -532,12 +532,15 @@ intelEmitImmediateColorExpandBlit(struct intel_context *intel,
 				  GLuint dst_offset,
 				  GLboolean dst_tiled,
 				  GLshort x, GLshort y, 
-				  GLshort w, GLshort h)
+				  GLshort w, GLshort h,
+				  GLenum logic_op)
 {
    struct xy_setup_blit setup;
    struct xy_text_immediate_blit text;
    int dwords = ((src_size + 7) & ~7) / 4;
 
+   assert( logic_op - GL_CLEAR >= 0 );
+   assert( logic_op - GL_CLEAR < 0x10 );
 
    if (w < 0 || h < 0) 
       return;
@@ -561,7 +564,7 @@ intelEmitImmediateColorExpandBlit(struct intel_context *intel,
    setup.br0.length = (sizeof(setup) / sizeof(int)) - 2;
       
    setup.br13.dest_pitch = dst_pitch;
-   setup.br13.rop = 0xcc;
+   setup.br13.rop = translate_raster_op(logic_op);
    setup.br13.color_depth = (cpp == 4) ? BR13_8888 : BR13_565;
    setup.br13.clipping_enable = 0;
    setup.br13.mono_source_transparency = 1;
diff --git a/src/mesa/drivers/dri/i965/intel_blit.h b/src/mesa/drivers/dri/i965/intel_blit.h
index 8b0cc65243..e361545c8f 100644
--- a/src/mesa/drivers/dri/i965/intel_blit.h
+++ b/src/mesa/drivers/dri/i965/intel_blit.h
@@ -72,6 +72,7 @@ intelEmitImmediateColorExpandBlit(struct intel_context *intel,
 				  GLuint dst_offset,
 				  GLboolean dst_tiled,
 				  GLshort dst_x, GLshort dst_y, 
-				  GLshort w, GLshort h);
+				  GLshort w, GLshort h,
+				  GLenum logic_op );
 
 #endif
diff --git a/src/mesa/drivers/dri/i965/intel_pixel_bitmap.c b/src/mesa/drivers/dri/i965/intel_pixel_bitmap.c
index 5841afaa3e..421fcc5e51 100644
--- a/src/mesa/drivers/dri/i965/intel_pixel_bitmap.c
+++ b/src/mesa/drivers/dri/i965/intel_pixel_bitmap.c
@@ -260,7 +260,9 @@ do_blit_bitmap( GLcontext *ctx,
 	       int h = MIN2(DY, box_h - py);
 	       int w = MIN2(DX, box_w - px); 
 	       GLuint sz = align(align(w,8) * h, 64)/8;
-	    
+	       GLenum logic_op = ctx->Color.ColorLogicOpEnabled ?
+		  ctx->Color.LogicOp : GL_COPY;
+
 	       assert(sz <= sizeof(stipple));
 	       memset(stipple, 0, sz);
 
@@ -288,7 +290,8 @@ do_blit_bitmap( GLcontext *ctx,
 						  dst->tiled,
 						  rect.x1 + px,
 						  rect.y2 - (py + h),
-						  w, h);
+						  w, h,
+						  logic_op);
 	    } 
 	 } 
       }
diff --git a/src/mesa/drivers/dri/i965/intel_tex_validate.c b/src/mesa/drivers/dri/i965/intel_tex_validate.c
index cb23b9dd87..44ee94614d 100644
--- a/src/mesa/drivers/dri/i965/intel_tex_validate.c
+++ b/src/mesa/drivers/dri/i965/intel_tex_validate.c
@@ -138,7 +138,7 @@ GLuint intel_finalize_mipmap_tree( struct intel_context *intel,
    
    /* We know/require this is true by now: 
     */
-   assert(intelObj->base.Complete);
+   assert(intelObj->base._Complete);
 
    /* What levels must the tree include at a minimum?
     */
diff --git a/src/mesa/drivers/dri/mach64/mach64_native_vb.c b/src/mesa/drivers/dri/mach64/mach64_native_vb.c
index 75cf0e2ed2..248fa2a9a2 100644
--- a/src/mesa/drivers/dri/mach64/mach64_native_vb.c
+++ b/src/mesa/drivers/dri/mach64/mach64_native_vb.c
@@ -103,7 +103,7 @@ void TAG(translate_vertex)(GLcontext *ctx,
 
    assert( p + 1 - (CARD32 *)src == 10 );
 	 
-   dst->pointSize = ctx->Point._Size;
+   dst->pointSize = ctx->Point.Size;
 }
 
 
diff --git a/src/mesa/drivers/dri/mach64/mach64_tris.c b/src/mesa/drivers/dri/mach64/mach64_tris.c
index 08cc1849a1..369f610442 100644
--- a/src/mesa/drivers/dri/mach64/mach64_tris.c
+++ b/src/mesa/drivers/dri/mach64/mach64_tris.c
@@ -673,7 +673,10 @@ static __inline void mach64_draw_line( mach64ContextPtr mmesa,
 #if MACH64_NATIVE_VTXFMT
    GLcontext *ctx = mmesa->glCtx;
    const GLuint vertsize = mmesa->vertex_size;
-   GLint width = (GLint)(mmesa->glCtx->Line._Width * 2.0); /* 2 fractional bits for hardware */
+   /* 2 fractional bits for hardware: */
+   const int width = (int) (2.0 * CLAMP(mmesa->glCtx->Line.Width,
+                                        mmesa->glCtx->Const.MinLineWidth,
+                                        mmesa->glCtx->Const.MaxLineWidth));
    GLfloat ooa;
    GLuint *pxy0, *pxy1;
    GLuint xy0old, xy0, xy1old, xy1;
@@ -691,9 +694,6 @@ static __inline void mach64_draw_line( mach64ContextPtr mmesa,
       mach64_print_vertex( ctx, v1 );
    }
   
-   if( !width )
-      width = 1;	/* round to the nearest supported width */
-      
    pxy0 = &v0->ui[xyoffset];
    xy0old = *pxy0;
    xy0 = LE32_IN( &xy0old );
@@ -961,7 +961,10 @@ static __inline void mach64_draw_point( mach64ContextPtr mmesa,
 #if MACH64_NATIVE_VTXFMT
    GLcontext *ctx = mmesa->glCtx;
    const GLuint vertsize = mmesa->vertex_size;
-   GLint sz = (GLint)(mmesa->glCtx->Point._Size * 2.0); /* 2 fractional bits for hardware */
+   /* 2 fractional bits for hardware: */
+   GLint sz = (GLint) (2.0 * CLAMP(mmesa->glCtx->Point.Size,
+                                   ctx->Const.MinPointSize,
+                                   ctx->Const.MaxPointSize));
    GLfloat ooa;
    GLuint *pxy;
    GLuint xyold, xy;
diff --git a/src/mesa/drivers/dri/mga/mgatris.c b/src/mesa/drivers/dri/mga/mgatris.c
index 2b7ea05b14..91b413ae76 100644
--- a/src/mesa/drivers/dri/mga/mgatris.c
+++ b/src/mesa/drivers/dri/mga/mgatris.c
@@ -104,8 +104,10 @@ static void __inline__ mga_draw_quad( mgaContextPtr mmesa,
 static __inline__ void mga_draw_point( mgaContextPtr mmesa,
 					mgaVertexPtr tmp )
 {
-   GLfloat sz = mmesa->glCtx->Point._Size * .5;
-   int vertex_size = mmesa->vertex_size;
+   const GLfloat sz = 0.5 * CLAMP(mmesa->glCtx->Point.Size,
+                                  mmesa->glCtx->Const.MinPointSize,
+                                  mmesa->glCtx->Const.MaxPointSize);
+   const int vertex_size = mmesa->vertex_size;
    GLuint *vb = mgaAllocDmaLow( mmesa, 6 * 4 * vertex_size );
    int j;
    
@@ -165,7 +167,9 @@ static __inline__ void mga_draw_line( mgaContextPtr mmesa,
    GLuint vertex_size = mmesa->vertex_size;
    GLuint *vb = mgaAllocDmaLow( mmesa, 6 * 4 * vertex_size );
    GLfloat dx, dy, ix, iy;
-   GLfloat width = mmesa->glCtx->Line._Width;
+   const GLfloat width = CLAMP(mmesa->glCtx->Line.Width,
+                               mmesa->glCtx->Const.MinLineWidth,
+                               mmesa->glCtx->Const.MaxLineWidth);
    GLint j;
 
 #if 0
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.c b/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.c
index 684ed7b017..fc14060c04 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.c
@@ -41,7 +41,7 @@ nouveau_bo_download_from_screen(GLcontext *ctx,	GLuint offset, GLuint size,
 		DEBUG("..sys_mem\n");
 		in_mem = nouveau_mem_alloc(ctx, NOUVEAU_MEM_AGP, size, 0);
 		if (in_mem) {
-			DEBUG("....via AGP\n");
+			DEBUG("....via GART\n");
 			/* otherwise, try blitting to faster memory and
 			 * copying from there
 			 */
@@ -86,7 +86,7 @@ nouveau_bo_upload_to_screen(GLcontext *ctx, GLuint offset, GLuint size,
 						 NOUVEAU_MEM_MAPPED,
 						 size, 0);
 		if (out_mem) {
-			DEBUG("....via AGP\n");
+			DEBUG("....via GART\n");
 			_mesa_memcpy(out_mem->map,
 					nbo->cpu_mem_sys + offset, size);
 			nouveau_memformat_flat_emit(ctx, nbo->gpu_mem, out_mem,
@@ -511,7 +511,7 @@ nouveauBufferData(GLcontext *ctx, GLenum target, GLsizeiptrARB size,
 		gpu_flags = 0;
 		break;
 	default:
-		gpu_flags = NOUVEAU_BO_VRAM_OK | NOUVEAU_BO_AGP_OK;
+		gpu_flags = NOUVEAU_BO_VRAM_OK | NOUVEAU_BO_GART_OK;
 		break;
 	}
 	nouveau_bo_init_storage(ctx, gpu_flags, size, data, usage, obj);
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.h b/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.h
index 932450fd87..3439a35e7c 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_bufferobj.h
@@ -5,7 +5,7 @@
 #include "nouveau_buffers.h"
 
 #define NOUVEAU_BO_VRAM_OK (NOUVEAU_MEM_FB | NOUVEAU_MEM_FB_ACCEPTABLE)
-#define NOUVEAU_BO_AGP_OK  (NOUVEAU_MEM_AGP | NOUVEAU_MEM_AGP_ACCEPTABLE)
+#define NOUVEAU_BO_GART_OK  (NOUVEAU_MEM_AGP | NOUVEAU_MEM_AGP_ACCEPTABLE)
 
 typedef struct nouveau_bufferobj_region_t {
 	uint32_t start;
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_buffers.c b/src/mesa/drivers/dri/nouveau/nouveau_buffers.c
index b54f68f402..d498f616c9 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_buffers.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_buffers.c
@@ -32,8 +32,8 @@ nouveau_memformat_flat_emit(GLcontext *ctx,
       return GL_FALSE;
    }
 
-   src_handle = (src->type & NOUVEAU_MEM_FB) ? NvDmaFB : NvDmaAGP;
-   dst_handle = (dst->type & NOUVEAU_MEM_FB) ? NvDmaFB : NvDmaAGP;
+   src_handle = (src->type & NOUVEAU_MEM_FB) ? NvDmaFB : NvDmaTT;
+   dst_handle = (dst->type & NOUVEAU_MEM_FB) ? NvDmaFB : NvDmaTT;
    src_offset += nouveau_mem_gpu_offset_get(ctx, src);
    dst_offset += nouveau_mem_gpu_offset_get(ctx, dst);
 
@@ -69,7 +69,7 @@ void
 nouveau_mem_free(GLcontext *ctx, nouveau_mem *mem)
 {
    nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
-   drm_nouveau_mem_free_t memf;
+   struct drm_nouveau_mem_free memf;
 
    if (NOUVEAU_DEBUG & DEBUG_MEM)  {
       fprintf(stderr, "%s: type=0x%x, offset=0x%x, size=0x%x\n",
@@ -78,8 +78,8 @@ nouveau_mem_free(GLcontext *ctx, nouveau_mem *mem)
 
    if (mem->map)
       drmUnmap(mem->map, mem->size);
-   memf.flags         = mem->type;
-   memf.region_offset = mem->offset;
+   memf.flags  = mem->type;
+   memf.offset = mem->offset;
    drmCommandWrite(nmesa->driFd, DRM_NOUVEAU_MEM_FREE, &memf, sizeof(memf));
    FREE(mem);
 }
@@ -88,7 +88,7 @@ nouveau_mem *
 nouveau_mem_alloc(GLcontext *ctx, int type, GLuint size, GLuint align)
 {
    nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
-   drm_nouveau_mem_alloc_t mema;
+   struct drm_nouveau_mem_alloc mema;
    nouveau_mem *mem;
    int ret;
 
@@ -111,7 +111,7 @@ nouveau_mem_alloc(GLcontext *ctx, int type, GLuint size, GLuint align)
       FREE(mem);
       return NULL;
    }
-   mem->offset = mema.region_offset;
+   mem->offset = mema.offset;
    mem->type   = mema.flags;
 
    if (NOUVEAU_DEBUG & DEBUG_MEM)  {
@@ -120,7 +120,7 @@ nouveau_mem_alloc(GLcontext *ctx, int type, GLuint size, GLuint align)
    }
 
    if (type & NOUVEAU_MEM_MAPPED)
-      ret = drmMap(nmesa->driFd, mem->offset, mem->size, &mem->map);
+      ret = drmMap(nmesa->driFd, mema.map_handle, mem->size, &mem->map);
    if (ret) {
       mem->map = NULL;
       nouveau_mem_free(ctx, mem);
@@ -135,12 +135,7 @@ nouveau_mem_gpu_offset_get(GLcontext *ctx, nouveau_mem *mem)
 {
    nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
 
-   if (mem->type & NOUVEAU_MEM_FB)
-      return (uint32_t)mem->offset - nmesa->vram_phys;
-   else if (mem->type & NOUVEAU_MEM_AGP)
-      return (uint32_t)mem->offset - nmesa->agp_phys;
-   else
-      return 0xDEADF00D;
+   return mem->offset;
 }
 
 static GLboolean
@@ -299,6 +294,8 @@ nouveau_cliprects_drawable_set(nouveauContextPtr nmesa,
    nmesa->pClipRects	= dPriv->pClipRects;
    nmesa->drawX		= dPriv->x;
    nmesa->drawY		= dPriv->y;
+   nmesa->drawW		= dPriv->w;
+   nmesa->drawH		= dPriv->h;
 }
 
 static void
@@ -313,6 +310,8 @@ nouveau_cliprects_renderbuffer_set(nouveauContextPtr nmesa,
    nmesa->osClipRect.y2	= nrb->mesa.Height;
    nmesa->drawX		= 0;
    nmesa->drawY		= 0;
+   nmesa->drawW		= nrb->mesa.Width;
+   nmesa->drawH		= nrb->mesa.Height;
 }
 
 void
@@ -391,9 +390,12 @@ nouveauNewRenderbuffer(GLcontext *ctx, GLuint name)
 }
 
 static void
-nouveauBindFramebuffer(GLcontext *ctx, GLenum target, struct gl_framebuffer *fb)
+nouveauBindFramebuffer(GLcontext *ctx, GLenum target,
+                       struct gl_framebuffer *fb, struct gl_framebuffer *fbread)
 {
-   nouveau_build_framebuffer(ctx, fb);
+   if (target == GL_FRAMEBUFFER_EXT || target == GL_DRAW_FRAMEBUFFER_EXT) {
+      nouveau_build_framebuffer(ctx, fb);
+   }
 }
 
 static void
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.c b/src/mesa/drivers/dri/nouveau/nouveau_context.c
index 8e11eb6134..44392c0267 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_context.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_context.c
@@ -145,10 +145,10 @@ GLboolean nouveauCreateContext( const __GLcontextModes *glVisual,
 		 		&nmesa->vram_size))
 	   return GL_FALSE;
 	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_AGP_PHYSICAL,
-		 		&nmesa->agp_phys))
+		 		&nmesa->gart_phys))
 	   return GL_FALSE;
 	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_AGP_SIZE,
-		 		&nmesa->agp_size))
+		 		&nmesa->gart_size))
 	   return GL_FALSE;
 	if (!nouveauFifoInit(nmesa))
 	   return GL_FALSE;
@@ -180,7 +180,7 @@ GLboolean nouveauCreateContext( const __GLcontextModes *glVisual,
 	driParseConfigFiles (&nmesa->optionCache, &screen->optionCache,
 			screen->driScreen->myNum, "nouveau");
 
-	nmesa->sarea = (drm_nouveau_sarea_t *)((char *)sPriv->pSAREA +
+	nmesa->sarea = (struct drm_nouveau_sarea *)((char *)sPriv->pSAREA +
 			screen->sarea_priv_offset);
 
 	/* Enable any supported extensions */
@@ -224,6 +224,8 @@ GLboolean nouveauCreateContext( const __GLcontextModes *glVisual,
 			nv04TriInitFunctions( ctx );
 			break;
 		case NV_10:
+		case NV_11:
+		case NV_17:
 		case NV_20:
 		case NV_30:
 		case NV_40:
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.h b/src/mesa/drivers/dri/nouveau/nouveau_context.h
index 87e4479da3..9a0be2cb2a 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_context.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_context.h
@@ -99,19 +99,22 @@ typedef struct nouveau_context {
 	/* The read-only regs */
 	volatile unsigned char* mmio;
 
+	/* The per-channel notifier block */
+	volatile void *notifier_block;
+
 	/* Physical addresses of AGP/VRAM apertures */
 	uint64_t vram_phys;
 	uint64_t vram_size;
-	uint64_t agp_phys;
-	uint64_t agp_size;
+	uint64_t gart_phys;
+	uint64_t gart_size;
 
 	/* Channel synchronisation */
-	nouveau_notifier *syncNotifier;
+	struct drm_nouveau_notifier_alloc *syncNotifier;
 
 	/* ARB_occlusion_query / EXT_timer_query */
 	GLuint		  query_object_max;
 	GLboolean *	  query_alloc;
-	nouveau_notifier *queryNotifier;
+	struct drm_nouveau_notifier_alloc *queryNotifier;
 
 	/* Additional hw-specific functions */
 	nouveau_hw_func hw_func;
@@ -150,7 +153,7 @@ typedef struct nouveau_context {
 	GLuint numClipRects;
 	drm_clip_rect_t *pClipRects;
 	drm_clip_rect_t osClipRect;
-	GLuint drawX, drawY;
+	GLuint drawX, drawY, drawW, drawH;
 
 	/* The rendering context information */
 	GLenum current_primitive; /* the current primitive enum */
@@ -165,7 +168,7 @@ typedef struct nouveau_context {
 	nouveauShader *passthrough_fp;
 
 	nouveauScreenRec *screen;
-	drm_nouveau_sarea_t *sarea;
+	struct drm_nouveau_sarea *sarea;
 
 	__DRIcontextPrivate  *driContext;    /* DRI context */
 	__DRIscreenPrivate   *driScreen;     /* DRI screen */
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_driver.c b/src/mesa/drivers/dri/nouveau/nouveau_driver.c
index 00956aa8f8..ddc9535624 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_driver.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_driver.c
@@ -41,7 +41,7 @@ GLboolean nouveauDRMGetParam(nouveauContextPtr nmesa,
 			     unsigned int      param,
 			     uint64_t*         value)
 {
-	drm_nouveau_getparam_t getp;
+	struct drm_nouveau_getparam getp;
 
 	getp.param = param;
 	if (!value || drmCommandWriteRead(nmesa->driFd, DRM_NOUVEAU_GETPARAM,
@@ -56,7 +56,7 @@ GLboolean nouveauDRMSetParam(nouveauContextPtr nmesa,
 			     unsigned int      param,
 			     uint64_t          value)
 {
-	drm_nouveau_setparam_t setp;
+	struct drm_nouveau_setparam setp;
 
 	setp.param = param;
 	setp.value = value;
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_fifo.c b/src/mesa/drivers/dri/nouveau/nouveau_fifo.c
index bd2b2eddd0..7b5e96b4c2 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_fifo.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_fifo.c
@@ -98,14 +98,15 @@ void nouveauWaitForIdle(nouveauContextPtr nmesa)
 // here we call the fifo initialization ioctl and fill in stuff accordingly
 GLboolean nouveauFifoInit(nouveauContextPtr nmesa)
 {
-	drm_nouveau_fifo_alloc_t fifo_init;
-	int i;
+	struct drm_nouveau_fifo_alloc fifo_init;
+	int i, ret;
 
 #ifdef NOUVEAU_RING_DEBUG
 	return GL_TRUE;
 #endif
 
-	int ret;
+	fifo_init.fb_ctxdma_handle = NvDmaFB;
+	fifo_init.tt_ctxdma_handle = NvDmaTT;
 	ret=drmCommandWriteRead(nmesa->driFd, DRM_NOUVEAU_FIFO_ALLOC, &fifo_init, sizeof(fifo_init));
 	if (ret) {
 		FATAL("Fifo initialization ioctl failed (returned %d)\n",ret);
@@ -117,12 +118,21 @@ GLboolean nouveauFifoInit(nouveauContextPtr nmesa)
 		FATAL("Unable to map the fifo (returned %d)\n",ret);
 		return GL_FALSE;
 	}
+
 	ret = drmMap(nmesa->driFd, fifo_init.ctrl, fifo_init.ctrl_size, &nmesa->fifo.mmio);
 	if (ret) {
 		FATAL("Unable to map the control regs (returned %d)\n",ret);
 		return GL_FALSE;
 	}
 
+	ret = drmMap(nmesa->driFd, fifo_init.notifier,
+				   fifo_init.notifier_size,
+				   &nmesa->notifier_block);
+	if (ret) {
+		FATAL("Unable to map the notifier block (returned %d)\n",ret);
+		return GL_FALSE;
+	}
+
 	/* Setup our initial FIFO tracking params */
 	nmesa->fifo.channel  = fifo_init.channel;
 	nmesa->fifo.put_base = fifo_init.put_base;
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_lock.c b/src/mesa/drivers/dri/nouveau/nouveau_lock.c
index c119d14dd7..aa86c9e783 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_lock.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_lock.c
@@ -44,7 +44,7 @@ void nouveauGetLock( nouveauContextPtr nmesa, GLuint flags )
 {
    __DRIdrawablePrivate *dPriv = nmesa->driDrawable;
    __DRIscreenPrivate *sPriv = nmesa->driScreen;
-   drm_nouveau_sarea_t *sarea = nmesa->sarea;
+   struct drm_nouveau_sarea *sarea = nmesa->sarea;
 
    drmGetLock( nmesa->driFd, nmesa->hHWContext, flags );
 
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_object.c b/src/mesa/drivers/dri/nouveau/nouveau_object.c
index b71acff430..a143488e8d 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_object.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_object.c
@@ -7,61 +7,18 @@
 GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa,
 				     uint32_t handle, int class)
 {
-	drm_nouveau_object_init_t cto;
+	struct drm_nouveau_grobj_alloc cto;
 	int ret;
 
 	cto.channel = nmesa->fifo.channel;
 	cto.handle  = handle;
 	cto.class   = class;
-	ret = drmCommandWrite(nmesa->driFd, DRM_NOUVEAU_OBJECT_INIT, &cto, sizeof(cto));
+	ret = drmCommandWrite(nmesa->driFd, DRM_NOUVEAU_GROBJ_ALLOC,
+			      &cto, sizeof(cto));
 
 	return ret == 0;
 }
 
-GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
-      				 uint32_t handle,
-				 int      class,
-				 uint32_t offset,
-				 uint32_t size,
-				 int	  target,
-				 int	  access)
-{
-	drm_nouveau_dma_object_init_t dma;
-	int ret;
-
-	dma.channel = nmesa->fifo.channel;
-	dma.class   = class;
-	dma.handle  = handle;
-	dma.target  = target;
-	dma.access  = access;
-	dma.offset  = offset;
-	dma.size    = size;
-	ret = drmCommandWriteRead(nmesa->driFd, DRM_NOUVEAU_DMA_OBJECT_INIT,
-				  &dma, sizeof(dma));
-	return ret == 0;
-}
-
-GLboolean nouveauCreateDmaObjectFromMem(nouveauContextPtr nmesa,
-					uint32_t handle, int class,
-					nouveau_mem *mem,
-					int access)
-{
-	uint32_t offset = mem->offset;
-	int target = mem->type & (NOUVEAU_MEM_FB | NOUVEAU_MEM_AGP);
-
-	if (!target)
-		return GL_FALSE;
-
-	if (target & NOUVEAU_MEM_FB)
-		offset -= nmesa->vram_phys;
-	else if (target & NOUVEAU_MEM_AGP)
-		offset -= nmesa->agp_phys;
-
-	return nouveauCreateDmaObject(nmesa, handle, class,
-				      offset, mem->size,
-				      target, access);
-}
-
 void nouveauObjectOnSubchannel(nouveauContextPtr nmesa, int subchannel, int handle)
 {
 	BEGIN_RING_SIZE(subchannel, 0, 1);
@@ -74,23 +31,16 @@ void nouveauObjectInit(nouveauContextPtr nmesa)
 	return;
 #endif
 
-/* We need to know vram size.. and AGP size (and even if the card is AGP..) */
-	nouveauCreateDmaObject( nmesa, NvDmaFB, NV_DMA_IN_MEMORY,
-				0, nmesa->vram_size,
-				NOUVEAU_MEM_FB,
-				NOUVEAU_MEM_ACCESS_RW);
-	nouveauCreateDmaObject( nmesa, NvDmaAGP, NV_DMA_IN_MEMORY,
-	      			0, nmesa->agp_size,
-				NOUVEAU_MEM_AGP,
-				NOUVEAU_MEM_ACCESS_RW);
-
 	nouveauCreateContextObject(nmesa, Nv3D, nmesa->screen->card->class_3d);
 	if (nmesa->screen->card->type>=NV_10) {
 		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV10_CONTEXT_SURFACES_2D);
-		nouveauCreateContextObject(nmesa, NvImageBlit, NV10_IMAGE_BLIT);
 	} else {
 		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV04_CONTEXT_SURFACES_2D);
 		nouveauCreateContextObject(nmesa, NvCtxSurf3D, NV04_CONTEXT_SURFACES_3D);
+	}
+	if (nmesa->screen->card->type>=NV_11) {
+		nouveauCreateContextObject(nmesa, NvImageBlit, NV10_IMAGE_BLIT);
+	} else {
 		nouveauCreateContextObject(nmesa, NvImageBlit, NV_IMAGE_BLIT);
 	}
 	nouveauCreateContextObject(nmesa, NvMemFormat, NV_MEMORY_TO_MEMORY_FORMAT);
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_object.h b/src/mesa/drivers/dri/nouveau/nouveau_object.h
index 0be9b4309c..8c72d014da 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_object.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_object.h
@@ -14,7 +14,7 @@ enum DMAObjects {
 	NvMemFormat		= 0x80000022,
 	NvCtxSurf3D		= 0x80000023,
 	NvDmaFB			= 0xD0FB0001,
-	NvDmaAGP		= 0xD0AA0001,
+	NvDmaTT			= 0xD0AA0001,
 	NvSyncNotify		= 0xD0000001,
 	NvQueryNotify		= 0xD0000002
 };
@@ -31,17 +31,5 @@ extern void nouveauObjectOnSubchannel(nouveauContextPtr nmesa, int subchannel, i
 
 extern GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa,
       					    uint32_t handle, int class);
-extern GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
-      					uint32_t handle,
-					int      class,
-					uint32_t offset,
-					uint32_t size,
-					int      target,
-					int      access);
-extern GLboolean nouveauCreateDmaObjectFromMem(nouveauContextPtr nmesa,
-					       uint32_t     handle,
-					       int          class,
-					       nouveau_mem *mem,
-					       int          access);
 
 #endif
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_query.c b/src/mesa/drivers/dri/nouveau/nouveau_query.c
index de3f5b0378..0154140069 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_query.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_query.c
@@ -68,7 +68,7 @@ nouveauBeginQuery(GLcontext *ctx, GLenum target, struct gl_query_object *q)
 	nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
 	nouveau_query_object *nq = (nouveau_query_object *)q;
 
-	nouveau_notifier_reset(nmesa->queryNotifier, nq->notifier_id);
+	nouveau_notifier_reset(ctx, nmesa->queryNotifier, nq->notifier_id);
 
 	switch (nmesa->screen->card->type) {
 	case NV_20:
@@ -105,12 +105,13 @@ nouveauUpdateQuery(GLcontext *ctx, GLenum target, struct gl_query_object *q)
 	nouveau_query_object *nq = (nouveau_query_object *)q;
 	int status;
 
-	status = nouveau_notifier_status(nmesa->queryNotifier,
+	status = nouveau_notifier_status(ctx, nmesa->queryNotifier,
 					 nq->notifier_id);
 
 	q->Ready = (status == NV_NOTIFY_STATE_STATUS_COMPLETED);
 	if (q->Ready)
-		q->Result = nouveau_notifier_return_val(nmesa->queryNotifier,
+		q->Result = nouveau_notifier_return_val(ctx,
+							nmesa->queryNotifier,
 							nq->notifier_id);
 }
 
@@ -120,7 +121,7 @@ nouveauWaitQueryResult(GLcontext *ctx, GLenum target, struct gl_query_object *q)
 	nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
 	nouveau_query_object *nq = (nouveau_query_object *)q;
 
-	nouveau_notifier_wait_status(nmesa->queryNotifier, nq->notifier_id,
+	nouveau_notifier_wait_status(ctx, nmesa->queryNotifier, nq->notifier_id,
 				     NV_NOTIFY_STATE_STATUS_COMPLETED, 0);
 	nouveauUpdateQuery(ctx, target, q);
 }
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_screen.c b/src/mesa/drivers/dri/nouveau/nouveau_screen.c
index 7a4b9f1cd0..69b0691bb7 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_screen.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_screen.c
@@ -328,7 +328,7 @@ void * __driCreateNewScreen_20050727( __DRInativeDisplay *dpy, int scrn, __DRIsc
 	static const __DRIversion ddx_expected = { 1, 2, 0 };
 	static const __DRIversion dri_expected = { 4, 0, 0 };
 	static const __DRIversion drm_expected = { 0, 0, NOUVEAU_DRM_HEADER_PATCHLEVEL };
-#if NOUVEAU_DRM_HEADER_PATCHLEVEL != 6
+#if NOUVEAU_DRM_HEADER_PATCHLEVEL != 9
 #error nouveau_drm.h version doesn't match expected version
 #endif
 	dri_interface = interface;
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_span.c b/src/mesa/drivers/dri/nouveau/nouveau_span.c
index 74dec66afc..6e3f9fadf4 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_span.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_span.c
@@ -37,6 +37,8 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define HAVE_HW_STENCIL_SPANS	0
 #define HAVE_HW_STENCIL_PIXELS	0
 
+static char *fake_span[1280*1024*4];
+
 #define HW_CLIPLOOP()							\
    do {									\
       int _nc = nmesa->numClipRects;					\
@@ -52,6 +54,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
    GLuint height = nrb->mesa.Height;					\
    GLubyte *map = (GLubyte *)(nrb->map ? nrb->map : nrb->mem->map) +    \
 	 (nmesa->drawY * nrb->pitch) + (nmesa->drawX * nrb->cpp);       \
+   map = fake_span; \
    GLuint p;								\
    (void) p;
 
@@ -120,6 +123,6 @@ nouveauSpanSetFunctions(nouveau_renderbuffer *nrb, const GLvisual *vis)
 {
    if (nrb->mesa._ActualFormat == GL_RGBA8)
       nouveauInitPointers_ARGB8888(&nrb->mesa);
-   else if (nrb->mesa._ActualFormat == GL_RGB5)
+   else // if (nrb->mesa._ActualFormat == GL_RGB5)
       nouveauInitPointers_RGB565(&nrb->mesa);
 }
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_state.c b/src/mesa/drivers/dri/nouveau/nouveau_state.c
index e9fd188d73..f618dcfc99 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_state.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_state.c
@@ -60,14 +60,14 @@ static void nouveauCalcViewport(GLcontext *ctx)
     nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
     const GLfloat *v = ctx->Viewport._WindowMap.m;
     GLfloat *m = nmesa->viewport.m;
-    GLfloat xoffset = nmesa->drawX, yoffset = nmesa->drawY;
+    GLfloat xoffset = nmesa->drawX, yoffset = nmesa->drawY + nmesa->drawH;
   
     nmesa->depth_scale = 1.0 / ctx->DrawBuffer->_DepthMaxF;
 
     m[MAT_SX] =   v[MAT_SX];
     m[MAT_TX] =   v[MAT_TX] + xoffset + SUBPIXEL_X;
     m[MAT_SY] = - v[MAT_SY];
-    m[MAT_TY] =   v[MAT_TY] + yoffset + SUBPIXEL_Y;
+    m[MAT_TY] = (-v[MAT_TY]) + yoffset + SUBPIXEL_Y;
     m[MAT_SZ] =   v[MAT_SZ] * nmesa->depth_scale;
     m[MAT_TZ] =   v[MAT_TZ] * nmesa->depth_scale;
 
@@ -162,6 +162,8 @@ void nouveauDDInitState(nouveauContextPtr nmesa)
             nv04InitStateFuncs(nmesa->glCtx, &nmesa->glCtx->Driver);
             break;
         case NV_10:
+        case NV_11:
+        case NV_17:
             nv10InitStateFuncs(nmesa->glCtx, &nmesa->glCtx->Driver);
             break;
         case NV_20:
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_sync.c b/src/mesa/drivers/dri/nouveau/nouveau_sync.c
index 30e6696269..8abc847e1e 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_sync.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_sync.c
@@ -35,53 +35,53 @@
 #include "nouveau_msg.h"
 #include "nouveau_sync.h"
 
-nouveau_notifier *
+#define NOTIFIER(__v) \
+	nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx); \
+	volatile uint32_t *__v = (void*)nmesa->notifier_block + notifier->offset
+
+struct drm_nouveau_notifier_alloc *
 nouveau_notifier_new(GLcontext *ctx, GLuint handle, GLuint count)
 {
 	nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
-	nouveau_notifier *notifier;
+	struct drm_nouveau_notifier_alloc *notifier;
+	int ret;
 
 #ifdef NOUVEAU_RING_DEBUG
 	return NULL;
 #endif
 
-	notifier = CALLOC_STRUCT(nouveau_notifier_t);
+	notifier = CALLOC_STRUCT(drm_nouveau_notifier_alloc);
 	if (!notifier)
 		return NULL;
 
-	notifier->mem = nouveau_mem_alloc(ctx,
-					  NOUVEAU_MEM_FB | NOUVEAU_MEM_MAPPED,
-					  count * NV_NOTIFIER_SIZE,
-					  0);
-	if (!notifier->mem) {
-		FREE(notifier);
-		return NULL;
-	}
-
-	if (!nouveauCreateDmaObjectFromMem(nmesa, handle, NV_DMA_IN_MEMORY,
-					   notifier->mem,
-					   NOUVEAU_MEM_ACCESS_RW)) {
-		nouveau_mem_free(ctx, notifier->mem);
+	notifier->channel = nmesa->fifo.channel;
+	notifier->handle  = handle;
+	notifier->count   = count;
+	ret = drmCommandWriteRead(nmesa->driFd, DRM_NOUVEAU_NOTIFIER_ALLOC,
+				  notifier, sizeof(*notifier));
+	if (ret) {
+		MESSAGE("Failed to create notifier 0x%08x: %d\n", handle, ret);
 		FREE(notifier);
 		return NULL;
 	}
 
-	notifier->handle = handle;
 	return notifier;
 }
 
 void
-nouveau_notifier_destroy(GLcontext *ctx, nouveau_notifier *notifier)
+nouveau_notifier_destroy(GLcontext *ctx,
+			 struct drm_nouveau_notifier_alloc *notifier)
 {
-	/*XXX: free DMA object.. */
-	nouveau_mem_free(ctx, notifier->mem);
+	/*XXX: free notifier object.. */
 	FREE(notifier);
 }
 
 void
-nouveau_notifier_reset(nouveau_notifier *notifier, GLuint id)
+nouveau_notifier_reset(GLcontext *ctx,
+		       struct drm_nouveau_notifier_alloc *notifier,
+		       GLuint id)
 {
-	volatile GLuint *n = notifier->mem->map + (id * NV_NOTIFIER_SIZE);
+	NOTIFIER(n);
 
 #ifdef NOUVEAU_RING_DEBUG
 	return;
@@ -95,26 +95,31 @@ nouveau_notifier_reset(nouveau_notifier *notifier, GLuint id)
 }
 
 GLuint
-nouveau_notifier_status(nouveau_notifier *notifier, GLuint id)
+nouveau_notifier_status(GLcontext *ctx,
+			struct drm_nouveau_notifier_alloc *notifier,
+			GLuint id)
 {
-	volatile GLuint *n = notifier->mem->map + (id * NV_NOTIFIER_SIZE);
+	NOTIFIER(n);
 
 	return n[NV_NOTIFY_STATE/4] >> NV_NOTIFY_STATE_STATUS_SHIFT;
 }
 
 GLuint
-nouveau_notifier_return_val(nouveau_notifier *notifier, GLuint id)
+nouveau_notifier_return_val(GLcontext *ctx,
+			    struct drm_nouveau_notifier_alloc *notifier,
+			    GLuint id)
 {
-	volatile GLuint *n = notifier->mem->map + (id * NV_NOTIFIER_SIZE);
+	NOTIFIER(n);
 
 	return n[NV_NOTIFY_RETURN_VALUE/4];
 }
 
 GLboolean
-nouveau_notifier_wait_status(nouveau_notifier *notifier, GLuint id,
-			     GLuint status, GLuint timeout)
+nouveau_notifier_wait_status(GLcontext *ctx,
+			     struct drm_nouveau_notifier_alloc *notifier,
+			     GLuint id, GLuint status, GLuint timeout)
 {
-	volatile GLuint *n = notifier->mem->map + (id * NV_NOTIFIER_SIZE);
+	NOTIFIER(n);
 	unsigned int time = 0;
 
 #ifdef NOUVEAU_RING_DEBUG
@@ -144,13 +149,14 @@ nouveau_notifier_wait_status(nouveau_notifier *notifier, GLuint id,
 }
 
 void
-nouveau_notifier_wait_nop(GLcontext *ctx, nouveau_notifier *notifier,
-					  GLuint subc)
+nouveau_notifier_wait_nop(GLcontext *ctx,
+			  struct drm_nouveau_notifier_alloc *notifier,
+			  GLuint subc)
 {
-	nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
+	NOTIFIER(n);
 	GLboolean ret;
 
-	nouveau_notifier_reset(notifier, 0);
+	nouveau_notifier_reset(ctx, notifier, 0);
 
 	BEGIN_RING_SIZE(subc, NV_NOTIFY, 1);
 	OUT_RING       (NV_NOTIFY_STYLE_WRITE_ONLY);
@@ -158,7 +164,7 @@ nouveau_notifier_wait_nop(GLcontext *ctx, nouveau_notifier *notifier,
 	OUT_RING       (0);
 	FIRE_RING();
 
-	ret = nouveau_notifier_wait_status(notifier, 0,
+	ret = nouveau_notifier_wait_status(ctx, notifier, 0,
 					   NV_NOTIFY_STATE_STATUS_COMPLETED,
 					   0 /* no timeout */);
 	if (ret == GL_FALSE) MESSAGE("wait on notifier failed\n");
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_sync.h b/src/mesa/drivers/dri/nouveau/nouveau_sync.h
index 019d5f6629..b76af17276 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_sync.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_sync.h
@@ -47,21 +47,25 @@
 #define NV_NOTIFY                                                     0x00000104
 #define NV_NOTIFY_STYLE_WRITE_ONLY                                             0
 
-typedef struct nouveau_notifier_t {
-	GLuint       handle;
-	nouveau_mem *mem;
-} nouveau_notifier;
-
-extern nouveau_notifier *nouveau_notifier_new(GLcontext *, GLuint handle,
-					      GLuint count);
-extern void nouveau_notifier_destroy(GLcontext *, nouveau_notifier *);
-extern void nouveau_notifier_reset(nouveau_notifier *, GLuint id);
-extern GLuint nouveau_notifier_status(nouveau_notifier *, GLuint id);
-extern GLuint nouveau_notifier_return_val(nouveau_notifier *, GLuint id);
-extern GLboolean nouveau_notifier_wait_status(nouveau_notifier *r, GLuint id,
-					      GLuint status, GLuint timeout);
-extern void nouveau_notifier_wait_nop(GLcontext *ctx,
-      				      nouveau_notifier *, GLuint subc);
+extern struct drm_nouveau_notifier_alloc *
+nouveau_notifier_new(GLcontext *, GLuint handle, GLuint count);
+extern void 
+nouveau_notifier_destroy(GLcontext *, struct drm_nouveau_notifier_alloc *);
+extern void
+nouveau_notifier_reset(GLcontext *, struct drm_nouveau_notifier_alloc *,
+		       GLuint id);
+extern GLuint
+nouveau_notifier_status(GLcontext *, struct drm_nouveau_notifier_alloc *,
+			GLuint id);
+extern GLuint
+nouveau_notifier_return_val(GLcontext *, struct drm_nouveau_notifier_alloc *,
+			    GLuint id);
+extern GLboolean
+nouveau_notifier_wait_status(GLcontext *, struct drm_nouveau_notifier_alloc *,
+			     GLuint id, GLuint status, GLuint timeout);
+extern void
+nouveau_notifier_wait_nop(GLcontext *ctx, struct drm_nouveau_notifier_alloc *,
+			  GLuint subc);
 
 extern GLboolean nouveauSyncInitFuncs(GLcontext *ctx);
 #endif
diff --git a/src/mesa/drivers/dri/nouveau/nv10_state.c b/src/mesa/drivers/dri/nouveau/nv10_state.c
index 5f304ccab9..47c4b14ba6 100644
--- a/src/mesa/drivers/dri/nouveau/nv10_state.c
+++ b/src/mesa/drivers/dri/nouveau/nv10_state.c
@@ -697,8 +697,7 @@ static GLboolean nv10InitCard(nouveauContextPtr nmesa)
 	BEGIN_RING_SIZE(NvSub3D, 0x03f4, 1);
 	OUT_RING(0);
 
-	/* not for nv10, only for >= nv11 */
-	if ((nmesa->screen->card->id>>4) >= 0x11) {
+	if (nmesa->screen->card->type >= NV_11) {
 	        BEGIN_RING_SIZE(NvSub3D, 0x120, 3);
         	OUT_RING(0);
 	        OUT_RING(1);
@@ -739,11 +738,11 @@ static GLboolean nv10BindBuffers(nouveauContextPtr nmesa, int num_color,
 	OUT_RING_CACHE(depth ? depth->offset : color[0]->offset);
 
 	/* Always set to bottom left of buffer */
-	BEGIN_RING_CACHE(NvSub3D, NV10_TCL_PRIMITIVE_3D_VIEWPORT_ORIGIN_X, 4);
+	/*BEGIN_RING_CACHE(NvSub3D, NV10_TCL_PRIMITIVE_3D_VIEWPORT_ORIGIN_X, 4);
 	OUT_RING_CACHEf (0.0);
 	OUT_RING_CACHEf ((GLfloat) h);
 	OUT_RING_CACHEf (0.0);
-	OUT_RING_CACHEf (0.0);
+	OUT_RING_CACHEf (0.0);*/
 
 	return GL_TRUE;
 }
diff --git a/src/mesa/drivers/dri/nouveau/nv10_swtcl.c b/src/mesa/drivers/dri/nouveau/nv10_swtcl.c
index 3bc84d862d..611469b6e4 100644
--- a/src/mesa/drivers/dri/nouveau/nv10_swtcl.c
+++ b/src/mesa/drivers/dri/nouveau/nv10_swtcl.c
@@ -58,7 +58,7 @@ static void nv10ResetLineStipple( GLcontext *ctx );
 
 static inline void nv10StartPrimitive(struct nouveau_context* nmesa,uint32_t primitive,uint32_t size)
 {
-	if (nmesa->screen->card->type==NV_10)
+	if ((nmesa->screen->card->type>=NV_10) && (nmesa->screen->card->type<=NV_17))
 		BEGIN_RING_SIZE(NvSub3D,NV10_TCL_PRIMITIVE_3D_BEGIN_END,1);
 	else if (nmesa->screen->card->type==NV_20)
 		BEGIN_RING_SIZE(NvSub3D,NV20_TCL_PRIMITIVE_3D_BEGIN_END,1);
@@ -66,7 +66,7 @@ static inline void nv10StartPrimitive(struct nouveau_context* nmesa,uint32_t pri
 		BEGIN_RING_SIZE(NvSub3D,NV30_TCL_PRIMITIVE_3D_BEGIN_END,1);
 	OUT_RING(primitive);
 
-	if (nmesa->screen->card->type==NV_10)
+	if ((nmesa->screen->card->type>=NV_10) && (nmesa->screen->card->type<=NV_17))
 		BEGIN_RING_SIZE(NvSub3D,NV10_TCL_PRIMITIVE_3D_VERTEX_ARRAY_DATA|NONINC_METHOD,size);
 	else if (nmesa->screen->card->type==NV_20)
 		BEGIN_RING_SIZE(NvSub3D,NV20_TCL_PRIMITIVE_3D_VERTEX_DATA|NONINC_METHOD,size);
@@ -76,7 +76,7 @@ static inline void nv10StartPrimitive(struct nouveau_context* nmesa,uint32_t pri
 
 inline void nv10FinishPrimitive(struct nouveau_context *nmesa)
 {
-	if (nmesa->screen->card->type==NV_10)
+	if ((nmesa->screen->card->type>=NV_10) && (nmesa->screen->card->type<=NV_17))
 		BEGIN_RING_SIZE(NvSub3D,NV10_TCL_PRIMITIVE_3D_BEGIN_END,1);
 	else if (nmesa->screen->card->type==NV_20)
 		BEGIN_RING_SIZE(NvSub3D,NV20_TCL_PRIMITIVE_3D_BEGIN_END,1);
@@ -392,15 +392,6 @@ static inline void nv10OutputVertexFormat(struct nouveau_context* nmesa)
 	int i;
 	int slots=0;
 	int total_size=0;
-	/* t_vertex_generic dereferences a NULL pointer if we
-	 * pass NULL as the vp transform...
-	 */
-	const GLfloat ident_vp[16] = {
-	   1.0, 0.0, 0.0, 0.0,
-	   0.0, 1.0, 0.0, 0.0,
-	   0.0, 0.0, 1.0, 0.0,
-	   0.0, 0.0, 0.0, 1.0
-	};
 
 	nmesa->vertex_attr_count = 0;
 	RENDERINPUTS_COPY(index, nmesa->render_inputs_bitset);
@@ -431,28 +422,20 @@ static inline void nv10OutputVertexFormat(struct nouveau_context* nmesa)
 		if (RENDERINPUTS_TEST(index, i))
 		{
 			slots=i+1;
-			if (i==_TNL_ATTRIB_POS)
-			{
-				/* special-case POS */
-				EMIT_ATTR(_TNL_ATTRIB_POS,EMIT_3F_VIEWPORT);
-			}
-			else
+			switch(attr_size[i])
 			{
-				switch(attr_size[i])
-				{
-					case 1:
-						EMIT_ATTR(i,EMIT_1F);
-						break;
-					case 2:
-						EMIT_ATTR(i,EMIT_2F);
-						break;
-					case 3:
-						EMIT_ATTR(i,EMIT_3F);
-						break;
-					case 4:
-						EMIT_ATTR(i,EMIT_4F);
-						break;
-				}
+				case 1:
+					EMIT_ATTR(i,EMIT_1F);
+					break;
+				case 2:
+					EMIT_ATTR(i,EMIT_2F);
+					break;
+				case 3:
+					EMIT_ATTR(i,EMIT_3F);
+					break;
+				case 4:
+					EMIT_ATTR(i,EMIT_4F);
+					break;
 			}
 			if (i==_TNL_ATTRIB_COLOR0)
 				nmesa->color_offset=total_size;
@@ -465,13 +448,13 @@ static inline void nv10OutputVertexFormat(struct nouveau_context* nmesa)
 	nmesa->vertex_size=_tnl_install_attrs( ctx,
 			nmesa->vertex_attrs, 
 			nmesa->vertex_attr_count,
-			ident_vp, 0 );
+			NULL, 0 );
 	assert(nmesa->vertex_size==total_size*4);
 
 	/* 
 	 * Tell the hardware about the vertex format
 	 */
-	if (nmesa->screen->card->type==NV_10) {
+	if ((nmesa->screen->card->type>=NV_10) && (nmesa->screen->card->type<=NV_17)) {
 		int size;
 
 #define NV_VERTEX_ATTRIBUTE_TYPE_FLOAT 2
diff --git a/src/mesa/drivers/dri/nouveau/nv20_state.c b/src/mesa/drivers/dri/nouveau/nv20_state.c
index 3d8d83a865..ccf2f6148b 100644
--- a/src/mesa/drivers/dri/nouveau/nv20_state.c
+++ b/src/mesa/drivers/dri/nouveau/nv20_state.c
@@ -568,10 +568,10 @@ static void nv20Scissor(GLcontext *ctx, GLint x, GLint y, GLsizei w, GLsizei h)
 	   y += nmesa->drawY;
 	}
 
-        BEGIN_RING_CACHE(NvSub3D, NV20_TCL_PRIMITIVE_3D_SCISSOR_X2_X1, 1);
+        /*BEGIN_RING_CACHE(NvSub3D, NV20_TCL_PRIMITIVE_3D_SCISSOR_X2_X1, 1);
         OUT_RING_CACHE((w << 16) | x );
         BEGIN_RING_CACHE(NvSub3D, NV20_TCL_PRIMITIVE_3D_SCISSOR_Y2_Y1, 1);
-        OUT_RING_CACHE((h << 16) | y );
+        OUT_RING_CACHE((h << 16) | y );*/
 
 }
 
@@ -764,11 +764,11 @@ static GLboolean nv20BindBuffers(nouveauContextPtr nmesa, int num_color,
 	}
 
 	/* Always set to bottom left of buffer */
-	BEGIN_RING_CACHE(NvSub3D, NV20_TCL_PRIMITIVE_3D_VIEWPORT_ORIGIN_X, 4);
+	/*BEGIN_RING_CACHE(NvSub3D, NV20_TCL_PRIMITIVE_3D_VIEWPORT_ORIGIN_X, 4);
 	OUT_RING_CACHEf (0.0);
 	OUT_RING_CACHEf ((GLfloat) h);
 	OUT_RING_CACHEf (0.0);
-	OUT_RING_CACHEf (0.0);
+	OUT_RING_CACHEf (0.0);*/
 
 	return GL_TRUE;
 }
diff --git a/src/mesa/drivers/dri/nouveau/nv30_state.c b/src/mesa/drivers/dri/nouveau/nv30_state.c
index ad21fa2730..9b010954b3 100644
--- a/src/mesa/drivers/dri/nouveau/nv30_state.c
+++ b/src/mesa/drivers/dri/nouveau/nv30_state.c
@@ -639,25 +639,45 @@ void (*ReadBuffer)( GLcontext *ctx, GLenum buffer );
 /** Set rasterization mode */
 void (*RenderMode)(GLcontext *ctx, GLenum mode );
 
+/* Translate GL coords to window coords, clamping w/h to the
+ * dimensions of the window.
+ */
+static void nv30WindowCoords(nouveauContextPtr nmesa,
+			     GLuint x, GLuint y, GLuint w, GLuint h,
+			     GLuint *wX, GLuint *wY, GLuint *wW, GLuint *wH)
+{
+	if ((x+w) > nmesa->drawW)
+		w = nmesa->drawW - x;
+	(*wX) = x + nmesa->drawX;
+	(*wW) = w;
+
+	if ((y+h) > nmesa->drawH)
+		h = nmesa->drawH - y;
+	(*wY) = (nmesa->drawH - y) - h + nmesa->drawY;
+	(*wH) = h;
+}
+
 /** Define the scissor box */
 static void nv30Scissor(GLcontext *ctx, GLint x, GLint y, GLsizei w, GLsizei h)
 {
         nouveauContextPtr nmesa = NOUVEAU_CONTEXT(ctx);
+	GLuint wX, wY, wW, wH;
 
 	/* There's no scissor enable bit, so adjust the scissor to cover the
 	 * maximum draw buffer bounds
 	 */
 	if (!ctx->Scissor.Enabled) {
-	   x = y = 0;
-	   w = h = 4095;
+	   wX = nmesa->drawX;
+	   wY = nmesa->drawY;
+	   wW = nmesa->drawW;
+	   wH = nmesa->drawH;
 	} else {
-	   x += nmesa->drawX;
-	   y += nmesa->drawY;
+	   nv30WindowCoords(nmesa, x, y, w, h, &wX, &wY, &wW, &wH);
 	}
 
         BEGIN_RING_CACHE(NvSub3D, NV30_TCL_PRIMITIVE_3D_SCISSOR_WIDTH_XPOS, 2);
-        OUT_RING_CACHE(((w) << 16) | x);
-        OUT_RING_CACHE(((h) << 16) | y);
+        OUT_RING_CACHE  ((wW << 16) | wX);
+        OUT_RING_CACHE  ((wH << 16) | wY);
 }
 
 /** Select flat or smooth shading */
@@ -751,19 +771,21 @@ static void nv30WindowMoved(nouveauContextPtr nmesa)
 {
 	GLcontext *ctx = nmesa->glCtx;
 	GLfloat *v = nmesa->viewport.m;
-	GLuint w = ctx->Viewport.Width;
-	GLuint h = ctx->Viewport.Height;
-	GLuint x = ctx->Viewport.X + nmesa->drawX;
-	GLuint y = ctx->Viewport.Y + nmesa->drawY;
+	GLuint wX, wY, wW, wH;
 
+	nv30WindowCoords(nmesa, ctx->Viewport.X, ctx->Viewport.Y, 
+				ctx->Viewport.Width, ctx->Viewport.Height,
+				&wX, &wY, &wW, &wH);
         BEGIN_RING_CACHE(NvSub3D, NV30_TCL_PRIMITIVE_3D_VIEWPORT_DIMS_0, 2);
-        OUT_RING_CACHE((w << 16) | x);
-        OUT_RING_CACHE((h << 16) | y);
+        OUT_RING_CACHE  ((wW << 16) | wX);
+        OUT_RING_CACHE  ((wH << 16) | wY);
+
 	/* something to do with clears, possibly doesn't belong here */
 	BEGIN_RING_CACHE(NvSub3D,
 	      NV30_TCL_PRIMITIVE_3D_VIEWPORT_COLOR_BUFFER_OFS0, 2);
-        OUT_RING_CACHE(((w+x) << 16) | x);
-        OUT_RING_CACHE(((h+y) << 16) | y);
+        OUT_RING_CACHE(((nmesa->drawX + nmesa->drawW) << 16) | nmesa->drawX);
+        OUT_RING_CACHE(((nmesa->drawY + nmesa->drawH) << 16) | nmesa->drawY);
+
 	/* viewport transform */
 	BEGIN_RING_CACHE(NvSub3D, NV30_TCL_PRIMITIVE_3D_VIEWPORT_XFRM_OX, 8);
 	OUT_RING_CACHEf (v[MAT_TX]);
@@ -786,7 +808,7 @@ static GLboolean nv30InitCard(nouveauContextPtr nmesa)
 
 	BEGIN_RING_SIZE(NvSub3D, NV30_TCL_PRIMITIVE_3D_SET_OBJECT1, 3);
 	OUT_RING(NvDmaFB);
-	OUT_RING(NvDmaAGP);
+	OUT_RING(NvDmaTT);
         OUT_RING(NvDmaFB);
 	BEGIN_RING_SIZE(NvSub3D, NV30_TCL_PRIMITIVE_3D_SET_OBJECT8, 1);
 	OUT_RING(NvDmaFB);
diff --git a/src/mesa/drivers/dri/r200/r200_state.c b/src/mesa/drivers/dri/r200/r200_state.c
index 2115799b9b..1d975ecd57 100644
--- a/src/mesa/drivers/dri/r200/r200_state.c
+++ b/src/mesa/drivers/dri/r200/r200_state.c
@@ -772,9 +772,11 @@ static void r200LineWidth( GLcontext *ctx, GLfloat widthf )
    R200_STATECHANGE( rmesa, set );
 
    /* Line width is stored in U6.4 format.
+    * Same min/max limits for AA, non-AA lines.
     */
    rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] &= ~0xffff;
-   rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] |= (GLuint)(ctx->Line._Width * 16.0);
+   rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] |= (GLuint)
+      (CLAMP(widthf, ctx->Const.MinLineWidth, ctx->Const.MaxLineWidth) * 16.0);
 
    if ( widthf > 1.0 ) {
       rmesa->hw.set.cmd[SET_SE_CNTL] |=  R200_WIDELINE_ENABLE;
diff --git a/src/mesa/drivers/dri/r300/Makefile b/src/mesa/drivers/dri/r300/Makefile
index c1d223c760..44248964fd 100644
--- a/src/mesa/drivers/dri/r300/Makefile
+++ b/src/mesa/drivers/dri/r300/Makefile
@@ -41,6 +41,7 @@ DRIVER_SOURCES = \
 		 r300_fragprog.c \
 		 r300_shader.c \
 		 r300_emit.c \
+		 r300_swtcl.c \
 		 $(EGL_SOURCES)
 
 C_SOURCES = $(COMMON_SOURCES) $(DRIVER_SOURCES)
diff --git a/src/mesa/drivers/dri/r300/r300_cmdbuf.c b/src/mesa/drivers/dri/r300/r300_cmdbuf.c
index 7055286ba9..9eca41fa38 100644
--- a/src/mesa/drivers/dri/r300/r300_cmdbuf.c
+++ b/src/mesa/drivers/dri/r300/r300_cmdbuf.c
@@ -321,8 +321,12 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	r300->hw.unk221C.cmd[0] = cmdpacket0(R300_VAP_UNKNOWN_221C, 1);
 	ALLOC_STATE(vap_clip, always, 5, 0);
 	r300->hw.vap_clip.cmd[0] = cmdpacket0(R300_VAP_CLIP_X_0, 4);
-	ALLOC_STATE(unk2288, always, 2, 0);
-	r300->hw.unk2288.cmd[0] = cmdpacket0(R300_VAP_UNKNOWN_2288, 1);
+
+	if (has_tcl) {
+		ALLOC_STATE(unk2288, always, 2, 0);
+		r300->hw.unk2288.cmd[0] = cmdpacket0(R300_VAP_UNKNOWN_2288, 1);
+	}
+
 	ALLOC_STATE(vof, always, R300_VOF_CMDSIZE, 0);
 	r300->hw.vof.cmd[R300_VOF_CMD_0] =
 	    cmdpacket0(R300_VAP_OUTPUT_VTX_FMT_0, 2);
diff --git a/src/mesa/drivers/dri/r300/r300_context.c b/src/mesa/drivers/dri/r300/r300_context.c
index 9ea14ab4c7..14e0f052fd 100644
--- a/src/mesa/drivers/dri/r300/r300_context.c
+++ b/src/mesa/drivers/dri/r300/r300_context.c
@@ -63,6 +63,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "r300_ioctl.h"
 #include "r300_tex.h"
 #include "r300_emit.h"
+#include "r300_swtcl.h"
 
 #ifdef USER_BUFFERS
 #include "r300_mem.h"
@@ -317,15 +318,17 @@ GLboolean r300CreateContext(const __GLcontextModes * glVisual,
 	_tnl_allow_vertex_fog(ctx, GL_TRUE);
 
 	/* currently bogus data */
-	ctx->Const.VertexProgram.MaxInstructions = VSF_MAX_FRAGMENT_LENGTH / 4;
-	ctx->Const.VertexProgram.MaxNativeInstructions =
-	    VSF_MAX_FRAGMENT_LENGTH / 4;
-	ctx->Const.VertexProgram.MaxNativeAttribs = 16;	/* r420 */
-	ctx->Const.VertexProgram.MaxTemps = 32;
-	ctx->Const.VertexProgram.MaxNativeTemps =
-	    /*VSF_MAX_FRAGMENT_TEMPS */ 32;
-	ctx->Const.VertexProgram.MaxNativeParameters = 256;	/* r420 */
-	ctx->Const.VertexProgram.MaxNativeAddressRegs = 1;
+	if (screen->chip_flags & RADEON_CHIPSET_TCL) {
+	        ctx->Const.VertexProgram.MaxInstructions = VSF_MAX_FRAGMENT_LENGTH / 4;
+		ctx->Const.VertexProgram.MaxNativeInstructions =
+		  VSF_MAX_FRAGMENT_LENGTH / 4;
+		ctx->Const.VertexProgram.MaxNativeAttribs = 16;	/* r420 */
+		ctx->Const.VertexProgram.MaxTemps = 32;
+		ctx->Const.VertexProgram.MaxNativeTemps =
+		  /*VSF_MAX_FRAGMENT_TEMPS */ 32;
+		ctx->Const.VertexProgram.MaxNativeParameters = 256;	/* r420 */
+		ctx->Const.VertexProgram.MaxNativeAddressRegs = 1;
+	}
 
 	ctx->Const.FragmentProgram.MaxNativeTemps = PFS_NUM_TEMP_REGS;
 	ctx->Const.FragmentProgram.MaxNativeAttribs = 11;	/* copy i915... */
@@ -363,6 +366,8 @@ GLboolean r300CreateContext(const __GLcontextModes * glVisual,
 	radeonInitSpanFuncs(ctx);
 	r300InitCmdBuf(r300);
 	r300InitState(r300);
+	if (!(screen->chip_flags & RADEON_CHIPSET_TCL))
+	        r300InitSwtcl(ctx);
 
 	TNL_CONTEXT(ctx)->Driver.RunPipeline = _tnl_run_pipeline;
 
diff --git a/src/mesa/drivers/dri/r300/r300_context.h b/src/mesa/drivers/dri/r300/r300_context.h
index 076bb49a00..be6909724a 100644
--- a/src/mesa/drivers/dri/r300/r300_context.h
+++ b/src/mesa/drivers/dri/r300/r300_context.h
@@ -568,38 +568,21 @@ struct r300_vertex_shader_fragment {
 	union {
 		GLuint d[VSF_MAX_FRAGMENT_LENGTH];
 		float f[VSF_MAX_FRAGMENT_LENGTH];
-		VERTEX_SHADER_INSTRUCTION i[VSF_MAX_FRAGMENT_LENGTH / 4];
+		GLuint i[VSF_MAX_FRAGMENT_LENGTH];
 	} body;
 };
 
-#define VSF_DEST_PROGRAM	0x0
-#define VSF_DEST_MATRIX0	0x200
-#define VSF_DEST_MATRIX1	0x204
-#define VSF_DEST_MATRIX2	0x208
-#define VSF_DEST_VECTOR0	0x20c
-#define VSF_DEST_VECTOR1	0x20d
-#define VSF_DEST_UNKNOWN1	0x400
-#define VSF_DEST_UNKNOWN2	0x406
-
 struct r300_vertex_shader_state {
 	struct r300_vertex_shader_fragment program;
-
-	struct r300_vertex_shader_fragment unknown1;
-	struct r300_vertex_shader_fragment unknown2;
-
-	int program_start;
-	int unknown_ptr1;	/* pointer within program space */
-	int program_end;
-
-	int param_offset;
-	int param_count;
-
-	int unknown_ptr2;	/* pointer within program space */
-	int unknown_ptr3;	/* pointer within program space */
 };
 
 extern int hw_tcl_on;
 
+#define COLOR_IS_RGBA
+#define TAG(x) r300##x
+#include "tnl_dd/t_dd_vertex.h"
+#undef TAG
+
 //#define CURRENT_VERTEX_SHADER(ctx) (ctx->VertexProgram._Current)
 #define CURRENT_VERTEX_SHADER(ctx) (R300_CONTEXT(ctx)->selected_vp)
 
@@ -796,13 +779,10 @@ struct r300_state {
 	GLuint *Elts;
 	struct r300_dma_region elt_dma;
 
-	 DECLARE_RENDERINPUTS(render_inputs_bitset);	/* actual render inputs that R300 was configured for.
+	struct r300_dma_region swtcl_dma;
+	DECLARE_RENDERINPUTS(render_inputs_bitset);	/* actual render inputs that R300 was configured for.
 							   They are the same as tnl->render_inputs for fixed pipeline */
 
-	struct {
-		int transform_offset;	/* Transform matrix offset, -1 if none */
-	} vap_param;		/* vertex processor parameter allocation - tells where to write parameters */
-
 	struct r300_stencilbuffer_state stencil;
 
 };
@@ -811,6 +791,62 @@ struct r300_state {
 #define R300_FALLBACK_TCL 1
 #define R300_FALLBACK_RAST 2
 
+/* r300_swtcl.c
+ */
+struct r300_swtcl_info {
+   GLuint RenderIndex;
+   
+   /**
+    * Size of a hardware vertex.  This is calculated when \c ::vertex_attrs is
+    * installed in the Mesa state vector.
+    */
+   GLuint vertex_size;
+
+   /**
+    * Attributes instructing the Mesa TCL pipeline where / how to put vertex
+    * data in the hardware buffer.
+    */
+   struct tnl_attr_map vertex_attrs[VERT_ATTRIB_MAX];
+
+   /**
+    * Number of elements of \c ::vertex_attrs that are actually used.
+    */
+   GLuint vertex_attr_count;
+
+   /**
+    * Cached pointer to the buffer where Mesa will store vertex data.
+    */
+   GLubyte *verts;
+
+   /* Fallback rasterization functions
+    */
+  //   r200_point_func draw_point;
+  //   r200_line_func draw_line;
+  //   r200_tri_func draw_tri;
+
+   GLuint hw_primitive;
+   GLenum render_primitive;
+   GLuint numverts;
+
+   /**
+    * Offset of the 4UB color data within a hardware (swtcl) vertex.
+    */
+   GLuint coloroffset;
+
+   /**
+    * Offset of the 3UB specular color data within a hardware (swtcl) vertex.
+    */
+   GLuint specoffset;
+
+   /**
+    * Should Mesa project vertex data or will the hardware do it?
+    */
+   GLboolean needproj;
+
+   struct r300_dma_region indexed_verts;
+};
+
+
 /**
  * \brief R300 context structure.
  */
@@ -849,6 +885,9 @@ struct r300_context {
 	GLvector4f *temp_attrib[_TNL_ATTRIB_MAX];
 
 	GLboolean disable_lowimpact_fallback;
+
+	DECLARE_RENDERINPUTS(tnl_index_bitset);	/* index of bits for last tnl_install_attrs */
+	struct r300_swtcl_info swtcl;
 };
 
 struct r300_buffer_object {
diff --git a/src/mesa/drivers/dri/r300/r300_emit.c b/src/mesa/drivers/dri/r300/r300_emit.c
index 4670c28a02..424bf44e59 100644
--- a/src/mesa/drivers/dri/r300/r300_emit.c
+++ b/src/mesa/drivers/dri/r300/r300_emit.c
@@ -217,14 +217,14 @@ static GLuint r300VAPInputRoute0(uint32_t * dst, GLvector4f ** attribptr,
 		dw = R300_INPUT_ROUTE_FLOAT | (inputs[tab[i]] << 8) | (attribptr[tab[i]]->size - 1);
 		dw |= (R300_INPUT_ROUTE_FLOAT | (inputs[tab[i + 1]] << 8) | (attribptr[tab[i + 1]]->size - 1)) << 16;
 		if (i + 2 == nr) {
-			dw |= (1 << (13 + 16));
+			dw |= (R300_VAP_INPUT_ROUTE_END << 16);
 		}
 		dst[i >> 1] = dw;
 	}
 
 	if (nr & 1) {
 		dw = R300_INPUT_ROUTE_FLOAT | (inputs[tab[nr - 1]] << 8) | (attribptr[tab[nr - 1]]->size - 1);
-		dw |= 1 << 13;
+		dw |= R300_VAP_INPUT_ROUTE_END;
 		dst[nr >> 1] = dw;
 	}
 
@@ -239,7 +239,7 @@ static GLuint r300VAPInputRoute1Swizzle(int swizzle[4])
 	    (swizzle[3] << R300_INPUT_ROUTE_W_SHIFT);
 }
 
-static GLuint r300VAPInputRoute1(uint32_t * dst, int swizzle[][4], GLuint nr)
+GLuint r300VAPInputRoute1(uint32_t * dst, int swizzle[][4], GLuint nr)
 {
 	GLuint i;
 
@@ -255,14 +255,14 @@ static GLuint r300VAPInputRoute1(uint32_t * dst, int swizzle[][4], GLuint nr)
 	return (nr + 1) >> 1;
 }
 
-static GLuint r300VAPInputCntl0(GLcontext * ctx, GLuint InputsRead)
+GLuint r300VAPInputCntl0(GLcontext * ctx, GLuint InputsRead)
 {
 	/* No idea what this value means. I have seen other values written to
 	 * this register... */
 	return 0x5555;
 }
 
-static GLuint r300VAPInputCntl1(GLcontext * ctx, GLuint InputsRead)
+GLuint r300VAPInputCntl1(GLcontext * ctx, GLuint InputsRead)
 {
 	r300ContextPtr rmesa = R300_CONTEXT(ctx);
 	GLuint i, vic_1 = 0;
@@ -286,7 +286,7 @@ static GLuint r300VAPInputCntl1(GLcontext * ctx, GLuint InputsRead)
 	return vic_1;
 }
 
-static GLuint r300VAPOutputCntl0(GLcontext * ctx, GLuint OutputsWritten)
+GLuint r300VAPOutputCntl0(GLcontext * ctx, GLuint OutputsWritten)
 {
 	GLuint ret = 0;
 
@@ -299,13 +299,14 @@ static GLuint r300VAPOutputCntl0(GLcontext * ctx, GLuint OutputsWritten)
 	if (OutputsWritten & (1 << VERT_RESULT_COL1))
 		ret |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_1_PRESENT;
 
-#if 0
-	if (OutputsWritten & (1 << VERT_RESULT_BFC0))
-		ret |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_2_PRESENT;
-
-	if (OutputsWritten & (1 << VERT_RESULT_BFC1))
-		ret |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_3_PRESENT;
+	if (OutputsWritten & (1 << VERT_RESULT_BFC0)
+	    || OutputsWritten & (1 << VERT_RESULT_BFC1))
+		ret |=
+		    R300_VAP_OUTPUT_VTX_FMT_0__COLOR_1_PRESENT |
+		    R300_VAP_OUTPUT_VTX_FMT_0__COLOR_2_PRESENT |
+		    R300_VAP_OUTPUT_VTX_FMT_0__COLOR_3_PRESENT;
 
+#if 0
 	if (OutputsWritten & (1 << VERT_RESULT_FOGC)) ;
 #endif
 
@@ -315,7 +316,7 @@ static GLuint r300VAPOutputCntl0(GLcontext * ctx, GLuint OutputsWritten)
 	return ret;
 }
 
-static GLuint r300VAPOutputCntl1(GLcontext * ctx, GLuint OutputsWritten)
+GLuint r300VAPOutputCntl1(GLcontext * ctx, GLuint OutputsWritten)
 {
 	GLuint i, ret = 0;
 
@@ -358,9 +359,11 @@ int r300EmitArrays(GLcontext * ctx)
 		DECLARE_RENDERINPUTS(render_inputs_bitset);
 		RENDERINPUTS_COPY(render_inputs_bitset, tnl->render_inputs_bitset);
 
+		vb->AttribPtr[VERT_ATTRIB_POS] = vb->ClipPtr;
+
 		assert(RENDERINPUTS_TEST(render_inputs_bitset, _TNL_ATTRIB_POS));
 		assert(RENDERINPUTS_TEST(render_inputs_bitset, _TNL_ATTRIB_NORMAL) == 0);
-		assert(RENDERINPUTS_TEST(render_inputs_bitset, _TNL_ATTRIB_COLOR0));
+		//assert(RENDERINPUTS_TEST(render_inputs_bitset, _TNL_ATTRIB_COLOR0));
 
 		if (RENDERINPUTS_TEST(render_inputs_bitset, _TNL_ATTRIB_POS)) {
 			InputsRead |= 1 << VERT_ATTRIB_POS;
@@ -392,20 +395,18 @@ int r300EmitArrays(GLcontext * ctx)
 			}
 		}
 
-		if (!(rmesa->radeon.radeonScreen->chip_flags & RADEON_CHIPSET_TCL)) {
-			/* Fixed, apply to vir0 only */
-			memcpy(vir_inputs, inputs, VERT_ATTRIB_MAX * sizeof(int));
-			inputs = vir_inputs;
-			if (InputsRead & VERT_ATTRIB_POS)
-				inputs[VERT_ATTRIB_POS] = 0;
-			if (InputsRead & (1 << VERT_ATTRIB_COLOR0))
-				inputs[VERT_ATTRIB_COLOR0] = 2;
-			if (InputsRead & (1 << VERT_ATTRIB_COLOR1))
-				inputs[VERT_ATTRIB_COLOR1] = 3;
-			for (i = VERT_ATTRIB_TEX0; i <= VERT_ATTRIB_TEX7; i++)
-				if (InputsRead & (1 << i))
-					inputs[i] = 6 + (i - VERT_ATTRIB_TEX0);
-		}
+		/* Fixed, apply to vir0 only */
+		memcpy(vir_inputs, inputs, VERT_ATTRIB_MAX * sizeof(int));
+		inputs = vir_inputs;
+		if (InputsRead & VERT_ATTRIB_POS)
+			inputs[VERT_ATTRIB_POS] = 0;
+		if (InputsRead & (1 << VERT_ATTRIB_COLOR0))
+			inputs[VERT_ATTRIB_COLOR0] = 2;
+		if (InputsRead & (1 << VERT_ATTRIB_COLOR1))
+			inputs[VERT_ATTRIB_COLOR1] = 3;
+		for (i = VERT_ATTRIB_TEX0; i <= VERT_ATTRIB_TEX7; i++)
+			if (InputsRead & (1 << i))
+				inputs[i] = 6 + (i - VERT_ATTRIB_TEX0);
 
 		RENDERINPUTS_COPY(rmesa->state.render_inputs_bitset, render_inputs_bitset);
 	}
@@ -532,3 +533,19 @@ void r300ReleaseArrays(GLcontext * ctx)
 		r300ReleaseDmaRegion(rmesa, &rmesa->state.aos[i], __FUNCTION__);
 	}
 }
+
+void r300EmitCacheFlush(r300ContextPtr rmesa)
+{
+        int cmd_reserved = 0;
+	int cmd_written = 0;
+
+	drm_radeon_cmd_header_t *cmd = NULL;
+
+	reg_start(R300_RB3D_DSTCACHE_CTLSTAT, 0);
+	e32(R300_RB3D_DSTCACHE_UNKNOWN_0A);
+
+	reg_start(R300_RB3D_ZCACHE_CTLSTAT, 0);
+	e32(R300_RB3D_ZCACHE_UNKNOWN_03);
+
+
+}
diff --git a/src/mesa/drivers/dri/r300/r300_emit.h b/src/mesa/drivers/dri/r300/r300_emit.h
index 2f79ee3a23..a6d69ec5ff 100644
--- a/src/mesa/drivers/dri/r300/r300_emit.h
+++ b/src/mesa/drivers/dri/r300/r300_emit.h
@@ -225,5 +225,15 @@ void r300UseArrays(GLcontext * ctx);
 #endif
 
 extern void r300ReleaseArrays(GLcontext * ctx);
+extern int r300PrimitiveType(r300ContextPtr rmesa, int prim);
+extern int r300NumVerts(r300ContextPtr rmesa, int num_verts, int prim);
+
+extern void r300EmitCacheFlush(r300ContextPtr rmesa);
+
+extern GLuint r300VAPInputRoute1(uint32_t * dst, int swizzle[][4], GLuint nr);
+extern GLuint r300VAPInputCntl0(GLcontext * ctx, GLuint InputsRead);
+extern GLuint r300VAPInputCntl1(GLcontext * ctx, GLuint InputsRead);
+extern GLuint r300VAPOutputCntl0(GLcontext * ctx, GLuint OutputsWritten);
+extern GLuint r300VAPOutputCntl1(GLcontext * ctx, GLuint OutputsWritten);
 
 #endif
diff --git a/src/mesa/drivers/dri/r300/r300_ioctl.c b/src/mesa/drivers/dri/r300/r300_ioctl.c
index 15c2cf3ad7..90f5027c9a 100644
--- a/src/mesa/drivers/dri/r300/r300_ioctl.c
+++ b/src/mesa/drivers/dri/r300/r300_ioctl.c
@@ -172,11 +172,7 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 	cmd2[7].u = r300PackFloat32(ctx->Color.ClearColor[2]);
 	cmd2[8].u = r300PackFloat32(ctx->Color.ClearColor[3]);
 
-	reg_start(R300_RB3D_DSTCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_DSTCACHE_UNKNOWN_0A);
-
-	reg_start(R300_RB3D_ZCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_ZCACHE_UNKNOWN_03);
+	r300EmitCacheFlush(rmesa);
 	cp_wait(rmesa, R300_WAIT_3D | R300_WAIT_3D_CLEAN);
 }
 
@@ -412,19 +408,22 @@ static void r300Clear(GLcontext * ctx, GLbitfield mask)
 
 void r300Flush(GLcontext * ctx)
 {
-	r300ContextPtr r300 = R300_CONTEXT(ctx);
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
 
 	if (RADEON_DEBUG & DEBUG_IOCTL)
 		fprintf(stderr, "%s\n", __FUNCTION__);
 
-	if (r300->cmdbuf.count_used > r300->cmdbuf.count_reemit)
-		r300FlushCmdBuf(r300, __FUNCTION__);
+	if (rmesa->dma.flush)
+		rmesa->dma.flush( rmesa );
+
+	if (rmesa->cmdbuf.count_used > rmesa->cmdbuf.count_reemit)
+		r300FlushCmdBuf(rmesa, __FUNCTION__);
 }
 
 #ifdef USER_BUFFERS
 #include "r300_mem.h"
 
-static void r300RefillCurrentDmaRegion(r300ContextPtr rmesa, int size)
+void r300RefillCurrentDmaRegion(r300ContextPtr rmesa, int size)
 {
 	struct r300_dma_buffer *dmabuf;
 	size = MAX2(size, RADEON_BUFFER_SIZE * 16);
@@ -436,9 +435,12 @@ static void r300RefillCurrentDmaRegion(r300ContextPtr rmesa, int size)
 		rmesa->dma.flush(rmesa);
 	}
 
-	if (rmesa->dma.current.buf)
+	if (rmesa->dma.current.buf) {
+#ifdef USER_BUFFERS
+		r300_mem_use(rmesa, rmesa->dma.current.buf->id);
+#endif
 		r300ReleaseDmaRegion(rmesa, &rmesa->dma.current, __FUNCTION__);
-
+	}
 	if (rmesa->dma.nr_released_bufs > 4)
 		r300FlushCmdBuf(rmesa, __FUNCTION__);
 
diff --git a/src/mesa/drivers/dri/r300/r300_ioctl.h b/src/mesa/drivers/dri/r300/r300_ioctl.h
index 7a19a2cf3f..e1143fb6c3 100644
--- a/src/mesa/drivers/dri/r300/r300_ioctl.h
+++ b/src/mesa/drivers/dri/r300/r300_ioctl.h
@@ -56,4 +56,5 @@ extern void r300AllocDmaRegion(r300ContextPtr rmesa,
 
 extern void r300InitIoctlFuncs(struct dd_function_table *functions);
 
+extern void r300RefillCurrentDmaRegion(r300ContextPtr rmesa, int size);
 #endif				/* __R300_IOCTL_H__ */
diff --git a/src/mesa/drivers/dri/r300/r300_reg.h b/src/mesa/drivers/dri/r300/r300_reg.h
index 3ce09c16d3..1baa74c526 100644
--- a/src/mesa/drivers/dri/r300/r300_reg.h
+++ b/src/mesa/drivers/dri/r300/r300_reg.h
@@ -282,9 +282,32 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 #define R300_VAP_PVS_UPLOAD_ADDRESS         0x2200
 #       define R300_PVS_UPLOAD_PROGRAM           0x00000000
+/* gap */
 #       define R300_PVS_UPLOAD_PARAMETERS        0x00000200
+/* gap */
+#       define R300_PVS_UPLOAD_CLIP_PLANE0       0x00000400
+#       define R300_PVS_UPLOAD_CLIP_PLANE1       0x00000401
+#       define R300_PVS_UPLOAD_CLIP_PLANE2       0x00000402
+#       define R300_PVS_UPLOAD_CLIP_PLANE3       0x00000403
+#       define R300_PVS_UPLOAD_CLIP_PLANE4       0x00000404
+#       define R300_PVS_UPLOAD_CLIP_PLANE5       0x00000405
 #       define R300_PVS_UPLOAD_POINTSIZE         0x00000406
 
+/*
+ * These are obsolete defines form r300_context.h, but they might give some
+ * clues when investigating the addresses further...
+ */
+#if 0
+#define VSF_DEST_PROGRAM        0x0
+#define VSF_DEST_MATRIX0        0x200
+#define VSF_DEST_MATRIX1        0x204
+#define VSF_DEST_MATRIX2        0x208
+#define VSF_DEST_VECTOR0        0x20c
+#define VSF_DEST_VECTOR1        0x20d
+#define VSF_DEST_UNKNOWN1       0x400
+#define VSF_DEST_UNKNOWN2       0x406
+#endif
+
 /* gap */
 
 #define R300_VAP_PVS_UPLOAD_DATA            0x2208
@@ -336,13 +359,15 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
  * The meaning of the two UNKNOWN fields is obviously not known. However,
  * experiments so far have shown that both *must* point to an instruction
  * inside the vertex program, otherwise the GPU locks up.
+ *
  * fglrx usually sets CNTL_3_UNKNOWN to the end of the program and
- * CNTL_1_UNKNOWN points to instruction where last write to position takes
- * place.
+ * R300_PVS_CNTL_1_POS_END_SHIFT points to instruction where last write to
+ * position takes place.
+ *
  * Most likely this is used to ignore rest of the program in cases
  * where group of verts arent visible. For some reason this "section"
  * is sometimes accepted other instruction that have no relationship with
- *position calculations.
+ * position calculations.
  */
 #define R300_VAP_PVS_CNTL_1                 0x22D0
 #       define R300_PVS_CNTL_1_PROGRAM_START_SHIFT   0
diff --git a/src/mesa/drivers/dri/r300/r300_render.c b/src/mesa/drivers/dri/r300/r300_render.c
index 1aaf43bfaf..eee1e803a0 100644
--- a/src/mesa/drivers/dri/r300/r300_render.c
+++ b/src/mesa/drivers/dri/r300/r300_render.c
@@ -79,7 +79,7 @@ extern int future_hw_tcl_on;
 /**
  * \brief Convert a OpenGL primitive type into a R300 primitive type.
  */
-static int r300PrimitiveType(r300ContextPtr rmesa, GLcontext * ctx, int prim)
+int r300PrimitiveType(r300ContextPtr rmesa, int prim)
 {
 	switch (prim & PRIM_MODE_MASK) {
 	case GL_POINTS:
@@ -119,7 +119,7 @@ static int r300PrimitiveType(r300ContextPtr rmesa, GLcontext * ctx, int prim)
 	}
 }
 
-static int r300NumVerts(r300ContextPtr rmesa, int num_verts, int prim)
+int r300NumVerts(r300ContextPtr rmesa, int num_verts, int prim)
 {
 	int verts_off = 0;
 
@@ -261,7 +261,7 @@ static void r300RunRenderPrimitive(r300ContextPtr rmesa, GLcontext * ctx,
 	TNLcontext *tnl = TNL_CONTEXT(ctx);
 	struct vertex_buffer *vb = &tnl->vb;
 
-	type = r300PrimitiveType(rmesa, ctx, prim);
+	type = r300PrimitiveType(rmesa, prim);
 	num_verts = r300NumVerts(rmesa, end - start, prim);
 
 	if (type < 0 || num_verts <= 0)
@@ -287,29 +287,20 @@ static GLboolean r300RunRender(GLcontext * ctx,
 {
 	r300ContextPtr rmesa = R300_CONTEXT(ctx);
 	int i;
-	int cmd_reserved = 0;
-	int cmd_written = 0;
-	drm_radeon_cmd_header_t *cmd = NULL;
 	TNLcontext *tnl = TNL_CONTEXT(ctx);
 	struct vertex_buffer *vb = &tnl->vb;
 
+
 	if (RADEON_DEBUG & DEBUG_PRIMS)
 		fprintf(stderr, "%s\n", __FUNCTION__);
 
-	if (hw_tcl_on == GL_FALSE)
-	  vb->AttribPtr[VERT_ATTRIB_POS] = vb->ClipPtr;
 	r300UpdateShaders(rmesa);
 	if (r300EmitArrays(ctx))
 		return GL_TRUE;
 
 	r300UpdateShaderStates(rmesa);
 
-	reg_start(R300_RB3D_DSTCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_DSTCACHE_UNKNOWN_0A);
-
-	reg_start(R300_RB3D_ZCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_ZCACHE_UNKNOWN_03);
-
+	r300EmitCacheFlush(rmesa);
 	r300EmitState(rmesa);
 
 	for (i = 0; i < vb->PrimitiveCount; i++) {
@@ -319,11 +310,7 @@ static GLboolean r300RunRender(GLcontext * ctx,
 		r300RunRenderPrimitive(rmesa, ctx, start, end, prim);
 	}
 
-	reg_start(R300_RB3D_DSTCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_DSTCACHE_UNKNOWN_0A);
-
-	reg_start(R300_RB3D_ZCACHE_CTLSTAT, 0);
-	e32(R300_RB3D_ZCACHE_UNKNOWN_03);
+	r300EmitCacheFlush(rmesa);
 
 #ifdef USER_BUFFERS
 	r300UseArrays(ctx);
@@ -371,8 +358,6 @@ static int r300Fallback(GLcontext * ctx)
 		FALLBACK_IF(ctx->Point.PointSprite);
 
 	if (!r300->disable_lowimpact_fallback) {
-		FALLBACK_IF(ctx->Polygon.OffsetPoint);
-		FALLBACK_IF(ctx->Polygon.OffsetLine);
 		FALLBACK_IF(ctx->Polygon.StippleFlag);
 		FALLBACK_IF(ctx->Multisample.Enabled);
 		FALLBACK_IF(ctx->Line.StippleFlag);
@@ -386,12 +371,17 @@ static int r300Fallback(GLcontext * ctx)
 static GLboolean r300RunNonTCLRender(GLcontext * ctx,
 				     struct tnl_pipeline_stage *stage)
 {
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
+
 	if (RADEON_DEBUG & DEBUG_PRIMS)
 		fprintf(stderr, "%s\n", __FUNCTION__);
 
 	if (r300Fallback(ctx) >= R300_FALLBACK_RAST)
 		return GL_TRUE;
 
+	if (!(rmesa->radeon.radeonScreen->chip_flags & RADEON_CHIPSET_TCL))
+ 	        return GL_TRUE;
+
 	return r300RunRender(ctx, stage);
 }
 
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index e616326274..088216c76e 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -82,6 +82,8 @@ static void r300BlendColor(GLcontext * ctx, const GLfloat cf[4])
 
 	rmesa->hw.blend_color.cmd[1] = PACK_COLOR_8888(color[3], color[0],
 						       color[1], color[2]);
+	rmesa->hw.blend_color.cmd[2] = 0;
+	rmesa->hw.blend_color.cmd[3] = 0;
 }
 
 /**
@@ -317,20 +319,34 @@ static void r300UpdateCulling(GLcontext * ctx)
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
 	uint32_t val = 0;
 
-	R300_STATECHANGE(r300, cul);
 	if (ctx->Polygon.CullFlag) {
-		if (ctx->Polygon.CullFaceMode == GL_FRONT_AND_BACK)
-			val = R300_CULL_FRONT | R300_CULL_BACK;
-		else if (ctx->Polygon.CullFaceMode == GL_FRONT)
+		switch (ctx->Polygon.CullFaceMode) {
+		case GL_FRONT:
 			val = R300_CULL_FRONT;
-		else
+			break;
+		case GL_BACK:
 			val = R300_CULL_BACK;
+			break;
+		case GL_FRONT_AND_BACK:
+			val = R300_CULL_FRONT | R300_CULL_BACK;
+			break;
+		default:
+			break;
+		}
+	}
 
-		if (ctx->Polygon.FrontFace == GL_CW)
-			val |= R300_FRONT_FACE_CW;
-		else
-			val |= R300_FRONT_FACE_CCW;
+	switch (ctx->Polygon.FrontFace) {
+	case GL_CW:
+		val |= R300_FRONT_FACE_CW;
+		break;
+	case GL_CCW:
+		val |= R300_FRONT_FACE_CCW;
+		break;
+	default:
+		break;
 	}
+
+	R300_STATECHANGE(r300, cul);
 	r300->hw.cul.cmd[R300_CUL_CULL] = val;
 }
 
@@ -344,6 +360,20 @@ static void r300SetEarlyZState(GLcontext * ctx)
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
 
 	R300_STATECHANGE(r300, zstencil_format);
+	switch (ctx->Visual.depthBits) {
+	case 16:
+		r300->hw.zstencil_format.cmd[1] = R300_DEPTH_FORMAT_16BIT_INT_Z;
+		break;
+	case 24:
+		r300->hw.zstencil_format.cmd[1] = R300_DEPTH_FORMAT_24BIT_INT_Z;
+		break;
+	default:
+		fprintf(stderr, "Error: Unsupported depth %d... exiting\n", ctx->Visual.depthBits);
+		_mesa_exit(-1);
+	}
+
+	// r300->hw.zstencil_format.cmd[1] |= R300_DEPTH_FORMAT_UNK32;
+
 	if (ctx->Color.AlphaEnabled && ctx->Color.AlphaFunc != GL_ALWAYS)
 		/* disable early Z */
 		r300->hw.zstencil_format.cmd[2] = R300_EARLY_Z_DISABLE;
@@ -355,6 +385,9 @@ static void r300SetEarlyZState(GLcontext * ctx)
 			/* disable early Z */
 			r300->hw.zstencil_format.cmd[2] = R300_EARLY_Z_DISABLE;
 	}
+
+	r300->hw.zstencil_format.cmd[3] = 0x00000003;
+	r300->hw.zstencil_format.cmd[4] = 0x00000000;
 }
 
 static void r300SetAlphaState(GLcontext * ctx)
@@ -403,6 +436,7 @@ static void r300SetAlphaState(GLcontext * ctx)
 
 	R300_STATECHANGE(r300, at);
 	r300->hw.at.cmd[R300_AT_ALPHA_TEST] = pp_misc;
+	r300->hw.at.cmd[R300_AT_UNKNOWN] = 0;
 
 	r300SetEarlyZState(ctx);
 }
@@ -513,6 +547,9 @@ static void r300UpdatePolygonMode(GLcontext * ctx)
 		R300_STATECHANGE(r300, polygon_mode);
 		r300->hw.polygon_mode.cmd[1] = hw_mode;
 	}
+
+	r300->hw.polygon_mode.cmd[2] = 0x00000001;
+	r300->hw.polygon_mode.cmd[3] = 0x00000000;
 }
 
 /**
@@ -696,8 +733,8 @@ static void r300Fogfv(GLcontext * ctx, GLenum pname, const GLfloat * param)
 static void r300PointSize(GLcontext * ctx, GLfloat size)
 {
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
-
-	size = ctx->Point._Size;
+        /* same size limits for AA, non-AA points */
+	size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize);
 
 	R300_STATECHANGE(r300, ps);
 	r300->hw.ps.cmd[R300_PS_POINTSIZE] =
@@ -712,10 +749,11 @@ static void r300LineWidth(GLcontext * ctx, GLfloat widthf)
 {
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
 
-	widthf = ctx->Line._Width;
-
+	widthf = CLAMP(widthf,
+                       ctx->Const.MinPointSize,
+                       ctx->Const.MaxPointSize);
 	R300_STATECHANGE(r300, lcntl);
-	r300->hw.lcntl.cmd[1] |=
+	r300->hw.lcntl.cmd[1] =
 	    R300_LINE_CNT_HO | R300_LINE_CNT_VE | (int)(widthf * 6.0);
 }
 
@@ -762,6 +800,7 @@ static void r300ShadeModel(GLcontext * ctx, GLenum mode)
 	r300ContextPtr rmesa = R300_CONTEXT(ctx);
 
 	R300_STATECHANGE(rmesa, shade);
+	rmesa->hw.shade.cmd[1] = 0x00000002;
 	switch (mode) {
 	case GL_FLAT:
 		rmesa->hw.shade.cmd[2] = R300_RE_SHADE_MODEL_FLAT;
@@ -772,6 +811,8 @@ static void r300ShadeModel(GLcontext * ctx, GLenum mode)
 	default:
 		return;
 	}
+	rmesa->hw.shade.cmd[3] = 0x00000000;
+	rmesa->hw.shade.cmd[4] = 0x00000000;
 }
 
 static void r300StencilFuncSeparate(GLcontext * ctx, GLenum face,
@@ -1477,9 +1518,7 @@ static void r300SetupRSUnit(GLcontext * ctx)
 	if(_nc>_p->vpu.count)_p->vpu.count=_nc;\
 	}while(0)
 
-void static inline setup_vertex_shader_fragment(r300ContextPtr r300, int dest, struct
-						r300_vertex_shader_fragment
-						*vsf)
+static inline void r300SetupVertexProgramFragment(r300ContextPtr r300, int dest, struct r300_vertex_shader_fragment *vsf)
 {
 	int i;
 
@@ -1487,8 +1526,7 @@ void static inline setup_vertex_shader_fragment(r300ContextPtr r300, int dest, s
 		return;
 
 	if (vsf->length & 0x3) {
-		fprintf(stderr,
-			"VERTEX_SHADER_FRAGMENT must have length divisible by 4\n");
+		fprintf(stderr, "VERTEX_SHADER_FRAGMENT must have length divisible by 4\n");
 		_mesa_exit(-1);
 	}
 
@@ -1496,147 +1534,101 @@ void static inline setup_vertex_shader_fragment(r300ContextPtr r300, int dest, s
 	case 0:
 		R300_STATECHANGE(r300, vpi);
 		for (i = 0; i < vsf->length; i++)
-			r300->hw.vpi.cmd[R300_VPI_INSTR_0 + i +
-					 4 * (dest & 0xff)] = (vsf->body.d[i]);
-		bump_vpu_count(r300->hw.vpi.cmd,
-			       vsf->length + 4 * (dest & 0xff));
+			r300->hw.vpi.cmd[R300_VPI_INSTR_0 + i + 4 * (dest & 0xff)] = (vsf->body.d[i]);
+		bump_vpu_count(r300->hw.vpi.cmd, vsf->length + 4 * (dest & 0xff));
 		break;
 
 	case 2:
 		R300_STATECHANGE(r300, vpp);
 		for (i = 0; i < vsf->length; i++)
-			r300->hw.vpp.cmd[R300_VPP_PARAM_0 + i +
-					 4 * (dest & 0xff)] = (vsf->body.d[i]);
-		bump_vpu_count(r300->hw.vpp.cmd,
-			       vsf->length + 4 * (dest & 0xff));
+			r300->hw.vpp.cmd[R300_VPP_PARAM_0 + i + 4 * (dest & 0xff)] = (vsf->body.d[i]);
+		bump_vpu_count(r300->hw.vpp.cmd, vsf->length + 4 * (dest & 0xff));
 		break;
 	case 4:
 		R300_STATECHANGE(r300, vps);
 		for (i = 0; i < vsf->length; i++)
-			r300->hw.vps.cmd[1 + i + 4 * (dest & 0xff)] =
-			    (vsf->body.d[i]);
-		bump_vpu_count(r300->hw.vps.cmd,
-			       vsf->length + 4 * (dest & 0xff));
+			r300->hw.vps.cmd[1 + i + 4 * (dest & 0xff)] = (vsf->body.d[i]);
+		bump_vpu_count(r300->hw.vps.cmd, vsf->length + 4 * (dest & 0xff));
 		break;
 	default:
-		fprintf(stderr,
-			"%s:%s don't know how to handle dest %04x\n",
-			__FILE__, __FUNCTION__, dest);
+		fprintf(stderr, "%s:%s don't know how to handle dest %04x\n", __FILE__, __FUNCTION__, dest);
 		_mesa_exit(-1);
 	}
 }
 
-/* just a skeleton for now.. */
-
-/* Generate a vertex shader that simply transforms vertex and texture coordinates,
-   while leaving colors intact. Nothing fancy (like lights)
-
-   If implementing lights make a copy first, so it is easy to switch between the two versions */
-static void r300GenerateSimpleVertexShader(r300ContextPtr r300)
+static void r300SetupDefaultVertexProgram(r300ContextPtr rmesa)
 {
-	int i;
+	struct r300_vertex_shader_state *prog = &(rmesa->state.vertex_shader);
 	GLuint o_reg = 0;
-
-	/* Allocate parameters */
-	r300->state.vap_param.transform_offset = 0x0;	/* transform matrix */
-	r300->state.vertex_shader.param_offset = 0x0;
-	r300->state.vertex_shader.param_count = 0x4;	/* 4 vector values - 4x4 matrix */
-
-	r300->state.vertex_shader.program_start = 0x0;
-	r300->state.vertex_shader.unknown_ptr1 = 0x4;	/* magic value ? */
-	r300->state.vertex_shader.program_end = 0x0;
-
-	r300->state.vertex_shader.unknown_ptr2 = 0x0;	/* magic value */
-	r300->state.vertex_shader.unknown_ptr3 = 0x4;	/* magic value */
-
-	r300->state.vertex_shader.unknown1.length = 0;
-	r300->state.vertex_shader.unknown2.length = 0;
-
-#define WRITE_OP(oper,source1,source2,source3)	{\
-	r300->state.vertex_shader.program.body.i[r300->state.vertex_shader.program_end].op=(oper); \
-	r300->state.vertex_shader.program.body.i[r300->state.vertex_shader.program_end].src[0]=(source1); \
-	r300->state.vertex_shader.program.body.i[r300->state.vertex_shader.program_end].src[1]=(source2); \
-	r300->state.vertex_shader.program.body.i[r300->state.vertex_shader.program_end].src[2]=(source3); \
-	r300->state.vertex_shader.program_end++; \
-	}
-
-	for (i = VERT_ATTRIB_POS; i < VERT_ATTRIB_MAX; i++)
-		if (r300->state.sw_tcl_inputs[i] != -1) {
-			WRITE_OP(EASY_VSF_OP(MUL, o_reg++, ALL, RESULT),
-				 VSF_REG(r300->state.sw_tcl_inputs[i]),
-				 VSF_ATTR_UNITY(r300->state.
-						sw_tcl_inputs[i]),
-				 VSF_UNITY(r300->state.sw_tcl_inputs[i])
-			    )
-
+	int i;
+	int inst_count = 0;
+	int param_count = 0;
+	int program_end = 0;
+
+	for (i = VERT_ATTRIB_POS; i < VERT_ATTRIB_MAX; i++) {
+		if (rmesa->state.sw_tcl_inputs[i] != -1) {
+			prog->program.body.i[program_end + 0] = EASY_VSF_OP(MUL, o_reg++, ALL, RESULT);
+			prog->program.body.i[program_end + 1] = VSF_REG(rmesa->state.sw_tcl_inputs[i]);
+			prog->program.body.i[program_end + 2] = VSF_ATTR_UNITY(rmesa->state.sw_tcl_inputs[i]);
+			prog->program.body.i[program_end + 3] = VSF_UNITY(rmesa->state.sw_tcl_inputs[i]);
+			program_end += 4;
 		}
+	}
 
-	r300->state.vertex_shader.program_end--;	/* r300 wants program length to be one more - no idea why */
-	r300->state.vertex_shader.program.length =
-	    (r300->state.vertex_shader.program_end + 1) * 4;
+	prog->program.length = program_end;
 
-	r300->state.vertex_shader.unknown_ptr1 = r300->state.vertex_shader.program_end;	/* magic value ? */
-	r300->state.vertex_shader.unknown_ptr2 = r300->state.vertex_shader.program_end;	/* magic value ? */
-	r300->state.vertex_shader.unknown_ptr3 = r300->state.vertex_shader.program_end;	/* magic value ? */
+	r300SetupVertexProgramFragment(rmesa, R300_PVS_UPLOAD_PROGRAM,
+				       &(prog->program));
+	inst_count = (prog->program.length / 4) - 1;
 
+	R300_STATECHANGE(rmesa, pvs);
+	rmesa->hw.pvs.cmd[R300_PVS_CNTL_1] =
+	    (0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT) |
+	    (inst_count << R300_PVS_CNTL_1_POS_END_SHIFT) |
+	    (inst_count << R300_PVS_CNTL_1_PROGRAM_END_SHIFT);
+	rmesa->hw.pvs.cmd[R300_PVS_CNTL_2] =
+	    (0 << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT) |
+	    (param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT);
+	rmesa->hw.pvs.cmd[R300_PVS_CNTL_3] =
+	    (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT) |
+	    (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN2_SHIFT);
 }
 
-static void r300SetupVertexProgram(r300ContextPtr rmesa)
+static void r300SetupRealVertexProgram(r300ContextPtr rmesa)
 {
 	GLcontext *ctx = rmesa->radeon.glCtx;
-	int inst_count;
-	int param_count;
-	struct r300_vertex_program *prog =
-	    (struct r300_vertex_program *)CURRENT_VERTEX_SHADER(ctx);
+	struct r300_vertex_program *prog = (struct r300_vertex_program *)CURRENT_VERTEX_SHADER(ctx);
+	int inst_count = 0;
+	int param_count = 0;
 
-	((drm_r300_cmd_header_t *) rmesa->hw.vpp.cmd)->vpu.count = 0;
+	/* FIXME: r300SetupVertexProgramFragment */
 	R300_STATECHANGE(rmesa, vpp);
 	param_count =
-	    r300VertexProgUpdateParams(ctx, (struct r300_vertex_program_cont *)
-				       ctx->VertexProgram._Current /*prog */ ,
+	    r300VertexProgUpdateParams(ctx,
+				       (struct r300_vertex_program_cont *)
+				       ctx->VertexProgram._Current,
 				       (float *)&rmesa->hw.vpp.
 				       cmd[R300_VPP_PARAM_0]);
 	bump_vpu_count(rmesa->hw.vpp.cmd, param_count);
 	param_count /= 4;
 
-	/* Reset state, in case we don't use something */
-	((drm_r300_cmd_header_t *) rmesa->hw.vpi.cmd)->vpu.count = 0;
-	((drm_r300_cmd_header_t *) rmesa->hw.vps.cmd)->vpu.count = 0;
-
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_PROGRAM, &(prog->program));
-
-#if 0
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_UNKNOWN1,
-				     &(rmesa->state.vertex_shader.unknown1));
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_UNKNOWN2,
-				     &(rmesa->state.vertex_shader.unknown2));
-#endif
-
-	inst_count = prog->program.length / 4 - 1;
+	r300SetupVertexProgramFragment(rmesa, R300_PVS_UPLOAD_PROGRAM, &(prog->program));
+	inst_count = (prog->program.length / 4) - 1;
 
 	R300_STATECHANGE(rmesa, pvs);
 	rmesa->hw.pvs.cmd[R300_PVS_CNTL_1] =
-	    (0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT)
-	    | (inst_count /*pos_end */  << R300_PVS_CNTL_1_POS_END_SHIFT)
-	    | (inst_count << R300_PVS_CNTL_1_PROGRAM_END_SHIFT);
+	  (0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT) |
+	  (inst_count << R300_PVS_CNTL_1_POS_END_SHIFT) |
+	  (inst_count << R300_PVS_CNTL_1_PROGRAM_END_SHIFT);
 	rmesa->hw.pvs.cmd[R300_PVS_CNTL_2] =
-	    (0 << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT)
-	    | (param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT);
+	  (0 << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT) |
+	  (param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT);
 	rmesa->hw.pvs.cmd[R300_PVS_CNTL_3] =
-	    (0 /*rmesa->state.vertex_shader.unknown_ptr2 */  <<
-	     R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT)
-	    | (inst_count /*rmesa->state.vertex_shader.unknown_ptr3 */  <<
-	       0);
-
-	/* This is done for vertex shader fragments, but also needs to be done for vap_pvs,
-	   so I leave it as a reminder */
-#if 0
-	reg_start(R300_VAP_PVS_WAITIDLE, 0);
-	e32(0x00000000);
-#endif
+	  (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT) |
+	  (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN2_SHIFT);
 }
 
-static void r300SetupVertexShader(r300ContextPtr rmesa)
+static void r300SetupVertexProgram(r300ContextPtr rmesa)
 {
 	GLcontext *ctx = rmesa->radeon.glCtx;
 
@@ -1649,46 +1641,16 @@ static void r300SetupVertexShader(r300ContextPtr rmesa)
 	   0x400 area might have something to do with pixel shaders as it appears right after pfs programming.
 	   0x406 is set to { 0.0, 0.0, 1.0, 0.0 } most of the time but should change with smooth points and in other rare cases. */
 	//setup_vertex_shader_fragment(rmesa, 0x406, &unk4);
-	if (hw_tcl_on
-	    && ((struct r300_vertex_program *)CURRENT_VERTEX_SHADER(ctx))->
-	    translated) {
-		r300SetupVertexProgram(rmesa);
-		return;
+	if (hw_tcl_on && ((struct r300_vertex_program *)CURRENT_VERTEX_SHADER(ctx))->translated) {
+		r300SetupRealVertexProgram(rmesa);
+	} else {
+		/* FIXME: This needs to be replaced by vertex shader generation code. */
+		r300SetupDefaultVertexProgram(rmesa);
 	}
 
-	/* This needs to be replaced by vertex shader generation code */
-	r300GenerateSimpleVertexShader(rmesa);
-
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_PROGRAM,
-				     &(rmesa->state.vertex_shader.program));
 
-#if 0
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_UNKNOWN1,
-				     &(rmesa->state.vertex_shader.unknown1));
-	setup_vertex_shader_fragment(rmesa, VSF_DEST_UNKNOWN2,
-				     &(rmesa->state.vertex_shader.unknown2));
-#endif
-
-	R300_STATECHANGE(rmesa, pvs);
-	rmesa->hw.pvs.cmd[R300_PVS_CNTL_1] =
-	    (rmesa->state.vertex_shader.
-	     program_start << R300_PVS_CNTL_1_PROGRAM_START_SHIFT)
-	    | (rmesa->state.vertex_shader.
-	       unknown_ptr1 << R300_PVS_CNTL_1_POS_END_SHIFT)
-	    | (rmesa->state.vertex_shader.
-	       program_end << R300_PVS_CNTL_1_PROGRAM_END_SHIFT);
-	rmesa->hw.pvs.cmd[R300_PVS_CNTL_2] =
-	    (rmesa->state.vertex_shader.
-	     param_offset << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT)
-	    | (rmesa->state.vertex_shader.
-	       param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT);
-	rmesa->hw.pvs.cmd[R300_PVS_CNTL_3] =
-	    (rmesa->state.vertex_shader.
-	     unknown_ptr2 << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT)
-	    | (rmesa->state.vertex_shader.unknown_ptr3 << 0);
-
-	/* This is done for vertex shader fragments, but also needs to be done for vap_pvs,
-	   so I leave it as a reminder */
+	/* FIXME: This is done for vertex shader fragments, but also needs to be
+	 * done for vap_pvs, so I leave it as a reminder. */
 #if 0
 	reg_start(R300_VAP_PVS_WAITIDLE, 0);
 	e32(0x00000000);
@@ -1769,8 +1731,6 @@ static void r300Enable(GLcontext * ctx, GLenum cap, GLboolean state)
 
 	case GL_POLYGON_OFFSET_POINT:
 	case GL_POLYGON_OFFSET_LINE:
-		break;
-
 	case GL_POLYGON_OFFSET_FILL:
 		R300_STATECHANGE(r300, occlusion_cntl);
 		if (state) {
@@ -1864,13 +1824,15 @@ static void r300ResetHwState(r300ContextPtr r300)
 	r300->hw.vap_clip.cmd[4] = r300PackFloat32(1.0); /* Y */
 
 	/* XXX: Other families? */
-	switch (r300->radeon.radeonScreen->chip_family) {
-	case CHIP_FAMILY_R300:
-		r300->hw.unk2288.cmd[1] = R300_2288_R300;
-		break;
-	default:
-		r300->hw.unk2288.cmd[1] = R300_2288_RV350;
-		break;
+	if (has_tcl) {
+		switch (r300->radeon.radeonScreen->chip_family) {
+		case CHIP_FAMILY_R300:
+			r300->hw.unk2288.cmd[1] = R300_2288_R300;
+			break;
+		default:
+			r300->hw.unk2288.cmd[1] = R300_2288_RV350;
+			break;
+		}
 	}
 
 	r300->hw.gb_enable.cmd[1] = R300_GB_POINT_STUFF_ENABLE
@@ -1928,19 +1890,16 @@ static void r300ResetHwState(r300ContextPtr r300)
 	r300->hw.unk4260.cmd[2] = r300PackFloat32(0.0);
 	r300->hw.unk4260.cmd[3] = r300PackFloat32(1.0);
 
-	r300->hw.shade.cmd[1] = 0x00000002;
 	r300ShadeModel(ctx, ctx->Light.ShadeModel);
-	r300->hw.shade.cmd[3] = 0x00000000;
-	r300->hw.shade.cmd[4] = 0x00000000;
 
 	r300PolygonMode(ctx, GL_FRONT, ctx->Polygon.FrontMode);
 	r300PolygonMode(ctx, GL_BACK, ctx->Polygon.BackMode);
-	r300->hw.polygon_mode.cmd[2] = 0x00000001;
-	r300->hw.polygon_mode.cmd[3] = 0x00000000;
 	r300->hw.zbias_cntl.cmd[1] = 0x00000000;
 
 	r300PolygonOffset(ctx, ctx->Polygon.OffsetFactor,
 			  ctx->Polygon.OffsetUnits);
+	r300Enable(ctx, GL_POLYGON_OFFSET_POINT, ctx->Polygon.OffsetPoint);
+	r300Enable(ctx, GL_POLYGON_OFFSET_LINE, ctx->Polygon.OffsetLine);
 	r300Enable(ctx, GL_POLYGON_OFFSET_FILL, ctx->Polygon.OffsetFill);
 
 	r300->hw.unk42C0.cmd[1] = 0x4B7FFFFF;
@@ -1965,14 +1924,11 @@ static void r300ResetHwState(r300ContextPtr r300)
 	r300Fogfv(ctx, GL_FOG_COLOR, ctx->Fog.Color);
 	r300Fogfv(ctx, GL_FOG_COORDINATE_SOURCE_EXT, NULL);
 
-	r300->hw.at.cmd[R300_AT_UNKNOWN] = 0;
 	r300->hw.unk4BD8.cmd[1] = 0;
 
 	r300->hw.unk4E00.cmd[1] = 0;
 
 	r300BlendColor(ctx, ctx->Color.BlendColor);
-	r300->hw.blend_color.cmd[2] = 0;
-	r300->hw.blend_color.cmd[3] = 0;
 
 	/* Again, r300ClearBuffer uses this */
 	r300->hw.cb.cmd[R300_CB_OFFSET] =
@@ -2003,25 +1959,6 @@ static void r300ResetHwState(r300ContextPtr r300)
 	r300->hw.unk4EA0.cmd[1] = 0x00000000;
 	r300->hw.unk4EA0.cmd[2] = 0xffffffff;
 
-	switch (ctx->Visual.depthBits) {
-	case 16:
-		r300->hw.zstencil_format.cmd[1] = R300_DEPTH_FORMAT_16BIT_INT_Z;
-		break;
-	case 24:
-		r300->hw.zstencil_format.cmd[1] = R300_DEPTH_FORMAT_24BIT_INT_Z;
-		break;
-	default:
-		fprintf(stderr, "Error: Unsupported depth %d... exiting\n",
-			ctx->Visual.depthBits);
-		_mesa_exit(-1);
-
-	}
-	/* z compress? */
-	//r300->hw.zstencil_format.cmd[1] |= R300_DEPTH_FORMAT_UNK32;
-
-	r300->hw.zstencil_format.cmd[3] = 0x00000003;
-	r300->hw.zstencil_format.cmd[4] = 0x00000000;
-
 	r300->hw.zb.cmd[R300_ZB_OFFSET] =
 	    r300->radeon.radeonScreen->depthOffset +
 	    r300->radeon.radeonScreen->fbLocation;
@@ -2094,7 +2031,6 @@ void r300UpdateShaders(r300ContextPtr rmesa)
 		}
 		r300UpdateStateParameters(ctx, _NEW_PROGRAM);
 	}
-
 }
 
 static void r300SetupPixelShader(r300ContextPtr rmesa)
@@ -2113,62 +2049,61 @@ static void r300SetupPixelShader(r300ContextPtr rmesa)
 			__FUNCTION__);
 		return;
 	}
-#define OUTPUT_FIELD(st, reg, field)  \
-		R300_STATECHANGE(rmesa, st); \
-		for(i=0;i<=fp->alu_end;i++) \
-			rmesa->hw.st.cmd[R300_FPI_INSTR_0+i]=fp->alu.inst[i].field;\
-		rmesa->hw.st.cmd[R300_FPI_CMD_0]=cmdpacket0(reg, fp->alu_end+1);
 
-	OUTPUT_FIELD(fpi[0], R300_PFS_INSTR0_0, inst0);
-	OUTPUT_FIELD(fpi[1], R300_PFS_INSTR1_0, inst1);
-	OUTPUT_FIELD(fpi[2], R300_PFS_INSTR2_0, inst2);
-	OUTPUT_FIELD(fpi[3], R300_PFS_INSTR3_0, inst3);
-#undef OUTPUT_FIELD
+	R300_STATECHANGE(rmesa, fpi[0]);
+	rmesa->hw.fpi[0].cmd[R300_FPI_CMD_0] = cmdpacket0(R300_PFS_INSTR0_0, fp->alu_end + 1);
+	for (i = 0; i <= fp->alu_end; i++) {
+		rmesa->hw.fpi[0].cmd[R300_FPI_INSTR_0 + i] = fp->alu.inst[i].inst0;
+	}
+
+	R300_STATECHANGE(rmesa, fpi[1]);
+	rmesa->hw.fpi[1].cmd[R300_FPI_CMD_0] = cmdpacket0(R300_PFS_INSTR1_0, fp->alu_end + 1);
+	for (i = 0; i <= fp->alu_end; i++) {
+		rmesa->hw.fpi[1].cmd[R300_FPI_INSTR_0 + i] = fp->alu.inst[i].inst1;
+	}
+
+	R300_STATECHANGE(rmesa, fpi[2]);
+	rmesa->hw.fpi[2].cmd[R300_FPI_CMD_0] = cmdpacket0(R300_PFS_INSTR2_0, fp->alu_end + 1);
+	for (i = 0; i <= fp->alu_end; i++) {
+		rmesa->hw.fpi[2].cmd[R300_FPI_INSTR_0 + i] = fp->alu.inst[i].inst2;
+	}
+
+	R300_STATECHANGE(rmesa, fpi[3]);
+	rmesa->hw.fpi[3].cmd[R300_FPI_CMD_0] = cmdpacket0(R300_PFS_INSTR3_0, fp->alu_end + 1);
+	for (i = 0; i <= fp->alu_end; i++) {
+		rmesa->hw.fpi[3].cmd[R300_FPI_INSTR_0 + i] = fp->alu.inst[i].inst3;
+	}
 
 	R300_STATECHANGE(rmesa, fp);
+	rmesa->hw.fp.cmd[R300_FP_CNTL0] = fp->cur_node | (fp->first_node_has_tex << 3);
+	rmesa->hw.fp.cmd[R300_FP_CNTL1] = fp->max_temp_idx;
+	rmesa->hw.fp.cmd[R300_FP_CNTL2] =
+	  (fp->alu_offset << R300_PFS_CNTL_ALU_OFFSET_SHIFT) |
+	  (fp->alu_end << R300_PFS_CNTL_ALU_END_SHIFT) |
+	  (fp->tex_offset << R300_PFS_CNTL_TEX_OFFSET_SHIFT) |
+	  (fp->tex_end << R300_PFS_CNTL_TEX_END_SHIFT);
 	/* I just want to say, the way these nodes are stored.. weird.. */
 	for (i = 0, k = (4 - (fp->cur_node + 1)); i < 4; i++, k++) {
 		if (i < (fp->cur_node + 1)) {
 			rmesa->hw.fp.cmd[R300_FP_NODE0 + k] =
-			    (fp->node[i].
-			     alu_offset << R300_PFS_NODE_ALU_OFFSET_SHIFT)
-			    | (fp->node[i].
-			       alu_end << R300_PFS_NODE_ALU_END_SHIFT)
-			    | (fp->node[i].
-			       tex_offset << R300_PFS_NODE_TEX_OFFSET_SHIFT)
-			    | (fp->node[i].
-			       tex_end << R300_PFS_NODE_TEX_END_SHIFT)
-			    | fp->node[i].flags;
+			  (fp->node[i].alu_offset << R300_PFS_NODE_ALU_OFFSET_SHIFT) |
+			  (fp->node[i].alu_end << R300_PFS_NODE_ALU_END_SHIFT) |
+			  (fp->node[i].tex_offset << R300_PFS_NODE_TEX_OFFSET_SHIFT) |
+			  (fp->node[i].tex_end << R300_PFS_NODE_TEX_END_SHIFT) |
+			  fp->node[i].flags;
 		} else {
 			rmesa->hw.fp.cmd[R300_FP_NODE0 + (3 - i)] = 0;
 		}
 	}
 
-	/*  PFS_CNTL_0 */
-	rmesa->hw.fp.cmd[R300_FP_CNTL0] =
-	    fp->cur_node | (fp->first_node_has_tex << 3);
-	/* PFS_CNTL_1 */
-	rmesa->hw.fp.cmd[R300_FP_CNTL1] = fp->max_temp_idx;
-	/* PFS_CNTL_2 */
-	rmesa->hw.fp.cmd[R300_FP_CNTL2] =
-	    (fp->alu_offset << R300_PFS_CNTL_ALU_OFFSET_SHIFT)
-	    | (fp->alu_end << R300_PFS_CNTL_ALU_END_SHIFT)
-	    | (fp->tex_offset << R300_PFS_CNTL_TEX_OFFSET_SHIFT)
-	    | (fp->tex_end << R300_PFS_CNTL_TEX_END_SHIFT);
-
 	R300_STATECHANGE(rmesa, fpp);
+	rmesa->hw.fpp.cmd[R300_FPP_CMD_0] = cmdpacket0(R300_PFS_PARAM_0_X, fp->const_nr * 4);
 	for (i = 0; i < fp->const_nr; i++) {
-		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 0] =
-		    r300PackFloat24(fp->constant[i][0]);
-		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 1] =
-		    r300PackFloat24(fp->constant[i][1]);
-		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 2] =
-		    r300PackFloat24(fp->constant[i][2]);
-		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 3] =
-		    r300PackFloat24(fp->constant[i][3]);
-	}
-	rmesa->hw.fpp.cmd[R300_FPP_CMD_0] =
-	    cmdpacket0(R300_PFS_PARAM_0_X, fp->const_nr * 4);
+		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 0] = r300PackFloat24(fp->constant[i][0]);
+		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 1] = r300PackFloat24(fp->constant[i][1]);
+		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 2] = r300PackFloat24(fp->constant[i][2]);
+		rmesa->hw.fpp.cmd[R300_FPP_PARAM_0 + 4 * i + 3] = r300PackFloat24(fp->constant[i][3]);
+	}
 }
 
 void r300UpdateShaderStates(r300ContextPtr rmesa)
@@ -2182,7 +2117,7 @@ void r300UpdateShaderStates(r300ContextPtr rmesa)
 	r300SetupTextures(ctx);
 
 	if ((rmesa->radeon.radeonScreen->chip_flags & RADEON_CHIPSET_TCL))
-		r300SetupVertexShader(rmesa);
+		r300SetupVertexProgram(rmesa);
 	r300SetupRSUnit(ctx);
 }
 
diff --git a/src/mesa/drivers/dri/r300/r300_state.h b/src/mesa/drivers/dri/r300/r300_state.h
index 21a49b7f36..365f7ecd0c 100644
--- a/src/mesa/drivers/dri/r300/r300_state.h
+++ b/src/mesa/drivers/dri/r300/r300_state.h
@@ -37,8 +37,15 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "r300_context.h"
 
+#define R300_NEWPRIM( rmesa )			\
+  do {						\
+    if ( rmesa->dma.flush )			\
+      rmesa->dma.flush( rmesa );		\
+  } while (0)
+
 #define R300_STATECHANGE(r300, atom) \
 	do {						\
+	  R300_NEWPRIM(r300);				\
 		r300->hw.atom.dirty = GL_TRUE;		\
 		r300->hw.is_dirty = GL_TRUE;		\
 	} while(0)
diff --git a/src/mesa/drivers/dri/r300/r300_swtcl.c b/src/mesa/drivers/dri/r300/r300_swtcl.c
new file mode 100644
index 0000000000..c949f33bf3
--- /dev/null
+++ b/src/mesa/drivers/dri/r300/r300_swtcl.c
@@ -0,0 +1,711 @@
+/**************************************************************************
+
+Copyright (C) 2007 Dave Airlie
+
+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
+on 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
+THE COPYRIGHT OWNER(S) 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.
+
+**************************************************************************/
+
+/*
+ * Authors:
+ *   Dave Airlie <airlied@linux.ie>
+ */
+
+/* derived from r200 swtcl path */
+
+
+
+#include "glheader.h"
+#include "mtypes.h"
+#include "colormac.h"
+#include "enums.h"
+#include "image.h"
+#include "imports.h"
+#include "macros.h"
+
+#include "swrast/s_context.h"
+#include "swrast/s_fog.h"
+#include "swrast_setup/swrast_setup.h"
+#include "math/m_translate.h"
+#include "tnl/tnl.h"
+#include "tnl/t_context.h"
+#include "tnl/t_pipeline.h"
+
+#include "r300_context.h"
+#include "r300_swtcl.h"
+#include "r300_state.h"
+#include "r300_ioctl.h"
+#include "r300_emit.h"
+#include "r300_mem.h"
+
+static void flush_last_swtcl_prim( r300ContextPtr rmesa  );
+
+
+void r300EmitVertexAOS(r300ContextPtr rmesa, GLuint vertex_size, GLuint offset);
+void r300EmitVbufPrim(r300ContextPtr rmesa, GLuint primitive, GLuint vertex_nr);
+#define EMIT_ATTR( ATTR, STYLE )					\
+do {									\
+   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].attrib = (ATTR);	\
+   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].format = (STYLE);	\
+   rmesa->swtcl.vertex_attr_count++;					\
+} while (0)
+
+#define EMIT_PAD( N )							\
+do {									\
+   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].attrib = 0;		\
+   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].format = EMIT_PAD;	\
+   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].offset = (N);		\
+   rmesa->swtcl.vertex_attr_count++;					\
+} while (0)
+
+/* this differs from the VIR0 in emit.c - TODO merge them using another option */
+static GLuint r300VAPInputRoute0(uint32_t * dst, GLvector4f ** attribptr,
+				 int *inputs, GLint * tab, GLuint nr)
+{
+	GLuint i, dw;
+
+	/* type, inputs, stop bit, size */
+	for (i = 0; i + 1 < nr; i += 2) {
+		dw = (inputs[tab[i]] << 8) | 0x3;
+		dw |= ((inputs[tab[i + 1]] << 8) | 0x3) << 16;
+		if (i + 2 == nr) {
+			dw |= (R300_VAP_INPUT_ROUTE_END << 16);
+		}
+		dst[i >> 1] = dw;
+	}
+
+	if (nr & 1) {
+		dw = (inputs[tab[nr - 1]] << 8) | 0x3;
+		dw |= R300_VAP_INPUT_ROUTE_END;
+		dst[nr >> 1] = dw;
+	}
+
+	return (nr + 1) >> 1;
+}
+
+static void r300SetVertexFormat( GLcontext *ctx )
+{
+	r300ContextPtr rmesa = R300_CONTEXT( ctx );
+	TNLcontext *tnl = TNL_CONTEXT(ctx);
+	struct vertex_buffer *VB = &tnl->vb;
+	DECLARE_RENDERINPUTS(index_bitset);
+	GLuint InputsRead = 0, OutputsWritten = 0;
+	int vap_fmt_0 = 0;
+	int vap_vte_cntl = 0;
+	int offset = 0;
+	int vte = 0;
+	GLint inputs[VERT_ATTRIB_MAX];
+	GLint tab[VERT_ATTRIB_MAX];
+	int swizzle[VERT_ATTRIB_MAX][4];
+	GLuint i, nr;
+
+	DECLARE_RENDERINPUTS(render_inputs_bitset);
+	RENDERINPUTS_COPY(render_inputs_bitset, tnl->render_inputs_bitset);
+	RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+	RENDERINPUTS_COPY(rmesa->state.render_inputs_bitset, render_inputs_bitset);
+
+	/* Important:
+	 */
+	if ( VB->NdcPtr != NULL ) {
+		VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
+	}
+	else {
+		VB->AttribPtr[VERT_ATTRIB_POS] = VB->ClipPtr;
+	}
+
+	assert( VB->AttribPtr[VERT_ATTRIB_POS] != NULL );
+	rmesa->swtcl.vertex_attr_count = 0;
+
+	/* EMIT_ATTR's must be in order as they tell t_vertex.c how to
+	 * build up a hardware vertex.
+	 */
+	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POS)) {
+		vap_vte_cntl |= R300_VTX_W0_FMT;
+		InputsRead |= 1 << VERT_ATTRIB_POS;
+		OutputsWritten |= 1 << VERT_RESULT_HPOS;
+		EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F );
+	} else
+		EMIT_PAD(4 * sizeof(float));
+
+	offset = 4;
+
+	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POINTSIZE )) {
+		EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F );
+		vap_fmt_0 |=  R300_VAP_OUTPUT_VTX_FMT_0__PT_SIZE_PRESENT;
+		offset += 1;
+	}
+
+	if (RENDERINPUTS_TEST(index_bitset, _TNL_ATTRIB_COLOR0)) {
+	        rmesa->swtcl.coloroffset = offset;
+		InputsRead |= 1 << VERT_ATTRIB_COLOR0;
+		OutputsWritten |= 1 << VERT_RESULT_COL0;
+		EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4F );
+	}
+
+	offset += 4;
+
+	rmesa->swtcl.specoffset = 0;
+	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
+		rmesa->swtcl.specoffset = offset;
+		EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_4F );
+		InputsRead |= 1 << VERT_ATTRIB_COLOR1;
+		OutputsWritten |= 1 << VERT_RESULT_COL1;
+	}
+
+	if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
+		int i;
+
+		for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
+			if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
+				InputsRead |= 1 << (VERT_ATTRIB_TEX0 + i);
+				OutputsWritten |= 1 << (VERT_RESULT_TEX0 + i);
+				EMIT_ATTR( _TNL_ATTRIB_TEX0+i, EMIT_4F );
+			}
+		}
+	}
+
+	for (i = 0, nr = 0; i < VERT_ATTRIB_MAX; i++) {
+		if (InputsRead & (1 << i)) {
+			inputs[i] = nr++;
+		} else {
+			inputs[i] = -1;
+		}
+	}
+	
+	/* Fixed, apply to vir0 only */
+	if (InputsRead & (1 << VERT_ATTRIB_POS))
+		inputs[VERT_ATTRIB_POS] = 0;
+	if (InputsRead & (1 << VERT_ATTRIB_COLOR0))
+		inputs[VERT_ATTRIB_COLOR0] = 2;
+	if (InputsRead & (1 << VERT_ATTRIB_COLOR1))
+		inputs[VERT_ATTRIB_COLOR1] = 3;
+	for (i = VERT_ATTRIB_TEX0; i <= VERT_ATTRIB_TEX7; i++)
+		if (InputsRead & (1 << i))
+			inputs[i] = 6 + (i - VERT_ATTRIB_TEX0);
+	
+	for (i = 0, nr = 0; i < VERT_ATTRIB_MAX; i++) {
+		if (InputsRead & (1 << i)) {
+			tab[nr++] = i;
+		}
+	}
+	
+	for (i = 0; i < nr; i++) {
+		int ci;
+		
+		swizzle[i][0] = SWIZZLE_ZERO;
+		swizzle[i][1] = SWIZZLE_ZERO;
+		swizzle[i][2] = SWIZZLE_ZERO;
+		swizzle[i][3] = SWIZZLE_ONE;
+
+		for (ci = 0; ci < VB->AttribPtr[tab[i]]->size; ci++) {
+			swizzle[i][ci] = ci;
+		}
+	}
+
+	R300_NEWPRIM(rmesa);
+	R300_STATECHANGE(rmesa, vir[0]);
+	((drm_r300_cmd_header_t *) rmesa->hw.vir[0].cmd)->packet0.count =
+		r300VAPInputRoute0(&rmesa->hw.vir[0].cmd[R300_VIR_CNTL_0],
+				   VB->AttribPtr, inputs, tab, nr);
+	R300_STATECHANGE(rmesa, vir[1]);
+	((drm_r300_cmd_header_t *) rmesa->hw.vir[1].cmd)->packet0.count =
+		r300VAPInputRoute1(&rmesa->hw.vir[1].cmd[R300_VIR_CNTL_0], swizzle,
+				   nr);
+   
+	R300_STATECHANGE(rmesa, vic);
+	rmesa->hw.vic.cmd[R300_VIC_CNTL_0] = r300VAPInputCntl0(ctx, InputsRead);
+	rmesa->hw.vic.cmd[R300_VIC_CNTL_1] = r300VAPInputCntl1(ctx, InputsRead);
+   
+	R300_STATECHANGE(rmesa, vof);
+	rmesa->hw.vof.cmd[R300_VOF_CNTL_0] = r300VAPOutputCntl0(ctx, OutputsWritten);
+	rmesa->hw.vof.cmd[R300_VOF_CNTL_1] = r300VAPOutputCntl1(ctx, OutputsWritten);
+   
+	rmesa->swtcl.vertex_size =
+		_tnl_install_attrs( ctx,
+				    rmesa->swtcl.vertex_attrs, 
+				    rmesa->swtcl.vertex_attr_count,
+				    NULL, 0 );
+	
+	rmesa->swtcl.vertex_size /= 4;
+
+	RENDERINPUTS_COPY( rmesa->tnl_index_bitset, index_bitset );
+
+	vte = rmesa->hw.vte.cmd[1];
+	R300_STATECHANGE(rmesa, vte);
+	rmesa->hw.vte.cmd[1] = vte;
+	rmesa->hw.vte.cmd[2] = rmesa->swtcl.vertex_size;
+}
+
+
+/* Flush vertices in the current dma region.
+ */
+static void flush_last_swtcl_prim( r300ContextPtr rmesa  )
+{
+	if (RADEON_DEBUG & DEBUG_IOCTL)
+		fprintf(stderr, "%s\n", __FUNCTION__);
+	
+	rmesa->dma.flush = NULL;
+
+	if (rmesa->dma.current.buf) {
+		struct r300_dma_region *current = &rmesa->dma.current;
+		GLuint current_offset = GET_START(current);
+
+		assert (current->start + 
+			rmesa->swtcl.numverts * rmesa->swtcl.vertex_size * 4 ==
+			current->ptr);
+
+		if (rmesa->dma.current.start != rmesa->dma.current.ptr) {
+
+			r300EnsureCmdBufSpace( rmesa, rmesa->hw.max_state_size + (12*sizeof(int)), __FUNCTION__);
+			
+			r300EmitState(rmesa);
+			
+			r300EmitVertexAOS( rmesa,
+					   rmesa->swtcl.vertex_size,
+					   current_offset);
+			
+			r300EmitVbufPrim( rmesa,
+					  rmesa->swtcl.hw_primitive,
+					  rmesa->swtcl.numverts);
+			
+			r300EmitCacheFlush(rmesa);
+		}
+		
+		rmesa->swtcl.numverts = 0;
+		current->start = current->ptr;
+	}
+}
+
+/* Alloc space in the current dma region.
+ */
+static void *
+r300AllocDmaLowVerts( r300ContextPtr rmesa, int nverts, int vsize )
+{
+	GLuint bytes = vsize * nverts;
+
+	if ( rmesa->dma.current.ptr + bytes > rmesa->dma.current.end ) 
+		r300RefillCurrentDmaRegion( rmesa, bytes);
+
+	if (!rmesa->dma.flush) {
+		rmesa->radeon.glCtx->Driver.NeedFlush |= FLUSH_STORED_VERTICES;
+		rmesa->dma.flush = flush_last_swtcl_prim;
+	}
+
+	ASSERT( vsize == rmesa->swtcl.vertex_size * 4 );
+	ASSERT( rmesa->dma.flush == flush_last_swtcl_prim );
+	ASSERT( rmesa->dma.current.start + 
+		rmesa->swtcl.numverts * rmesa->swtcl.vertex_size * 4 ==
+		rmesa->dma.current.ptr );
+
+	{
+		GLubyte *head = (GLubyte *) (rmesa->dma.current.address + rmesa->dma.current.ptr);
+		rmesa->dma.current.ptr += bytes;
+		rmesa->swtcl.numverts += nverts;
+		return head;
+	}
+}
+
+static GLuint reduced_prim[] = {
+  GL_POINTS,
+  GL_LINES,
+  GL_LINES,
+  GL_LINES,
+  GL_TRIANGLES,
+  GL_TRIANGLES,
+  GL_TRIANGLES,
+  GL_TRIANGLES,
+  GL_TRIANGLES,
+  GL_TRIANGLES,
+};
+
+static void r300RasterPrimitive( GLcontext *ctx, GLuint prim );
+static void r300RenderPrimitive( GLcontext *ctx, GLenum prim );
+//static void r300ResetLineStipple( GLcontext *ctx );
+
+/***********************************************************************
+ *                    Emit primitives as inline vertices               *
+ ***********************************************************************/
+
+
+#define HAVE_POINTS      1
+#define HAVE_LINES       1
+#define HAVE_LINE_STRIPS 1
+#define HAVE_TRIANGLES   1
+#define HAVE_TRI_STRIPS  1
+#define HAVE_TRI_STRIP_1 0
+#define HAVE_TRI_FANS    1
+#define HAVE_QUADS       0
+#define HAVE_QUAD_STRIPS 0
+#define HAVE_POLYGONS    1
+#define HAVE_ELTS        1
+
+#undef LOCAL_VARS
+#undef ALLOC_VERTS
+#define CTX_ARG r300ContextPtr rmesa
+#define GET_VERTEX_DWORDS() rmesa->swtcl.vertex_size
+#define ALLOC_VERTS( n, size ) r300AllocDmaLowVerts( rmesa, n, size * 4 )
+#define LOCAL_VARS						\
+   r300ContextPtr rmesa = R300_CONTEXT(ctx);		\
+   const char *r300verts = (char *)rmesa->swtcl.verts;
+#define VERT(x) (r300Vertex *)(r300verts + ((x) * vertsize * sizeof(int)))
+#define VERTEX r300Vertex 
+#define DO_DEBUG_VERTS (1 && (RADEON_DEBUG & DEBUG_VERTS))
+#define PRINT_VERTEX(x)
+#undef TAG
+#define TAG(x) r300_##x
+#include "tnl_dd/t_dd_triemit.h"
+
+
+
+/***********************************************************************
+ *          Macros for t_dd_tritmp.h to draw basic primitives          *
+ ***********************************************************************/
+
+#define QUAD( a, b, c, d ) r300_quad( rmesa, a, b, c, d )
+#define TRI( a, b, c )     r300_triangle( rmesa, a, b, c )
+#define LINE( a, b )       r300_line( rmesa, a, b )
+#define POINT( a )         r300_point( rmesa, a )
+
+/***********************************************************************
+ *              Build render functions from dd templates               *
+ ***********************************************************************/
+
+#define R300_TWOSIDE_BIT	0x01
+#define R300_UNFILLED_BIT	0x02
+#define R300_MAX_TRIFUNC	0x04
+
+static struct {
+   tnl_points_func	        points;
+   tnl_line_func		line;
+   tnl_triangle_func	triangle;
+   tnl_quad_func		quad;
+} rast_tab[R300_MAX_TRIFUNC];
+
+#define DO_FALLBACK  0
+#define DO_UNFILLED (IND & R300_UNFILLED_BIT)
+#define DO_TWOSIDE  (IND & R300_TWOSIDE_BIT)
+#define DO_FLAT      0
+#define DO_OFFSET     0
+#define DO_TRI       1
+#define DO_QUAD      1
+#define DO_LINE      1
+#define DO_POINTS    1
+#define DO_FULL_QUAD 1
+
+#define HAVE_RGBA   1
+#define HAVE_SPEC   1
+#define HAVE_BACK_COLORS  0
+#define HAVE_HW_FLATSHADE 1
+#define TAB rast_tab
+
+#define DEPTH_SCALE 1.0
+#define UNFILLED_TRI unfilled_tri
+#define UNFILLED_QUAD unfilled_quad
+#define VERT_X(_v) _v->v.x
+#define VERT_Y(_v) _v->v.y
+#define VERT_Z(_v) _v->v.z
+#define AREA_IS_CCW( a ) (a < 0)
+#define GET_VERTEX(e) (rmesa->swtcl.verts + (e*rmesa->swtcl.vertex_size*sizeof(int)))
+
+/* Only used to pull back colors into vertices (ie, we know color is
+ * floating point).
+ */
+#define R300_COLOR( dst, src )				\
+do {							\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[0], (src)[2]);	\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[1], (src)[1]);	\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[2], (src)[0]);	\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[3], (src)[3]);	\
+} while (0)
+
+#define VERT_SET_RGBA( v, c )    if (coloroffset) R300_COLOR( v->ub4[coloroffset], c )
+#define VERT_COPY_RGBA( v0, v1 ) if (coloroffset) v0->ui[coloroffset] = v1->ui[coloroffset]
+#define VERT_SAVE_RGBA( idx )    if (coloroffset) color[idx] = v[idx]->ui[coloroffset]
+#define VERT_RESTORE_RGBA( idx ) if (coloroffset) v[idx]->ui[coloroffset] = color[idx]
+
+#define R300_SPEC( dst, src )				\
+do {							\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[0], (src)[2]);	\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[1], (src)[1]);	\
+   UNCLAMPED_FLOAT_TO_UBYTE((dst)[2], (src)[0]);	\
+} while (0)
+
+#define VERT_SET_SPEC( v, c )    if (specoffset) R300_SPEC( v->ub4[specoffset], c )
+#define VERT_COPY_SPEC( v0, v1 ) if (specoffset) COPY_3V(v0->ub4[specoffset], v1->ub4[specoffset])
+#define VERT_SAVE_SPEC( idx )    if (specoffset) spec[idx] = v[idx]->ui[specoffset]
+#define VERT_RESTORE_SPEC( idx ) if (specoffset) v[idx]->ui[specoffset] = spec[idx]
+
+#undef LOCAL_VARS
+#undef TAG
+#undef INIT
+
+#define LOCAL_VARS(n)							\
+   r300ContextPtr rmesa = R300_CONTEXT(ctx);			\
+   GLuint color[n], spec[n];						\
+   GLuint coloroffset = rmesa->swtcl.coloroffset;	\
+   GLuint specoffset = rmesa->swtcl.specoffset;			\
+   (void) color; (void) spec; (void) coloroffset; (void) specoffset;
+
+/***********************************************************************
+ *                Helpers for rendering unfilled primitives            *
+ ***********************************************************************/
+
+#define RASTERIZE(x) r300RasterPrimitive( ctx, reduced_prim[x] )
+#define RENDER_PRIMITIVE rmesa->swtcl.render_primitive
+#undef TAG
+#define TAG(x) x
+#include "tnl_dd/t_dd_unfilled.h"
+#undef IND
+
+
+/***********************************************************************
+ *                      Generate GL render functions                   *
+ ***********************************************************************/
+
+
+#define IND (0)
+#define TAG(x) x
+#include "tnl_dd/t_dd_tritmp.h"
+
+#define IND (R300_TWOSIDE_BIT)
+#define TAG(x) x##_twoside
+#include "tnl_dd/t_dd_tritmp.h"
+
+#define IND (R300_UNFILLED_BIT)
+#define TAG(x) x##_unfilled
+#include "tnl_dd/t_dd_tritmp.h"
+
+#define IND (R300_TWOSIDE_BIT|R300_UNFILLED_BIT)
+#define TAG(x) x##_twoside_unfilled
+#include "tnl_dd/t_dd_tritmp.h"
+
+
+
+static void init_rast_tab( void )
+{
+   init();
+   init_twoside();
+   init_unfilled();
+   init_twoside_unfilled();
+}
+
+/**********************************************************************/
+/*               Render unclipped begin/end objects                   */
+/**********************************************************************/
+
+#define RENDER_POINTS( start, count )		\
+   for ( ; start < count ; start++)		\
+      r300_point( rmesa, VERT(start) )
+#define RENDER_LINE( v0, v1 ) \
+   r300_line( rmesa, VERT(v0), VERT(v1) )
+#define RENDER_TRI( v0, v1, v2 )  \
+   r300_triangle( rmesa, VERT(v0), VERT(v1), VERT(v2) )
+#define RENDER_QUAD( v0, v1, v2, v3 ) \
+   r300_quad( rmesa, VERT(v0), VERT(v1), VERT(v2), VERT(v3) )
+#define INIT(x) do {					\
+   r300RenderPrimitive( ctx, x );			\
+} while (0)
+#undef LOCAL_VARS
+#define LOCAL_VARS						\
+   r300ContextPtr rmesa = R300_CONTEXT(ctx);		\
+   const GLuint vertsize = rmesa->swtcl.vertex_size;		\
+   const char *r300verts = (char *)rmesa->swtcl.verts;		\
+   const GLuint * const elt = TNL_CONTEXT(ctx)->vb.Elts;	\
+   const GLboolean stipple = ctx->Line.StippleFlag;		\
+   (void) elt; (void) stipple;
+#define RESET_STIPPLE	//if ( stipple ) r200ResetLineStipple( ctx );
+#define RESET_OCCLUSION
+#define PRESERVE_VB_DEFS
+#define ELT(x) (x)
+#define TAG(x) r300_##x##_verts
+#include "tnl/t_vb_rendertmp.h"
+#undef ELT
+#undef TAG
+#define TAG(x) r300_##x##_elts
+#define ELT(x) elt[x]
+#include "tnl/t_vb_rendertmp.h"
+
+
+
+
+/**********************************************************************/
+/*                    Choose render functions                         */
+/**********************************************************************/
+static void r300ChooseRenderState( GLcontext *ctx )
+{
+	TNLcontext *tnl = TNL_CONTEXT(ctx);
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
+	GLuint index = 0;
+	GLuint flags = ctx->_TriangleCaps;
+
+	if (flags & DD_TRI_LIGHT_TWOSIDE) index |= R300_TWOSIDE_BIT;
+	if (flags & DD_TRI_UNFILLED)      index |= R300_UNFILLED_BIT;
+
+	if (index != rmesa->swtcl.RenderIndex) {
+		tnl->Driver.Render.Points = rast_tab[index].points;
+		tnl->Driver.Render.Line = rast_tab[index].line;
+		tnl->Driver.Render.ClippedLine = rast_tab[index].line;
+		tnl->Driver.Render.Triangle = rast_tab[index].triangle;
+		tnl->Driver.Render.Quad = rast_tab[index].quad;
+
+		if (index == 0) {
+			tnl->Driver.Render.PrimTabVerts = r300_render_tab_verts;
+			tnl->Driver.Render.PrimTabElts = r300_render_tab_elts;
+			tnl->Driver.Render.ClippedPolygon = r300_fast_clipped_poly;
+		} else {
+			tnl->Driver.Render.PrimTabVerts = _tnl_render_tab_verts;
+			tnl->Driver.Render.PrimTabElts = _tnl_render_tab_elts;
+			tnl->Driver.Render.ClippedPolygon = _tnl_RenderClippedPolygon;
+		}
+
+		rmesa->swtcl.RenderIndex = index;
+	}
+}
+
+
+static void r300RenderStart(GLcontext *ctx)
+{
+        r300ContextPtr rmesa = R300_CONTEXT( ctx );
+	//	fprintf(stderr, "%s\n", __FUNCTION__);
+
+	r300ChooseRenderState(ctx);	
+	r300SetVertexFormat(ctx);
+
+	r300UpdateShaderStates(rmesa);
+
+	r300EmitCacheFlush(rmesa);
+	
+	if (rmesa->dma.flush != 0 && 
+	    rmesa->dma.flush != flush_last_swtcl_prim)
+		rmesa->dma.flush( rmesa );
+
+}
+
+static void r300RenderFinish(GLcontext *ctx)
+{
+}
+
+static void r300RasterPrimitive( GLcontext *ctx, GLuint hwprim )
+{
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
+	
+	if (rmesa->swtcl.hw_primitive != hwprim) {
+	        R300_NEWPRIM( rmesa );
+		rmesa->swtcl.hw_primitive = hwprim;
+	}
+}
+
+static void r300RenderPrimitive(GLcontext *ctx, GLenum prim)
+{
+
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
+	rmesa->swtcl.render_primitive = prim;
+
+	if ((prim == GL_TRIANGLES) && (ctx->_TriangleCaps & DD_TRI_UNFILLED))
+	  return;
+
+	r300RasterPrimitive( ctx, reduced_prim[prim] );
+	//	fprintf(stderr, "%s\n", __FUNCTION__);
+	
+}
+
+static void r300ResetLineStipple(GLcontext *ctx)
+{
+
+
+}
+
+void r300InitSwtcl(GLcontext *ctx)
+{
+	TNLcontext *tnl = TNL_CONTEXT(ctx);
+	r300ContextPtr rmesa = R300_CONTEXT(ctx);
+	static int firsttime = 1;
+	
+	if (firsttime) {
+		init_rast_tab();
+		firsttime = 0;
+	}
+	
+	tnl->Driver.Render.Start = r300RenderStart;
+	tnl->Driver.Render.Finish = r300RenderFinish;
+	tnl->Driver.Render.PrimitiveNotify = r300RenderPrimitive;
+	tnl->Driver.Render.ResetLineStipple = r300ResetLineStipple;
+	tnl->Driver.Render.BuildVertices = _tnl_build_vertices;
+	tnl->Driver.Render.CopyPV = _tnl_copy_pv;
+	tnl->Driver.Render.Interp = _tnl_interp;
+	
+	/* FIXME: what are these numbers? */
+	_tnl_init_vertices( ctx, ctx->Const.MaxArrayLockSize + 12, 
+			    48 * sizeof(GLfloat) );
+	
+	rmesa->swtcl.verts = (GLubyte *)tnl->clipspace.vertex_buf;
+	rmesa->swtcl.RenderIndex = ~0;
+	rmesa->swtcl.render_primitive = GL_TRIANGLES;
+	rmesa->swtcl.hw_primitive = 0;	
+
+	_tnl_invalidate_vertex_state( ctx, ~0 );
+	_tnl_invalidate_vertices( ctx, ~0 );
+	RENDERINPUTS_ZERO( rmesa->tnl_index_bitset );
+
+	_tnl_need_projected_coords( ctx, GL_FALSE );
+	r300ChooseRenderState(ctx);
+
+	_mesa_validate_all_lighting_tables( ctx ); 
+
+	tnl->Driver.NotifyMaterialChange = 
+	  _mesa_validate_all_lighting_tables;
+}
+
+void r300DestroySwtcl(GLcontext *ctx)
+{
+}
+
+void r300EmitVertexAOS(r300ContextPtr rmesa, GLuint vertex_size, GLuint offset)
+{
+	int cmd_reserved = 0;
+	int cmd_written = 0;
+
+	drm_radeon_cmd_header_t *cmd = NULL;
+	if (RADEON_DEBUG & DEBUG_VERTS)
+	  fprintf(stderr, "%s:  vertex_size %d, offset 0x%x \n",
+		  __FUNCTION__, vertex_size, offset);
+
+	start_packet3(CP_PACKET3(R300_PACKET3_3D_LOAD_VBPNTR, 2), 2);
+	e32(1);
+	e32(vertex_size | (vertex_size << 8));
+	e32(offset);
+}
+
+void r300EmitVbufPrim(r300ContextPtr rmesa, GLuint primitive, GLuint vertex_nr)
+{
+
+	int cmd_reserved = 0;
+	int cmd_written = 0;
+	int type, num_verts;
+	drm_radeon_cmd_header_t *cmd = NULL;
+
+	type = r300PrimitiveType(rmesa, primitive);
+	num_verts = r300NumVerts(rmesa, vertex_nr, primitive);
+	
+	start_packet3(CP_PACKET3(R300_PACKET3_3D_DRAW_VBUF_2, 0), 0);
+	e32(R300_VAP_VF_CNTL__PRIM_WALK_VERTEX_LIST | (num_verts << 16) | type);
+}
diff --git a/src/mesa/drivers/dri/r300/r300_swtcl.h b/src/mesa/drivers/dri/r300/r300_swtcl.h
new file mode 100644
index 0000000000..2ea6ceded7
--- /dev/null
+++ b/src/mesa/drivers/dri/r300/r300_swtcl.h
@@ -0,0 +1,45 @@
+/*
+Copyright (C) The Weather Channel, Inc.  2002.  All Rights Reserved.
+
+The Weather Channel (TM) funded Tungsten Graphics to develop the
+initial release of the Radeon 8500 driver under the XFree86 license.
+This notice must be preserved.
+
+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 (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 NONINFRINGEMENT.
+IN NO EVENT SHALL THE COPYRIGHT OWNER(S) 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.
+*/
+
+/*
+ * Authors:
+ *   Keith Whitwell <keith@tungstengraphics.com> - original r200 code
+ *   Dave Airlie <airlied@linux.ie>
+ */
+
+#ifndef __R300_SWTCL_H__
+#define __R300_SWTCL_H__
+
+#include "mtypes.h"
+#include "swrast/swrast.h"
+#include "r300_context.h"
+
+extern void r300InitSwtcl( GLcontext *ctx );
+extern void r300DestroySwtcl( GLcontext *ctx );
+
+#endif
diff --git a/src/mesa/drivers/dri/r300/r300_vertprog.c b/src/mesa/drivers/dri/r300/r300_vertprog.c
index 16dddf6557..7d4e8c9511 100644
--- a/src/mesa/drivers/dri/r300/r300_vertprog.c
+++ b/src/mesa/drivers/dri/r300/r300_vertprog.c
@@ -29,6 +29,10 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
  * \file
  *
  * \author Aapo Tahkola <aet@rasterburn.org>
+ *
+ * \author Oliver McFadden <z3ro.geek@gmail.com>
+ *
+ * For a description of the vertex program instruction set see r300_reg.h.
  */
 
 #include "glheader.h"
@@ -55,54 +59,58 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #error Cannot change these!
 #endif
 
-#define SCALAR_FLAG (1<<31)
-#define FLAG_MASK (1<<31)
-#define OP_MASK	(0xf)		/* we are unlikely to have more than 15 */
-#define OPN(operator, ip) {#operator, OPCODE_##operator, ip}
-
-static struct {
-	char *name;
-	int opcode;
-	unsigned long ip;	/* number of input operands and flags */
-} op_names[] = {
-	/* *INDENT-OFF* */
-	OPN(ABS, 1),
-	OPN(ADD, 2),
-	OPN(ARL, 1 | SCALAR_FLAG),
-	OPN(DP3, 2),
-	OPN(DP4, 2),
-	OPN(DPH, 2),
-	OPN(DST, 2),
-	OPN(EX2, 1 | SCALAR_FLAG),
-	OPN(EXP, 1 | SCALAR_FLAG),
-	OPN(FLR, 1),
-	OPN(FRC, 1),
-	OPN(LG2, 1 | SCALAR_FLAG),
-	OPN(LIT, 1),
-	OPN(LOG, 1 | SCALAR_FLAG),
-	OPN(MAD, 3),
-	OPN(MAX, 2),
-	OPN(MIN, 2),
-	OPN(MOV, 1),
-	OPN(MUL, 2),
-	OPN(POW, 2 | SCALAR_FLAG),
-	OPN(RCP, 1 | SCALAR_FLAG),
-	OPN(RSQ, 1 | SCALAR_FLAG),
-	OPN(SGE, 2),
-	OPN(SLT, 2),
-	OPN(SUB, 2),
-	OPN(SWZ, 1),
-	OPN(XPD, 2),
-	OPN(RCC, 0),	//extra
-	OPN(PRINT, 0),
-	OPN(END, 0)
-	/* *INDENT-ON* */
-};
-
-#undef OPN
+/* TODO: Get rid of t_src_class call */
+#define CMP_SRCS(a, b) ((a.RelAddr != b.RelAddr) || (a.Index != b.Index && \
+		       ((t_src_class(a.File) == VSF_IN_CLASS_PARAM && \
+			 t_src_class(b.File) == VSF_IN_CLASS_PARAM) || \
+			(t_src_class(a.File) == VSF_IN_CLASS_ATTR && \
+			 t_src_class(b.File) == VSF_IN_CLASS_ATTR)))) \
+
+#define ZERO_SRC_0 (MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    t_src_class(src[0].File), VSF_FLAG_NONE) | (src[0].RelAddr << 4))
+
+#define ZERO_SRC_1 (MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    t_src_class(src[1].File), VSF_FLAG_NONE) | (src[1].RelAddr << 4))
+
+#define ZERO_SRC_2 (MAKE_VSF_SOURCE(t_src_index(vp, &src[2]), \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
+				    t_src_class(src[2].File), VSF_FLAG_NONE) | (src[2].RelAddr << 4))
+
+#define ONE_SRC_0 (MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    t_src_class(src[0].File), VSF_FLAG_NONE) | (src[0].RelAddr << 4))
+
+#define ONE_SRC_1 (MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    t_src_class(src[1].File), VSF_FLAG_NONE) | (src[1].RelAddr << 4))
+
+#define ONE_SRC_2 (MAKE_VSF_SOURCE(t_src_index(vp, &src[2]), \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    SWIZZLE_ONE, SWIZZLE_ONE, \
+				    t_src_class(src[2].File), VSF_FLAG_NONE) | (src[2].RelAddr << 4))
+
+/* DP4 version seems to trigger some hw peculiarity */
+//#define PREFER_DP4
+
+#define FREE_TEMPS() \
+	do { \
+		if(u_temp_i < vp->num_temporaries) { \
+			WARN_ONCE("Ran out of temps, num temps %d, us %d\n", vp->num_temporaries, u_temp_i); \
+			vp->native = GL_FALSE; \
+		} \
+		u_temp_i=VSF_MAX_FRAGMENT_TEMPS-1; \
+	} while (0)
 
 int r300VertexProgUpdateParams(GLcontext * ctx,
-			       struct r300_vertex_program_cont *vp, float *dst)
+			       struct r300_vertex_program_cont *vp,
+			       float *dst)
 {
 	int pi;
 	struct gl_vertex_program *mesa_vp = &vp->mesa_program;
@@ -234,8 +242,8 @@ static void vp_dump_inputs(struct r300_vertex_program *vp, char *caller)
 	int i;
 
 	if (vp == NULL) {
-		fprintf(stderr, "vp null in call to %s from %s\n", __FUNCTION__,
-			caller);
+		fprintf(stderr, "vp null in call to %s from %s\n",
+			__FUNCTION__, caller);
 		return;
 	}
 
@@ -276,6 +284,8 @@ static unsigned long t_src_index(struct r300_vertex_program *vp,
 	}
 }
 
+/* these two functions should probably be merged... */
+
 static unsigned long t_src(struct r300_vertex_program *vp,
 			   struct prog_src_register *src)
 {
@@ -294,7 +304,9 @@ static unsigned long t_src(struct r300_vertex_program *vp,
 static unsigned long t_src_scalar(struct r300_vertex_program *vp,
 				  struct prog_src_register *src)
 {
-
+	/* src->NegateBase uses the NEGATE_ flags from program_instruction.h,
+	 * which equal our VSF_FLAGS_ values, so it's safe to just pass it here.
+	 */
 	return MAKE_VSF_SOURCE(t_src_index(vp, src),
 			       t_swizzle(GET_SWZ(src->Swizzle, 0)),
 			       t_swizzle(GET_SWZ(src->Swizzle, 0)),
@@ -306,128 +318,741 @@ static unsigned long t_src_scalar(struct r300_vertex_program *vp,
 	    (src->RelAddr << 4);
 }
 
-static unsigned long t_opcode(enum prog_opcode opcode)
+static GLboolean valid_dst(struct r300_vertex_program *vp,
+			   struct prog_dst_register *dst)
 {
+	if (dst->File == PROGRAM_OUTPUT && vp->outputs[dst->Index] == -1) {
+		return GL_FALSE;
+	} else if (dst->File == PROGRAM_ADDRESS) {
+		assert(dst->Index == 0);
+	}
 
-	switch (opcode) {
-	/* *INDENT-OFF* */
-	case OPCODE_ARL: return R300_VPI_OUT_OP_ARL;
-	case OPCODE_DST: return R300_VPI_OUT_OP_DST;
-	case OPCODE_EX2: return R300_VPI_OUT_OP_EX2;
-	case OPCODE_EXP: return R300_VPI_OUT_OP_EXP;
-	case OPCODE_FRC: return R300_VPI_OUT_OP_FRC;
-	case OPCODE_LG2: return R300_VPI_OUT_OP_LG2;
-	case OPCODE_LOG: return R300_VPI_OUT_OP_LOG;
-	case OPCODE_MAX: return R300_VPI_OUT_OP_MAX;
-	case OPCODE_MIN: return R300_VPI_OUT_OP_MIN;
-	case OPCODE_MUL: return R300_VPI_OUT_OP_MUL;
-	case OPCODE_RCP: return R300_VPI_OUT_OP_RCP;
-	case OPCODE_RSQ: return R300_VPI_OUT_OP_RSQ;
-	case OPCODE_SGE: return R300_VPI_OUT_OP_SGE;
-	case OPCODE_SLT: return R300_VPI_OUT_OP_SLT;
-	case OPCODE_DP4: return R300_VPI_OUT_OP_DOT;
-	/* *INDENT-ON* */
+	return GL_TRUE;
+}
 
-	default:
-		fprintf(stderr, "%s: Should not be called with opcode %d!",
-			__FUNCTION__, opcode);
-	}
-	_mesa_exit(-1);
-	return 0;
+/*
+ * Instruction    Inputs  Output   Description
+ * -----------    ------  ------   --------------------------------
+ * ABS            v       v        absolute value
+ * ADD            v,v     v        add
+ * ARL            s       a        address register load
+ * DP3            v,v     ssss     3-component dot product
+ * DP4            v,v     ssss     4-component dot product
+ * DPH            v,v     ssss     homogeneous dot product
+ * DST            v,v     v        distance vector
+ * EX2            s       ssss     exponential base 2
+ * EXP            s       v        exponential base 2 (approximate)
+ * FLR            v       v        floor
+ * FRC            v       v        fraction
+ * LG2            s       ssss     logarithm base 2
+ * LIT            v       v        compute light coefficients
+ * LOG            s       v        logarithm base 2 (approximate)
+ * MAD            v,v,v   v        multiply and add
+ * MAX            v,v     v        maximum
+ * MIN            v,v     v        minimum
+ * MOV            v       v        move
+ * MUL            v,v     v        multiply
+ * POW            s,s     ssss     exponentiate
+ * RCP            s       ssss     reciprocal
+ * RSQ            s       ssss     reciprocal square root
+ * SGE            v,v     v        set on greater than or equal
+ * SLT            v,v     v        set on less than
+ * SUB            v,v     v        subtract
+ * SWZ            v       v        extended swizzle
+ * XPD            v,v     v        cross product
+ *
+ * Table X.5:  Summary of vertex program instructions.  "v" indicates a
+ * floating-point vector input or output, "s" indicates a floating-point
+ * scalar input, "ssss" indicates a scalar output replicated across a
+ * 4-component result vector, and "a" indicates a single address register
+ * component.
+ */
+
+static GLuint *t_opcode_abs(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//MAX RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_MAX, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 3)),
+			    t_src_class(src[0].File),
+			    (!src[0].
+			     NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+	    (src[0].RelAddr << 4);
+	inst[3] = 0;
+
+	return inst;
 }
 
-static unsigned long op_operands(enum prog_opcode opcode)
+static GLuint *t_opcode_add(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
 {
-	int i;
+	unsigned long hw_op;
 
-	/* Can we trust mesas opcodes to be in order ? */
-	for (i = 0; i < sizeof(op_names) / sizeof(*op_names); i++)
-		if (op_names[i].opcode == opcode)
-			return op_names[i].ip;
+#if 1
+	hw_op = (src[0].File == PROGRAM_TEMPORARY
+		 && src[1].File ==
+		 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = ONE_SRC_0;
+	inst[2] = t_src(vp, &src[0]);
+	inst[3] = t_src(vp, &src[1]);
+#else
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
 
-	fprintf(stderr, "op %d not found in op_names\n", opcode);
-	_mesa_exit(-1);
-	return 0;
+#endif
+
+	return inst;
 }
 
-static GLboolean valid_dst(struct r300_vertex_program *vp,
-			   struct prog_dst_register *dst)
+static GLuint *t_opcode_arl(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
 {
-	if (dst->File == PROGRAM_OUTPUT && vp->outputs[dst->Index] == -1) {
-		return GL_FALSE;
-	} else if (dst->File == PROGRAM_ADDRESS) {
-		assert(dst->Index == 0);
-	}
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ARL, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
 
-	return GL_TRUE;
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
 }
 
-/* TODO: Get rid of t_src_class call */
-#define CMP_SRCS(a, b) ((a.RelAddr != b.RelAddr) || (a.Index != b.Index && \
-		       ((t_src_class(a.File) == VSF_IN_CLASS_PARAM && \
-			 t_src_class(b.File) == VSF_IN_CLASS_PARAM) || \
-			(t_src_class(a.File) == VSF_IN_CLASS_ATTR && \
-			 t_src_class(b.File) == VSF_IN_CLASS_ATTR)))) \
+static GLuint *t_opcode_dp3(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ZERO} PARAM 0{} {X Y Z ZERO}
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
+			    SWIZZLE_ZERO, t_src_class(src[0].File),
+			    src[0].
+			    NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+	    (src[0].RelAddr << 4);
+
+	inst[2] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 2)),
+			    SWIZZLE_ZERO, t_src_class(src[1].File),
+			    src[1].
+			    NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+	    (src[1].RelAddr << 4);
+
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
 
-#define ZERO_SRC_0 (MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    t_src_class(src[0].File), VSF_FLAG_NONE) | (src[0].RelAddr << 4))
+static GLuint *t_opcode_dp4(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
 
-#define ZERO_SRC_1 (MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    t_src_class(src[1].File), VSF_FLAG_NONE) | (src[1].RelAddr << 4))
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
 
-#define ZERO_SRC_2 (MAKE_VSF_SOURCE(t_src_index(vp, &src[2]), \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    SWIZZLE_ZERO, SWIZZLE_ZERO, \
-				    t_src_class(src[2].File), VSF_FLAG_NONE) | (src[2].RelAddr << 4))
+	return inst;
+}
 
-#define ONE_SRC_0 (MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    t_src_class(src[0].File), VSF_FLAG_NONE) | (src[0].RelAddr << 4))
+static GLuint *t_opcode_dph(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ONE} PARAM 0{} {X Y Z W}
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 2)),
+			    VSF_IN_COMPONENT_ONE, t_src_class(src[0].File),
+			    src[0].
+			    NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) |
+	    (src[0].RelAddr << 4);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
 
-#define ONE_SRC_1 (MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    t_src_class(src[1].File), VSF_FLAG_NONE) | (src[1].RelAddr << 4))
+static GLuint *t_opcode_dst(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_DST, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
 
-#define ONE_SRC_2 (MAKE_VSF_SOURCE(t_src_index(vp, &src[2]), \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    SWIZZLE_ONE, SWIZZLE_ONE, \
-				    t_src_class(src[2].File), VSF_FLAG_NONE) | (src[2].RelAddr << 4))
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
 
-/* DP4 version seems to trigger some hw peculiarity */
-//#define PREFER_DP4
+	return inst;
+}
 
-#define FREE_TEMPS() \
-	do { \
-		if(u_temp_i < vp->num_temporaries) { \
-			WARN_ONCE("Ran out of temps, num temps %d, us %d\n", vp->num_temporaries, u_temp_i); \
-			vp->native = GL_FALSE; \
-		} \
-		u_temp_i=VSF_MAX_FRAGMENT_TEMPS-1; \
-	} while (0)
+static GLuint *t_opcode_ex2(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_EX2, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
 
-static void r300TranslateVertexShader(struct r300_vertex_program *vp,
-				      struct prog_instruction *vpi)
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_exp(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_EXP, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_flr(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3], int *u_temp_i)
+{
+	/* FRC TMP 0.X Y Z W PARAM 0{} {X Y Z W}
+	   ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} TMP 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W */
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_FRC, *u_temp_i,
+			t_dst_mask(vpi->DstReg.WriteMask),
+			VSF_OUT_CLASS_TMP);
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+	inst += 4;
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] =
+	    MAKE_VSF_SOURCE(*u_temp_i, VSF_IN_COMPONENT_X,
+			    VSF_IN_COMPONENT_Y, VSF_IN_COMPONENT_Z,
+			    VSF_IN_COMPONENT_W, VSF_IN_CLASS_TMP,
+			    /* Not 100% sure about this */
+			    (!src[0].
+			     NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE
+			    /*VSF_FLAG_ALL */ );
+
+	inst[3] = ZERO_SRC_0;
+	(*u_temp_i)--;
+
+	return inst;
+}
+
+static GLuint *t_opcode_frc(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_FRC, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_lg2(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	// LG2 RESULT 1.X Y Z W PARAM 0{} {X X X X}
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_LG2, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[0].Swizzle, 0)),
+			    t_src_class(src[0].File),
+			    src[0].
+			    NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+	    (src[0].RelAddr << 4);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_lit(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//LIT TMP 1.Y Z TMP 1{} {X W Z Y} TMP 1{} {Y W Z X} TMP 1{} {Y X Z W}
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_LIT, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	/* NOTE: Users swizzling might not work. */
+	inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
+				  VSF_IN_COMPONENT_ZERO,	// z
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
+				  t_src_class(src[0].File),
+				  src[0].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+	inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
+				  VSF_IN_COMPONENT_ZERO,	// z
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
+				  t_src_class(src[0].File),
+				  src[0].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+	inst[3] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
+				  VSF_IN_COMPONENT_ZERO,	// z
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
+				  t_src_class(src[0].File),
+				  src[0].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+
+	return inst;
+}
+
+static GLuint *t_opcode_log(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_LOG, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_mad(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
 {
-	int i, cur_reg = 0;
-	VERTEX_SHADER_INSTRUCTION *o_inst;
-	unsigned long operands;
-	int are_srcs_scalar;
 	unsigned long hw_op;
-	/* Initial value should be last tmp reg that hw supports.
-	   Strangely enough r300 doesnt mind even though these would be out of range.
-	   Smart enough to realize that it doesnt need it? */
-	int u_temp_i = VSF_MAX_FRAGMENT_TEMPS - 1;
-	struct prog_src_register src[3];
 
-	vp->pos_end = 0;	/* Not supported yet */
-	vp->program.length = 0;
-	/*vp->num_temporaries=mesa_vp->Base.NumTemporaries; */
+	hw_op = (src[0].File == PROGRAM_TEMPORARY
+		 && src[1].File == PROGRAM_TEMPORARY
+		 && src[2].File ==
+		 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = t_src(vp, &src[2]);
+
+	return inst;
+}
+
+static GLuint *t_opcode_max(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_MAX, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
+
+static GLuint *t_opcode_min(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_MIN, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
+
+static GLuint *t_opcode_mov(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{} {ZERO ZERO ZERO ZERO}
+
+#if 1
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+#else
+	hw_op =
+	    (src[0].File ==
+	     PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ONE_SRC_0;
+	inst[3] = ZERO_SRC_0;
+#endif
+
+	return inst;
+}
+
+static GLuint *t_opcode_mul(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	unsigned long hw_op;
+
+	// HW mul can take third arg but appears to have some other limitations.
+
+	hw_op = (src[0].File == PROGRAM_TEMPORARY
+		 && src[1].File ==
+		 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
+
+static GLuint *t_opcode_pow(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_POW, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = t_src_scalar(vp, &src[1]);
+
+	return inst;
+}
+
+static GLuint *t_opcode_rcp(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_RCP, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_rsq(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_RSQ, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src_scalar(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+
+	return inst;
+}
+
+static GLuint *t_opcode_sge(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_SGE, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
+
+static GLuint *t_opcode_slt(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_SLT, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = t_src(vp, &src[1]);
+	inst[3] = ZERO_SRC_1;
+
+	return inst;
+}
+
+static GLuint *t_opcode_sub(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	unsigned long hw_op;
+
+	//ADD RESULT 1.X Y Z W TMP 0{} {X Y Z W} PARAM 1{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W
+
+#if 1
+	hw_op = (src[0].File == PROGRAM_TEMPORARY
+		 && src[1].File ==
+		 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ONE_SRC_0;
+	inst[3] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 2)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 3)),
+			    t_src_class(src[1].File),
+			    (!src[1].
+			     NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+	    (src[1].RelAddr << 4);
+#else
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] =
+	    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 0)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 1)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 2)),
+			    t_swizzle(GET_SWZ(src[1].Swizzle, 3)),
+			    t_src_class(src[1].File),
+			    (!src[1].
+			     NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) |
+	    (src[1].RelAddr << 4);
+	inst[3] = 0;
+#endif
+
+	return inst;
+}
+
+static GLuint *t_opcode_swz(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3])
+{
+	//ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{} {ZERO ZERO ZERO ZERO}
+
+#if 1
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ZERO_SRC_0;
+	inst[3] = ZERO_SRC_0;
+#else
+	hw_op =
+	    (src[0].File ==
+	     PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
+	    R300_VPI_OUT_OP_MAD;
+
+	inst[0] =
+	    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+	inst[1] = t_src(vp, &src[0]);
+	inst[2] = ONE_SRC_0;
+	inst[3] = ZERO_SRC_0;
+#endif
+
+	return inst;
+}
+
+static GLuint *t_opcode_xpd(struct r300_vertex_program *vp,
+			    struct prog_instruction *vpi, GLuint * inst,
+			    struct prog_src_register src[3], int *u_temp_i)
+{
+	/* mul r0, r1.yzxw, r2.zxyw
+	   mad r0, -r2.yzxw, r1.zxyw, r0
+	   NOTE: might need MAD_2
+	 */
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_MAD, *u_temp_i,
+			t_dst_mask(vpi->DstReg.WriteMask),
+			VSF_OUT_CLASS_TMP);
+
+	inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 2)),	// z
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
+				  t_src_class(src[0].File),
+				  src[0].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+
+	inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 2)),	// z
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 0)),	// x
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// w
+				  t_src_class(src[1].File),
+				  src[1].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[1].RelAddr << 4);
+
+	inst[3] = ZERO_SRC_1;
+	inst += 4;
+	(*u_temp_i)--;
+
+	inst[0] =
+	    MAKE_VSF_OP(R300_VPI_OUT_OP_MAD, t_dst_index(vp, &vpi->DstReg),
+			t_dst_mask(vpi->DstReg.WriteMask),
+			t_dst_class(vpi->DstReg.File));
+
+	inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 2)),	// z
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 0)),	// x
+				  t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// w
+				  t_src_class(src[1].File),
+				  (!src[1].
+				   NegateBase) ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[1].RelAddr << 4);
+
+	inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 2)),	// z
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
+				  t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
+				  t_src_class(src[0].File),
+				  src[0].
+				  NegateBase ? VSF_FLAG_ALL :
+				  VSF_FLAG_NONE) | (src[0].RelAddr << 4);
+
+	inst[3] =
+	    MAKE_VSF_SOURCE(*u_temp_i + 1, VSF_IN_COMPONENT_X,
+			    VSF_IN_COMPONENT_Y, VSF_IN_COMPONENT_Z,
+			    VSF_IN_COMPONENT_W, VSF_IN_CLASS_TMP,
+			    VSF_FLAG_NONE);
+
+	return inst;
+}
+
+static void t_inputs_outputs(struct r300_vertex_program *vp)
+{
+	int i;
+	int cur_reg = 0;
 
 	for (i = 0; i < VERT_ATTRIB_MAX; i++)
 		vp->inputs[i] = -1;
@@ -437,39 +1062,71 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp,
 
 	assert(vp->key.OutputsWritten & (1 << VERT_RESULT_HPOS));
 
-	/* Assign outputs */
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_HPOS))
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_HPOS)) {
 		vp->outputs[VERT_RESULT_HPOS] = cur_reg++;
+	}
 
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_PSIZ))
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_PSIZ)) {
 		vp->outputs[VERT_RESULT_PSIZ] = cur_reg++;
+	}
 
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_COL0))
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_COL0)) {
 		vp->outputs[VERT_RESULT_COL0] = cur_reg++;
+	}
 
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_COL1))
-		vp->outputs[VERT_RESULT_COL1] = cur_reg++;
-
-#if 0				/* Not supported yet */
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC0))
-		vp->outputs[VERT_RESULT_BFC0] = cur_reg++;
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_COL1)) {
+		vp->outputs[VERT_RESULT_COL1] =
+		    vp->outputs[VERT_RESULT_COL0] + 1;
+		cur_reg = vp->outputs[VERT_RESULT_COL1] + 1;
+	}
 
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC1))
-		vp->outputs[VERT_RESULT_BFC1] = cur_reg++;
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC0)) {
+		vp->outputs[VERT_RESULT_BFC0] =
+		    vp->outputs[VERT_RESULT_COL0] + 2;
+		cur_reg = vp->outputs[VERT_RESULT_BFC0] + 2;
+	}
 
-	if (vp->key.OutputsWritten & (1 << VERT_RESULT_FOGC))
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC1)) {
+		vp->outputs[VERT_RESULT_BFC1] =
+		    vp->outputs[VERT_RESULT_COL0] + 3;
+		cur_reg = vp->outputs[VERT_RESULT_BFC1] + 1;
+	}
+#if 0
+	if (vp->key.OutputsWritten & (1 << VERT_RESULT_FOGC)) {
 		vp->outputs[VERT_RESULT_FOGC] = cur_reg++;
+	}
 #endif
 
-	for (i = VERT_RESULT_TEX0; i <= VERT_RESULT_TEX7; i++)
-		if (vp->key.OutputsWritten & (1 << i))
+	for (i = VERT_RESULT_TEX0; i <= VERT_RESULT_TEX7; i++) {
+		if (vp->key.OutputsWritten & (1 << i)) {
 			vp->outputs[i] = cur_reg++;
+		}
+	}
+}
+
+static void r300TranslateVertexShader(struct r300_vertex_program *vp,
+				      struct prog_instruction *vpi)
+{
+	int i;
+	GLuint *inst;
+	unsigned long num_operands;
+	/* Initial value should be last tmp reg that hw supports.
+	   Strangely enough r300 doesnt mind even though these would be out of range.
+	   Smart enough to realize that it doesnt need it? */
+	int u_temp_i = VSF_MAX_FRAGMENT_TEMPS - 1;
+	struct prog_src_register src[3];
 
+	vp->pos_end = 0;	/* Not supported yet */
+	vp->program.length = 0;
+	/*vp->num_temporaries=mesa_vp->Base.NumTemporaries; */
 	vp->translated = GL_TRUE;
 	vp->native = GL_TRUE;
 
-	o_inst = vp->program.body.i;
-	for (; vpi->Opcode != OPCODE_END; vpi++, o_inst++) {
+	t_inputs_outputs(vp);
+
+	for (inst = vp->program.body.i; vpi->Opcode != OPCODE_END;
+	     vpi++, inst += 4) {
+
 		FREE_TEMPS();
 
 		if (!valid_dst(vp, &vpi->DstReg)) {
@@ -478,61 +1135,62 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp,
 			vpi->DstReg.Index = u_temp_i;
 		}
 
-		operands = op_operands(vpi->Opcode);
-		are_srcs_scalar = operands & SCALAR_FLAG;
-		operands &= OP_MASK;
+		num_operands = _mesa_num_inst_src_regs(vpi->Opcode);
 
-		for (i = 0; i < operands; i++)
+		/* copy the sources (src) from mesa into a local variable... is this needed? */
+		for (i = 0; i < num_operands; i++) {
 			src[i] = vpi->SrcReg[i];
+		}
 
-		if (operands == 3) {	/* TODO: scalars */
+		if (num_operands == 3) {	/* TODO: scalars */
 			if (CMP_SRCS(src[1], src[2])
 			    || CMP_SRCS(src[0], src[2])) {
-				o_inst->op =
-				    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, u_temp_i,
-						VSF_FLAG_ALL,
+				inst[0] =
+				    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
+						u_temp_i, VSF_FLAG_ALL,
 						VSF_OUT_CLASS_TMP);
 
-				o_inst->src[0] =
-				    MAKE_VSF_SOURCE(t_src_index(vp, &src[2]),
+				inst[1] =
+				    MAKE_VSF_SOURCE(t_src_index
+						    (vp, &src[2]),
 						    SWIZZLE_X, SWIZZLE_Y,
 						    SWIZZLE_Z, SWIZZLE_W,
-						    t_src_class(src[2].File),
-						    VSF_FLAG_NONE) | (src[2].
-								      RelAddr <<
-								      4);
+						    t_src_class(src[2].
+								File),
+						    VSF_FLAG_NONE) |
+				    (src[2].RelAddr << 4);
 
-				o_inst->src[1] = ZERO_SRC_2;
-				o_inst->src[2] = ZERO_SRC_2;
-				o_inst++;
+				inst[2] = ZERO_SRC_2;
+				inst[3] = ZERO_SRC_2;
+				inst += 4;
 
 				src[2].File = PROGRAM_TEMPORARY;
 				src[2].Index = u_temp_i;
 				src[2].RelAddr = 0;
 				u_temp_i--;
 			}
-
 		}
 
-		if (operands >= 2) {
+		if (num_operands >= 2) {
 			if (CMP_SRCS(src[1], src[0])) {
-				o_inst->op =
-				    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, u_temp_i,
-						VSF_FLAG_ALL,
+				inst[0] =
+				    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
+						u_temp_i, VSF_FLAG_ALL,
 						VSF_OUT_CLASS_TMP);
 
-				o_inst->src[0] =
-				    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
+				inst[1] =
+				    MAKE_VSF_SOURCE(t_src_index
+						    (vp, &src[0]),
 						    SWIZZLE_X, SWIZZLE_Y,
 						    SWIZZLE_Z, SWIZZLE_W,
-						    t_src_class(src[0].File),
-						    VSF_FLAG_NONE) | (src[0].
-								      RelAddr <<
-								      4);
+						    t_src_class(src[0].
+								File),
+						    VSF_FLAG_NONE) |
+				    (src[0].RelAddr << 4);
 
-				o_inst->src[1] = ZERO_SRC_0;
-				o_inst->src[2] = ZERO_SRC_0;
-				o_inst++;
+				inst[2] = ZERO_SRC_0;
+				inst[3] = ZERO_SRC_0;
+				inst += 4;
 
 				src[0].File = PROGRAM_TEMPORARY;
 				src[0].Index = u_temp_i;
@@ -541,517 +1199,103 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp,
 			}
 		}
 
-		/* These ops need special handling. */
 		switch (vpi->Opcode) {
-		case OPCODE_POW:
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_POW,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src_scalar(vp, &src[0]);
-			o_inst->src[1] = ZERO_SRC_0;
-			o_inst->src[2] = t_src_scalar(vp, &src[1]);
-			goto next;
-
-		case OPCODE_MOV:	//ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{} {ZERO ZERO ZERO ZERO}
-		case OPCODE_SWZ:
-#if 1
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = ZERO_SRC_0;
-			o_inst->src[2] = ZERO_SRC_0;
-#else
-			hw_op =
-			    (src[0].File ==
-			     PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
-			    R300_VPI_OUT_OP_MAD;
-
-			o_inst->op =
-			    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = ONE_SRC_0;
-			o_inst->src[2] = ZERO_SRC_0;
-#endif
-
-			goto next;
-
+		case OPCODE_ABS:
+			inst = t_opcode_abs(vp, vpi, inst, src);
+			break;
 		case OPCODE_ADD:
-#if 1
-			hw_op = (src[0].File == PROGRAM_TEMPORARY &&
-				 src[1].File ==
-				 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
-			    R300_VPI_OUT_OP_MAD;
-
-			o_inst->op =
-			    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = ONE_SRC_0;
-			o_inst->src[1] = t_src(vp, &src[0]);
-			o_inst->src[2] = t_src(vp, &src[1]);
-#else
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = t_src(vp, &src[1]);
-			o_inst->src[2] = ZERO_SRC_1;
-
-#endif
-			goto next;
-
-		case OPCODE_MAD:
-			hw_op = (src[0].File == PROGRAM_TEMPORARY &&
-				 src[1].File == PROGRAM_TEMPORARY &&
-				 src[2].File ==
-				 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
-			    R300_VPI_OUT_OP_MAD;
-
-			o_inst->op =
-			    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = t_src(vp, &src[1]);
-			o_inst->src[2] = t_src(vp, &src[2]);
-			goto next;
-
-		case OPCODE_MUL:	/* HW mul can take third arg but appears to have some other limitations. */
-			hw_op = (src[0].File == PROGRAM_TEMPORARY &&
-				 src[1].File ==
-				 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
-			    R300_VPI_OUT_OP_MAD;
-
-			o_inst->op =
-			    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = t_src(vp, &src[1]);
-
-			o_inst->src[2] = ZERO_SRC_1;
-			goto next;
-
-		case OPCODE_DP3:	//DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ZERO} PARAM 0{} {X Y Z ZERO}
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_DOT,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 2)),
-					    SWIZZLE_ZERO,
-					    t_src_class(src[0].File),
-					    src[0].
-					    NegateBase ? VSF_FLAG_XYZ :
-					    VSF_FLAG_NONE) | (src[0].
-							      RelAddr << 4);
-
-			o_inst->src[1] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 2)),
-					    SWIZZLE_ZERO,
-					    t_src_class(src[1].File),
-					    src[1].
-					    NegateBase ? VSF_FLAG_XYZ :
-					    VSF_FLAG_NONE) | (src[1].
-							      RelAddr << 4);
-
-			o_inst->src[2] = ZERO_SRC_1;
-			goto next;
-
-		case OPCODE_SUB:	//ADD RESULT 1.X Y Z W TMP 0{} {X Y Z W} PARAM 1{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W
-#if 1
-			hw_op = (src[0].File == PROGRAM_TEMPORARY &&
-				 src[1].File ==
-				 PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 :
-			    R300_VPI_OUT_OP_MAD;
-
-			o_inst->op =
-			    MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = ONE_SRC_0;
-			o_inst->src[2] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 2)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 3)),
-					    t_src_class(src[1].File),
-					    (!src[1].
-					     NegateBase) ? VSF_FLAG_ALL :
-					    VSF_FLAG_NONE) | (src[1].
-							      RelAddr << 4);
-#else
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[1]),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 2)),
-					    t_swizzle(GET_SWZ
-						      (src[1].Swizzle, 3)),
-					    t_src_class(src[1].File),
-					    (!src[1].
-					     NegateBase) ? VSF_FLAG_ALL :
-					    VSF_FLAG_NONE) | (src[1].
-							      RelAddr << 4);
-			o_inst->src[2] = 0;
-#endif
-			goto next;
-
-		case OPCODE_ABS:	//MAX RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_MAX,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 2)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 3)),
-					    t_src_class(src[0].File),
-					    (!src[0].
-					     NegateBase) ? VSF_FLAG_ALL :
-					    VSF_FLAG_NONE) | (src[0].
-							      RelAddr << 4);
-			o_inst->src[2] = 0;
-			goto next;
-
+			inst = t_opcode_add(vp, vpi, inst, src);
+			break;
+		case OPCODE_ARL:
+			inst = t_opcode_arl(vp, vpi, inst, src);
+			break;
+		case OPCODE_DP3:
+			inst = t_opcode_dp3(vp, vpi, inst, src);
+			break;
+		case OPCODE_DP4:
+			inst = t_opcode_dp4(vp, vpi, inst, src);
+			break;
+		case OPCODE_DPH:
+			inst = t_opcode_dph(vp, vpi, inst, src);
+			break;
+		case OPCODE_DST:
+			inst = t_opcode_dst(vp, vpi, inst, src);
+			break;
+		case OPCODE_EX2:
+			inst = t_opcode_ex2(vp, vpi, inst, src);
+			break;
+		case OPCODE_EXP:
+			inst = t_opcode_exp(vp, vpi, inst, src);
+			break;
 		case OPCODE_FLR:
-			/* FRC TMP 0.X Y Z W PARAM 0{} {X Y Z W}
-			   ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} TMP 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W */
-
-			o_inst->op = MAKE_VSF_OP(R300_VPI_OUT_OP_FRC, u_temp_i,
-						 t_dst_mask(vpi->DstReg.
-							    WriteMask),
-						 VSF_OUT_CLASS_TMP);
-
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = ZERO_SRC_0;
-			o_inst->src[2] = ZERO_SRC_0;
-			o_inst++;
-
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_ADD,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] = t_src(vp, &src[0]);
-			o_inst->src[1] = MAKE_VSF_SOURCE(u_temp_i,
-							 VSF_IN_COMPONENT_X,
-							 VSF_IN_COMPONENT_Y,
-							 VSF_IN_COMPONENT_Z,
-							 VSF_IN_COMPONENT_W,
-							 VSF_IN_CLASS_TMP,
-							 /* Not 100% sure about this */
-							 (!src[0].
-							  NegateBase) ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE
-							 /*VSF_FLAG_ALL */ );
-
-			o_inst->src[2] = ZERO_SRC_0;
-			u_temp_i--;
-			goto next;
-
-		case OPCODE_LG2:	// LG2 RESULT 1.X Y Z W PARAM 0{} {X X X X}
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_LG2,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_src_class(src[0].File),
-					    src[0].
-					    NegateBase ? VSF_FLAG_ALL :
-					    VSF_FLAG_NONE) | (src[0].
-							      RelAddr << 4);
-			o_inst->src[1] = ZERO_SRC_0;
-			o_inst->src[2] = ZERO_SRC_0;
-			goto next;
-
-		case OPCODE_LIT:	//LIT TMP 1.Y Z TMP 1{} {X W Z Y} TMP 1{} {Y W Z X} TMP 1{} {Y X Z W}
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_LIT,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-			/* NOTE: Users swizzling might not work. */
-			o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
-							 VSF_IN_COMPONENT_ZERO,	// z
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
-							 t_src_class(src[0].
-								     File),
-							 src[0].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[0].RelAddr << 4);
-			o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
-							 VSF_IN_COMPONENT_ZERO,	// z
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
-							 t_src_class(src[0].
-								     File),
-							 src[0].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[0].RelAddr << 4);
-			o_inst->src[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
-							 VSF_IN_COMPONENT_ZERO,	// z
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
-							 t_src_class(src[0].
-								     File),
-							 src[0].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[0].RelAddr << 4);
-			goto next;
-
-		case OPCODE_DPH:	//DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ONE} PARAM 0{} {X Y Z W}
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_DOT,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] =
-			    MAKE_VSF_SOURCE(t_src_index(vp, &src[0]),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 0)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 1)),
-					    t_swizzle(GET_SWZ
-						      (src[0].Swizzle, 2)),
-					    VSF_IN_COMPONENT_ONE,
-					    t_src_class(src[0].File),
-					    src[0].
-					    NegateBase ? VSF_FLAG_XYZ :
-					    VSF_FLAG_NONE) | (src[0].
-							      RelAddr << 4);
-			o_inst->src[1] = t_src(vp, &src[1]);
-			o_inst->src[2] = ZERO_SRC_1;
-			goto next;
-
-		case OPCODE_XPD:
-			/* mul r0, r1.yzxw, r2.zxyw
-			   mad r0, -r2.yzxw, r1.zxyw, r0
-			   NOTE: might need MAD_2
-			 */
-
-			o_inst->op = MAKE_VSF_OP(R300_VPI_OUT_OP_MAD, u_temp_i,
-						 t_dst_mask(vpi->DstReg.
-							    WriteMask),
-						 VSF_OUT_CLASS_TMP);
-
-			o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 2)),	// z
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
-							 t_src_class(src[0].
-								     File),
-							 src[0].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[0].RelAddr << 4);
-
-			o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 2)),	// z
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 0)),	// x
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// w
-							 t_src_class(src[1].
-								     File),
-							 src[1].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[1].RelAddr << 4);
-
-			o_inst->src[2] = ZERO_SRC_1;
-			o_inst++;
-			u_temp_i--;
-
-			o_inst->op =
-			    MAKE_VSF_OP(R300_VPI_OUT_OP_MAD,
-					t_dst_index(vp, &vpi->DstReg),
-					t_dst_mask(vpi->DstReg.WriteMask),
-					t_dst_class(vpi->DstReg.File));
-
-			o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 2)),	// z
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 0)),	// x
-							 t_swizzle(GET_SWZ(src[1].Swizzle, 3)),	// w
-							 t_src_class(src[1].
-								     File),
-							 (!src[1].
-							  NegateBase) ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[1].RelAddr << 4);
-
-			o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 2)),	// z
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 0)),	// x
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 1)),	// y
-							 t_swizzle(GET_SWZ(src[0].Swizzle, 3)),	// w
-							 t_src_class(src[0].
-								     File),
-							 src[0].
-							 NegateBase ?
-							 VSF_FLAG_ALL :
-							 VSF_FLAG_NONE) |
-			    (src[0].RelAddr << 4);
-
-			o_inst->src[2] = MAKE_VSF_SOURCE(u_temp_i + 1,
-							 VSF_IN_COMPONENT_X,
-							 VSF_IN_COMPONENT_Y,
-							 VSF_IN_COMPONENT_Z,
-							 VSF_IN_COMPONENT_W,
-							 VSF_IN_CLASS_TMP,
-							 VSF_FLAG_NONE);
-
-			goto next;
-
-		case OPCODE_RCC:
-			fprintf(stderr, "Dont know how to handle op %d yet\n",
-				vpi->Opcode);
-			_mesa_exit(-1);
+			inst =
+			    t_opcode_flr(vp, vpi, inst, src, /* FIXME */
+					 &u_temp_i);
+			break;
+		case OPCODE_FRC:
+			inst = t_opcode_frc(vp, vpi, inst, src);
+			break;
+		case OPCODE_LG2:
+			inst = t_opcode_lg2(vp, vpi, inst, src);
+			break;
+		case OPCODE_LIT:
+			inst = t_opcode_lit(vp, vpi, inst, src);
+			break;
+		case OPCODE_LOG:
+			inst = t_opcode_log(vp, vpi, inst, src);
+			break;
+		case OPCODE_MAD:
+			inst = t_opcode_mad(vp, vpi, inst, src);
+			break;
+		case OPCODE_MAX:
+			inst = t_opcode_max(vp, vpi, inst, src);
+			break;
+		case OPCODE_MIN:
+			inst = t_opcode_min(vp, vpi, inst, src);
+			break;
+		case OPCODE_MOV:
+			inst = t_opcode_mov(vp, vpi, inst, src);
 			break;
-		case OPCODE_END:
+		case OPCODE_MUL:
+			inst = t_opcode_mul(vp, vpi, inst, src);
+			break;
+		case OPCODE_POW:
+			inst = t_opcode_pow(vp, vpi, inst, src);
+			break;
+		case OPCODE_RCP:
+			inst = t_opcode_rcp(vp, vpi, inst, src);
+			break;
+		case OPCODE_RSQ:
+			inst = t_opcode_rsq(vp, vpi, inst, src);
+			break;
+		case OPCODE_SGE:
+			inst = t_opcode_sge(vp, vpi, inst, src);
+			break;
+		case OPCODE_SLT:
+			inst = t_opcode_slt(vp, vpi, inst, src);
+			break;
+		case OPCODE_SUB:
+			inst = t_opcode_sub(vp, vpi, inst, src);
+			break;
+		case OPCODE_SWZ:
+			inst = t_opcode_swz(vp, vpi, inst, src);
+			break;
+		case OPCODE_XPD:
+			inst =
+			    t_opcode_xpd(vp, vpi, inst, src, /* FIXME */
+					 &u_temp_i);
 			break;
 		default:
+			assert(0);
 			break;
 		}
-
-		o_inst->op =
-		    MAKE_VSF_OP(t_opcode(vpi->Opcode),
-				t_dst_index(vp, &vpi->DstReg),
-				t_dst_mask(vpi->DstReg.WriteMask),
-				t_dst_class(vpi->DstReg.File));
-
-		if (are_srcs_scalar) {
-			switch (operands) {
-			case 1:
-				o_inst->src[0] = t_src_scalar(vp, &src[0]);
-				o_inst->src[1] = ZERO_SRC_0;
-				o_inst->src[2] = ZERO_SRC_0;
-				break;
-
-			case 2:
-				o_inst->src[0] = t_src_scalar(vp, &src[0]);
-				o_inst->src[1] = t_src_scalar(vp, &src[1]);
-				o_inst->src[2] = ZERO_SRC_1;
-				break;
-
-			case 3:
-				o_inst->src[0] = t_src_scalar(vp, &src[0]);
-				o_inst->src[1] = t_src_scalar(vp, &src[1]);
-				o_inst->src[2] = t_src_scalar(vp, &src[2]);
-				break;
-
-			default:
-				fprintf(stderr,
-					"scalars and op RCC not handled yet");
-				_mesa_exit(-1);
-				break;
-			}
-		} else {
-			switch (operands) {
-			case 1:
-				o_inst->src[0] = t_src(vp, &src[0]);
-				o_inst->src[1] = ZERO_SRC_0;
-				o_inst->src[2] = ZERO_SRC_0;
-				break;
-
-			case 2:
-				o_inst->src[0] = t_src(vp, &src[0]);
-				o_inst->src[1] = t_src(vp, &src[1]);
-				o_inst->src[2] = ZERO_SRC_1;
-				break;
-
-			case 3:
-				o_inst->src[0] = t_src(vp, &src[0]);
-				o_inst->src[1] = t_src(vp, &src[1]);
-				o_inst->src[2] = t_src(vp, &src[2]);
-				break;
-
-			default:
-				fprintf(stderr,
-					"scalars and op RCC not handled yet");
-				_mesa_exit(-1);
-				break;
-			}
-		}
-	      next:;
 	}
 
-	/* Will most likely segfault before we get here... fix later. */
-	if (o_inst - vp->program.body.i >= VSF_MAX_FRAGMENT_LENGTH / 4) {
+	vp->program.length = (inst - vp->program.body.i);
+	if (vp->program.length >= VSF_MAX_FRAGMENT_LENGTH) {
 		vp->program.length = 0;
 		vp->native = GL_FALSE;
-		return;
 	}
-	vp->program.length = (o_inst - vp->program.body.i) * 4;
 #if 0
 	fprintf(stderr, "hw program:\n");
 	for (i = 0; i < vp->program.length; i++)
@@ -1065,7 +1309,8 @@ static void position_invariant(struct gl_program *prog)
 	struct gl_program_parameter_list *paramList;
 	int i;
 
-	gl_state_index tokens[STATE_LENGTH] = { STATE_MVP_MATRIX, 0, 0, 0, 0 };
+	gl_state_index tokens[STATE_LENGTH] =
+	    { STATE_MVP_MATRIX, 0, 0, 0, 0 };
 
 	/* tokens[4] = matrix modifier */
 #ifdef PREFER_DP4
@@ -1159,8 +1404,8 @@ static void insert_wpos(struct r300_vertex_program *vp,
 				prog->NumInstructions - 1);
 	/* END */
 	_mesa_copy_instructions(&vpi[prog->NumInstructions + 1],
-				&prog->Instructions[prog->NumInstructions - 1],
-				1);
+				&prog->Instructions[prog->NumInstructions -
+						    1], 1);
 	vpi_insert = &vpi[prog->NumInstructions - 1];
 
 	vpi_insert[i].Opcode = OPCODE_MOV;
@@ -1206,8 +1451,8 @@ static void pos_as_texcoord(struct r300_vertex_program *vp,
 	prog->NumTemporaries++;
 
 	for (vpi = prog->Instructions; vpi->Opcode != OPCODE_END; vpi++) {
-		if (vpi->DstReg.File == PROGRAM_OUTPUT &&
-		    vpi->DstReg.Index == VERT_RESULT_HPOS) {
+		if (vpi->DstReg.File == PROGRAM_OUTPUT
+		    && vpi->DstReg.Index == VERT_RESULT_HPOS) {
 			vpi->DstReg.File = PROGRAM_TEMPORARY;
 			vpi->DstReg.Index = tempregi;
 		}
@@ -1223,20 +1468,18 @@ static struct r300_vertex_program *build_program(struct r300_vertex_program_key
 
 	vp = _mesa_calloc(sizeof(*vp));
 	_mesa_memcpy(&vp->key, wanted_key, sizeof(vp->key));
-
 	vp->wpos_idx = wpos_idx;
 
 	if (mesa_vp->IsPositionInvariant) {
 		position_invariant(&mesa_vp->Base);
 	}
 
-	if (wpos_idx > -1)
+	if (wpos_idx > -1) {
 		pos_as_texcoord(vp, &mesa_vp->Base);
+	}
 
 	assert(mesa_vp->Base.NumInstructions);
-
 	vp->num_temporaries = mesa_vp->Base.NumTemporaries;
-
 	r300TranslateVertexShader(vp, mesa_vp->Base.Instructions);
 
 	return vp;
@@ -1252,11 +1495,10 @@ void r300SelectVertexShader(r300ContextPtr r300)
 	struct r300_vertex_program *vp;
 	GLint wpos_idx;
 
-	vpc = (struct r300_vertex_program_cont *)ctx->VertexProgram._Current;
+	vpc =
+	    (struct r300_vertex_program_cont *)ctx->VertexProgram._Current;
 	InputsRead = ctx->FragmentProgram._Current->Base.InputsRead;
 
-	wanted_key.OutputsWritten |= 1 << VERT_RESULT_HPOS;
-
 	wpos_idx = -1;
 	if (InputsRead & FRAG_BIT_WPOS) {
 		for (i = 0; i < ctx->Const.MaxTextureUnits; i++)
@@ -1271,28 +1513,35 @@ void r300SelectVertexShader(r300ContextPtr r300)
 		InputsRead |= (FRAG_BIT_TEX0 << i);
 		wpos_idx = i;
 	}
+	wanted_key.InputsRead = vpc->mesa_program.Base.InputsRead;
+	wanted_key.OutputsWritten = vpc->mesa_program.Base.OutputsWritten;
+
+	wanted_key.OutputsWritten |= 1 << VERT_RESULT_HPOS;
 
-	if (InputsRead & FRAG_BIT_COL0)
+	if (InputsRead & FRAG_BIT_COL0) {
 		wanted_key.OutputsWritten |= 1 << VERT_RESULT_COL0;
+	}
 
-	if ((InputsRead & FRAG_BIT_COL1)	/*||
-						   (InputsRead & FRAG_BIT_FOGC) */ )
+	if ((InputsRead & FRAG_BIT_COL1)) {
 		wanted_key.OutputsWritten |= 1 << VERT_RESULT_COL1;
+	}
 
-	for (i = 0; i < ctx->Const.MaxTextureUnits; i++)
-		if (InputsRead & (FRAG_BIT_TEX0 << i))
+	for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
+		if (InputsRead & (FRAG_BIT_TEX0 << i)) {
 			wanted_key.OutputsWritten |=
 			    1 << (VERT_RESULT_TEX0 + i);
+		}
+	}
 
-	wanted_key.InputsRead = vpc->mesa_program.Base.InputsRead;
 	if (vpc->mesa_program.IsPositionInvariant) {
 		/* we wan't position don't we ? */
 		wanted_key.InputsRead |= (1 << VERT_ATTRIB_POS);
+		wanted_key.OutputsWritten |= (1 << VERT_RESULT_HPOS);
 	}
 
 	for (vp = vpc->progs; vp; vp = vp->next)
-		if (_mesa_memcmp(&vp->key, &wanted_key, sizeof(wanted_key)) ==
-		    0) {
+		if (_mesa_memcmp(&vp->key, &wanted_key, sizeof(wanted_key))
+		    == 0) {
 			r300->selected_vp = vp;
 			return;
 		}
diff --git a/src/mesa/drivers/dri/r300/r300_vertprog.h b/src/mesa/drivers/dri/r300/r300_vertprog.h
index 252d5a901f..3df0eee799 100644
--- a/src/mesa/drivers/dri/r300/r300_vertprog.h
+++ b/src/mesa/drivers/dri/r300/r300_vertprog.h
@@ -3,11 +3,6 @@
 
 #include "r300_reg.h"
 
-typedef struct {
-	GLuint op;
-	GLuint src[3];
-} VERTEX_SHADER_INSTRUCTION;
-
 #define VSF_FLAG_X	1
 #define VSF_FLAG_Y	2
 #define VSF_FLAG_Z	4
diff --git a/src/mesa/drivers/dri/r300/radeon_lock.h b/src/mesa/drivers/dri/r300/radeon_lock.h
index c47adc9575..a344837f47 100644
--- a/src/mesa/drivers/dri/r300/radeon_lock.h
+++ b/src/mesa/drivers/dri/r300/radeon_lock.h
@@ -42,9 +42,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #ifndef __RADEON_LOCK_H__
 #define __RADEON_LOCK_H__
 
-#if 0
-#include "r200_ioctl.h"
-#endif
 #include "radeon_context.h"
 
 extern void radeonGetLock(radeonContextPtr rmesa, GLuint flags);
diff --git a/src/mesa/drivers/dri/savage/savagetris.c b/src/mesa/drivers/dri/savage/savagetris.c
index 3dd821a4d3..4ce2f60b4f 100644
--- a/src/mesa/drivers/dri/savage/savagetris.c
+++ b/src/mesa/drivers/dri/savage/savagetris.c
@@ -131,7 +131,9 @@ static __inline__ void savage_draw_point (savageContextPtr imesa,
    u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize);
    const GLfloat x = tmp->v.x;
    const GLfloat y = tmp->v.y;
-   const GLfloat sz = imesa->glCtx->Point._Size * .5;
+   const GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size,
+                                  imesa->glCtx->Const.MinPointSize,
+                                  imesa->glCtx->Const.MaxPointSize);
    GLuint j;
 
    *(float *)&vb[0] = x - sz;
@@ -164,7 +166,9 @@ static __inline__ void savage_draw_line (savageContextPtr imesa,
 					 savageVertexPtr v1 ) {
    GLuint vertsize = imesa->HwVertexSize;
    u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize);
-   GLfloat width = imesa->glCtx->Line._Width;
+   const GLfloat width = CLAMP(imesa->glCtx->Line.Width,
+                               imesa->glCtx->Const.MinLineWidth,
+                               imesa->glCtx->Const.MaxLineWidth);
    GLfloat dx, dy, ix, iy;
    GLuint j;
 
@@ -234,7 +238,9 @@ static __inline__ void savage_ptex_line (savageContextPtr imesa,
 					 savageVertexPtr v1 ) {
    GLuint vertsize = imesa->HwVertexSize;
    u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize);
-   GLfloat width = imesa->glCtx->Line._Width;
+   const GLfloat width = CLAMP(imesa->glCtx->Line.Width,
+                               imesa->glCtx->Const.MinLineWidth,
+                               imesa->glCtx->Const.MaxLineWidth);
    GLfloat dx, dy, ix, iy;
    savageVertex tmp0, tmp1;
    GLuint j;
@@ -281,7 +287,9 @@ static __inline__ void savage_ptex_point (savageContextPtr imesa,
    u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize);
    const GLfloat x = v0->v.x;
    const GLfloat y = v0->v.y;
-   const GLfloat sz = imesa->glCtx->Point._Size * .5;
+   const GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size,
+                                  imesa->glCtx->Const.MinPointSize,
+                                  imesa->glCtx->Const.MaxPointSize);
    savageVertex tmp;
    GLuint j;
 
diff --git a/src/mesa/drivers/dri/tdfx/tdfx_tris.c b/src/mesa/drivers/dri/tdfx/tdfx_tris.c
index 96f9ae27fc..7252a7e7dc 100644
--- a/src/mesa/drivers/dri/tdfx/tdfx_tris.c
+++ b/src/mesa/drivers/dri/tdfx/tdfx_tris.c
@@ -184,7 +184,7 @@ tdfx_translate_vertex( GLcontext *ctx, const tdfxVertex *src, SWvertex *dst)
       }
    }
 
-   dst->pointSize = ctx->Point._Size;
+   dst->pointSize = ctx->Point.Size;
 }
 
 
diff --git a/src/mesa/drivers/fbdev/glfbdev.c b/src/mesa/drivers/fbdev/glfbdev.c
index 6c6511b7e5..e95a424698 100644
--- a/src/mesa/drivers/fbdev/glfbdev.c
+++ b/src/mesa/drivers/fbdev/glfbdev.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.1
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -63,6 +63,9 @@
 #include "drivers/common/driverfuncs.h"
 
 
+/**
+ * Pixel formats we support:
+ */
 #define PF_B8G8R8     1
 #define PF_B8G8R8A8   2
 #define PF_B5G6R5     3
@@ -70,7 +73,7 @@
 #define PF_CI8        5
 
 
-/*
+/**
  * Derived from Mesa's GLvisual class.
  */
 struct GLFBDevVisualRec {
@@ -80,7 +83,7 @@ struct GLFBDevVisualRec {
    int pixelFormat;
 };
 
-/*
+/**
  * Derived from Mesa's GLframebuffer class.
  */
 struct GLFBDevBufferRec {
@@ -92,7 +95,7 @@ struct GLFBDevBufferRec {
    GLuint bytesPerPixel;
 };
 
-/*
+/**
  * Derived from Mesa's GLcontext class.
  */
 struct GLFBDevContextRec {
@@ -103,7 +106,7 @@ struct GLFBDevContextRec {
    GLFBDevBufferPtr curBuffer;
 };
 
-/*
+/**
  * Derived from Mesa's gl_renderbuffer class.
  */
 struct GLFBDevRenderbufferRec {
@@ -114,11 +117,6 @@ struct GLFBDevRenderbufferRec {
 };
 
 
-
-#define GLFBDEV_CONTEXT(CTX)  ((GLFBDevContextPtr) (CTX))
-#define GLFBDEV_BUFFER(BUF)  ((GLFBDevBufferPtr) (BUF))
-
-
 /**********************************************************************/
 /* Internal device driver functions                                   */
 /**********************************************************************/
@@ -151,7 +149,7 @@ update_state( GLcontext *ctx, GLuint new_state )
 static void
 get_buffer_size( GLframebuffer *buffer, GLuint *width, GLuint *height )
 {
-   const GLFBDevBufferPtr fbdevbuffer = GLFBDEV_BUFFER(buffer);
+   const GLFBDevBufferPtr fbdevbuffer = (GLFBDevBufferPtr) buffer;
    *width = fbdevbuffer->var.xres;
    *height = fbdevbuffer->var.yres;
 }
@@ -389,8 +387,8 @@ glFBDevCreateVisual( const struct fb_fix_screeninfo *fixInfo,
          /* ignored for now */
          break;
       case GLFBDEV_MULTISAMPLE:
-	 numSamples = attrib[1];
-	 attrib++;
+         numSamples = attrib[1];
+         attrib++;
          break;
       default:
          /* unexpected token */
@@ -406,36 +404,36 @@ glFBDevCreateVisual( const struct fb_fix_screeninfo *fixInfo,
       alphaBits = varInfo->transp.length;
 
       if (fixInfo->visual == FB_VISUAL_TRUECOLOR ||
-	  fixInfo->visual == FB_VISUAL_DIRECTCOLOR) {
-	 if(varInfo->bits_per_pixel == 24
-	    && varInfo->red.offset == 16
-	    && varInfo->green.offset == 8
-	    && varInfo->blue.offset == 0)
-	    vis->pixelFormat = PF_B8G8R8;
-
-	 else if(varInfo->bits_per_pixel == 32
-		 && varInfo->red.offset == 16
-		 && varInfo->green.offset == 8
-		 && varInfo->blue.offset == 0)
-	    vis->pixelFormat = PF_B8G8R8A8;
-
-	 else if(varInfo->bits_per_pixel == 16
-		 && varInfo->red.offset == 11
-		 && varInfo->green.offset == 5
-		 && varInfo->blue.offset == 0)
-	    vis->pixelFormat = PF_B5G6R5;
-
-	 else if(varInfo->bits_per_pixel == 16
-		 && varInfo->red.offset == 10
-		 && varInfo->green.offset == 5
-		 && varInfo->blue.offset == 0)
-	    vis->pixelFormat = PF_B5G5R5;
-
-	 else {
-	    _mesa_problem(NULL, "Unsupported fbdev RGB visual/bitdepth!\n");
-	    _mesa_free(vis);
-	    return NULL;
-	 }
+          fixInfo->visual == FB_VISUAL_DIRECTCOLOR) {
+         if (varInfo->bits_per_pixel == 24
+             && varInfo->red.offset == 16
+             && varInfo->green.offset == 8
+             && varInfo->blue.offset == 0) {
+            vis->pixelFormat = PF_B8G8R8;
+         }
+         else if (varInfo->bits_per_pixel == 32
+                  && varInfo->red.offset == 16
+                  && varInfo->green.offset == 8
+                  && varInfo->blue.offset == 0) {
+            vis->pixelFormat = PF_B8G8R8A8;
+         }
+         else if (varInfo->bits_per_pixel == 16
+                  && varInfo->red.offset == 11
+                  && varInfo->green.offset == 5
+                  && varInfo->blue.offset == 0) {
+            vis->pixelFormat = PF_B5G6R5;
+         }
+         else if (varInfo->bits_per_pixel == 16
+                  && varInfo->red.offset == 10
+                  && varInfo->green.offset == 5
+                  && varInfo->blue.offset == 0) {
+            vis->pixelFormat = PF_B5G5R5;
+         }
+         else {
+            _mesa_problem(NULL, "Unsupported fbdev RGB visual/bitdepth!\n");
+            _mesa_free(vis);
+            return NULL;
+         }
       }
    }
    else {
@@ -578,7 +576,7 @@ new_glfbdev_renderbuffer(void *bufferStart, const GLFBDevVisualPtr visual)
 
       rb->rowStride = visual->var.xres_virtual * visual->var.bits_per_pixel / 8;
       rb->bottom = (GLubyte *) bufferStart
-	  + (visual->var.yres - 1) * rb->rowStride;
+                 + (visual->var.yres - 1) * rb->rowStride;
 
       rb->Base.Width = visual->var.xres;
       rb->Base.Height = visual->var.yres;
@@ -635,7 +633,7 @@ glFBDevCreateBuffer( const struct fb_fix_screeninfo *fixInfo,
                           &frontrb->Base);
    /* add back renderbuffer */
    if (visual->glvisual.doubleBufferMode) {
-      int malloced = !backBuffer;
+      const int malloced = !backBuffer;
       if (malloced) {
          /* malloc a back buffer */
          backBuffer = _mesa_malloc(size);
@@ -647,8 +645,11 @@ glFBDevCreateBuffer( const struct fb_fix_screeninfo *fixInfo,
       }
 
       backrb = new_glfbdev_renderbuffer(backBuffer, visual);
-      if(malloced)
-	 backrb->mallocedBuffer = GL_TRUE;
+      if (!backrb) {
+         /* out of mem */
+         return NULL;
+      }
+      backrb->mallocedBuffer = malloced;
 
       _mesa_add_renderbuffer(&buf->glframebuffer, BUFFER_BACK_LEFT,
                              &backrb->Base);
@@ -682,16 +683,10 @@ glFBDevDestroyBuffer( GLFBDevBufferPtr buffer )
       if (buffer == curDraw || buffer == curRead) {
          glFBDevMakeCurrent( NULL, NULL, NULL);
       }
-#if 0
-      /* free the software depth, stencil, accum buffers */
-      _mesa_free_framebuffer_data(&buffer->glframebuffer);
-      _mesa_free(buffer);
-#else
       {
          struct gl_framebuffer *fb = &buffer->glframebuffer;
          _mesa_unreference_framebuffer(&fb);
       }
-#endif
    }
 }
 
diff --git a/src/mesa/drivers/windows/gdi/wgl.c b/src/mesa/drivers/windows/gdi/wgl.c
index 197de0743c..dad3dc1160 100644
--- a/src/mesa/drivers/windows/gdi/wgl.c
+++ b/src/mesa/drivers/windows/gdi/wgl.c
@@ -32,10 +32,31 @@
 
 /* We're essentially building part of GDI here, so define this so that
  * we get the right export linkage. */
+#ifdef __MINGW32__
+
+#include <stdarg.h>
+#include <windef.h>
+#include <wincon.h>
+#include <winbase.h>
+
+#  if defined(BUILD_GL32)
+#    define WINGDIAPI __declspec(dllexport)	
+#  else
+#    define __W32API_USE_DLLIMPORT__
+#  endif
+
+#include <wingdi.h>
+#include "GL/mesa_wgl.h"
+#include <stdlib.h>
+
+#else
+
 #define _GDI32_
 #include <windows.h>
-#include "glapi.h"
 
+#endif
+
+#include "glapi.h"
 #include "GL/wmesa.h"   /* protos for wmesa* functions */
 
 /*
@@ -333,7 +354,7 @@ WINGDIAPI int GLAPIENTRY wglGetPixelFormat(HDC hdc)
 }
 
 WINGDIAPI BOOL GLAPIENTRY wglSetPixelFormat(HDC hdc,int iPixelFormat,
-					    PIXELFORMATDESCRIPTOR *ppfd)
+					const PIXELFORMATDESCRIPTOR *ppfd)
 {
     (void) hdc;
     
@@ -386,12 +407,12 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar,
     
     bitDevice = CreateCompatibleDC(fontDevice);
     
-    // Swap fore and back colors so the bitmap has the right polarity
+    /* Swap fore and back colors so the bitmap has the right polarity */
     tempColor = GetBkColor(bitDevice);
     SetBkColor(bitDevice, GetTextColor(bitDevice));
     SetTextColor(bitDevice, tempColor);
     
-    // Place chars based on base line
+    /* Place chars based on base line */
     VERIFY(SetTextAlign(bitDevice, TA_BASELINE) != GDI_ERROR ? 1 : 0);
     
     for(i = 0; i < (int)numChars; i++) {
@@ -404,36 +425,36 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar,
 	
 	curChar = (char)(i + firstChar);
 	
-	// Find how high/wide this character is
+	/* Find how high/wide this character is */
 	VERIFY(GetTextExtentPoint32(bitDevice, &curChar, 1, &size));
 	
-	// Create the output bitmap
+	/* Create the output bitmap */
 	charWidth = size.cx;
 	charHeight = size.cy;
-	// Round up to the next multiple of 32 bits
+	/* Round up to the next multiple of 32 bits */
 	bmapWidth = ((charWidth + 31) / 32) * 32;   
 	bmapHeight = charHeight;
 	bitObject = CreateCompatibleBitmap(bitDevice,
 					   bmapWidth,
 					   bmapHeight);
-	//VERIFY(bitObject);
+	/* VERIFY(bitObject); */
 	
-	// Assign the output bitmap to the device
+	/* Assign the output bitmap to the device */
 	origBmap = SelectObject(bitDevice, bitObject);
 	(void) VERIFY(origBmap);
 	
 	VERIFY( PatBlt( bitDevice, 0, 0, bmapWidth, bmapHeight,BLACKNESS ) );
 	
-	// Use our source font on the device
+	/* Use our source font on the device */
 	VERIFY(SelectObject(bitDevice, GetCurrentObject(fontDevice,OBJ_FONT)));
 	
-	// Draw the character
+	/* Draw the character */
 	VERIFY(TextOut(bitDevice, 0, metric.tmAscent, &curChar, 1));
 	
-	// Unselect our bmap object
+	/* Unselect our bmap object */
 	VERIFY(SelectObject(bitDevice, origBmap));
 	
-	// Convert the display dependant representation to a 1 bit deep DIB
+	/* Convert the display dependant representation to a 1 bit deep DIB */
 	numBytes = (bmapWidth * bmapHeight) / 8;
 	bmap = malloc(numBytes);
 	dibInfo->bmiHeader.biWidth = bmapWidth;
@@ -441,24 +462,24 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar,
 	res = GetDIBits(bitDevice, bitObject, 0, bmapHeight, bmap,
 			dibInfo,
 			DIB_RGB_COLORS);
-	//VERIFY(res);
+	/* VERIFY(res); */
 	
-	// Create the GL object
+	/* Create the GL object */
 	glNewList(i + listBase, GL_COMPILE);
 	glBitmap(bmapWidth, bmapHeight, 0.0, (GLfloat)metric.tmDescent,
 		 (GLfloat)charWidth, 0.0,
 		 bmap);
 	glEndList();
-	// CheckGL();
+	/* CheckGL(); */
 	
-	// Destroy the bmap object
+	/* Destroy the bmap object */
 	DeleteObject(bitObject);
 	
-	// Deallocate the bitmap data
+	/* Deallocate the bitmap data */
 	free(bmap);
     }
     
-    // Destroy the DC
+    /* Destroy the DC */
     VERIFY(DeleteDC(bitDevice));
     
     free(dibInfo);
diff --git a/src/mesa/drivers/windows/gdi/wmesa.c b/src/mesa/drivers/windows/gdi/wmesa.c
index c5cd2c615b..5b67439f0f 100644
--- a/src/mesa/drivers/windows/gdi/wmesa.c
+++ b/src/mesa/drivers/windows/gdi/wmesa.c
@@ -6,6 +6,7 @@
 #include "wmesadef.h"
 #include "colors.h"
 #include <GL/wmesa.h>
+#include <winuser.h>
 #include "context.h"
 #include "extensions.h"
 #include "framebuffer.h"
@@ -114,7 +115,7 @@ static void wmSetPixelFormat(WMesaFramebuffer pwfb, HDC hDC)
 {
     pwfb->cColorBits = GetDeviceCaps(hDC, BITSPIXEL);
 
-    // Only 16 and 32 bit targets are supported now
+    /* Only 16 and 32 bit targets are supported now */
     assert(pwfb->cColorBits == 0 ||
 	   pwfb->cColorBits == 16 || 
 	   pwfb->cColorBits == 32);
@@ -1171,7 +1172,7 @@ WMesaContext WMesaCreateContext(HDC hDC,
     /* I do not understand this contributed code */
     /* Support memory and device contexts */
     if(WindowFromDC(hDC) != NULL) {
-	c->hDC = GetDC(WindowFromDC(hDC)); // huh ????
+	c->hDC = GetDC(WindowFromDC(hDC)); /* huh ???? */
     }
     else {
 	c->hDC = hDC;
@@ -1404,70 +1405,72 @@ void WMesaSwapBuffers( HDC hdc )
  * table entries.  Hopefully, I'll find a better solution.  The
  * dispatch table generation scripts ought to be making these dummy
  * stubs as well. */
-void gl_dispatch_stub_543(void){};
-void gl_dispatch_stub_544(void){};
-void gl_dispatch_stub_545(void){};
-void gl_dispatch_stub_546(void){};
-void gl_dispatch_stub_547(void){};
-void gl_dispatch_stub_548(void){};
-void gl_dispatch_stub_549(void){};
-void gl_dispatch_stub_550(void){};
-void gl_dispatch_stub_551(void){};
-void gl_dispatch_stub_552(void){};
-void gl_dispatch_stub_553(void){};
-void gl_dispatch_stub_554(void){};
-void gl_dispatch_stub_555(void){};
-void gl_dispatch_stub_556(void){};
-void gl_dispatch_stub_557(void){};
-void gl_dispatch_stub_558(void){};
-void gl_dispatch_stub_559(void){};
-void gl_dispatch_stub_560(void){};
-void gl_dispatch_stub_561(void){};
-void gl_dispatch_stub_565(void){};
-void gl_dispatch_stub_566(void){};
-void gl_dispatch_stub_577(void){};
-void gl_dispatch_stub_578(void){};
-void gl_dispatch_stub_603(void){};
-void gl_dispatch_stub_645(void){};
-void gl_dispatch_stub_646(void){};
-void gl_dispatch_stub_647(void){};
-void gl_dispatch_stub_648(void){};
-void gl_dispatch_stub_649(void){};
-void gl_dispatch_stub_650(void){};
-void gl_dispatch_stub_651(void){};
-void gl_dispatch_stub_652(void){};
-void gl_dispatch_stub_653(void){};
-void gl_dispatch_stub_734(void){};
-void gl_dispatch_stub_735(void){};
-void gl_dispatch_stub_736(void){};
-void gl_dispatch_stub_737(void){};
-void gl_dispatch_stub_738(void){};
-void gl_dispatch_stub_745(void){};
-void gl_dispatch_stub_746(void){};
-void gl_dispatch_stub_760(void){};
-void gl_dispatch_stub_761(void){};
-void gl_dispatch_stub_766(void){};
-void gl_dispatch_stub_767(void){};
-void gl_dispatch_stub_768(void){};
-
-void gl_dispatch_stub_562(void){};
-void gl_dispatch_stub_563(void){};
-void gl_dispatch_stub_564(void){};
-void gl_dispatch_stub_567(void){};
-void gl_dispatch_stub_568(void){};
-void gl_dispatch_stub_569(void){};
-void gl_dispatch_stub_580(void){};
-void gl_dispatch_stub_581(void){};
-void gl_dispatch_stub_606(void){};
-void gl_dispatch_stub_654(void){};
-void gl_dispatch_stub_655(void){};
-void gl_dispatch_stub_656(void){};
-void gl_dispatch_stub_739(void){};
-void gl_dispatch_stub_740(void){};
-void gl_dispatch_stub_741(void){};
-void gl_dispatch_stub_748(void){};
-void gl_dispatch_stub_749(void){};
-void gl_dispatch_stub_769(void){};
-void gl_dispatch_stub_770(void){};
-void gl_dispatch_stub_771(void){};
+#if !defined(__MINGW32__) || !defined(GL_NO_STDCALL)
+void gl_dispatch_stub_543(void){}
+void gl_dispatch_stub_544(void){}
+void gl_dispatch_stub_545(void){}
+void gl_dispatch_stub_546(void){}
+void gl_dispatch_stub_547(void){}
+void gl_dispatch_stub_548(void){}
+void gl_dispatch_stub_549(void){}
+void gl_dispatch_stub_550(void){}
+void gl_dispatch_stub_551(void){}
+void gl_dispatch_stub_552(void){}
+void gl_dispatch_stub_553(void){}
+void gl_dispatch_stub_554(void){}
+void gl_dispatch_stub_555(void){}
+void gl_dispatch_stub_556(void){}
+void gl_dispatch_stub_557(void){}
+void gl_dispatch_stub_558(void){}
+void gl_dispatch_stub_559(void){}
+void gl_dispatch_stub_560(void){}
+void gl_dispatch_stub_561(void){}
+void gl_dispatch_stub_565(void){}
+void gl_dispatch_stub_566(void){}
+void gl_dispatch_stub_577(void){}
+void gl_dispatch_stub_578(void){}
+void gl_dispatch_stub_603(void){}
+void gl_dispatch_stub_645(void){}
+void gl_dispatch_stub_646(void){}
+void gl_dispatch_stub_647(void){}
+void gl_dispatch_stub_648(void){}
+void gl_dispatch_stub_649(void){}
+void gl_dispatch_stub_650(void){}
+void gl_dispatch_stub_651(void){}
+void gl_dispatch_stub_652(void){}
+void gl_dispatch_stub_653(void){}
+void gl_dispatch_stub_734(void){}
+void gl_dispatch_stub_735(void){}
+void gl_dispatch_stub_736(void){}
+void gl_dispatch_stub_737(void){}
+void gl_dispatch_stub_738(void){}
+void gl_dispatch_stub_745(void){}
+void gl_dispatch_stub_746(void){}
+void gl_dispatch_stub_760(void){}
+void gl_dispatch_stub_761(void){}
+void gl_dispatch_stub_766(void){}
+void gl_dispatch_stub_767(void){}
+void gl_dispatch_stub_768(void){}
+
+void gl_dispatch_stub_562(void){}
+void gl_dispatch_stub_563(void){}
+void gl_dispatch_stub_564(void){}
+void gl_dispatch_stub_567(void){}
+void gl_dispatch_stub_568(void){}
+void gl_dispatch_stub_569(void){}
+void gl_dispatch_stub_580(void){}
+void gl_dispatch_stub_581(void){}
+void gl_dispatch_stub_606(void){}
+void gl_dispatch_stub_654(void){}
+void gl_dispatch_stub_655(void){}
+void gl_dispatch_stub_656(void){}
+void gl_dispatch_stub_739(void){}
+void gl_dispatch_stub_740(void){}
+void gl_dispatch_stub_741(void){}
+void gl_dispatch_stub_748(void){}
+void gl_dispatch_stub_749(void){}
+void gl_dispatch_stub_769(void){}
+void gl_dispatch_stub_770(void){}
+void gl_dispatch_stub_771(void){}
 
+#endif
diff --git a/src/mesa/drivers/windows/gdi/wmesadef.h b/src/mesa/drivers/windows/gdi/wmesadef.h
index 97b063a8ba..83a42e6082 100644
--- a/src/mesa/drivers/windows/gdi/wmesadef.h
+++ b/src/mesa/drivers/windows/gdi/wmesadef.h
@@ -1,6 +1,8 @@
 #ifndef WMESADEF_H
 #define WMESADEF_H
-
+#ifdef __MINGW32__
+#include <windows.h>
+#endif
 #include "context.h"
 
 
diff --git a/src/mesa/drivers/x11/fakeglx.c b/src/mesa/drivers/x11/fakeglx.c
index 1587df66bc..7a170b4d3d 100644
--- a/src/mesa/drivers/x11/fakeglx.c
+++ b/src/mesa/drivers/x11/fakeglx.c
@@ -296,6 +296,11 @@ save_glx_visual( Display *dpy, XVisualInfo *vinfo,
       }
    }
 
+   if (stereoFlag) {
+      /* stereo not supported */
+      return NULL;
+   }
+
    /* Comparing IDs uses less memory but sometimes fails. */
    /* XXX revisit this after 3.0 is finished. */
    if (_mesa_getenv("MESA_GLX_VISUAL_HACK"))
@@ -1079,7 +1084,7 @@ choose_visual( Display *dpy, int screen, const int *list, GLboolean fbConfig )
             else {
                stereo_flag = GL_TRUE;
             }
-            return NULL; /* stereo not supported */
+            break;
 	 case GLX_AUX_BUFFERS:
 	    parselist++;
             numAux = *parselist++;
diff --git a/src/mesa/drivers/x11/glxapi.c b/src/mesa/drivers/x11/glxapi.c
index 5f11c90c13..309a0008d7 100644
--- a/src/mesa/drivers/x11/glxapi.c
+++ b/src/mesa/drivers/x11/glxapi.c
@@ -141,7 +141,7 @@ static void
 SetCurrentContext(GLXContext c)
 {
 #if defined(GLX_USE_TLS)
-   CurrentContext = context;
+   CurrentContext = c;
 #elif defined(THREADS)
    _glthread_SetTSD(&ContextTSD, c);
 #else
@@ -1169,7 +1169,7 @@ _glxapi_get_extensions(void)
 #ifdef GLX_SGIX_pbuffer
       "GLX_SGIX_pbuffer",
 #endif
-#ifdef GLX_EXT_texture_from_pixmap,
+#ifdef GLX_EXT_texture_from_pixmap
       "GLX_EXT_texture_from_pixmap",
 #endif
       NULL
diff --git a/src/mesa/glapi/gl_API.xml b/src/mesa/glapi/gl_API.xml
index c7b48c45b5..4bd3b2f0fb 100644
--- a/src/mesa/glapi/gl_API.xml
+++ b/src/mesa/glapi/gl_API.xml
@@ -5605,7 +5605,10 @@
         <param name="index" type="GLuint"/>
         <param name="v" type="const GLushort *"/>
     </function>
-
+    <function name="VertexAttrib4bv" alias="VertexAttrib4bvARB">
+        <param name="index" type="GLuint"/>
+        <param name="v" type="const GLbyte *" count="4"/>
+    </function>
     <function name="VertexAttrib4d" alias="VertexAttrib4dARB">
         <param name="index" type="GLuint"/>
         <param name="x" type="GLdouble"/>
@@ -5628,6 +5631,10 @@
         <param name="index" type="GLuint"/>
         <param name="v" type="const GLfloat *"/>
     </function>
+    <function name="VertexAttrib4iv" alias="VertexAttrib4ivARB">
+        <param name="index" type="GLuint"/>
+        <param name="v" type="const GLint *" count="4"/>
+    </function>
     <function name="VertexAttrib4s" alias="VertexAttrib4sARB">
         <param name="index" type="GLuint"/>
         <param name="x" type="GLshort"/>
@@ -5639,6 +5646,18 @@
         <param name="index" type="GLuint"/>
         <param name="v" type="const GLshort *"/>
     </function>
+    <function name="VertexAttrib4ubv" alias="VertexAttrib4ubvARB">
+        <param name="index" type="GLuint"/>
+        <param name="v" type="const GLubyte *" count="4"/>
+    </function>
+    <function name="VertexAttrib4uiv" alias="VertexAttrib4uivARB">
+        <param name="index" type="GLuint"/>
+        <param name="v" type="const GLuint *" count="4"/>
+    </function>
+    <function name="VertexAttrib4usv" alias="VertexAttrib4usvARB">
+        <param name="index" type="GLuint"/>
+        <param name="v" type="const GLushort *" count="4"/>
+    </function>
 
     <function name="VertexAttribPointer" alias="VertexAttribPointerARB">
         <param name="index" type="GLuint"/>
diff --git a/src/mesa/glapi/gl_x86_asm.py b/src/mesa/glapi/gl_x86_asm.py
index 650331a0c5..403e87261b 100644
--- a/src/mesa/glapi/gl_x86_asm.py
+++ b/src/mesa/glapi/gl_x86_asm.py
@@ -82,12 +82,18 @@ class PrintGenericStubs(gl_XML.gl_print_base):
 		print ''
 		print '#ifdef GLX_USE_TLS'
 		print ''
+		print '#ifdef GLX_X86_READONLY_TEXT'
+		print '# define CTX_INSNS MOV_L(GS:(EAX), EAX)'
+		print '#else'
+		print '# define CTX_INSNS NOP /* Pad for init_glapi_relocs() */'
+		print '#endif'
+		print ''
 		print '#  define GL_STUB(fn,off,fn_alt)\t\t\t\\'
 		print 'ALIGNTEXT16;\t\t\t\t\t\t\\'
 		print 'GLOBL_FN(GL_PREFIX(fn, fn_alt));\t\t\t\\'
 		print 'GL_PREFIX(fn, fn_alt):\t\t\t\t\t\\'
 		print '\tCALL(_x86_get_dispatch) ;\t\t\t\\'
-		print '\tNOP ;\t\t\t\t\t\t\\'
+		print '\tCTX_INSNS ;					\\'
 		print '\tJMP(GL_OFFSET(off))'
 		print ''
 		print '#elif defined(PTHREADS)'
@@ -138,7 +144,10 @@ class PrintGenericStubs(gl_XML.gl_print_base):
 		print '\tHIDDEN(GLNAME(_x86_get_dispatch))'
 		print 'ALIGNTEXT16'
 		print 'GLNAME(_x86_get_dispatch):'
-		print '\tmovl\t%gs:_glapi_tls_Dispatch@NTPOFF, %eax'
+		print '\tcall	1f'
+		print '1:\tpopl	%eax'
+		print '\taddl	$_GLOBAL_OFFSET_TABLE_+[.-1b], %eax'
+		print '\tmovl	_glapi_tls_Dispatch@GOTNTPOFF(%eax), %eax'
 		print '\tret'
 		print ''
 		print '#elif defined(PTHREADS)'
@@ -158,7 +167,7 @@ class PrintGenericStubs(gl_XML.gl_print_base):
 		print '#endif'
 		print ''
 
-		print '#if defined( GLX_USE_TLS )'
+		print '#if defined( GLX_USE_TLS ) && !defined( GLX_X86_READONLY_TEXT )'
 		print '\t\t.section\twtext, "awx", @progbits'
 		print '#endif /* defined( GLX_USE_TLS ) */'
 
diff --git a/src/mesa/glapi/glapi.c b/src/mesa/glapi/glapi.c
index 5815dbff84..47c5782273 100644
--- a/src/mesa/glapi/glapi.c
+++ b/src/mesa/glapi/glapi.c
@@ -1028,22 +1028,24 @@ _glapi_check_table(const struct _glapi_table *table)
 #if defined(PTHREADS) || defined(GLX_USE_TLS)
 /**
  * Perform platform-specific GL API entry-point fixups.
- * 
- * 
  */
 static void
 init_glapi_relocs( void )
 {
-#if defined( USE_X86_ASM ) && defined( GLX_USE_TLS )
-    extern void * _x86_get_dispatch(void);
-    const GLubyte * const get_disp = (const GLubyte *) _x86_get_dispatch;
+#if defined(USE_X86_ASM) && defined(GLX_USE_TLS) && !defined(GLX_X86_READONLY_TEXT)
+    extern unsigned long _x86_get_dispatch(void);
+    char run_time_patch[] = {
+       0x65, 0xa1, 0, 0, 0, 0 /* movl %gs:0,%eax */
+    };
+    GLuint *offset = (GLuint *) &run_time_patch[2]; /* 32-bits for x86/32 */
+    const GLubyte * const get_disp = (const GLubyte *) run_time_patch;
     GLubyte * curr_func = (GLubyte *) gl_dispatch_functions_start;
 
-
+    *offset = _x86_get_dispatch();
     while ( curr_func != (GLubyte *) gl_dispatch_functions_end ) {
-	(void) memcpy( curr_func, get_disp, 6 );
+	(void) memcpy( curr_func, get_disp, sizeof(run_time_patch));
 	curr_func += DISPATCH_FUNCTION_SIZE;
     }
-#endif /* defined( USE_X86_ASM ) && defined( GLX_USE_TLS ) */
-}
 #endif
+}
+#endif /* defined(PTHREADS) || defined(GLX_USE_TLS) */
diff --git a/src/mesa/glapi/glapitemp.h b/src/mesa/glapi/glapitemp.h
index 2a2fe2a35d..10af9b3085 100644
--- a/src/mesa/glapi/glapitemp.h
+++ b/src/mesa/glapi/glapitemp.h
@@ -3284,6 +3284,11 @@ KEYWORD1 void KEYWORD2 NAME(VertexAttrib4NusvARB)(GLuint index, const GLushort *
    DISPATCH(VertexAttrib4NusvARB, (index, v), (F, "glVertexAttrib4NusvARB(%d, %p);\n", index, (const void *) v));
 }
 
+KEYWORD1 void KEYWORD2 NAME(VertexAttrib4bv)(GLuint index, const GLbyte * v)
+{
+   DISPATCH(VertexAttrib4bvARB, (index, v), (F, "glVertexAttrib4bv(%d, %p);\n", index, (const void *) v));
+}
+
 KEYWORD1 void KEYWORD2 NAME(VertexAttrib4bvARB)(GLuint index, const GLbyte * v)
 {
    DISPATCH(VertexAttrib4bvARB, (index, v), (F, "glVertexAttrib4bvARB(%d, %p);\n", index, (const void *) v));
@@ -3329,6 +3334,11 @@ KEYWORD1 void KEYWORD2 NAME(VertexAttrib4fvARB)(GLuint index, const GLfloat * v)
    DISPATCH(VertexAttrib4fvARB, (index, v), (F, "glVertexAttrib4fvARB(%d, %p);\n", index, (const void *) v));
 }
 
+KEYWORD1 void KEYWORD2 NAME(VertexAttrib4iv)(GLuint index, const GLint * v)
+{
+   DISPATCH(VertexAttrib4ivARB, (index, v), (F, "glVertexAttrib4iv(%d, %p);\n", index, (const void *) v));
+}
+
 KEYWORD1 void KEYWORD2 NAME(VertexAttrib4ivARB)(GLuint index, const GLint * v)
 {
    DISPATCH(VertexAttrib4ivARB, (index, v), (F, "glVertexAttrib4ivARB(%d, %p);\n", index, (const void *) v));
@@ -3354,16 +3364,31 @@ KEYWORD1 void KEYWORD2 NAME(VertexAttrib4svARB)(GLuint index, const GLshort * v)
    DISPATCH(VertexAttrib4svARB, (index, v), (F, "glVertexAttrib4svARB(%d, %p);\n", index, (const void *) v));
 }
 
+KEYWORD1 void KEYWORD2 NAME(VertexAttrib4ubv)(GLuint index, const GLubyte * v)
+{
+   DISPATCH(VertexAttrib4ubvARB, (index, v), (F, "glVertexAttrib4ubv(%d, %p);\n", index, (const void *) v));
+}
+
 KEYWORD1 void KEYWORD2 NAME(VertexAttrib4ubvARB)(GLuint index, const GLubyte * v)
 {
    DISPATCH(VertexAttrib4ubvARB, (index, v), (F, "glVertexAttrib4ubvARB(%d, %p);\n", index, (const void *) v));
 }
 
+KEYWORD1 void KEYWORD2 NAME(VertexAttrib4uiv)(GLuint index, const GLuint * v)
+{
+   DISPATCH(VertexAttrib4uivARB, (index, v), (F, "glVertexAttrib4uiv(%d, %p);\n", index, (const void *) v));
+}
+
 KEYWORD1 void KEYWORD2 NAME(VertexAttrib4uivARB)(GLuint index, const GLuint * v)
 {
    DISPATCH(VertexAttrib4uivARB, (index, v), (F, "glVertexAttrib4uivARB(%d, %p);\n", index, (const void *) v));
 }
 
+KEYWORD1 void KEYWORD2 NAME(VertexAttrib4usv)(GLuint index, const GLushort * v)
+{
+   DISPATCH(VertexAttrib4usvARB, (index, v), (F, "glVertexAttrib4usv(%d, %p);\n", index, (const void *) v));
+}
+
 KEYWORD1 void KEYWORD2 NAME(VertexAttrib4usvARB)(GLuint index, const GLushort * v)
 {
    DISPATCH(VertexAttrib4usvARB, (index, v), (F, "glVertexAttrib4usvARB(%d, %p);\n", index, (const void *) v));
@@ -6464,12 +6489,17 @@ static _glapi_proc UNUSED_TABLE_NAME[] = {
    TABLE_ENTRY(VertexAttrib4Nubv),
    TABLE_ENTRY(VertexAttrib4Nuiv),
    TABLE_ENTRY(VertexAttrib4Nusv),
+   TABLE_ENTRY(VertexAttrib4bv),
    TABLE_ENTRY(VertexAttrib4d),
    TABLE_ENTRY(VertexAttrib4dv),
    TABLE_ENTRY(VertexAttrib4f),
    TABLE_ENTRY(VertexAttrib4fv),
+   TABLE_ENTRY(VertexAttrib4iv),
    TABLE_ENTRY(VertexAttrib4s),
    TABLE_ENTRY(VertexAttrib4sv),
+   TABLE_ENTRY(VertexAttrib4ubv),
+   TABLE_ENTRY(VertexAttrib4uiv),
+   TABLE_ENTRY(VertexAttrib4usv),
    TABLE_ENTRY(VertexAttribPointer),
    TABLE_ENTRY(BindBuffer),
    TABLE_ENTRY(BufferData),
diff --git a/src/mesa/glapi/glprocs.h b/src/mesa/glapi/glprocs.h
index b6651a6bf4..c461333c51 100644
--- a/src/mesa/glapi/glprocs.h
+++ b/src/mesa/glapi/glprocs.h
@@ -958,12 +958,17 @@ static const char gl_string_table[] =
     "glVertexAttrib4Nubv\0"
     "glVertexAttrib4Nuiv\0"
     "glVertexAttrib4Nusv\0"
+    "glVertexAttrib4bv\0"
     "glVertexAttrib4d\0"
     "glVertexAttrib4dv\0"
     "glVertexAttrib4f\0"
     "glVertexAttrib4fv\0"
+    "glVertexAttrib4iv\0"
     "glVertexAttrib4s\0"
     "glVertexAttrib4sv\0"
+    "glVertexAttrib4ubv\0"
+    "glVertexAttrib4uiv\0"
+    "glVertexAttrib4usv\0"
     "glVertexAttribPointer\0"
     "glBindBuffer\0"
     "glBufferData\0"
@@ -2102,142 +2107,147 @@ static const glprocs_table_t static_functions[] = {
     NAME_FUNC_OFFSET(16297, glVertexAttrib4NubvARB, glVertexAttrib4NubvARB, NULL, _gloffset_VertexAttrib4NubvARB),
     NAME_FUNC_OFFSET(16317, glVertexAttrib4NuivARB, glVertexAttrib4NuivARB, NULL, _gloffset_VertexAttrib4NuivARB),
     NAME_FUNC_OFFSET(16337, glVertexAttrib4NusvARB, glVertexAttrib4NusvARB, NULL, _gloffset_VertexAttrib4NusvARB),
-    NAME_FUNC_OFFSET(16357, glVertexAttrib4dARB, glVertexAttrib4dARB, NULL, _gloffset_VertexAttrib4dARB),
-    NAME_FUNC_OFFSET(16374, glVertexAttrib4dvARB, glVertexAttrib4dvARB, NULL, _gloffset_VertexAttrib4dvARB),
-    NAME_FUNC_OFFSET(16392, glVertexAttrib4fARB, glVertexAttrib4fARB, NULL, _gloffset_VertexAttrib4fARB),
-    NAME_FUNC_OFFSET(16409, glVertexAttrib4fvARB, glVertexAttrib4fvARB, NULL, _gloffset_VertexAttrib4fvARB),
-    NAME_FUNC_OFFSET(16427, glVertexAttrib4sARB, glVertexAttrib4sARB, NULL, _gloffset_VertexAttrib4sARB),
-    NAME_FUNC_OFFSET(16444, glVertexAttrib4svARB, glVertexAttrib4svARB, NULL, _gloffset_VertexAttrib4svARB),
-    NAME_FUNC_OFFSET(16462, glVertexAttribPointerARB, glVertexAttribPointerARB, NULL, _gloffset_VertexAttribPointerARB),
-    NAME_FUNC_OFFSET(16484, glBindBufferARB, glBindBufferARB, NULL, _gloffset_BindBufferARB),
-    NAME_FUNC_OFFSET(16497, glBufferDataARB, glBufferDataARB, NULL, _gloffset_BufferDataARB),
-    NAME_FUNC_OFFSET(16510, glBufferSubDataARB, glBufferSubDataARB, NULL, _gloffset_BufferSubDataARB),
-    NAME_FUNC_OFFSET(16526, glDeleteBuffersARB, glDeleteBuffersARB, NULL, _gloffset_DeleteBuffersARB),
-    NAME_FUNC_OFFSET(16542, glGenBuffersARB, glGenBuffersARB, NULL, _gloffset_GenBuffersARB),
-    NAME_FUNC_OFFSET(16555, glGetBufferParameterivARB, glGetBufferParameterivARB, NULL, _gloffset_GetBufferParameterivARB),
-    NAME_FUNC_OFFSET(16578, glGetBufferPointervARB, glGetBufferPointervARB, NULL, _gloffset_GetBufferPointervARB),
-    NAME_FUNC_OFFSET(16598, glGetBufferSubDataARB, glGetBufferSubDataARB, NULL, _gloffset_GetBufferSubDataARB),
-    NAME_FUNC_OFFSET(16617, glIsBufferARB, glIsBufferARB, NULL, _gloffset_IsBufferARB),
-    NAME_FUNC_OFFSET(16628, glMapBufferARB, glMapBufferARB, NULL, _gloffset_MapBufferARB),
-    NAME_FUNC_OFFSET(16640, glUnmapBufferARB, glUnmapBufferARB, NULL, _gloffset_UnmapBufferARB),
-    NAME_FUNC_OFFSET(16654, glBeginQueryARB, glBeginQueryARB, NULL, _gloffset_BeginQueryARB),
-    NAME_FUNC_OFFSET(16667, glDeleteQueriesARB, glDeleteQueriesARB, NULL, _gloffset_DeleteQueriesARB),
-    NAME_FUNC_OFFSET(16683, glEndQueryARB, glEndQueryARB, NULL, _gloffset_EndQueryARB),
-    NAME_FUNC_OFFSET(16694, glGenQueriesARB, glGenQueriesARB, NULL, _gloffset_GenQueriesARB),
-    NAME_FUNC_OFFSET(16707, glGetQueryObjectivARB, glGetQueryObjectivARB, NULL, _gloffset_GetQueryObjectivARB),
-    NAME_FUNC_OFFSET(16726, glGetQueryObjectuivARB, glGetQueryObjectuivARB, NULL, _gloffset_GetQueryObjectuivARB),
-    NAME_FUNC_OFFSET(16746, glGetQueryivARB, glGetQueryivARB, NULL, _gloffset_GetQueryivARB),
-    NAME_FUNC_OFFSET(16759, glIsQueryARB, glIsQueryARB, NULL, _gloffset_IsQueryARB),
-    NAME_FUNC_OFFSET(16769, glCompileShaderARB, glCompileShaderARB, NULL, _gloffset_CompileShaderARB),
-    NAME_FUNC_OFFSET(16785, glGetActiveUniformARB, glGetActiveUniformARB, NULL, _gloffset_GetActiveUniformARB),
-    NAME_FUNC_OFFSET(16804, glGetShaderSourceARB, glGetShaderSourceARB, NULL, _gloffset_GetShaderSourceARB),
-    NAME_FUNC_OFFSET(16822, glGetUniformLocationARB, glGetUniformLocationARB, NULL, _gloffset_GetUniformLocationARB),
-    NAME_FUNC_OFFSET(16843, glGetUniformfvARB, glGetUniformfvARB, NULL, _gloffset_GetUniformfvARB),
-    NAME_FUNC_OFFSET(16858, glGetUniformivARB, glGetUniformivARB, NULL, _gloffset_GetUniformivARB),
-    NAME_FUNC_OFFSET(16873, glLinkProgramARB, glLinkProgramARB, NULL, _gloffset_LinkProgramARB),
-    NAME_FUNC_OFFSET(16887, glShaderSourceARB, glShaderSourceARB, NULL, _gloffset_ShaderSourceARB),
-    NAME_FUNC_OFFSET(16902, glUniform1fARB, glUniform1fARB, NULL, _gloffset_Uniform1fARB),
-    NAME_FUNC_OFFSET(16914, glUniform1fvARB, glUniform1fvARB, NULL, _gloffset_Uniform1fvARB),
-    NAME_FUNC_OFFSET(16927, glUniform1iARB, glUniform1iARB, NULL, _gloffset_Uniform1iARB),
-    NAME_FUNC_OFFSET(16939, glUniform1ivARB, glUniform1ivARB, NULL, _gloffset_Uniform1ivARB),
-    NAME_FUNC_OFFSET(16952, glUniform2fARB, glUniform2fARB, NULL, _gloffset_Uniform2fARB),
-    NAME_FUNC_OFFSET(16964, glUniform2fvARB, glUniform2fvARB, NULL, _gloffset_Uniform2fvARB),
-    NAME_FUNC_OFFSET(16977, glUniform2iARB, glUniform2iARB, NULL, _gloffset_Uniform2iARB),
-    NAME_FUNC_OFFSET(16989, glUniform2ivARB, glUniform2ivARB, NULL, _gloffset_Uniform2ivARB),
-    NAME_FUNC_OFFSET(17002, glUniform3fARB, glUniform3fARB, NULL, _gloffset_Uniform3fARB),
-    NAME_FUNC_OFFSET(17014, glUniform3fvARB, glUniform3fvARB, NULL, _gloffset_Uniform3fvARB),
-    NAME_FUNC_OFFSET(17027, glUniform3iARB, glUniform3iARB, NULL, _gloffset_Uniform3iARB),
-    NAME_FUNC_OFFSET(17039, glUniform3ivARB, glUniform3ivARB, NULL, _gloffset_Uniform3ivARB),
-    NAME_FUNC_OFFSET(17052, glUniform4fARB, glUniform4fARB, NULL, _gloffset_Uniform4fARB),
-    NAME_FUNC_OFFSET(17064, glUniform4fvARB, glUniform4fvARB, NULL, _gloffset_Uniform4fvARB),
-    NAME_FUNC_OFFSET(17077, glUniform4iARB, glUniform4iARB, NULL, _gloffset_Uniform4iARB),
-    NAME_FUNC_OFFSET(17089, glUniform4ivARB, glUniform4ivARB, NULL, _gloffset_Uniform4ivARB),
-    NAME_FUNC_OFFSET(17102, glUniformMatrix2fvARB, glUniformMatrix2fvARB, NULL, _gloffset_UniformMatrix2fvARB),
-    NAME_FUNC_OFFSET(17121, glUniformMatrix3fvARB, glUniformMatrix3fvARB, NULL, _gloffset_UniformMatrix3fvARB),
-    NAME_FUNC_OFFSET(17140, glUniformMatrix4fvARB, glUniformMatrix4fvARB, NULL, _gloffset_UniformMatrix4fvARB),
-    NAME_FUNC_OFFSET(17159, glUseProgramObjectARB, glUseProgramObjectARB, NULL, _gloffset_UseProgramObjectARB),
-    NAME_FUNC_OFFSET(17172, glValidateProgramARB, glValidateProgramARB, NULL, _gloffset_ValidateProgramARB),
-    NAME_FUNC_OFFSET(17190, glBindAttribLocationARB, glBindAttribLocationARB, NULL, _gloffset_BindAttribLocationARB),
-    NAME_FUNC_OFFSET(17211, glGetActiveAttribARB, glGetActiveAttribARB, NULL, _gloffset_GetActiveAttribARB),
-    NAME_FUNC_OFFSET(17229, glGetAttribLocationARB, glGetAttribLocationARB, NULL, _gloffset_GetAttribLocationARB),
-    NAME_FUNC_OFFSET(17249, glDrawBuffersARB, glDrawBuffersARB, NULL, _gloffset_DrawBuffersARB),
-    NAME_FUNC_OFFSET(17263, glDrawBuffersARB, glDrawBuffersARB, NULL, _gloffset_DrawBuffersARB),
-    NAME_FUNC_OFFSET(17280, gl_dispatch_stub_568, gl_dispatch_stub_568, NULL, _gloffset_SampleMaskSGIS),
-    NAME_FUNC_OFFSET(17296, gl_dispatch_stub_569, gl_dispatch_stub_569, NULL, _gloffset_SamplePatternSGIS),
-    NAME_FUNC_OFFSET(17315, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
-    NAME_FUNC_OFFSET(17333, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
-    NAME_FUNC_OFFSET(17354, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
-    NAME_FUNC_OFFSET(17376, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
-    NAME_FUNC_OFFSET(17395, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
-    NAME_FUNC_OFFSET(17417, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
-    NAME_FUNC_OFFSET(17440, glSecondaryColor3bEXT, glSecondaryColor3bEXT, NULL, _gloffset_SecondaryColor3bEXT),
-    NAME_FUNC_OFFSET(17459, glSecondaryColor3bvEXT, glSecondaryColor3bvEXT, NULL, _gloffset_SecondaryColor3bvEXT),
-    NAME_FUNC_OFFSET(17479, glSecondaryColor3dEXT, glSecondaryColor3dEXT, NULL, _gloffset_SecondaryColor3dEXT),
-    NAME_FUNC_OFFSET(17498, glSecondaryColor3dvEXT, glSecondaryColor3dvEXT, NULL, _gloffset_SecondaryColor3dvEXT),
-    NAME_FUNC_OFFSET(17518, glSecondaryColor3fEXT, glSecondaryColor3fEXT, NULL, _gloffset_SecondaryColor3fEXT),
-    NAME_FUNC_OFFSET(17537, glSecondaryColor3fvEXT, glSecondaryColor3fvEXT, NULL, _gloffset_SecondaryColor3fvEXT),
-    NAME_FUNC_OFFSET(17557, glSecondaryColor3iEXT, glSecondaryColor3iEXT, NULL, _gloffset_SecondaryColor3iEXT),
-    NAME_FUNC_OFFSET(17576, glSecondaryColor3ivEXT, glSecondaryColor3ivEXT, NULL, _gloffset_SecondaryColor3ivEXT),
-    NAME_FUNC_OFFSET(17596, glSecondaryColor3sEXT, glSecondaryColor3sEXT, NULL, _gloffset_SecondaryColor3sEXT),
-    NAME_FUNC_OFFSET(17615, glSecondaryColor3svEXT, glSecondaryColor3svEXT, NULL, _gloffset_SecondaryColor3svEXT),
-    NAME_FUNC_OFFSET(17635, glSecondaryColor3ubEXT, glSecondaryColor3ubEXT, NULL, _gloffset_SecondaryColor3ubEXT),
-    NAME_FUNC_OFFSET(17655, glSecondaryColor3ubvEXT, glSecondaryColor3ubvEXT, NULL, _gloffset_SecondaryColor3ubvEXT),
-    NAME_FUNC_OFFSET(17676, glSecondaryColor3uiEXT, glSecondaryColor3uiEXT, NULL, _gloffset_SecondaryColor3uiEXT),
-    NAME_FUNC_OFFSET(17696, glSecondaryColor3uivEXT, glSecondaryColor3uivEXT, NULL, _gloffset_SecondaryColor3uivEXT),
-    NAME_FUNC_OFFSET(17717, glSecondaryColor3usEXT, glSecondaryColor3usEXT, NULL, _gloffset_SecondaryColor3usEXT),
-    NAME_FUNC_OFFSET(17737, glSecondaryColor3usvEXT, glSecondaryColor3usvEXT, NULL, _gloffset_SecondaryColor3usvEXT),
-    NAME_FUNC_OFFSET(17758, glSecondaryColorPointerEXT, glSecondaryColorPointerEXT, NULL, _gloffset_SecondaryColorPointerEXT),
-    NAME_FUNC_OFFSET(17782, glMultiDrawArraysEXT, glMultiDrawArraysEXT, NULL, _gloffset_MultiDrawArraysEXT),
-    NAME_FUNC_OFFSET(17800, glMultiDrawElementsEXT, glMultiDrawElementsEXT, NULL, _gloffset_MultiDrawElementsEXT),
-    NAME_FUNC_OFFSET(17820, glFogCoordPointerEXT, glFogCoordPointerEXT, NULL, _gloffset_FogCoordPointerEXT),
-    NAME_FUNC_OFFSET(17838, glFogCoorddEXT, glFogCoorddEXT, NULL, _gloffset_FogCoorddEXT),
-    NAME_FUNC_OFFSET(17850, glFogCoorddvEXT, glFogCoorddvEXT, NULL, _gloffset_FogCoorddvEXT),
-    NAME_FUNC_OFFSET(17863, glFogCoordfEXT, glFogCoordfEXT, NULL, _gloffset_FogCoordfEXT),
-    NAME_FUNC_OFFSET(17875, glFogCoordfvEXT, glFogCoordfvEXT, NULL, _gloffset_FogCoordfvEXT),
-    NAME_FUNC_OFFSET(17888, glBlendFuncSeparateEXT, glBlendFuncSeparateEXT, NULL, _gloffset_BlendFuncSeparateEXT),
-    NAME_FUNC_OFFSET(17908, glBlendFuncSeparateEXT, glBlendFuncSeparateEXT, NULL, _gloffset_BlendFuncSeparateEXT),
-    NAME_FUNC_OFFSET(17932, glWindowPos2dMESA, glWindowPos2dMESA, NULL, _gloffset_WindowPos2dMESA),
-    NAME_FUNC_OFFSET(17946, glWindowPos2dMESA, glWindowPos2dMESA, NULL, _gloffset_WindowPos2dMESA),
-    NAME_FUNC_OFFSET(17963, glWindowPos2dvMESA, glWindowPos2dvMESA, NULL, _gloffset_WindowPos2dvMESA),
-    NAME_FUNC_OFFSET(17978, glWindowPos2dvMESA, glWindowPos2dvMESA, NULL, _gloffset_WindowPos2dvMESA),
-    NAME_FUNC_OFFSET(17996, glWindowPos2fMESA, glWindowPos2fMESA, NULL, _gloffset_WindowPos2fMESA),
-    NAME_FUNC_OFFSET(18010, glWindowPos2fMESA, glWindowPos2fMESA, NULL, _gloffset_WindowPos2fMESA),
-    NAME_FUNC_OFFSET(18027, glWindowPos2fvMESA, glWindowPos2fvMESA, NULL, _gloffset_WindowPos2fvMESA),
-    NAME_FUNC_OFFSET(18042, glWindowPos2fvMESA, glWindowPos2fvMESA, NULL, _gloffset_WindowPos2fvMESA),
-    NAME_FUNC_OFFSET(18060, glWindowPos2iMESA, glWindowPos2iMESA, NULL, _gloffset_WindowPos2iMESA),
-    NAME_FUNC_OFFSET(18074, glWindowPos2iMESA, glWindowPos2iMESA, NULL, _gloffset_WindowPos2iMESA),
-    NAME_FUNC_OFFSET(18091, glWindowPos2ivMESA, glWindowPos2ivMESA, NULL, _gloffset_WindowPos2ivMESA),
-    NAME_FUNC_OFFSET(18106, glWindowPos2ivMESA, glWindowPos2ivMESA, NULL, _gloffset_WindowPos2ivMESA),
-    NAME_FUNC_OFFSET(18124, glWindowPos2sMESA, glWindowPos2sMESA, NULL, _gloffset_WindowPos2sMESA),
-    NAME_FUNC_OFFSET(18138, glWindowPos2sMESA, glWindowPos2sMESA, NULL, _gloffset_WindowPos2sMESA),
-    NAME_FUNC_OFFSET(18155, glWindowPos2svMESA, glWindowPos2svMESA, NULL, _gloffset_WindowPos2svMESA),
-    NAME_FUNC_OFFSET(18170, glWindowPos2svMESA, glWindowPos2svMESA, NULL, _gloffset_WindowPos2svMESA),
-    NAME_FUNC_OFFSET(18188, glWindowPos3dMESA, glWindowPos3dMESA, NULL, _gloffset_WindowPos3dMESA),
-    NAME_FUNC_OFFSET(18202, glWindowPos3dMESA, glWindowPos3dMESA, NULL, _gloffset_WindowPos3dMESA),
-    NAME_FUNC_OFFSET(18219, glWindowPos3dvMESA, glWindowPos3dvMESA, NULL, _gloffset_WindowPos3dvMESA),
-    NAME_FUNC_OFFSET(18234, glWindowPos3dvMESA, glWindowPos3dvMESA, NULL, _gloffset_WindowPos3dvMESA),
-    NAME_FUNC_OFFSET(18252, glWindowPos3fMESA, glWindowPos3fMESA, NULL, _gloffset_WindowPos3fMESA),
-    NAME_FUNC_OFFSET(18266, glWindowPos3fMESA, glWindowPos3fMESA, NULL, _gloffset_WindowPos3fMESA),
-    NAME_FUNC_OFFSET(18283, glWindowPos3fvMESA, glWindowPos3fvMESA, NULL, _gloffset_WindowPos3fvMESA),
-    NAME_FUNC_OFFSET(18298, glWindowPos3fvMESA, glWindowPos3fvMESA, NULL, _gloffset_WindowPos3fvMESA),
-    NAME_FUNC_OFFSET(18316, glWindowPos3iMESA, glWindowPos3iMESA, NULL, _gloffset_WindowPos3iMESA),
-    NAME_FUNC_OFFSET(18330, glWindowPos3iMESA, glWindowPos3iMESA, NULL, _gloffset_WindowPos3iMESA),
-    NAME_FUNC_OFFSET(18347, glWindowPos3ivMESA, glWindowPos3ivMESA, NULL, _gloffset_WindowPos3ivMESA),
-    NAME_FUNC_OFFSET(18362, glWindowPos3ivMESA, glWindowPos3ivMESA, NULL, _gloffset_WindowPos3ivMESA),
-    NAME_FUNC_OFFSET(18380, glWindowPos3sMESA, glWindowPos3sMESA, NULL, _gloffset_WindowPos3sMESA),
-    NAME_FUNC_OFFSET(18394, glWindowPos3sMESA, glWindowPos3sMESA, NULL, _gloffset_WindowPos3sMESA),
-    NAME_FUNC_OFFSET(18411, glWindowPos3svMESA, glWindowPos3svMESA, NULL, _gloffset_WindowPos3svMESA),
-    NAME_FUNC_OFFSET(18426, glWindowPos3svMESA, glWindowPos3svMESA, NULL, _gloffset_WindowPos3svMESA),
-    NAME_FUNC_OFFSET(18444, glBindProgramNV, glBindProgramNV, NULL, _gloffset_BindProgramNV),
-    NAME_FUNC_OFFSET(18461, glDeleteProgramsNV, glDeleteProgramsNV, NULL, _gloffset_DeleteProgramsNV),
-    NAME_FUNC_OFFSET(18481, glGenProgramsNV, glGenProgramsNV, NULL, _gloffset_GenProgramsNV),
-    NAME_FUNC_OFFSET(18498, glGetVertexAttribPointervNV, glGetVertexAttribPointervNV, NULL, _gloffset_GetVertexAttribPointervNV),
-    NAME_FUNC_OFFSET(18524, glGetVertexAttribPointervNV, glGetVertexAttribPointervNV, NULL, _gloffset_GetVertexAttribPointervNV),
-    NAME_FUNC_OFFSET(18553, glIsProgramNV, glIsProgramNV, NULL, _gloffset_IsProgramNV),
-    NAME_FUNC_OFFSET(18568, glPointParameteriNV, glPointParameteriNV, NULL, _gloffset_PointParameteriNV),
-    NAME_FUNC_OFFSET(18586, glPointParameterivNV, glPointParameterivNV, NULL, _gloffset_PointParameterivNV),
-    NAME_FUNC_OFFSET(18605, gl_dispatch_stub_749, gl_dispatch_stub_749, NULL, _gloffset_BlendEquationSeparateEXT),
-    NAME_FUNC_OFFSET(18629, gl_dispatch_stub_749, gl_dispatch_stub_749, NULL, _gloffset_BlendEquationSeparateEXT),
+    NAME_FUNC_OFFSET(16357, glVertexAttrib4bvARB, glVertexAttrib4bvARB, NULL, _gloffset_VertexAttrib4bvARB),
+    NAME_FUNC_OFFSET(16375, glVertexAttrib4dARB, glVertexAttrib4dARB, NULL, _gloffset_VertexAttrib4dARB),
+    NAME_FUNC_OFFSET(16392, glVertexAttrib4dvARB, glVertexAttrib4dvARB, NULL, _gloffset_VertexAttrib4dvARB),
+    NAME_FUNC_OFFSET(16410, glVertexAttrib4fARB, glVertexAttrib4fARB, NULL, _gloffset_VertexAttrib4fARB),
+    NAME_FUNC_OFFSET(16427, glVertexAttrib4fvARB, glVertexAttrib4fvARB, NULL, _gloffset_VertexAttrib4fvARB),
+    NAME_FUNC_OFFSET(16445, glVertexAttrib4ivARB, glVertexAttrib4ivARB, NULL, _gloffset_VertexAttrib4ivARB),
+    NAME_FUNC_OFFSET(16463, glVertexAttrib4sARB, glVertexAttrib4sARB, NULL, _gloffset_VertexAttrib4sARB),
+    NAME_FUNC_OFFSET(16480, glVertexAttrib4svARB, glVertexAttrib4svARB, NULL, _gloffset_VertexAttrib4svARB),
+    NAME_FUNC_OFFSET(16498, glVertexAttrib4ubvARB, glVertexAttrib4ubvARB, NULL, _gloffset_VertexAttrib4ubvARB),
+    NAME_FUNC_OFFSET(16517, glVertexAttrib4uivARB, glVertexAttrib4uivARB, NULL, _gloffset_VertexAttrib4uivARB),
+    NAME_FUNC_OFFSET(16536, glVertexAttrib4usvARB, glVertexAttrib4usvARB, NULL, _gloffset_VertexAttrib4usvARB),
+    NAME_FUNC_OFFSET(16555, glVertexAttribPointerARB, glVertexAttribPointerARB, NULL, _gloffset_VertexAttribPointerARB),
+    NAME_FUNC_OFFSET(16577, glBindBufferARB, glBindBufferARB, NULL, _gloffset_BindBufferARB),
+    NAME_FUNC_OFFSET(16590, glBufferDataARB, glBufferDataARB, NULL, _gloffset_BufferDataARB),
+    NAME_FUNC_OFFSET(16603, glBufferSubDataARB, glBufferSubDataARB, NULL, _gloffset_BufferSubDataARB),
+    NAME_FUNC_OFFSET(16619, glDeleteBuffersARB, glDeleteBuffersARB, NULL, _gloffset_DeleteBuffersARB),
+    NAME_FUNC_OFFSET(16635, glGenBuffersARB, glGenBuffersARB, NULL, _gloffset_GenBuffersARB),
+    NAME_FUNC_OFFSET(16648, glGetBufferParameterivARB, glGetBufferParameterivARB, NULL, _gloffset_GetBufferParameterivARB),
+    NAME_FUNC_OFFSET(16671, glGetBufferPointervARB, glGetBufferPointervARB, NULL, _gloffset_GetBufferPointervARB),
+    NAME_FUNC_OFFSET(16691, glGetBufferSubDataARB, glGetBufferSubDataARB, NULL, _gloffset_GetBufferSubDataARB),
+    NAME_FUNC_OFFSET(16710, glIsBufferARB, glIsBufferARB, NULL, _gloffset_IsBufferARB),
+    NAME_FUNC_OFFSET(16721, glMapBufferARB, glMapBufferARB, NULL, _gloffset_MapBufferARB),
+    NAME_FUNC_OFFSET(16733, glUnmapBufferARB, glUnmapBufferARB, NULL, _gloffset_UnmapBufferARB),
+    NAME_FUNC_OFFSET(16747, glBeginQueryARB, glBeginQueryARB, NULL, _gloffset_BeginQueryARB),
+    NAME_FUNC_OFFSET(16760, glDeleteQueriesARB, glDeleteQueriesARB, NULL, _gloffset_DeleteQueriesARB),
+    NAME_FUNC_OFFSET(16776, glEndQueryARB, glEndQueryARB, NULL, _gloffset_EndQueryARB),
+    NAME_FUNC_OFFSET(16787, glGenQueriesARB, glGenQueriesARB, NULL, _gloffset_GenQueriesARB),
+    NAME_FUNC_OFFSET(16800, glGetQueryObjectivARB, glGetQueryObjectivARB, NULL, _gloffset_GetQueryObjectivARB),
+    NAME_FUNC_OFFSET(16819, glGetQueryObjectuivARB, glGetQueryObjectuivARB, NULL, _gloffset_GetQueryObjectuivARB),
+    NAME_FUNC_OFFSET(16839, glGetQueryivARB, glGetQueryivARB, NULL, _gloffset_GetQueryivARB),
+    NAME_FUNC_OFFSET(16852, glIsQueryARB, glIsQueryARB, NULL, _gloffset_IsQueryARB),
+    NAME_FUNC_OFFSET(16862, glCompileShaderARB, glCompileShaderARB, NULL, _gloffset_CompileShaderARB),
+    NAME_FUNC_OFFSET(16878, glGetActiveUniformARB, glGetActiveUniformARB, NULL, _gloffset_GetActiveUniformARB),
+    NAME_FUNC_OFFSET(16897, glGetShaderSourceARB, glGetShaderSourceARB, NULL, _gloffset_GetShaderSourceARB),
+    NAME_FUNC_OFFSET(16915, glGetUniformLocationARB, glGetUniformLocationARB, NULL, _gloffset_GetUniformLocationARB),
+    NAME_FUNC_OFFSET(16936, glGetUniformfvARB, glGetUniformfvARB, NULL, _gloffset_GetUniformfvARB),
+    NAME_FUNC_OFFSET(16951, glGetUniformivARB, glGetUniformivARB, NULL, _gloffset_GetUniformivARB),
+    NAME_FUNC_OFFSET(16966, glLinkProgramARB, glLinkProgramARB, NULL, _gloffset_LinkProgramARB),
+    NAME_FUNC_OFFSET(16980, glShaderSourceARB, glShaderSourceARB, NULL, _gloffset_ShaderSourceARB),
+    NAME_FUNC_OFFSET(16995, glUniform1fARB, glUniform1fARB, NULL, _gloffset_Uniform1fARB),
+    NAME_FUNC_OFFSET(17007, glUniform1fvARB, glUniform1fvARB, NULL, _gloffset_Uniform1fvARB),
+    NAME_FUNC_OFFSET(17020, glUniform1iARB, glUniform1iARB, NULL, _gloffset_Uniform1iARB),
+    NAME_FUNC_OFFSET(17032, glUniform1ivARB, glUniform1ivARB, NULL, _gloffset_Uniform1ivARB),
+    NAME_FUNC_OFFSET(17045, glUniform2fARB, glUniform2fARB, NULL, _gloffset_Uniform2fARB),
+    NAME_FUNC_OFFSET(17057, glUniform2fvARB, glUniform2fvARB, NULL, _gloffset_Uniform2fvARB),
+    NAME_FUNC_OFFSET(17070, glUniform2iARB, glUniform2iARB, NULL, _gloffset_Uniform2iARB),
+    NAME_FUNC_OFFSET(17082, glUniform2ivARB, glUniform2ivARB, NULL, _gloffset_Uniform2ivARB),
+    NAME_FUNC_OFFSET(17095, glUniform3fARB, glUniform3fARB, NULL, _gloffset_Uniform3fARB),
+    NAME_FUNC_OFFSET(17107, glUniform3fvARB, glUniform3fvARB, NULL, _gloffset_Uniform3fvARB),
+    NAME_FUNC_OFFSET(17120, glUniform3iARB, glUniform3iARB, NULL, _gloffset_Uniform3iARB),
+    NAME_FUNC_OFFSET(17132, glUniform3ivARB, glUniform3ivARB, NULL, _gloffset_Uniform3ivARB),
+    NAME_FUNC_OFFSET(17145, glUniform4fARB, glUniform4fARB, NULL, _gloffset_Uniform4fARB),
+    NAME_FUNC_OFFSET(17157, glUniform4fvARB, glUniform4fvARB, NULL, _gloffset_Uniform4fvARB),
+    NAME_FUNC_OFFSET(17170, glUniform4iARB, glUniform4iARB, NULL, _gloffset_Uniform4iARB),
+    NAME_FUNC_OFFSET(17182, glUniform4ivARB, glUniform4ivARB, NULL, _gloffset_Uniform4ivARB),
+    NAME_FUNC_OFFSET(17195, glUniformMatrix2fvARB, glUniformMatrix2fvARB, NULL, _gloffset_UniformMatrix2fvARB),
+    NAME_FUNC_OFFSET(17214, glUniformMatrix3fvARB, glUniformMatrix3fvARB, NULL, _gloffset_UniformMatrix3fvARB),
+    NAME_FUNC_OFFSET(17233, glUniformMatrix4fvARB, glUniformMatrix4fvARB, NULL, _gloffset_UniformMatrix4fvARB),
+    NAME_FUNC_OFFSET(17252, glUseProgramObjectARB, glUseProgramObjectARB, NULL, _gloffset_UseProgramObjectARB),
+    NAME_FUNC_OFFSET(17265, glValidateProgramARB, glValidateProgramARB, NULL, _gloffset_ValidateProgramARB),
+    NAME_FUNC_OFFSET(17283, glBindAttribLocationARB, glBindAttribLocationARB, NULL, _gloffset_BindAttribLocationARB),
+    NAME_FUNC_OFFSET(17304, glGetActiveAttribARB, glGetActiveAttribARB, NULL, _gloffset_GetActiveAttribARB),
+    NAME_FUNC_OFFSET(17322, glGetAttribLocationARB, glGetAttribLocationARB, NULL, _gloffset_GetAttribLocationARB),
+    NAME_FUNC_OFFSET(17342, glDrawBuffersARB, glDrawBuffersARB, NULL, _gloffset_DrawBuffersARB),
+    NAME_FUNC_OFFSET(17356, glDrawBuffersARB, glDrawBuffersARB, NULL, _gloffset_DrawBuffersARB),
+    NAME_FUNC_OFFSET(17373, gl_dispatch_stub_568, gl_dispatch_stub_568, NULL, _gloffset_SampleMaskSGIS),
+    NAME_FUNC_OFFSET(17389, gl_dispatch_stub_569, gl_dispatch_stub_569, NULL, _gloffset_SamplePatternSGIS),
+    NAME_FUNC_OFFSET(17408, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
+    NAME_FUNC_OFFSET(17426, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
+    NAME_FUNC_OFFSET(17447, glPointParameterfEXT, glPointParameterfEXT, NULL, _gloffset_PointParameterfEXT),
+    NAME_FUNC_OFFSET(17469, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
+    NAME_FUNC_OFFSET(17488, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
+    NAME_FUNC_OFFSET(17510, glPointParameterfvEXT, glPointParameterfvEXT, NULL, _gloffset_PointParameterfvEXT),
+    NAME_FUNC_OFFSET(17533, glSecondaryColor3bEXT, glSecondaryColor3bEXT, NULL, _gloffset_SecondaryColor3bEXT),
+    NAME_FUNC_OFFSET(17552, glSecondaryColor3bvEXT, glSecondaryColor3bvEXT, NULL, _gloffset_SecondaryColor3bvEXT),
+    NAME_FUNC_OFFSET(17572, glSecondaryColor3dEXT, glSecondaryColor3dEXT, NULL, _gloffset_SecondaryColor3dEXT),
+    NAME_FUNC_OFFSET(17591, glSecondaryColor3dvEXT, glSecondaryColor3dvEXT, NULL, _gloffset_SecondaryColor3dvEXT),
+    NAME_FUNC_OFFSET(17611, glSecondaryColor3fEXT, glSecondaryColor3fEXT, NULL, _gloffset_SecondaryColor3fEXT),
+    NAME_FUNC_OFFSET(17630, glSecondaryColor3fvEXT, glSecondaryColor3fvEXT, NULL, _gloffset_SecondaryColor3fvEXT),
+    NAME_FUNC_OFFSET(17650, glSecondaryColor3iEXT, glSecondaryColor3iEXT, NULL, _gloffset_SecondaryColor3iEXT),
+    NAME_FUNC_OFFSET(17669, glSecondaryColor3ivEXT, glSecondaryColor3ivEXT, NULL, _gloffset_SecondaryColor3ivEXT),
+    NAME_FUNC_OFFSET(17689, glSecondaryColor3sEXT, glSecondaryColor3sEXT, NULL, _gloffset_SecondaryColor3sEXT),
+    NAME_FUNC_OFFSET(17708, glSecondaryColor3svEXT, glSecondaryColor3svEXT, NULL, _gloffset_SecondaryColor3svEXT),
+    NAME_FUNC_OFFSET(17728, glSecondaryColor3ubEXT, glSecondaryColor3ubEXT, NULL, _gloffset_SecondaryColor3ubEXT),
+    NAME_FUNC_OFFSET(17748, glSecondaryColor3ubvEXT, glSecondaryColor3ubvEXT, NULL, _gloffset_SecondaryColor3ubvEXT),
+    NAME_FUNC_OFFSET(17769, glSecondaryColor3uiEXT, glSecondaryColor3uiEXT, NULL, _gloffset_SecondaryColor3uiEXT),
+    NAME_FUNC_OFFSET(17789, glSecondaryColor3uivEXT, glSecondaryColor3uivEXT, NULL, _gloffset_SecondaryColor3uivEXT),
+    NAME_FUNC_OFFSET(17810, glSecondaryColor3usEXT, glSecondaryColor3usEXT, NULL, _gloffset_SecondaryColor3usEXT),
+    NAME_FUNC_OFFSET(17830, glSecondaryColor3usvEXT, glSecondaryColor3usvEXT, NULL, _gloffset_SecondaryColor3usvEXT),
+    NAME_FUNC_OFFSET(17851, glSecondaryColorPointerEXT, glSecondaryColorPointerEXT, NULL, _gloffset_SecondaryColorPointerEXT),
+    NAME_FUNC_OFFSET(17875, glMultiDrawArraysEXT, glMultiDrawArraysEXT, NULL, _gloffset_MultiDrawArraysEXT),
+    NAME_FUNC_OFFSET(17893, glMultiDrawElementsEXT, glMultiDrawElementsEXT, NULL, _gloffset_MultiDrawElementsEXT),
+    NAME_FUNC_OFFSET(17913, glFogCoordPointerEXT, glFogCoordPointerEXT, NULL, _gloffset_FogCoordPointerEXT),
+    NAME_FUNC_OFFSET(17931, glFogCoorddEXT, glFogCoorddEXT, NULL, _gloffset_FogCoorddEXT),
+    NAME_FUNC_OFFSET(17943, glFogCoorddvEXT, glFogCoorddvEXT, NULL, _gloffset_FogCoorddvEXT),
+    NAME_FUNC_OFFSET(17956, glFogCoordfEXT, glFogCoordfEXT, NULL, _gloffset_FogCoordfEXT),
+    NAME_FUNC_OFFSET(17968, glFogCoordfvEXT, glFogCoordfvEXT, NULL, _gloffset_FogCoordfvEXT),
+    NAME_FUNC_OFFSET(17981, glBlendFuncSeparateEXT, glBlendFuncSeparateEXT, NULL, _gloffset_BlendFuncSeparateEXT),
+    NAME_FUNC_OFFSET(18001, glBlendFuncSeparateEXT, glBlendFuncSeparateEXT, NULL, _gloffset_BlendFuncSeparateEXT),
+    NAME_FUNC_OFFSET(18025, glWindowPos2dMESA, glWindowPos2dMESA, NULL, _gloffset_WindowPos2dMESA),
+    NAME_FUNC_OFFSET(18039, glWindowPos2dMESA, glWindowPos2dMESA, NULL, _gloffset_WindowPos2dMESA),
+    NAME_FUNC_OFFSET(18056, glWindowPos2dvMESA, glWindowPos2dvMESA, NULL, _gloffset_WindowPos2dvMESA),
+    NAME_FUNC_OFFSET(18071, glWindowPos2dvMESA, glWindowPos2dvMESA, NULL, _gloffset_WindowPos2dvMESA),
+    NAME_FUNC_OFFSET(18089, glWindowPos2fMESA, glWindowPos2fMESA, NULL, _gloffset_WindowPos2fMESA),
+    NAME_FUNC_OFFSET(18103, glWindowPos2fMESA, glWindowPos2fMESA, NULL, _gloffset_WindowPos2fMESA),
+    NAME_FUNC_OFFSET(18120, glWindowPos2fvMESA, glWindowPos2fvMESA, NULL, _gloffset_WindowPos2fvMESA),
+    NAME_FUNC_OFFSET(18135, glWindowPos2fvMESA, glWindowPos2fvMESA, NULL, _gloffset_WindowPos2fvMESA),
+    NAME_FUNC_OFFSET(18153, glWindowPos2iMESA, glWindowPos2iMESA, NULL, _gloffset_WindowPos2iMESA),
+    NAME_FUNC_OFFSET(18167, glWindowPos2iMESA, glWindowPos2iMESA, NULL, _gloffset_WindowPos2iMESA),
+    NAME_FUNC_OFFSET(18184, glWindowPos2ivMESA, glWindowPos2ivMESA, NULL, _gloffset_WindowPos2ivMESA),
+    NAME_FUNC_OFFSET(18199, glWindowPos2ivMESA, glWindowPos2ivMESA, NULL, _gloffset_WindowPos2ivMESA),
+    NAME_FUNC_OFFSET(18217, glWindowPos2sMESA, glWindowPos2sMESA, NULL, _gloffset_WindowPos2sMESA),
+    NAME_FUNC_OFFSET(18231, glWindowPos2sMESA, glWindowPos2sMESA, NULL, _gloffset_WindowPos2sMESA),
+    NAME_FUNC_OFFSET(18248, glWindowPos2svMESA, glWindowPos2svMESA, NULL, _gloffset_WindowPos2svMESA),
+    NAME_FUNC_OFFSET(18263, glWindowPos2svMESA, glWindowPos2svMESA, NULL, _gloffset_WindowPos2svMESA),
+    NAME_FUNC_OFFSET(18281, glWindowPos3dMESA, glWindowPos3dMESA, NULL, _gloffset_WindowPos3dMESA),
+    NAME_FUNC_OFFSET(18295, glWindowPos3dMESA, glWindowPos3dMESA, NULL, _gloffset_WindowPos3dMESA),
+    NAME_FUNC_OFFSET(18312, glWindowPos3dvMESA, glWindowPos3dvMESA, NULL, _gloffset_WindowPos3dvMESA),
+    NAME_FUNC_OFFSET(18327, glWindowPos3dvMESA, glWindowPos3dvMESA, NULL, _gloffset_WindowPos3dvMESA),
+    NAME_FUNC_OFFSET(18345, glWindowPos3fMESA, glWindowPos3fMESA, NULL, _gloffset_WindowPos3fMESA),
+    NAME_FUNC_OFFSET(18359, glWindowPos3fMESA, glWindowPos3fMESA, NULL, _gloffset_WindowPos3fMESA),
+    NAME_FUNC_OFFSET(18376, glWindowPos3fvMESA, glWindowPos3fvMESA, NULL, _gloffset_WindowPos3fvMESA),
+    NAME_FUNC_OFFSET(18391, glWindowPos3fvMESA, glWindowPos3fvMESA, NULL, _gloffset_WindowPos3fvMESA),
+    NAME_FUNC_OFFSET(18409, glWindowPos3iMESA, glWindowPos3iMESA, NULL, _gloffset_WindowPos3iMESA),
+    NAME_FUNC_OFFSET(18423, glWindowPos3iMESA, glWindowPos3iMESA, NULL, _gloffset_WindowPos3iMESA),
+    NAME_FUNC_OFFSET(18440, glWindowPos3ivMESA, glWindowPos3ivMESA, NULL, _gloffset_WindowPos3ivMESA),
+    NAME_FUNC_OFFSET(18455, glWindowPos3ivMESA, glWindowPos3ivMESA, NULL, _gloffset_WindowPos3ivMESA),
+    NAME_FUNC_OFFSET(18473, glWindowPos3sMESA, glWindowPos3sMESA, NULL, _gloffset_WindowPos3sMESA),
+    NAME_FUNC_OFFSET(18487, glWindowPos3sMESA, glWindowPos3sMESA, NULL, _gloffset_WindowPos3sMESA),
+    NAME_FUNC_OFFSET(18504, glWindowPos3svMESA, glWindowPos3svMESA, NULL, _gloffset_WindowPos3svMESA),
+    NAME_FUNC_OFFSET(18519, glWindowPos3svMESA, glWindowPos3svMESA, NULL, _gloffset_WindowPos3svMESA),
+    NAME_FUNC_OFFSET(18537, glBindProgramNV, glBindProgramNV, NULL, _gloffset_BindProgramNV),
+    NAME_FUNC_OFFSET(18554, glDeleteProgramsNV, glDeleteProgramsNV, NULL, _gloffset_DeleteProgramsNV),
+    NAME_FUNC_OFFSET(18574, glGenProgramsNV, glGenProgramsNV, NULL, _gloffset_GenProgramsNV),
+    NAME_FUNC_OFFSET(18591, glGetVertexAttribPointervNV, glGetVertexAttribPointervNV, NULL, _gloffset_GetVertexAttribPointervNV),
+    NAME_FUNC_OFFSET(18617, glGetVertexAttribPointervNV, glGetVertexAttribPointervNV, NULL, _gloffset_GetVertexAttribPointervNV),
+    NAME_FUNC_OFFSET(18646, glIsProgramNV, glIsProgramNV, NULL, _gloffset_IsProgramNV),
+    NAME_FUNC_OFFSET(18661, glPointParameteriNV, glPointParameteriNV, NULL, _gloffset_PointParameteriNV),
+    NAME_FUNC_OFFSET(18679, glPointParameterivNV, glPointParameterivNV, NULL, _gloffset_PointParameterivNV),
+    NAME_FUNC_OFFSET(18698, gl_dispatch_stub_749, gl_dispatch_stub_749, NULL, _gloffset_BlendEquationSeparateEXT),
+    NAME_FUNC_OFFSET(18722, gl_dispatch_stub_749, gl_dispatch_stub_749, NULL, _gloffset_BlendEquationSeparateEXT),
     NAME_FUNC_OFFSET(-1, NULL, NULL, NULL, 0)
 };
 
diff --git a/src/mesa/main/api_arrayelt.c b/src/mesa/main/api_arrayelt.c
index 1899975213..72091b0789 100644
--- a/src/mesa/main/api_arrayelt.c
+++ b/src/mesa/main/api_arrayelt.c
@@ -31,8 +31,8 @@
 #include "context.h"
 #include "imports.h"
 #include "macros.h"
-#include "glapioffsets.h"
-#include "dispatch.h"
+#include "glapi/glapioffsets.h"
+#include "glapi/dispatch.h"
 
 typedef void (GLAPIENTRY *array_func)( const void * );
 
diff --git a/src/mesa/main/api_loopback.c b/src/mesa/main/api_loopback.c
index efe5a77d58..924d7134a2 100644
--- a/src/mesa/main/api_loopback.c
+++ b/src/mesa/main/api_loopback.c
@@ -30,14 +30,14 @@
 
 
 #include "glheader.h"
-#include "glapi.h"
-#include "glapitable.h"
 #include "macros.h"
 #include "colormac.h"
 #include "api_loopback.h"
-#include "glthread.h"
 #include "mtypes.h"
-#include "dispatch.h"
+#include "glapi/glapi.h"
+#include "glapi/glapitable.h"
+#include "glapi/glthread.h"
+#include "glapi/dispatch.h"
 
 /* KW: A set of functions to convert unusual Color/Normal/Vertex/etc
  * calls to a smaller set of driver-provided formats.  Currently just
diff --git a/src/mesa/main/api_noop.c b/src/mesa/main/api_noop.c
index 0c1a35361f..3df64362ea 100644
--- a/src/mesa/main/api_noop.c
+++ b/src/mesa/main/api_noop.c
@@ -31,7 +31,7 @@
 #include "light.h"
 #include "macros.h"
 #include "dlist.h"
-#include "dispatch.h"
+#include "glapi/dispatch.h"
 
 
 /**
diff --git a/src/mesa/main/api_validate.c b/src/mesa/main/api_validate.c
index 3d20ba7d14..841c6a5302 100644
--- a/src/mesa/main/api_validate.c
+++ b/src/mesa/main/api_validate.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.1
+ * Version:  7.0.1
  *
- * Copyright (C) 1999-2003  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -100,6 +100,11 @@ _mesa_validate_DrawElements(GLcontext *ctx,
                          (const GLubyte *) indices);
       }
    }
+   else {
+      /* not using a VBO */
+      if (!indices)
+         return GL_FALSE;
+   }
 
    if (ctx->Const.CheckArrayBounds) {
       /* find max array index */
@@ -170,6 +175,16 @@ _mesa_validate_DrawRangeElements(GLcontext *ctx, GLenum mode,
        && !(ctx->VertexProgram._Enabled && ctx->Array.ArrayObj->VertexAttrib[0].Enabled))
       return GL_FALSE;
 
+   /* Vertex buffer object tests */
+   if (ctx->Array.ElementArrayBufferObj->Name) {
+      /* XXX re-use code from above? */
+   }
+   else {
+      /* not using VBO */
+      if (!indices)
+         return GL_FALSE;
+   }
+
    if (ctx->Const.CheckArrayBounds) {
       /* Find max array index.
        * We don't trust the user's start and end values.
diff --git a/src/mesa/main/arrayobj.c b/src/mesa/main/arrayobj.c
index d601ee461e..f08f99d8e1 100644
--- a/src/mesa/main/arrayobj.c
+++ b/src/mesa/main/arrayobj.c
@@ -46,7 +46,7 @@
 #include "bufferobj.h"
 #endif
 #include "arrayobj.h"
-#include "dispatch.h"
+#include "glapi/dispatch.h"
 
 
 /**
diff --git a/src/mesa/main/attrib.c b/src/mesa/main/attrib.c
index 1aa0a02fc7..b422198f92 100644
--- a/src/mesa/main/attrib.c
+++ b/src/mesa/main/attrib.c
@@ -98,9 +98,13 @@ _mesa_PushAttrib(GLbitfield mask)
    }
 
    if (mask & GL_COLOR_BUFFER_BIT) {
+      GLuint i;
       struct gl_colorbuffer_attrib *attr;
       attr = MALLOC_STRUCT( gl_colorbuffer_attrib );
       MEMCPY( attr, &ctx->Color, sizeof(struct gl_colorbuffer_attrib) );
+      /* push the Draw FBO's DrawBuffer[] state, not ctx->Color.DrawBuffer[] */
+      for (i = 0; i < ctx->Const.MaxDrawBuffers; i ++)
+         attr->DrawBuffer[i] = ctx->DrawBuffer->ColorDrawBuffer[i];
       newnode = new_attrib_node( GL_COLOR_BUFFER_BIT );
       newnode->data = attr;
       newnode->next = head;
diff --git a/src/mesa/main/buffers.c b/src/mesa/main/buffers.c
index 11bd173e35..0e6ca8ea1c 100644
--- a/src/mesa/main/buffers.c
+++ b/src/mesa/main/buffers.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.2
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -184,17 +184,19 @@ _mesa_Clear( GLbitfield mask )
 
 /**
  * Return bitmask of BUFFER_BIT_* flags indicating which color buffers are
- * available to the rendering context.
- * This depends on the framebuffer we're writing to.  For window system
- * framebuffers we look at the framebuffer's visual.  But for user-
- * create framebuffers we look at the number of supported color attachments.
+ * available to the rendering context (for drawing or reading).
+ * This depends on the type of framebuffer.  For window system framebuffers
+ * we look at the framebuffer's visual.  But for user-create framebuffers we
+ * look at the number of supported color attachments.
+ * \param fb  the framebuffer to draw to, or read from
+ * \return  bitmask of BUFFER_BIT_* flags
  */
 static GLbitfield
-supported_buffer_bitmask(const GLcontext *ctx, GLuint framebufferID)
+supported_buffer_bitmask(const GLcontext *ctx, const struct gl_framebuffer *fb)
 {
    GLbitfield mask = 0x0;
 
-   if (framebufferID > 0) {
+   if (fb->Name > 0) {
       /* A user-created renderbuffer */
       GLuint i;
       ASSERT(ctx->Extensions.EXT_framebuffer_object);
@@ -203,20 +205,20 @@ supported_buffer_bitmask(const GLcontext *ctx, GLuint framebufferID)
       }
    }
    else {
-      /* A window system renderbuffer */
+      /* A window system framebuffer */
       GLint i;
       mask = BUFFER_BIT_FRONT_LEFT; /* always have this */
-      if (ctx->Visual.stereoMode) {
+      if (fb->Visual.stereoMode) {
          mask |= BUFFER_BIT_FRONT_RIGHT;
-         if (ctx->Visual.doubleBufferMode) {
+         if (fb->Visual.doubleBufferMode) {
             mask |= BUFFER_BIT_BACK_LEFT | BUFFER_BIT_BACK_RIGHT;
          }
       }
-      else if (ctx->Visual.doubleBufferMode) {
+      else if (fb->Visual.doubleBufferMode) {
          mask |= BUFFER_BIT_BACK_LEFT;
       }
 
-      for (i = 0; i < ctx->Visual.numAuxBuffers; i++) {
+      for (i = 0; i < fb->Visual.numAuxBuffers; i++) {
          mask |= (BUFFER_BIT_AUX0 << i);
       }
    }
@@ -338,7 +340,6 @@ read_buffer_enum_to_index(GLenum buffer)
 void GLAPIENTRY
 _mesa_DrawBuffer(GLenum buffer)
 {
-   GLuint bufferID;
    GLbitfield destMask;
    GET_CURRENT_CONTEXT(ctx);
    ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); /* too complex... */
@@ -347,13 +348,12 @@ _mesa_DrawBuffer(GLenum buffer)
       _mesa_debug(ctx, "glDrawBuffer %s\n", _mesa_lookup_enum_by_nr(buffer));
    }
 
-   bufferID = ctx->DrawBuffer->Name;
-
    if (buffer == GL_NONE) {
       destMask = 0x0;
    }
    else {
-      const GLbitfield supportedMask = supported_buffer_bitmask(ctx, bufferID);
+      const GLbitfield supportedMask
+         = supported_buffer_bitmask(ctx, ctx->DrawBuffer);
       destMask = draw_buffer_enum_to_bitmask(buffer);
       if (destMask == BAD_MASK) {
          /* totally bogus buffer */
@@ -370,6 +370,14 @@ _mesa_DrawBuffer(GLenum buffer)
 
    /* if we get here, there's no error so set new state */
    _mesa_drawbuffers(ctx, 1, &buffer, &destMask);
+
+   /*
+    * Call device driver function.
+    */
+   if (ctx->Driver.DrawBuffers)
+      ctx->Driver.DrawBuffers(ctx, 1, &buffer);
+   else if (ctx->Driver.DrawBuffer)
+      ctx->Driver.DrawBuffer(ctx, buffer);
 }
 
 
@@ -386,7 +394,6 @@ void GLAPIENTRY
 _mesa_DrawBuffersARB(GLsizei n, const GLenum *buffers)
 {
    GLint output;
-   GLuint bufferID;
    GLbitfield usedBufferMask, supportedMask;
    GLbitfield destMask[MAX_DRAW_BUFFERS];
    GET_CURRENT_CONTEXT(ctx);
@@ -401,9 +408,7 @@ _mesa_DrawBuffersARB(GLsizei n, const GLenum *buffers)
       return;
    }
 
-   bufferID = ctx->DrawBuffer->Name;
-
-   supportedMask = supported_buffer_bitmask(ctx, bufferID);
+   supportedMask = supported_buffer_bitmask(ctx, ctx->DrawBuffer);
    usedBufferMask = 0x0;
 
    /* complicated error checking... */
@@ -438,6 +443,14 @@ _mesa_DrawBuffersARB(GLsizei n, const GLenum *buffers)
 
    /* OK, if we get here, there were no errors so set the new state */
    _mesa_drawbuffers(ctx, n, buffers, destMask);
+
+   /*
+    * Call device driver function.
+    */
+   if (ctx->Driver.DrawBuffers)
+      ctx->Driver.DrawBuffers(ctx, n, buffers);
+   else if (ctx->Driver.DrawBuffer)
+      ctx->Driver.DrawBuffer(ctx, buffers[0]);
 }
 
 
@@ -466,14 +479,15 @@ set_color_output(GLcontext *ctx, GLuint output, GLenum buffer,
    /* not really needed, will be set later */
    fb->_NumColorDrawBuffers[output] = 0;
 
+   if (fb->Name == 0)
    /* Set traditional state var */
-   ctx->Color.DrawBuffer[output] = buffer;
+      ctx->Color.DrawBuffer[output] = buffer;
 }
 
 
 /**
  * Helper routine used by _mesa_DrawBuffer, _mesa_DrawBuffersARB and
- * _mesa_PopAttrib to set drawbuffer state.
+ * other places (window fbo fixup) to set fbo (and the old ctx) fields.
  * All error checking will have been done prior to calling this function
  * so nothing should go wrong at this point.
  * \param ctx  current context
@@ -492,8 +506,8 @@ _mesa_drawbuffers(GLcontext *ctx, GLuint n, const GLenum *buffers,
 
    if (!destMask) {
       /* compute destMask values now */
-      const GLuint bufferID = ctx->DrawBuffer->Name;
-      const GLbitfield supportedMask = supported_buffer_bitmask(ctx, bufferID);
+      const GLbitfield supportedMask
+         = supported_buffer_bitmask(ctx, ctx->DrawBuffer);
       for (output = 0; output < n; output++) {
          mask[output] = draw_buffer_enum_to_bitmask(buffers[output]);
          ASSERT(mask[output] != BAD_MASK);
@@ -512,39 +526,22 @@ _mesa_drawbuffers(GLcontext *ctx, GLuint n, const GLenum *buffers,
    }
 
    ctx->NewState |= _NEW_COLOR;
-
-   /*
-    * Call device driver function.
-    */
-   if (ctx->Driver.DrawBuffers)
-      ctx->Driver.DrawBuffers(ctx, n, buffers);
-   else if (ctx->Driver.DrawBuffer)
-      ctx->Driver.DrawBuffer(ctx, buffers[0]);
 }
 
 
-
-/**
- * Called by glReadBuffer to set the source renderbuffer for reading pixels.
- * \param mode color buffer such as GL_FRONT, GL_BACK, etc.
- */
-void GLAPIENTRY
-_mesa_ReadBuffer(GLenum buffer)
+GLboolean
+_mesa_readbuffer_update_fields(GLcontext *ctx, GLenum buffer)
 {
    struct gl_framebuffer *fb;
    GLbitfield supportedMask;
    GLint srcBuffer;
-   GLuint bufferID;
-   GET_CURRENT_CONTEXT(ctx);
-   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx);
 
    fb = ctx->ReadBuffer;
-   bufferID = fb->Name;
 
    if (MESA_VERBOSE & VERBOSE_API)
       _mesa_debug(ctx, "glReadBuffer %s\n", _mesa_lookup_enum_by_nr(buffer));
 
-   if (bufferID > 0 && buffer == GL_NONE) {
+   if (fb->Name > 0 && buffer == GL_NONE) {
       /* This is legal for user-created framebuffer objects */
       srcBuffer = -1;
    }
@@ -553,21 +550,42 @@ _mesa_ReadBuffer(GLenum buffer)
       srcBuffer = read_buffer_enum_to_index(buffer);
       if (srcBuffer == -1) {
          _mesa_error(ctx, GL_INVALID_ENUM, "glReadBuffer(buffer=0x%x)", buffer);
-         return;
+         return GL_FALSE;
       }
-      supportedMask = supported_buffer_bitmask(ctx, bufferID);
+      supportedMask = supported_buffer_bitmask(ctx, fb);
       if (((1 << srcBuffer) & supportedMask) == 0) {
          _mesa_error(ctx, GL_INVALID_OPERATION, "glReadBuffer(buffer=0x%x)", buffer);
-         return;
+         return GL_FALSE;
       }
    }
 
-   if (bufferID == 0) {
+   if (fb->Name == 0) {
       ctx->Pixel.ReadBuffer = buffer;
    }
    fb->ColorReadBuffer = buffer;
    fb->_ColorReadBufferIndex = srcBuffer;
 
+   return GL_TRUE;
+}
+
+
+
+/**
+ * Called by glReadBuffer to set the source renderbuffer for reading pixels.
+ * \param mode color buffer such as GL_FRONT, GL_BACK, etc.
+ */
+void GLAPIENTRY
+_mesa_ReadBuffer(GLenum buffer)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx);
+
+   if (MESA_VERBOSE & VERBOSE_API)
+      _mesa_debug(ctx, "glReadBuffer %s\n", _mesa_lookup_enum_by_nr(buffer));
+
+   if (!_mesa_readbuffer_update_fields(ctx, buffer))
+      return;
+
    ctx->NewState |= _NEW_PIXEL;
 
    /*
diff --git a/src/mesa/main/buffers.h b/src/mesa/main/buffers.h
index fcc2152342..208e7af2b9 100644
--- a/src/mesa/main/buffers.h
+++ b/src/mesa/main/buffers.h
@@ -56,6 +56,9 @@ extern void
 _mesa_drawbuffers(GLcontext *ctx, GLuint n, const GLenum *buffers,
                   const GLbitfield *destMask);
 
+extern GLboolean
+_mesa_readbuffer_update_fields(GLcontext *ctx, GLenum buffer);
+
 extern void GLAPIENTRY
 _mesa_ReadBuffer( GLenum mode );
 
diff --git a/src/mesa/main/config.h b/src/mesa/main/config.h
index 9e4d1838ad..cebef1c383 100644
--- a/src/mesa/main/config.h
+++ b/src/mesa/main/config.h
@@ -5,9 +5,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.2
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -82,7 +82,7 @@
 /** Minimum point size */
 #define MIN_POINT_SIZE 1.0
 /** Maximum point size */
-#define MAX_POINT_SIZE 20.0
+#define MAX_POINT_SIZE 60.0
 /** Point size granularity */
 #define POINT_SIZE_GRANULARITY 0.1
 
@@ -243,10 +243,7 @@
 
 
 /**
- * Bits per depth buffer value.  
- *
- * Any reasonable value up to 31 will work.  32 doesn't work because of integer
- * overflow problems in the rasterizer code.
+ * Bits per depth buffer value (max is 32).
  */
 #ifndef DEFAULT_SOFTWARE_DEPTH_BITS
 #define DEFAULT_SOFTWARE_DEPTH_BITS 16
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index ccaf6f6428..00e4c8328e 100644
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -97,12 +97,9 @@
 #include "fog.h"
 #include "framebuffer.h"
 #include "get.h"
-#include "glthread.h"
-#include "glapioffsets.h"
 #include "histogram.h"
 #include "hint.h"
 #include "hash.h"
-#include "atifragshader.h"
 #include "light.h"
 #include "lines.h"
 #include "macros.h"
@@ -110,9 +107,6 @@
 #include "pixel.h"
 #include "points.h"
 #include "polygon.h"
-#if FEATURE_NV_vertex_program || FEATURE_NV_fragment_program
-#include "program.h"
-#endif
 #include "queryobj.h"
 #include "rastpos.h"
 #include "simple_list.h"
@@ -126,13 +120,19 @@
 #include "varray.h"
 #include "version.h"
 #include "vtxfmt.h"
+#include "glapi/glthread.h"
+#include "glapi/glapioffsets.h"
+#if FEATURE_NV_vertex_program || FEATURE_NV_fragment_program
+#include "shader/program.h"
+#endif
+#include "shader/shader_api.h"
+#include "shader/atifragshader.h"
 #if _HAVE_FULL_GL
 #include "math/m_translate.h"
 #include "math/m_matrix.h"
 #include "math/m_xform.h"
 #include "math/mathmod.h"
 #endif
-#include "shader_api.h"
 
 #ifdef USE_SPARC_ASM
 #include "sparc/sparc.h"
@@ -978,7 +978,7 @@ init_attrib_groups(GLcontext *ctx)
 static int
 generic_nop(void)
 {
-   _mesa_problem(NULL, "User called no-op dispatch function (an unsupported extension function?)");
+   _mesa_warning(NULL, "User called no-op dispatch function (an unsupported extension function?)");
    return 0;
 }
 
@@ -1357,9 +1357,9 @@ _mesa_copy_context( const GLcontext *src, GLcontext *dst, GLuint mask )
  * Check if the given context can render into the given framebuffer
  * by checking visual attributes.
  *
- * XXX this may go away someday because we're moving toward more freedom
- * in binding contexts to drawables with different visual attributes.
- * The GL_EXT_f_b_o extension is prompting some of that.
+ * Most of these tests could go away because Mesa is now pretty flexible
+ * in terms of mixing rendering contexts with framebuffers.  As long
+ * as RGB vs. CI mode agree, we're probably good.
  *
  * \return GL_TRUE if compatible, GL_FALSE otherwise.
  */
@@ -1393,8 +1393,11 @@ check_compatible(const GLcontext *ctx, const GLframebuffer *buffer)
       return GL_FALSE;
    if (ctxvis->blueMask && ctxvis->blueMask != bufvis->blueMask)
       return GL_FALSE;
+#if 0
+   /* disabled (see bug 11161) */
    if (ctxvis->depthBits && ctxvis->depthBits != bufvis->depthBits)
       return GL_FALSE;
+#endif
    if (ctxvis->stencilBits && ctxvis->stencilBits != bufvis->stencilBits)
       return GL_FALSE;
 
@@ -1492,9 +1495,19 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
           */
          if (!newCtx->DrawBuffer || newCtx->DrawBuffer->Name == 0) {
             _mesa_reference_framebuffer(&newCtx->DrawBuffer, drawBuffer);
+         /* fix up the fb fields - these will end up wrong otherwise
+            if the DRIdrawable changes, and everything relies on them.
+            This is a bit messy (same as needed in _mesa_BindFramebufferEXT) */
+            int i;
+            GLenum buffers[MAX_DRAW_BUFFERS];
+            for(i = 0; i < newCtx->Const.MaxDrawBuffers; i++) {
+               buffers[i] = newCtx->Color.DrawBuffer[i];
+            }
+            _mesa_drawbuffers(newCtx, newCtx->Const.MaxDrawBuffers, buffers, NULL);
          }
          if (!newCtx->ReadBuffer || newCtx->ReadBuffer->Name == 0) {
             _mesa_reference_framebuffer(&newCtx->ReadBuffer, readBuffer);
+            _mesa_readbuffer_update_fields(newCtx, newCtx->Pixel.ReadBuffer);
          }
 
 	 newCtx->NewState |= _NEW_BUFFERS;
diff --git a/src/mesa/main/context.h b/src/mesa/main/context.h
index b6a9d13149..099912aa15 100644
--- a/src/mesa/main/context.h
+++ b/src/mesa/main/context.h
@@ -49,7 +49,7 @@
 #define CONTEXT_H
 
 
-#include "glapi.h"
+#include "glapi/glapi.h"
 #include "imports.h"
 #include "mtypes.h"
 
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index 88f33943b3..caa50dd682 100644
--- a/src/mesa/main/dd.h
+++ b/src/mesa/main/dd.h
@@ -782,7 +782,7 @@ struct dd_function_table {
    struct gl_framebuffer * (*NewFramebuffer)(GLcontext *ctx, GLuint name);
    struct gl_renderbuffer * (*NewRenderbuffer)(GLcontext *ctx, GLuint name);
    void (*BindFramebuffer)(GLcontext *ctx, GLenum target,
-                           struct gl_framebuffer *fb);
+                           struct gl_framebuffer *fb, struct gl_framebuffer *fbread);
    void (*FramebufferRenderbuffer)(GLcontext *ctx, 
                                    struct gl_framebuffer *fb,
                                    GLenum attachment,
diff --git a/src/mesa/main/depth.c b/src/mesa/main/depth.c
index f5511ce2fb..91c036ef96 100644
--- a/src/mesa/main/depth.c
+++ b/src/mesa/main/depth.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5
+ * Version:  7.1
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -28,7 +28,6 @@
 #include "context.h"
 #include "depth.h"
 #include "enums.h"
-#include "hash.h"
 #include "macros.h"
 #include "mtypes.h"
 
@@ -153,36 +152,11 @@ _mesa_DepthBoundsEXT( GLclampd zmin, GLclampd zmax )
 /**
  * Initialize the depth buffer attribute group in the given context.
  */
-void _mesa_init_depth( GLcontext * ctx )
+void
+_mesa_init_depth(GLcontext *ctx)
 {
-   /* Depth buffer group */
    ctx->Depth.Test = GL_FALSE;
    ctx->Depth.Clear = 1.0;
    ctx->Depth.Func = GL_LESS;
    ctx->Depth.Mask = GL_TRUE;
-
-   /* XXX this is now per-framebuffer state */
-#if 00
-   /* Z buffer stuff */
-   if (ctx->Visual.depthBits == 0) {
-      /* Special case.  Even if we don't have a depth buffer we need
-       * good values for DepthMax for Z vertex transformation purposes
-       * and for per-fragment fog computation.
-       */
-      ctx->DepthMax = (1 << 16) - 1;
-      ctx->DepthMaxF = (GLfloat) ctx->DepthMax;
-   }
-   else if (ctx->Visual.depthBits < 32) {
-      ctx->DepthMax = (1 << ctx->Visual.depthBits) - 1;
-      ctx->DepthMaxF = (GLfloat) ctx->DepthMax;
-   }
-   else {
-      /* Special case since shift values greater than or equal to the
-       * number of bits in the left hand expression's type are undefined.
-       */
-      ctx->DepthMax = 0xffffffff;
-      ctx->DepthMaxF = (GLfloat) ctx->DepthMax;
-   }
-   ctx->MRD = 1.0;  /* Minimum resolvable depth value, for polygon offset */
-#endif
 }
diff --git a/src/mesa/main/dlist.c b/src/mesa/main/dlist.c
index 91ddcbf9ed..293ee5fa34 100644
--- a/src/mesa/main/dlist.c
+++ b/src/mesa/main/dlist.c
@@ -33,10 +33,6 @@
 #include "api_arrayelt.h"
 #include "api_loopback.h"
 #include "config.h"
-#if FEATURE_ARB_vertex_program || FEATURE_ARB_fragment_program
-#include "arbprogram.h"
-#include "program.h"
-#endif
 #include "attrib.h"
 #include "blend.h"
 #include "buffers.h"
@@ -57,7 +53,7 @@
 #include "extensions.h"
 #include "feedback.h"
 #include "get.h"
-#include "glapi.h"
+#include "glapi/glapi.h"
 #include "hash.h"
 #include "histogram.h"
 #include "image.h"
@@ -76,18 +72,22 @@
 #include "texstate.h"
 #include "mtypes.h"
 #include "varray.h"
+#if FEATURE_ARB_vertex_program || FEATURE_ARB_fragment_program
+#include "shader/arbprogram.h"
+#include "shader/program.h"
+#endif
 #if FEATURE_NV_vertex_program || FEATURE_NV_fragment_program
-#include "nvprogram.h"
-#include "program.h"
+#include "shader/nvprogram.h"
+#include "shader/program.h"
 #endif
 #if FEATURE_ATI_fragment_shader
-#include "atifragshader.h"
+#include "shader/atifragshader.h"
 #endif
 
 #include "math/m_matrix.h"
 #include "math/m_xform.h"
 
-#include "dispatch.h"
+#include "glapi/dispatch.h"
 
 
 /**
diff --git a/src/mesa/main/execmem.c b/src/mesa/main/execmem.c
index df3095232d..40f66d7da2 100644
--- a/src/mesa/main/execmem.c
+++ b/src/mesa/main/execmem.c
@@ -32,7 +32,7 @@
 
 
 #include "imports.h"
-#include "glthread.h"
+#include "glapi/glthread.h"
 
 
 
diff --git a/src/mesa/main/fbobject.c b/src/mesa/main/fbobject.c
index e3bada5ae8..6f7effcce7 100644
--- a/src/mesa/main/fbobject.c
+++ b/src/mesa/main/fbobject.c
@@ -29,6 +29,7 @@
  */
 
 
+#include "buffers.h"
 #include "context.h"
 #include "fbobject.h"
 #include "framebuffer.h"
@@ -924,7 +925,7 @@ check_end_texture_render(GLcontext *ctx, struct gl_framebuffer *fb)
 void GLAPIENTRY
 _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
 {
-   struct gl_framebuffer *newFb;
+   struct gl_framebuffer *newFb, *newFbread;
    GLboolean bindReadBuf, bindDrawBuf;
    GET_CURRENT_CONTEXT(ctx);
 
@@ -984,12 +985,14 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
 	 }
          _mesa_HashInsert(ctx->Shared->FrameBuffers, framebuffer, newFb);
       }
+      newFbread = newFb;
    }
    else {
       /* Binding the window system framebuffer (which was originally set
        * with MakeCurrent).
        */
       newFb = ctx->WinSysDrawBuffer;
+      newFbread = ctx->WinSysReadBuffer;
    }
 
    ASSERT(newFb);
@@ -999,8 +1002,16 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
     * XXX check if re-binding same buffer and skip some of this code.
     */
 
+   /* for window-framebuffers, re-initialize the fbo values, as they
+      could be wrong (makecurrent with a new drawable while still a fbo
+      was bound will lead to default init fbo values).
+      note that therefore the context ReadBuffer/DrawBuffer values are not
+      valid while fbo's are bound!!! */
    if (bindReadBuf) {
-      _mesa_reference_framebuffer(&ctx->ReadBuffer, newFb);
+      _mesa_reference_framebuffer(&ctx->ReadBuffer, newFbread);
+      if (!newFbread->Name) {
+         _mesa_readbuffer_update_fields(ctx, ctx->Pixel.ReadBuffer);
+      }
    }
 
    if (bindDrawBuf) {
@@ -1008,14 +1019,22 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
       check_end_texture_render(ctx, ctx->DrawBuffer);
       /* check if time to delete this framebuffer */
       _mesa_reference_framebuffer(&ctx->DrawBuffer, newFb);
-      if (newFb->Name != 0) {
+      if (!newFb->Name) {
+         GLuint i;
+         GLenum buffers[MAX_DRAW_BUFFERS];
+         for(i = 0; i < ctx->Const.MaxDrawBuffers; i++) {
+            buffers[i] = ctx->Color.DrawBuffer[i];
+         }
+         _mesa_drawbuffers(ctx, ctx->Const.MaxDrawBuffers, buffers, NULL);
+      }
+      else {
          /* check if newly bound framebuffer has any texture attachments */
          check_begin_texture_render(ctx, newFb);
       }
    }
 
    if (ctx->Driver.BindFramebuffer) {
-      ctx->Driver.BindFramebuffer(ctx, target, newFb);
+      ctx->Driver.BindFramebuffer(ctx, target, newFb, newFbread);
    }
 }
 
diff --git a/src/mesa/main/framebuffer.c b/src/mesa/main/framebuffer.c
index 1fd31a5321..c9b30d3252 100644
--- a/src/mesa/main/framebuffer.c
+++ b/src/mesa/main/framebuffer.c
@@ -65,7 +65,9 @@ compute_depth_max(struct gl_framebuffer *fb)
       fb->_DepthMax = 0xffffffff;
    }
    fb->_DepthMaxF = (GLfloat) fb->_DepthMax;
-   fb->_MRD = 1.0;  /* Minimum resolvable depth value, for polygon offset */
+
+   /* Minimum resolvable depth value, for polygon offset */
+   fb->_MRD = 1.0 / fb->_DepthMaxF;
 }
 
 
@@ -652,6 +654,27 @@ update_color_read_buffer(GLcontext *ctx, struct gl_framebuffer *fb)
 }
 
 
+static void
+update_framebuffer(GLcontext *ctx, struct gl_framebuffer *fb)
+{
+   /* Completeness only matters for user-created framebuffers */
+   if (fb->Name != 0) {
+      /* XXX: EXT_framebuffer_blit:
+         framebuffer must still be complete wrt read/draw? */
+      _mesa_test_framebuffer_completeness(ctx, fb);
+      _mesa_update_framebuffer_visual(fb);
+   }
+
+   /* update_color_draw/read_buffers not needed for
+      read/draw only fb, but shouldn't hurt ??? */
+   update_color_draw_buffers(ctx, fb);
+   update_color_read_buffer(ctx, fb);
+   _mesa_update_depth_buffer(ctx, fb, BUFFER_DEPTH);
+   _mesa_update_stencil_buffer(ctx, fb, BUFFER_STENCIL);
+
+   compute_depth_max(fb);
+}
+
 /**
  * Update state related to the current draw/read framebuffers.
  * Specifically, update these framebuffer fields:
@@ -669,19 +692,11 @@ void
 _mesa_update_framebuffer(GLcontext *ctx)
 {
    struct gl_framebuffer *fb = ctx->DrawBuffer;
+   struct gl_framebuffer *fbread = ctx->ReadBuffer;
 
-   /* Completeness only matters for user-created framebuffers */
-   if (fb->Name != 0) {
-      _mesa_test_framebuffer_completeness(ctx, fb);
-      _mesa_update_framebuffer_visual(fb);
-   }
-
-   update_color_draw_buffers(ctx, fb);
-   update_color_read_buffer(ctx, fb);
-   _mesa_update_depth_buffer(ctx, fb, BUFFER_DEPTH);
-   _mesa_update_stencil_buffer(ctx, fb, BUFFER_STENCIL);
-
-   compute_depth_max(fb);
+   update_framebuffer(ctx, fb);
+   if (fbread != fb)
+      update_framebuffer(ctx, fbread);
 }
 
 
diff --git a/src/mesa/main/glheader.h b/src/mesa/main/glheader.h
index 5abea137d7..fd4127558a 100644
--- a/src/mesa/main/glheader.h
+++ b/src/mesa/main/glheader.h
@@ -91,6 +91,9 @@
 #  pragma disable_message(201) /* Disable unreachable code warnings */
 #endif
 
+#ifdef WGLAPI
+#	undef WGLAPI
+#endif
 
 #if !defined(OPENSTEP) && (defined(__WIN32__) && !defined(__CYGWIN__)) && !defined(BUILD_FOR_SNAP)
 #  if !defined(__GNUC__) /* mingw environment */
diff --git a/src/mesa/main/hash.c b/src/mesa/main/hash.c
index 2d5bcc3e01..ffb2c4d946 100644
--- a/src/mesa/main/hash.c
+++ b/src/mesa/main/hash.c
@@ -37,7 +37,7 @@
 
 #include "glheader.h"
 #include "imports.h"
-#include "glthread.h"
+#include "glapi/glthread.h"
 #include "hash.h"
 
 
diff --git a/src/mesa/main/image.c b/src/mesa/main/image.c
index caaf281657..ba46cdc1b1 100644
--- a/src/mesa/main/image.c
+++ b/src/mesa/main/image.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  7.0
+ * Version:  7.1
  *
  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
@@ -3648,11 +3648,13 @@ _mesa_unpack_stencil_span( const GLcontext *ctx, GLuint n,
     * Try simple cases first
     */
    if (transferOps == 0 &&
+       !ctx->Pixel.MapStencilFlag &&
        srcType == GL_UNSIGNED_BYTE &&
        dstType == GL_UNSIGNED_BYTE) {
       _mesa_memcpy(dest, source, n * sizeof(GLubyte));
    }
    else if (transferOps == 0 &&
+            !ctx->Pixel.MapStencilFlag &&
             srcType == GL_UNSIGNED_INT &&
             dstType == GL_UNSIGNED_INT &&
             !srcPacking->SwapBytes) {
@@ -3668,19 +3670,17 @@ _mesa_unpack_stencil_span( const GLcontext *ctx, GLuint n,
       extract_uint_indexes(n, indexes, GL_STENCIL_INDEX, srcType, source,
                            srcPacking);
 
-      if (transferOps) {
-         if (transferOps & IMAGE_SHIFT_OFFSET_BIT) {
-            /* shift and offset indexes */
-            shift_and_offset_ci(ctx, n, indexes);
-         }
+      if (transferOps & IMAGE_SHIFT_OFFSET_BIT) {
+         /* shift and offset indexes */
+         shift_and_offset_ci(ctx, n, indexes);
+      }
 
-         if (ctx->Pixel.MapStencilFlag) {
-            /* Apply stencil lookup table */
-            GLuint mask = ctx->PixelMaps.StoS.Size - 1;
-            GLuint i;
-            for (i=0;i<n;i++) {
-               indexes[i] = ctx->PixelMaps.StoS.Map[ indexes[i] & mask ];
-            }
+      if (ctx->Pixel.MapStencilFlag) {
+         /* Apply stencil lookup table */
+         const GLuint mask = ctx->PixelMaps.StoS.Size - 1;
+         GLuint i;
+         for (i = 0; i < n; i++) {
+            indexes[i] = ctx->PixelMaps.StoS.Map[ indexes[i] & mask ];
          }
       }
 
@@ -3878,17 +3878,28 @@ _mesa_pack_stencil_span( const GLcontext *ctx, GLuint n,
                     SWAP4BYTE(value);                                   \
                 }                                                       \
             }                                                           \
-            depthValues[i] = CLAMP(GLTYPE2FLOAT(value), 0.0F, 1.0F);    \
+            depthValues[i] = GLTYPE2FLOAT(value);                       \
         }                                                               \
     } while (0)
 
+
+/**
+ * Unpack a row of depth/z values from memory, returning GLushort, GLuint
+ * or GLfloat values.
+ * The glPixelTransfer (scale/bias) params will be applied.
+ *
+ * \param dstType  one of GL_UNSIGNED_SHORT, GL_UNSIGNED_INT, GL_FLOAT
+ * \param depthMax  max value for returned GLushort or GLuint values
+ *                  (ignored for GLfloat).
+ */
 void
 _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
-                         GLenum dstType, GLvoid *dest, GLfloat depthScale,
+                         GLenum dstType, GLvoid *dest, GLuint depthMax,
                          GLenum srcType, const GLvoid *source,
                          const struct gl_pixelstore_attrib *srcPacking )
 {
    GLfloat depthTemp[MAX_WIDTH], *depthValues;
+   GLboolean needClamp = GL_FALSE;
 
    /* Look for special cases first.
     * Not only are these faster, they're less prone to numeric conversion
@@ -3906,7 +3917,9 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
          }
          return;
       }
-      if (srcType == GL_UNSIGNED_SHORT && dstType == GL_UNSIGNED_INT) {
+      if (srcType == GL_UNSIGNED_SHORT
+          && dstType == GL_UNSIGNED_INT
+          && depthMax == 0xffffffff) {
          const GLushort *src = (const GLushort *) source;
          GLuint *dst = (GLuint *) dest;
          GLuint i;
@@ -3918,7 +3931,7 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
       /* XXX may want to add additional cases here someday */
    }
 
-   /* general case path */
+   /* general case path follows */
 
    if (dstType == GL_FLOAT) {
       depthValues = (GLfloat *) dest;
@@ -3927,32 +3940,34 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
       depthValues = depthTemp;
    }
 
-   /* XXX we need to obey srcPacking->SwapBytes here!!! */
-   (void) srcPacking;
-
-   /* convert incoming values to GLfloat */
+   /* Convert incoming values to GLfloat.  Some conversions will require
+    * clamping, below.
+    */
    switch (srcType) {
       case GL_BYTE:
-          DEPTH_VALUES(GLbyte, BYTE_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLbyte, BYTE_TO_FLOAT);
+         needClamp = GL_TRUE;
+         break;
       case GL_UNSIGNED_BYTE:
-          DEPTH_VALUES(GLubyte, UBYTE_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLubyte, UBYTE_TO_FLOAT);
+         break;
       case GL_SHORT:
-          DEPTH_VALUES(GLshort, SHORT_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLshort, SHORT_TO_FLOAT);
+         needClamp = GL_TRUE;
+         break;
       case GL_UNSIGNED_SHORT:
-          DEPTH_VALUES(GLushort, USHORT_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLushort, USHORT_TO_FLOAT);
+         break;
       case GL_INT:
-          DEPTH_VALUES(GLint, INT_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLint, INT_TO_FLOAT);
+         needClamp = GL_TRUE;
+         break;
       case GL_UNSIGNED_INT:
-          DEPTH_VALUES(GLuint, UINT_TO_FLOAT);
-          break;
+         DEPTH_VALUES(GLuint, UINT_TO_FLOAT);
+         break;
       case GL_UNSIGNED_INT_24_8_EXT: /* GL_EXT_packed_depth_stencil */
          if (dstType == GL_UNSIGNED_INT &&
-             depthScale == (GLfloat) 0xffffff &&
+             depthMax == 0xffffff &&
              ctx->Pixel.DepthScale == 1.0 &&
              ctx->Pixel.DepthBias == 0.0) {
             const GLuint *src = (const GLuint *) source;
@@ -3981,19 +3996,21 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
          }
          break;
       case GL_FLOAT:
-          DEPTH_VALUES(GLfloat, 1*);
-          break;
+         DEPTH_VALUES(GLfloat, 1*);
+         needClamp = GL_TRUE;
+         break;
       case GL_HALF_FLOAT_ARB:
          {
             GLuint i;
             const GLhalfARB *src = (const GLhalfARB *) source;
             for (i = 0; i < n; i++) {
-                GLhalfARB value = src[i];
-                if (srcPacking->SwapBytes) {
-                    SWAP2BYTE(value);
-                }
+               GLhalfARB value = src[i];
+               if (srcPacking->SwapBytes) {
+                  SWAP2BYTE(value);
+               }
                depthValues[i] = _mesa_half_to_float(value);
             }
+            needClamp = GL_TRUE;
          }
          break;
       default:
@@ -4001,25 +4018,43 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
          return;
    }
 
+   /* apply depth scale and bias */
+   {
+      const GLfloat scale = ctx->Pixel.DepthScale;
+      const GLfloat bias = ctx->Pixel.DepthBias;
+      if (scale != 1.0 || bias != 0.0) {
+         GLuint i;
+         for (i = 0; i < n; i++) {
+            depthValues[i] = depthValues[i] * scale + bias;
+         }
+         needClamp = GL_TRUE;
+      }
+   }
 
-   /* apply depth scale and bias and clamp to [0,1] */
-   if (ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0) {
-      _mesa_scale_and_bias_depth(ctx, n, depthValues);
+   /* clamp to [0, 1] */
+   if (needClamp) {
+      GLuint i;
+      for (i = 0; i < n; i++) {
+         depthValues[i] = CLAMP(depthValues[i], 0.0, 1.0);
+      }
    }
 
+   /*
+    * Convert values to dstType
+    */
    if (dstType == GL_UNSIGNED_INT) {
       GLuint *zValues = (GLuint *) dest;
       GLuint i;
-      if (depthScale <= (GLfloat) 0xffffff) {
+      if (depthMax <= 0xffffff) {
          /* no overflow worries */
          for (i = 0; i < n; i++) {
-            zValues[i] = (GLuint) (depthValues[i] * depthScale);
+            zValues[i] = (GLuint) (depthValues[i] * (GLfloat) depthMax);
          }
       }
       else {
          /* need to use double precision to prevent overflow problems */
          for (i = 0; i < n; i++) {
-            GLdouble z = depthValues[i] * depthScale;
+            GLdouble z = depthValues[i] * (GLfloat) depthMax;
             if (z >= (GLdouble) 0xffffffff)
                zValues[i] = 0xffffffff;
             else
@@ -4030,14 +4065,14 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
    else if (dstType == GL_UNSIGNED_SHORT) {
       GLushort *zValues = (GLushort *) dest;
       GLuint i;
-      ASSERT(depthScale <= 65535.0);
+      ASSERT(depthMax <= 0xffff);
       for (i = 0; i < n; i++) {
-         zValues[i] = (GLushort) (depthValues[i] * depthScale);
+         zValues[i] = (GLushort) (depthValues[i] * (GLfloat) depthMax);
       }
    }
    else {
       ASSERT(dstType == GL_FLOAT);
-      ASSERT(depthScale == 1.0F);
+      /*ASSERT(depthMax == 1.0F);*/
    }
 }
 
diff --git a/src/mesa/main/image.h b/src/mesa/main/image.h
index 990398a7c4..2a16989fa7 100644
--- a/src/mesa/main/image.h
+++ b/src/mesa/main/image.h
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.2
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -181,7 +181,7 @@ _mesa_pack_stencil_span( const GLcontext *ctx, GLuint n,
 
 extern void
 _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n,
-                         GLenum dstType, GLvoid *dest, GLfloat depthScale,
+                         GLenum dstType, GLvoid *dest, GLuint depthMax,
                          GLenum srcType, const GLvoid *source,
                          const struct gl_pixelstore_attrib *srcPacking );
 
diff --git a/src/mesa/main/imports.c b/src/mesa/main/imports.c
index ca5f079f72..3ae56c8b0b 100644
--- a/src/mesa/main/imports.c
+++ b/src/mesa/main/imports.c
@@ -20,7 +20,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  7.0
+ * Version:  7.1
  *
  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
@@ -254,7 +254,7 @@ _mesa_memset16( unsigned short *dst, unsigned short val, size_t n )
       *dst++ = val;
 }
 
-/** Wrapper around either memcpy() or bzero() */
+/** Wrapper around either memset() or bzero() */
 void
 _mesa_bzero( void *dst, size_t n )
 {
@@ -575,7 +575,11 @@ _mesa_ffs(int i)
  *          if no bits set.
  */
 int
+#ifdef __MINGW32__
+_mesa_ffsll(long val)
+#else
 _mesa_ffsll(long long val)
+#endif
 {
 #ifdef ffsll
    return ffsll(val);
diff --git a/src/mesa/main/imports.h b/src/mesa/main/imports.h
index 9be8014a13..ebdfc452a7 100644
--- a/src/mesa/main/imports.h
+++ b/src/mesa/main/imports.h
@@ -700,7 +700,11 @@ extern int
 _mesa_ffs(int i);
 
 extern int
+#ifdef __MINGW32__
+_mesa_ffsll(long i);
+#else
 _mesa_ffsll(long long i);
+#endif
 
 extern unsigned int
 _mesa_bitcount(unsigned int n);
diff --git a/src/mesa/main/lines.c b/src/mesa/main/lines.c
index dc7195d4eb..0c2dbf915a 100644
--- a/src/mesa/main/lines.c
+++ b/src/mesa/main/lines.c
@@ -55,9 +55,6 @@ _mesa_LineWidth( GLfloat width )
 
    FLUSH_VERTICES(ctx, _NEW_LINE);
    ctx->Line.Width = width;
-   ctx->Line._Width = CLAMP(width,
-			    ctx->Const.MinLineWidth,
-			    ctx->Const.MaxLineWidth);
 
    if (ctx->Driver.LineWidth)
       ctx->Driver.LineWidth(ctx, width);
@@ -105,13 +102,12 @@ _mesa_LineStipple( GLint factor, GLushort pattern )
  * Initializes __GLcontextRec::Line and line related constants in
  * __GLcontextRec::Const.
  */
-void GLAPIENTRY _mesa_init_line( GLcontext * ctx )
+void GLAPIENTRY
+_mesa_init_line( GLcontext * ctx )
 {
-   /* Line group */
    ctx->Line.SmoothFlag = GL_FALSE;
    ctx->Line.StippleFlag = GL_FALSE;
    ctx->Line.Width = 1.0;
-   ctx->Line._Width = 1.0;
    ctx->Line.StipplePattern = 0xffff;
    ctx->Line.StippleFactor = 1;
 }
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 49332501d2..3e4c49249a 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -38,8 +38,8 @@
 #include "glheader.h"
 #include <GL/internal/glcore.h>	/* __GLcontextModes (GLvisual) */
 #include "config.h"		/* Hardwired parameters */
-#include "glapitable.h"
-#include "glthread.h"
+#include "glapi/glapitable.h"
+#include "glapi/glthread.h"
 #include "math/m_matrix.h"	/* GLmatrix */
 #include "bitset.h"
 
@@ -917,7 +917,6 @@ struct gl_line_attrib
    GLushort StipplePattern;	/**< Stipple pattern */
    GLint StippleFactor;		/**< Stipple repeat factor */
    GLfloat Width;		/**< Line width */
-   GLfloat _Width;		/**< Clamped Line width */
 };
 
 
@@ -1063,7 +1062,6 @@ struct gl_point_attrib
 {
    GLboolean SmoothFlag;	/**< True if GL_POINT_SMOOTH is enabled */
    GLfloat Size;		/**< User-specified point size */
-   GLfloat _Size;		/**< Size clamped to Const.Min/MaxPointSize */
    GLfloat Params[3];		/**< GL_EXT_point_parameters */
    GLfloat MinSize, MaxSize;	/**< GL_EXT_point_parameters */
    GLfloat Threshold;		/**< GL_EXT_point_parameters */
@@ -1133,13 +1131,13 @@ struct gl_stencil_attrib
  * An index for each type of texture object
  */
 /*@{*/
-#define TEXTURE_1D_INDEX    0
-#define TEXTURE_2D_INDEX    1
-#define TEXTURE_3D_INDEX    2
-#define TEXTURE_CUBE_INDEX  3
-#define TEXTURE_RECT_INDEX  4
-#define TEXTURE_1D_ARRAY_INDEX    5
-#define TEXTURE_2D_ARRAY_INDEX    6
+#define TEXTURE_1D_INDEX       0
+#define TEXTURE_2D_INDEX       1
+#define TEXTURE_3D_INDEX       2
+#define TEXTURE_CUBE_INDEX     3
+#define TEXTURE_RECT_INDEX     4
+#define TEXTURE_1D_ARRAY_INDEX 5
+#define TEXTURE_2D_ARRAY_INDEX 6
 /*@}*/
 
 /**
@@ -1147,13 +1145,13 @@ struct gl_stencil_attrib
  * Used for Texture.Unit[]._ReallyEnabled flags.
  */
 /*@{*/
-#define TEXTURE_1D_BIT   (1 << TEXTURE_1D_INDEX)
-#define TEXTURE_2D_BIT   (1 << TEXTURE_2D_INDEX)
-#define TEXTURE_3D_BIT   (1 << TEXTURE_3D_INDEX)
-#define TEXTURE_CUBE_BIT (1 << TEXTURE_CUBE_INDEX)
-#define TEXTURE_RECT_BIT (1 << TEXTURE_RECT_INDEX)
-#define TEXTURE_1D_ARRAY_BIT   (1 << TEXTURE_1D_ARRAY_INDEX)
-#define TEXTURE_2D_ARRAY_BIT   (1 << TEXTURE_2D_ARRAY_INDEX)
+#define TEXTURE_1D_BIT       (1 << TEXTURE_1D_INDEX)
+#define TEXTURE_2D_BIT       (1 << TEXTURE_2D_INDEX)
+#define TEXTURE_3D_BIT       (1 << TEXTURE_3D_INDEX)
+#define TEXTURE_CUBE_BIT     (1 << TEXTURE_CUBE_INDEX)
+#define TEXTURE_RECT_BIT     (1 << TEXTURE_RECT_INDEX)
+#define TEXTURE_1D_ARRAY_BIT (1 << TEXTURE_1D_ARRAY_INDEX)
+#define TEXTURE_2D_ARRAY_BIT (1 << TEXTURE_2D_ARRAY_INDEX)
 /*@}*/
 
 
@@ -1327,8 +1325,6 @@ struct gl_texture_format
 };
 
 
-#define MAX_3D_TEXTURE_SIZE (1 << (MAX_3D_TEXTURE_LEVELS - 1))
-
 /**
  * Texture image state.  Describes the dimensions of a texture image,
  * the texel format and pointers to Texel Fetch functions.
@@ -1393,7 +1389,7 @@ struct gl_texture_image
 #define FACE_NEG_Y   3
 #define FACE_POS_Z   4
 #define FACE_NEG_Z   5
-#define MAX_FACES  6
+#define MAX_FACES    6
 /*@}*/
 
 
@@ -1426,7 +1422,7 @@ struct gl_texture_object
    GLfloat ShadowAmbient;       /**< GL_ARB_shadow_ambient */
    GLenum CompareMode;		/**< GL_ARB_shadow */
    GLenum CompareFunc;		/**< GL_ARB_shadow */
-   GLenum _Function;		/**< Comparison function derrived from 
+   GLenum _Function;		/**< Comparison function derived from 
 				 * \c CompareOperator, \c CompareMode, and
 				 * \c CompareFunc.
 				 */
@@ -1434,7 +1430,7 @@ struct gl_texture_object
    GLint _MaxLevel;		/**< actual max mipmap level (q in the spec) */
    GLfloat _MaxLambda;		/**< = _MaxLevel - BaseLevel (q - b in spec) */
    GLboolean GenerateMipmap;    /**< GL_SGIS_generate_mipmap */
-   GLboolean Complete;		/**< Is texture object complete? */
+   GLboolean _Complete;		/**< Is texture object complete? */
 
    /** Actual texture images, indexed by [cube face] and [mipmap level] */
    struct gl_texture_image *Image[MAX_FACES][MAX_TEXTURE_LEVELS];
@@ -2193,12 +2189,11 @@ struct gl_shared_state
     * \todo Improve the granularity of locking.
     */
    /*@{*/
-   _glthread_Mutex TexMutex;		   /**< texobj thread safety */
-   GLuint TextureStateStamp;	           /**< state notification for shared tex  */
+   _glthread_Mutex TexMutex;		/**< texobj thread safety */
+   GLuint TextureStateStamp;	        /**< state notification for shared tex */
    /*@}*/
 
 
-
    /**
     * \name Vertex/fragment programs
     */
diff --git a/src/mesa/main/points.c b/src/mesa/main/points.c
index 8674c7299c..e83db5de78 100644
--- a/src/mesa/main/points.c
+++ b/src/mesa/main/points.c
@@ -57,9 +57,6 @@ _mesa_PointSize( GLfloat size )
 
    FLUSH_VERTICES(ctx, _NEW_POINT);
    ctx->Point.Size = size;
-   ctx->Point._Size = CLAMP(ctx->Point.Size,
-			    ctx->Point.MinSize,
-			    ctx->Point.MaxSize);
 
    if (ctx->Driver.PointSize)
       ctx->Driver.PointSize(ctx, size);
@@ -252,7 +249,6 @@ _mesa_init_point(GLcontext *ctx)
 
    ctx->Point.SmoothFlag = GL_FALSE;
    ctx->Point.Size = 1.0;
-   ctx->Point._Size = 1.0;
    ctx->Point.Params[0] = 1.0;
    ctx->Point.Params[1] = 0.0;
    ctx->Point.Params[2] = 0.0;
diff --git a/src/mesa/main/shaders.c b/src/mesa/main/shaders.c
index 58be1f46e5..7bf8808767 100644
--- a/src/mesa/main/shaders.c
+++ b/src/mesa/main/shaders.c
@@ -83,7 +83,7 @@ _mesa_CreateShader(GLenum type)
 }
 
 
-GLhandleARB APIENTRY
+GLhandleARB GLAPIENTRY
 _mesa_CreateShaderObjectARB(GLenum type)
 {
    GET_CURRENT_CONTEXT(ctx);
@@ -99,7 +99,7 @@ _mesa_CreateProgram(void)
 }
 
 
-GLhandleARB APIENTRY
+GLhandleARB GLAPIENTRY
 _mesa_CreateProgramObjectARB(void)
 {
    GET_CURRENT_CONTEXT(ctx);
@@ -319,7 +319,7 @@ _mesa_GetUniformivARB(GLhandleARB program, GLint location, GLint * params)
 
 
 #if 0
-GLint APIENTRY
+GLint GLAPIENTRY
 _mesa_GetUniformLocation(GLuint program, const GLcharARB *name)
 {
    GET_CURRENT_CONTEXT(ctx);
@@ -336,7 +336,7 @@ _mesa_GetHandleARB(GLenum pname)
 }
 
 
-GLint APIENTRY
+GLint GLAPIENTRY
 _mesa_GetUniformLocationARB(GLhandleARB programObj, const GLcharARB *name)
 {
    GET_CURRENT_CONTEXT(ctx);
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 0a83abc7dd..444f227760 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -35,10 +35,10 @@
 #include "accum.h"
 #include "api_loopback.h"
 #if FEATURE_ARB_vertex_program || FEATURE_ARB_fragment_program
-#include "arbprogram.h"
+#include "shader/arbprogram.h"
 #endif
 #if FEATURE_ATI_fragment_shader
-#include "atifragshader.h"
+#include "shader/atifragshader.h"
 #endif
 #include "attrib.h"
 #include "blend.h"
@@ -85,18 +85,18 @@
 #include "mtypes.h"
 #include "varray.h"
 #if FEATURE_NV_vertex_program
-#include "nvprogram.h"
+#include "shader/nvprogram.h"
 #endif
 #if FEATURE_NV_fragment_program
-#include "nvprogram.h"
-#include "program.h"
+#include "shader/nvprogram.h"
+#include "shader/program.h"
 #include "texenvprogram.h"
 #endif
 #if FEATURE_ARB_shader_objects
 #include "shaders.h"
 #endif
 #include "debug.h"
-#include "dispatch.h"
+#include "glapi/dispatch.h"
 
 
 
@@ -1068,7 +1068,7 @@ update_tricaps(GLcontext *ctx, GLbitfield new_state)
    if (1/*new_state & _NEW_POINT*/) {
       if (ctx->Point.SmoothFlag)
          ctx->_TriangleCaps |= DD_POINT_SMOOTH;
-      if (ctx->Point._Size != 1.0F)
+      if (ctx->Point.Size != 1.0F)
          ctx->_TriangleCaps |= DD_POINT_SIZE;
       if (ctx->Point._Attenuated)
          ctx->_TriangleCaps |= DD_POINT_ATTEN;
@@ -1082,7 +1082,7 @@ update_tricaps(GLcontext *ctx, GLbitfield new_state)
          ctx->_TriangleCaps |= DD_LINE_SMOOTH;
       if (ctx->Line.StippleFlag)
          ctx->_TriangleCaps |= DD_LINE_STIPPLE;
-      if (ctx->Line._Width != 1.0)
+      if (ctx->Line.Width != 1.0)
          ctx->_TriangleCaps |= DD_LINE_WIDTH;
    }
 
diff --git a/src/mesa/main/texcompress_fxt1.c b/src/mesa/main/texcompress_fxt1.c
index 411d51cfcc..b6991f45ed 100644
--- a/src/mesa/main/texcompress_fxt1.c
+++ b/src/mesa/main/texcompress_fxt1.c
@@ -302,7 +302,12 @@ const struct gl_texture_format _mesa_texformat_rgba_fxt1 = {
 
 #define FX64_NATIVE 1
 
+#ifdef __MINGW32__
+typedef unsigned long Fx64;
+#else
 typedef unsigned long long Fx64;
+#endif
+
 
 #define FX64_MOV32(a, b) a = b
 #define FX64_OR32(a, b)  a |= b
diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogram.c
index 0c6fa82f11..72b54b27d9 100644
--- a/src/mesa/main/texenvprogram.c
+++ b/src/mesa/main/texenvprogram.c
@@ -28,10 +28,10 @@
 #include "glheader.h"
 #include "macros.h"
 #include "enums.h"
-#include "prog_parameter.h"
-#include "prog_instruction.h"
-#include "prog_print.h"
-#include "prog_statevars.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_print.h"
+#include "shader/prog_statevars.h"
 #include "texenvprogram.h"
 
 /**
@@ -871,7 +871,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
 			  key->unit[unit].OptRGB);
    }
    else if (key->unit[unit].ModeRGB == MODE_DOT3_RGBA_EXT ||
-	    key->unit[unit].ModeA == MODE_DOT3_RGBA) {
+	    key->unit[unit].ModeRGB == MODE_DOT3_RGBA) {
 
       out = emit_combine( p, dest, WRITEMASK_XYZW, saturate,
 			  unit,
diff --git a/src/mesa/main/teximage.c b/src/mesa/main/teximage.c
index 1f4c9f4722..3420d8e2ba 100644
--- a/src/mesa/main/teximage.c
+++ b/src/mesa/main/teximage.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.1
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -1264,6 +1264,10 @@ _mesa_init_teximage_fields(GLcontext *ctx, GLenum target,
  * A hardware driver might override this function if, for example, the
  * max 3D texture size is 512x512x64 (i.e. not a cube).
  *
+ * Note that width, height, depth == 0 is not an error.  However, a
+ * texture with zero width/height/depth will be considered "incomplete"
+ * and texturing will effectively be disabled.
+ *
  * \param target  one of GL_PROXY_TEXTURE_1D, GL_PROXY_TEXTURE_2D,
  *                GL_PROXY_TEXTURE_3D, GL_PROXY_TEXTURE_RECTANGLE_NV,
  *                GL_PROXY_TEXTURE_CUBE_MAP_ARB.
@@ -1293,7 +1297,7 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.MaxTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width >0 && _mesa_bitcount(width - 2 * border) != 1) ||
           level >= ctx->Const.MaxTextureLevels) {
          /* bad width or level */
          return GL_FALSE;
@@ -1303,10 +1307,10 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.MaxTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width > 0 && _mesa_bitcount(width - 2 * border) != 1) ||
           height < 2 * border || height > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(height - 2 * border) != 1) ||
+           height > 0 && _mesa_bitcount(height - 2 * border) != 1) ||
           level >= ctx->Const.MaxTextureLevels) {
          /* bad width or height or level */
          return GL_FALSE;
@@ -1316,21 +1320,21 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.Max3DTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width > 0 && _mesa_bitcount(width - 2 * border) != 1) ||
           height < 2 * border || height > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(height - 2 * border) != 1) ||
+           height > 0 && _mesa_bitcount(height - 2 * border) != 1) ||
           depth < 2 * border || depth > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(depth - 2 * border) != 1) ||
+           depth > 0 && _mesa_bitcount(depth - 2 * border) != 1) ||
           level >= ctx->Const.Max3DTextureLevels) {
          /* bad width or height or depth or level */
          return GL_FALSE;
       }
       return GL_TRUE;
    case GL_PROXY_TEXTURE_RECTANGLE_NV:
-      if (width < 1 || width > ctx->Const.MaxTextureRectSize ||
-          height < 1 || height > ctx->Const.MaxTextureRectSize ||
+      if (width < 0 || width > ctx->Const.MaxTextureRectSize ||
+          height < 0 || height > ctx->Const.MaxTextureRectSize ||
           level != 0) {
          /* bad width or height or level */
          return GL_FALSE;
@@ -1340,10 +1344,10 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.MaxCubeTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width > 0 && _mesa_bitcount(width - 2 * border) != 1) ||
           height < 2 * border || height > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(height - 2 * border) != 1) ||
+           height > 0 && _mesa_bitcount(height - 2 * border) != 1) ||
           level >= ctx->Const.MaxCubeTextureLevels) {
          /* bad width or height */
          return GL_FALSE;
@@ -1353,7 +1357,7 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.MaxTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width > 0 && _mesa_bitcount(width - 2 * border) != 1) ||
           level >= ctx->Const.MaxTextureLevels) {
          /* bad width or level */
          return GL_FALSE;
@@ -1367,10 +1371,10 @@ _mesa_test_proxy_teximage(GLcontext *ctx, GLenum target, GLint level,
       maxSize = 1 << (ctx->Const.MaxTextureLevels - 1);
       if (width < 2 * border || width > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(width - 2 * border) != 1) ||
+           width > 0 && _mesa_bitcount(width - 2 * border) != 1) ||
           height < 2 * border || height > 2 + maxSize ||
           (!ctx->Extensions.ARB_texture_non_power_of_two &&
-           _mesa_bitcount(height - 2 * border) != 1) ||
+           height > 0 && _mesa_bitcount(height - 2 * border) != 1) ||
           level >= ctx->Const.MaxTextureLevels) {
          /* bad width or height or level */
          return GL_FALSE;
@@ -1472,7 +1476,7 @@ texture_error_check( GLcontext *ctx, GLenum target,
       if (target == GL_PROXY_TEXTURE_1D || target == GL_TEXTURE_1D) {
          proxy_target = GL_PROXY_TEXTURE_1D;
          height = 1;
-         width = 1;
+         depth = 1;
       }
       else {
          _mesa_error( ctx, GL_INVALID_ENUM, "glTexImage1D(target)" );
@@ -2463,7 +2467,7 @@ _mesa_TexImage1D( GLenum target, GLint level, GLint internalFormat,
 	 update_fbo_texture(ctx, texObj, face, level);
 	 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -2566,7 +2570,7 @@ _mesa_TexImage2D( GLenum target, GLint level, GLint internalFormat,
 	 update_fbo_texture(ctx, texObj, face, level);
 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -2667,7 +2671,7 @@ _mesa_TexImage3D( GLenum target, GLint level, GLint internalFormat,
 	 update_fbo_texture(ctx, texObj, face, level);
 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -2938,7 +2942,7 @@ _mesa_CopyTexImage1D( GLenum target, GLint level,
       update_fbo_texture(ctx, texObj, face, level);
 
       /* state update */
-      texObj->Complete = GL_FALSE;
+      texObj->_Complete = GL_FALSE;
       ctx->NewState |= _NEW_TEXTURE;
    }
  out:
@@ -3004,7 +3008,7 @@ _mesa_CopyTexImage2D( GLenum target, GLint level, GLenum internalFormat,
       update_fbo_texture(ctx, texObj, face, level);
 
       /* state update */
-      texObj->Complete = GL_FALSE;
+      texObj->_Complete = GL_FALSE;
       ctx->NewState |= _NEW_TEXTURE;
    }
  out:
@@ -3398,7 +3402,7 @@ _mesa_CompressedTexImage1DARB(GLenum target, GLint level,
 					     texObj, texImage);
 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -3495,7 +3499,7 @@ _mesa_CompressedTexImage2DARB(GLenum target, GLint level,
 					     texObj, texImage);
 	 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -3591,7 +3595,7 @@ _mesa_CompressedTexImage3DARB(GLenum target, GLint level,
 					     texObj, texImage);
 	 
 	 /* state update */
-	 texObj->Complete = GL_FALSE;
+	 texObj->_Complete = GL_FALSE;
 	 ctx->NewState |= _NEW_TEXTURE;
       }
    out:
@@ -3768,7 +3772,7 @@ _mesa_CompressedTexSubImage3DARB(GLenum target, GLint level, GLint xoffset,
                                              width, height, depth, /*size*/
                                              format, imageSize);
    if (error) {
-      _mesa_error(ctx, error, "glCompressedTexSubImage2D");
+      _mesa_error(ctx, error, "glCompressedTexSubImage3D");
       return;
    }
 
@@ -3819,7 +3823,6 @@ _mesa_GetCompressedTexImageARB(GLenum target, GLint level, GLvoid *img)
    GET_CURRENT_CONTEXT(ctx);
    ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx);
 
-
    texUnit = &ctx->Texture.Unit[ctx->Texture.CurrentUnit];
    texObj = _mesa_select_tex_object(ctx, texUnit, target);
    if (!texObj) {
@@ -3840,24 +3843,25 @@ _mesa_GetCompressedTexImageARB(GLenum target, GLint level, GLvoid *img)
       return;
    }
 
-
    _mesa_lock_texture(ctx, texObj);
    {
       texImage = _mesa_select_tex_image(ctx, texObj, target, level);
-      if (!texImage) {
-	 /* probably invalid mipmap level */
-	 _mesa_error(ctx, GL_INVALID_VALUE, "glGetCompressedTexImageARB(level)");
-	 goto out;
+      if (texImage) {
+         if (texImage->IsCompressed) {
+            /* this typically calls _mesa_get_compressed_teximage() */
+            ctx->Driver.GetCompressedTexImage(ctx, target, level, img,
+                                              texObj, texImage);
+         }
+         else {
+            _mesa_error(ctx, GL_INVALID_OPERATION,
+                        "glGetCompressedTexImageARB");
+         }
       }
-
-      if (!texImage->IsCompressed) {
-	 _mesa_error(ctx, GL_INVALID_OPERATION, "glGetCompressedTexImageARB");
-	 goto out;
+      else {
+         /* probably invalid mipmap level */
+         _mesa_error(ctx, GL_INVALID_VALUE,
+                     "glGetCompressedTexImageARB(level)");
       }
-
-      /* this typically calls _mesa_get_compressed_teximage() */
-      ctx->Driver.GetCompressedTexImage(ctx, target, level, img, texObj,texImage);
    }
- out:
    _mesa_unlock_texture(ctx, texObj);
 }
diff --git a/src/mesa/main/texobj.c b/src/mesa/main/texobj.c
index 0d2946e95a..df64002f99 100644
--- a/src/mesa/main/texobj.c
+++ b/src/mesa/main/texobj.c
@@ -209,7 +209,7 @@ _mesa_copy_texture_object( struct gl_texture_object *dest,
    dest->_MaxLambda = src->_MaxLambda;
    dest->GenerateMipmap = src->GenerateMipmap;
    dest->Palette = src->Palette;
-   dest->Complete = src->Complete;
+   dest->_Complete = src->_Complete;
 }
 
 
@@ -251,7 +251,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
    const GLint baseLevel = t->BaseLevel;
    GLint maxLog2 = 0, maxLevels = 0;
 
-   t->Complete = GL_TRUE;  /* be optimistic */
+   t->_Complete = GL_TRUE;  /* be optimistic */
 
    /* Always need the base level image */
    if (!t->Image[0][baseLevel]) {
@@ -259,7 +259,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
       _mesa_sprintf(s, "obj %p (%d) Image[baseLevel=%d] == NULL",
               (void *) t, t->Name, baseLevel);
       incomplete(t, s);
-      t->Complete = GL_FALSE;
+      t->_Complete = GL_FALSE;
       return;
    }
 
@@ -268,7 +268,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
        t->Image[0][baseLevel]->Height == 0 ||
        t->Image[0][baseLevel]->Depth == 0) {
       incomplete(t, "texture width = 0");
-      t->Complete = GL_FALSE;
+      t->_Complete = GL_FALSE;
       return;
    }
 
@@ -322,7 +322,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
 	 if (t->Image[face][baseLevel] == NULL ||
 	     t->Image[face][baseLevel]->Width2 != w ||
 	     t->Image[face][baseLevel]->Height2 != h) {
-	    t->Complete = GL_FALSE;
+	    t->_Complete = GL_FALSE;
 	    incomplete(t, "Non-quare cubemap image");
 	    return;
 	 }
@@ -339,7 +339,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
       GLint maxLevel = t->_MaxLevel;
 
       if (minLevel > maxLevel) {
-         t->Complete = GL_FALSE;
+         t->_Complete = GL_FALSE;
          incomplete(t, "minLevel > maxLevel");
          return;
       }
@@ -348,12 +348,12 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
       for (i = minLevel; i <= maxLevel; i++) {
          if (t->Image[0][i]) {
             if (t->Image[0][i]->TexFormat != t->Image[0][baseLevel]->TexFormat) {
-               t->Complete = GL_FALSE;
+               t->_Complete = GL_FALSE;
                incomplete(t, "Format[i] != Format[baseLevel]");
                return;
             }
             if (t->Image[0][i]->Border != t->Image[0][baseLevel]->Border) {
-               t->Complete = GL_FALSE;
+               t->_Complete = GL_FALSE;
                incomplete(t, "Border[i] != Border[baseLevel]");
                return;
             }
@@ -371,12 +371,12 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             }
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[0][i]) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "1D Image[0][i] == NULL");
                   return;
                }
                if (t->Image[0][i]->Width2 != width ) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "1D Image[0][i] bad width");
                   return;
                }
@@ -400,17 +400,17 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             }
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[0][i]) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "2D Image[0][i] == NULL");
                   return;
                }
                if (t->Image[0][i]->Width2 != width) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "2D Image[0][i] bad width");
                   return;
                }
                if (t->Image[0][i]->Height2 != height) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "2D Image[0][i] bad height");
                   return;
                }
@@ -438,26 +438,26 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[0][i]) {
                   incomplete(t, "3D Image[0][i] == NULL");
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   return;
                }
                if (t->Image[0][i]->_BaseFormat == GL_DEPTH_COMPONENT) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "GL_DEPTH_COMPONENT only works with 1/2D tex");
                   return;
                }
                if (t->Image[0][i]->Width2 != width) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "3D Image[0][i] bad width");
                   return;
                }
                if (t->Image[0][i]->Height2 != height) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "3D Image[0][i] bad height");
                   return;
                }
                if (t->Image[0][i]->Depth2 != depth) {
-                  t->Complete = GL_FALSE;
+                  t->_Complete = GL_FALSE;
                   incomplete(t, "3D Image[0][i] bad depth");
                   return;
                }
@@ -483,20 +483,20 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
 	       for (face = 0; face < 6; face++) {
 		  /* check that we have images defined */
 		  if (!t->Image[face][i]) {
-		     t->Complete = GL_FALSE;
+		     t->_Complete = GL_FALSE;
 		     incomplete(t, "CubeMap Image[n][i] == NULL");
 		     return;
 		  }
 		  /* Don't support GL_DEPTH_COMPONENT for cube maps */
 		  if (t->Image[face][i]->_BaseFormat == GL_DEPTH_COMPONENT) {
-		     t->Complete = GL_FALSE;
+		     t->_Complete = GL_FALSE;
 		     incomplete(t, "GL_DEPTH_COMPONENT only works with 1/2D tex");
 		     return;
 		  }
 		  /* check that all six images have same size */
 		  if (t->Image[face][i]->Width2!=width || 
 		      t->Image[face][i]->Height2!=height) {
-		     t->Complete = GL_FALSE;
+		     t->_Complete = GL_FALSE;
 		     incomplete(t, "CubeMap Image[n][i] bad size");
 		     return;
 		  }
diff --git a/src/mesa/main/texstate.c b/src/mesa/main/texstate.c
index 18afaf4edd..c9f8a0656e 100644
--- a/src/mesa/main/texstate.c
+++ b/src/mesa/main/texstate.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.3
+ * Version:  7.1
  *
  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
@@ -1141,26 +1141,29 @@ _mesa_GetTexEnviv( GLenum target, GLenum pname, GLint *params )
 /*                       Texture Parameters                           */
 /**********************************************************************/
 
+/**
+ * Check if a coordinate wrap mode is supported for the texture target.
+ * \return GL_TRUE if legal, GL_FALSE otherwise
+ */
 static GLboolean 
-_mesa_validate_texture_wrap_mode(GLcontext * ctx,
-				 GLenum target, GLenum eparam)
+validate_texture_wrap_mode(GLcontext * ctx, GLenum target, GLenum wrap)
 {
    const struct gl_extensions * const e = & ctx->Extensions;
 
-   if (eparam == GL_CLAMP || eparam == GL_CLAMP_TO_EDGE ||
-       (eparam == GL_CLAMP_TO_BORDER && e->ARB_texture_border_clamp)) {
+   if (wrap == GL_CLAMP || wrap == GL_CLAMP_TO_EDGE ||
+       (wrap == GL_CLAMP_TO_BORDER && e->ARB_texture_border_clamp)) {
       /* any texture target */
       return GL_TRUE;
    }
    else if (target != GL_TEXTURE_RECTANGLE_NV &&
-	    (eparam == GL_REPEAT ||
-	     (eparam == GL_MIRRORED_REPEAT &&
+	    (wrap == GL_REPEAT ||
+	     (wrap == GL_MIRRORED_REPEAT &&
 	      e->ARB_texture_mirrored_repeat) ||
-	     (eparam == GL_MIRROR_CLAMP_EXT &&
+	     (wrap == GL_MIRROR_CLAMP_EXT &&
 	      (e->ATI_texture_mirror_once || e->EXT_texture_mirror_clamp)) ||
-	     (eparam == GL_MIRROR_CLAMP_TO_EDGE_EXT &&
+	     (wrap == GL_MIRROR_CLAMP_TO_EDGE_EXT &&
 	      (e->ATI_texture_mirror_once || e->EXT_texture_mirror_clamp)) ||
-	     (eparam == GL_MIRROR_CLAMP_TO_BORDER_EXT &&
+	     (wrap == GL_MIRROR_CLAMP_TO_BORDER_EXT &&
 	      (e->EXT_texture_mirror_clamp)))) {
       /* non-rectangle texture */
       return GL_TRUE;
@@ -1178,36 +1181,6 @@ _mesa_TexParameterf( GLenum target, GLenum pname, GLfloat param )
 }
 
 
-/**
- * Update derrived compare function state.
- */
-void
-_mesa_update_texture_compare_function(struct gl_texture_object *tObj,
-				      GLboolean in_frag_prog)
-{
-   if (in_frag_prog) {
-      tObj->_Function = GL_NONE;
-   }
-   else if (tObj->CompareFlag) {
-      /* GL_SGIX_shadow */
-      if (tObj->CompareOperator == GL_TEXTURE_LEQUAL_R_SGIX) {
-         tObj->_Function = GL_LEQUAL;
-      }
-      else {
-         ASSERT(tObj->CompareOperator == GL_TEXTURE_GEQUAL_R_SGIX);
-         tObj->_Function = GL_GEQUAL;
-      }
-   }
-   else if (tObj->CompareMode == GL_COMPARE_R_TO_TEXTURE_ARB) {
-      /* GL_ARB_shadow */
-      tObj->_Function = tObj->CompareFunc;
-   }
-   else {
-      tObj->_Function = GL_NONE;  /* pass depth through as grayscale */
-   }
-}
-
-
 void GLAPIENTRY
 _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
 {
@@ -1313,7 +1286,7 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
       case GL_TEXTURE_WRAP_S:
          if (texObj->WrapS == eparam)
             return;
-         if (_mesa_validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
+         if (validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
             FLUSH_VERTICES(ctx, _NEW_TEXTURE);
             texObj->WrapS = eparam;
          }
@@ -1324,7 +1297,7 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
       case GL_TEXTURE_WRAP_T:
          if (texObj->WrapT == eparam)
             return;
-         if (_mesa_validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
+         if (validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
             FLUSH_VERTICES(ctx, _NEW_TEXTURE);
             texObj->WrapT = eparam;
          }
@@ -1335,7 +1308,7 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
       case GL_TEXTURE_WRAP_R:
          if (texObj->WrapR == eparam)
             return;
-         if (_mesa_validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
+         if (validate_texture_wrap_mode(ctx, texObj->Target, eparam)) {
             FLUSH_VERTICES(ctx, _NEW_TEXTURE);
             texObj->WrapR = eparam;
          }
@@ -1415,7 +1388,6 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
          if (ctx->Extensions.SGIX_shadow) {
             FLUSH_VERTICES(ctx, _NEW_TEXTURE);
             texObj->CompareFlag = params[0] ? GL_TRUE : GL_FALSE;
-	    _mesa_update_texture_compare_function(texObj, GL_FALSE);
          }
          else {
             _mesa_error(ctx, GL_INVALID_ENUM,
@@ -1430,7 +1402,6 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
                 op == GL_TEXTURE_GEQUAL_R_SGIX) {
                FLUSH_VERTICES(ctx, _NEW_TEXTURE);
                texObj->CompareOperator = op;
-	       _mesa_update_texture_compare_function(texObj, GL_FALSE);
             }
             else {
                _mesa_error(ctx, GL_INVALID_ENUM, "glTexParameter(param)");
@@ -1469,7 +1440,6 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
             if (mode == GL_NONE || mode == GL_COMPARE_R_TO_TEXTURE_ARB) {
                FLUSH_VERTICES(ctx, _NEW_TEXTURE);
                texObj->CompareMode = mode;
-	       _mesa_update_texture_compare_function(texObj, GL_FALSE);
             }
             else {
                _mesa_error(ctx, GL_INVALID_ENUM,
@@ -1505,8 +1475,6 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
                            "glTexParameter(bad GL_TEXTURE_COMPARE_FUNC_ARB)");
                return;
             }
-
-	    _mesa_update_texture_compare_function(texObj, GL_FALSE);
          }
          else {
             _mesa_error(ctx, GL_INVALID_ENUM,
@@ -1527,8 +1495,6 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
                           "glTexParameter(bad GL_DEPTH_TEXTURE_MODE_ARB)");
                return;
             }
-
-	    _mesa_update_texture_compare_function(texObj, GL_FALSE);
          }
          else {
             _mesa_error(ctx, GL_INVALID_ENUM,
@@ -1552,7 +1518,7 @@ _mesa_TexParameterfv( GLenum target, GLenum pname, const GLfloat *params )
          return;
    }
 
-   texObj->Complete = GL_FALSE;
+   texObj->_Complete = GL_FALSE;
 
    if (ctx->Driver.TexParameter) {
       (*ctx->Driver.TexParameter)( ctx, target, texObj, pname, params );
@@ -2830,6 +2796,47 @@ update_texture_matrices( GLcontext *ctx )
 
 
 /**
+ * Update texture object's _Function field.  We need to do this
+ * whenever any of the texture object's shadow-related fields change
+ * or when we start/stop using a fragment program.
+ *
+ * This function could be expanded someday to update additional per-object
+ * fields that depend on assorted state changes.
+ */
+static void
+update_texture_compare_function(GLcontext *ctx,
+                                struct gl_texture_object *tObj)
+{
+   /* XXX temporarily disable this test since it breaks the GLSL
+    * shadow2D(), etc. functions.
+    */
+   if (0 /*ctx->FragmentProgram._Current*/) {
+      /* Texel/coordinate comparison is ignored for programs.
+       * See GL_ARB_fragment_program/shader spec for details.
+       */
+      tObj->_Function = GL_NONE;
+   }
+   else if (tObj->CompareFlag) {
+      /* GL_SGIX_shadow */
+      if (tObj->CompareOperator == GL_TEXTURE_LEQUAL_R_SGIX) {
+         tObj->_Function = GL_LEQUAL;
+      }
+      else {
+         ASSERT(tObj->CompareOperator == GL_TEXTURE_GEQUAL_R_SGIX);
+         tObj->_Function = GL_GEQUAL;
+      }
+   }
+   else if (tObj->CompareMode == GL_COMPARE_R_TO_TEXTURE_ARB) {
+      /* GL_ARB_shadow */
+      tObj->_Function = tObj->CompareFunc;
+   }
+   else {
+      tObj->_Function = GL_NONE;  /* pass depth through as grayscale */
+   }
+}
+
+
+/**
  * Helper function for determining which texture object (1D, 2D, cube, etc)
  * should actually be used.
  */
@@ -2839,12 +2846,13 @@ texture_override(GLcontext *ctx,
                  struct gl_texture_object *texObj, GLuint textureBit)
 {
    if (!texUnit->_ReallyEnabled && (enableBits & textureBit)) {
-      if (!texObj->Complete) {
+      if (!texObj->_Complete) {
          _mesa_test_texobj_completeness(ctx, texObj);
       }
-      if (texObj->Complete) {
+      if (texObj->_Complete) {
          texUnit->_ReallyEnabled = textureBit;
          texUnit->_Current = texObj;
+         update_texture_compare_function(ctx, texObj);
       }
    }
 }
diff --git a/src/mesa/main/texstate.h b/src/mesa/main/texstate.h
index df468ecf9b..60145691b8 100644
--- a/src/mesa/main/texstate.h
+++ b/src/mesa/main/texstate.h
@@ -5,9 +5,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  6.5
+ * Version:  7.1
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -41,9 +41,6 @@ _mesa_copy_texture_state( const GLcontext *src, GLcontext *dst );
 extern void
 _mesa_print_texunit_state( GLcontext *ctx, GLuint unit );
 
-extern void
-_mesa_update_texture_compare_function(struct gl_texture_object *tObj,
-				      GLboolean in_frag_prog);
 
 
 /**
diff --git a/src/mesa/main/texstore.c b/src/mesa/main/texstore.c
index 9b8a06df14..3b5151ed17 100644
--- a/src/mesa/main/texstore.c
+++ b/src/mesa/main/texstore.c
@@ -1077,7 +1077,7 @@ _mesa_texstore_rgba(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_z32(TEXSTORE_PARAMS)
 {
-   const GLfloat depthScale = (GLfloat) 0xffffffff;
+   const GLuint depthScale = 0xffffffff;
    (void) dims;
    ASSERT(dstFormat == &_mesa_texformat_z32);
    ASSERT(dstFormat->TexelBytes == sizeof(GLuint));
@@ -1124,7 +1124,7 @@ _mesa_texstore_z32(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_z16(TEXSTORE_PARAMS)
 {
-   const GLfloat depthScale = 65535.0f;
+   const GLuint depthScale = 0xffff;
    (void) dims;
    ASSERT(dstFormat == &_mesa_texformat_z16);
    ASSERT(dstFormat->TexelBytes == sizeof(GLushort));
@@ -2319,6 +2319,8 @@ _mesa_texstore_ycbcr(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_z24_s8(TEXSTORE_PARAMS)
 {
+   const GLuint depthScale = 0xffffff;
+
    ASSERT(dstFormat == &_mesa_texformat_z24_s8);
    ASSERT(srcFormat == GL_DEPTH_STENCIL_EXT);
    ASSERT(srcType == GL_UNSIGNED_INT_24_8_EXT);
@@ -2357,7 +2359,7 @@ _mesa_texstore_z24_s8(TEXSTORE_PARAMS)
             _mesa_unpack_depth_span(ctx, srcWidth,
                                     GL_UNSIGNED_INT, /* dst type */
                                     dstRow, /* dst addr */
-                                    (GLfloat) 0xffffff, /* depthScale */
+                                    depthScale,
                                     srcType, src, srcPacking);
             /* get the 8-bit stencil values */
             _mesa_unpack_stencil_span(ctx, srcWidth,
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index bf1ad0165e..fe4a7c684f 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@@ -32,7 +32,7 @@
 #include "mtypes.h"
 #include "varray.h"
 #include "arrayobj.h"
-#include "dispatch.h"
+#include "glapi/dispatch.h"
 
 
 /**
diff --git a/src/mesa/main/vtxfmt_tmp.h b/src/mesa/main/vtxfmt_tmp.h
index 783b06558d..6f5d01e40f 100644
--- a/src/mesa/main/vtxfmt_tmp.h
+++ b/src/mesa/main/vtxfmt_tmp.h
@@ -29,8 +29,8 @@
 #define PRE_LOOPBACK( FUNC )
 #endif
 
-#include "dispatch.h"
-#include "glapioffsets.h"
+#include "glapi/dispatch.h"
+#include "glapi/glapioffsets.h"
 
 static void GLAPIENTRY TAG(ArrayElement)( GLint i )
 {
diff --git a/src/mesa/math/m_eval.c b/src/mesa/math/m_eval.c
index 42ffd4133d..d324673c5d 100644
--- a/src/mesa/math/m_eval.c
+++ b/src/mesa/math/m_eval.c
@@ -37,8 +37,8 @@
  */
 
 
-#include "glheader.h"
-#include "config.h"
+#include "main/glheader.h"
+#include "main/config.h"
 #include "m_eval.h"
 
 static GLfloat inv_tab[MAX_EVAL_ORDER];
diff --git a/src/mesa/math/m_eval.h b/src/mesa/math/m_eval.h
index a23cbd402e..d73ecaafb2 100644
--- a/src/mesa/math/m_eval.h
+++ b/src/mesa/math/m_eval.h
@@ -26,7 +26,7 @@
 #ifndef _M_EVAL_H
 #define _M_EVAL_H
 
-#include "glheader.h"
+#include "main/glheader.h"
 
 void _math_init_eval( void );
 
diff --git a/src/mesa/math/m_translate.h b/src/mesa/math/m_translate.h
index 0bcf96005c..c677682d50 100644
--- a/src/mesa/math/m_translate.h
+++ b/src/mesa/math/m_translate.h
@@ -26,8 +26,8 @@
 #ifndef _M_TRANSLATE_H_
 #define _M_TRANSLATE_H_
 
-#include "config.h"
-#include "mtypes.h"		/* hack for GLchan */
+#include "main/config.h"
+#include "main/mtypes.h"		/* hack for GLchan */
 
 
 /**
diff --git a/src/mesa/math/m_xform.c b/src/mesa/math/m_xform.c
index fa3f57a8e5..901ae5b416 100644
--- a/src/mesa/math/m_xform.c
+++ b/src/mesa/math/m_xform.c
@@ -33,8 +33,8 @@
  * 3. Transformation of a point p by a matrix M is: p' = M * p
  */
 
-#include "glheader.h"
-#include "macros.h"
+#include "main/glheader.h"
+#include "main/macros.h"
 
 #include "m_eval.h"
 #include "m_matrix.h"
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index 5d8f763741..9a5290d920 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -30,10 +30,10 @@
  * \author Karl Rasche
  */
 
-#include "glheader.h"
-#include "imports.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "shader/grammar/grammar_mesa.h"
 #include "arbprogparse.h"
-#include "grammar_mesa.h"
 #include "program.h"
 #include "prog_parameter.h"
 #include "prog_statevars.h"
@@ -3573,7 +3573,7 @@ parse_instructions(GLcontext * ctx, const GLubyte * inst,
 
 /* XXX temporary */
 LONGSTRING static char core_grammar_text[] =
-#include "grammar_syn.h"
+#include "shader/grammar/grammar_syn.h"
 ;
 
 
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index 9faf9d8613..28d195d0ee 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -43,7 +43,7 @@
 #include "prog_instruction.h"
 #include "prog_parameter.h"
 #include "prog_print.h"
-#include "slang_library_noise.h"
+#include "shader/slang/slang_library_noise.h"
 
 
 /* See comments below for info about this */
diff --git a/src/mesa/shader/program.c b/src/mesa/shader/program.c
index 4205919828..1f227390af 100644
--- a/src/mesa/shader/program.c
+++ b/src/mesa/shader/program.c
@@ -333,7 +333,7 @@ _mesa_clone_program(GLcontext *ctx, const struct gl_program *prog)
 {
    struct gl_program *clone;
 
-   clone = _mesa_new_program(ctx, prog->Target, prog->Id);
+   clone = ctx->Driver.NewProgram(ctx, prog->Target, prog->Id);
    if (!clone)
       return NULL;
 
diff --git a/src/mesa/shader/shader_api.c b/src/mesa/shader/shader_api.c
index 3a54e68d0d..1a931326af 100644
--- a/src/mesa/shader/shader_api.c
+++ b/src/mesa/shader/shader_api.c
@@ -43,10 +43,9 @@
 #include "prog_parameter.h"
 #include "prog_print.h"
 #include "prog_statevars.h"
-#include "shader_api.h"
-
-#include "slang_compile.h"
-#include "slang_link.h"
+#include "shader/shader_api.h"
+#include "shader/slang/slang_compile.h"
+#include "shader/slang/slang_link.h"
 
 
 
@@ -780,7 +779,7 @@ _mesa_get_programiv(GLcontext *ctx, GLuint program,
       *params = shProg->Validated;
       break;
    case GL_INFO_LOG_LENGTH:
-      *params = shProg->InfoLog ? strlen(shProg->InfoLog) : 0;
+      *params = shProg->InfoLog ? strlen(shProg->InfoLog) + 1 : 0;
       break;
    case GL_ATTACHED_SHADERS:
       *params = shProg->NumShaders;
@@ -832,10 +831,10 @@ _mesa_get_shaderiv(GLcontext *ctx, GLuint name, GLenum pname, GLint *params)
       *params = shader->CompileStatus;
       break;
    case GL_INFO_LOG_LENGTH:
-      *params = shader->InfoLog ? strlen(shader->InfoLog) : 0;
+      *params = shader->InfoLog ? strlen(shader->InfoLog) + 1 : 0;
       break;
    case GL_SHADER_SOURCE_LENGTH:
-      *params = shader->Source ? strlen((char *) shader->Source) : 0;
+      *params = shader->Source ? strlen((char *) shader->Source) + 1 : 0;
       break;
    default:
       _mesa_error(ctx, GL_INVALID_ENUM, "glGetShaderiv(pname)");
diff --git a/src/mesa/shader/slang/library/slang_core.gc b/src/mesa/shader/slang/library/slang_core.gc
index e398eabb15..2e7ebc347a 100644
--- a/src/mesa/shader/slang/library/slang_core.gc
+++ b/src/mesa/shader/slang/library/slang_core.gc
@@ -181,6 +181,11 @@ vec2 __constructor(const vec3 v)
    __retVal.xy = v.xy;
 }
 
+vec2 __constructor(const vec4 v)
+{
+   __retVal.st = v.xy;
+}
+
 
 //// vec3 constructors
 
diff --git a/src/mesa/shader/slang/library/slang_core_gc.h b/src/mesa/shader/slang/library/slang_core_gc.h
index dcd39a0e28..f2b4fd6464 100644
--- a/src/mesa/shader/slang/library/slang_core_gc.h
+++ b/src/mesa/shader/slang/library/slang_core_gc.h
@@ -18,59 +18,60 @@
 20,0,0,1,0,10,1,1,1,0,5,105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,105,0,59,
 120,120,0,20,0,0,1,0,10,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,98,
 0,59,120,120,0,20,0,0,1,0,10,1,1,1,0,11,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,
-0,18,118,0,59,120,121,0,20,0,0,1,0,11,1,1,1,0,9,120,0,0,1,1,0,9,121,0,0,1,1,0,9,122,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,
-18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,0,1,0,11,1,1,1,0,9,102,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,102,0,59,120,120,120,0,20,0,0,1,0,11,1,
-1,1,0,5,105,0,0,0,1,4,105,110,116,95,116,111,95,102,108,111,97,116,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,122,0,0,18,105,0,59,120,120,120,0,0,0,0,1,0,11,1,1,1,0,1,98,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,11,1,1,1,0,12,118,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,118,0,59,120,121,122,0,20,0,0,1,0,12,1,
-1,1,0,9,120,0,0,1,1,0,9,121,0,0,1,1,0,9,122,0,0,1,1,0,9,119,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,119,
-0,20,0,0,1,0,12,1,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,102,0,59,120,120,120,
-120,0,20,0,0,1,0,12,1,1,1,0,5,105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,105,0,59,120,120,
-120,120,0,20,0,0,1,0,12,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,98,0,59,120,120,
-120,120,0,20,0,0,1,0,12,1,1,1,0,11,118,51,0,0,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,59,120,121,122,0,18,118,51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,102,0,20,0,0,1,
-0,6,1,1,1,0,5,105,0,0,1,1,0,5,106,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,105,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,0,20,0,0,1,0,6,1,1,1,0,5,105,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,59,120,121,0,18,105,0,59,120,120,0,20,0,0,1,0,6,1,1,1,0,9,102,0,0,0,1,4,
-102,108,111,97,116,95,116,111,95,105,110,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,
-102,0,59,120,120,0,0,0,0,1,0,6,1,1,1,0,1,98,0,0,0,1,4,102,108,111,97,116,95,116,111,95,105,110,116,
-0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,98,0,59,120,120,0,0,0,0,1,0,7,1,1,1,0,5,105,0,
-0,1,1,0,5,106,0,0,1,1,0,5,107,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,105,0,20,0,9,
-18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,
-0,18,107,0,20,0,0,1,0,7,1,1,1,0,5,105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,
-18,105,0,59,120,120,120,0,20,0,0,1,0,7,1,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,
-120,121,122,0,18,102,0,59,120,120,120,0,20,0,0,1,0,7,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,8,1,1,1,0,5,120,0,0,1,1,0,5,121,0,
-0,1,1,0,5,122,0,0,1,1,0,5,119,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,
-18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,
-0,18,122,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,119,0,20,0,0,1,0,8,1,1,1,0,5,105,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,105,0,59,120,120,120,120,0,20,0,0,1,0,8,1,1,1,0,9,102,
-0,0,0,1,4,102,108,111,97,116,95,116,111,95,105,110,116,0,18,95,95,114,101,116,86,97,108,0,0,18,102,
-0,59,120,120,120,120,0,0,0,0,1,0,8,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,98,0,
-59,120,120,120,120,0,20,0,0,1,0,2,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,98,50,0,20,0,0,
-1,0,2,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,98,0,59,120,120,0,20,
-0,0,1,0,2,1,1,1,0,9,102,0,0,0,1,3,2,1,10,1,122,101,114,111,0,2,58,118,101,99,50,0,17,48,0,48,0,0,0,
-17,48,0,48,0,0,0,0,0,0,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,0,0,18,102,0,59,120,120,0,0,18,122,101,114,111,0,0,0,0,1,0,2,1,1,1,0,5,105,0,0,0,1,3,2,1,6,1,
-122,101,114,111,0,2,58,105,118,101,99,50,0,16,8,48,0,0,16,8,48,0,0,0,0,0,4,118,101,99,52,95,115,
-101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,105,0,59,120,120,0,0,18,122,101,114,
-111,0,0,0,0,1,0,3,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,1,1,0,1,98,51,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,98,50,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,98,51,0,20,0,0,1,0,3,1,1,1,0,1,98,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,3,1,1,1,0,9,102,0,
-0,0,1,3,2,1,11,1,122,101,114,111,0,2,58,118,101,99,51,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,48,0,
-48,0,0,0,0,0,0,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,
-0,18,102,0,59,120,120,120,0,0,18,122,101,114,111,0,0,0,0,1,0,3,1,1,1,0,5,105,0,0,0,1,3,2,1,7,1,122,
-101,114,111,0,2,58,105,118,101,99,51,0,16,8,48,0,0,16,8,48,0,0,16,8,48,0,0,0,0,0,4,118,101,99,52,
-95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,105,0,59,120,120,120,0,0,
-18,122,101,114,111,0,0,0,0,1,0,4,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,1,1,0,1,98,51,0,0,1,1,0,1,
-98,52,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,
+0,18,118,0,59,120,121,0,20,0,0,1,0,10,1,1,1,0,12,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,
+115,116,0,18,118,0,59,120,121,0,20,0,0,1,0,11,1,1,1,0,9,120,0,0,1,1,0,9,121,0,0,1,1,0,9,122,0,0,0,
+1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
+121,0,18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,0,1,0,11,1,1,1,0,9,
+102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,102,0,59,120,120,120,0,20,0,0,1,
+0,11,1,1,1,0,5,105,0,0,0,1,4,105,110,116,95,116,111,95,102,108,111,97,116,0,18,95,95,114,101,116,
+86,97,108,0,59,120,121,122,0,0,18,105,0,59,120,120,120,0,0,0,0,1,0,11,1,1,1,0,1,98,0,0,0,1,9,18,95,
+95,114,101,116,86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,11,1,1,1,0,12,118,
+0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,118,0,59,120,121,122,0,20,0,0,1,0,
+12,1,1,1,0,9,120,0,0,1,1,0,9,121,0,0,1,1,0,9,122,0,0,1,1,0,9,119,0,0,0,1,9,18,95,95,114,101,116,86,
+97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,0,9,18,95,
+95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,
+119,0,20,0,0,1,0,12,1,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,102,0,59,120,120,
+120,120,0,20,0,0,1,0,12,1,1,1,0,5,105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,105,0,59,120,
+120,120,120,0,20,0,0,1,0,12,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,98,0,59,120,
+120,120,120,0,20,0,0,1,0,12,1,1,1,0,11,118,51,0,0,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,
+108,0,59,120,121,122,0,18,118,51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,102,0,20,0,
+0,1,0,6,1,1,1,0,5,105,0,0,1,1,0,5,106,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,105,0,
+20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,0,20,0,0,1,0,6,1,1,1,0,5,105,0,0,0,1,9,18,
+95,95,114,101,116,86,97,108,0,59,120,121,0,18,105,0,59,120,120,0,20,0,0,1,0,6,1,1,1,0,9,102,0,0,0,
+1,4,102,108,111,97,116,95,116,111,95,105,110,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,
+18,102,0,59,120,120,0,0,0,0,1,0,6,1,1,1,0,1,98,0,0,0,1,4,102,108,111,97,116,95,116,111,95,105,110,
+116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,98,0,59,120,120,0,0,0,0,1,0,7,1,1,1,0,5,
+105,0,0,1,1,0,5,106,0,0,1,1,0,5,107,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,105,0,
+20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
+59,122,0,18,107,0,20,0,0,1,0,7,1,1,1,0,5,105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,
+122,0,18,105,0,59,120,120,120,0,20,0,0,1,0,7,1,1,1,0,9,102,0,0,0,1,9,18,95,95,114,101,116,86,97,
+108,0,59,120,121,122,0,18,102,0,59,120,120,120,0,20,0,0,1,0,7,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,
+101,116,86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,8,1,1,1,0,5,120,0,0,1,1,0,
+5,121,0,0,1,1,0,5,122,0,0,1,1,0,5,119,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,
+20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
+59,122,0,18,122,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,119,0,20,0,0,1,0,8,1,1,1,0,5,
+105,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,105,0,59,120,120,120,120,0,20,0,0,1,0,8,1,1,1,0,
+9,102,0,0,0,1,4,102,108,111,97,116,95,116,111,95,105,110,116,0,18,95,95,114,101,116,86,97,108,0,0,
+18,102,0,59,120,120,120,120,0,0,0,0,1,0,8,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
+18,98,0,59,120,120,120,120,0,20,0,0,1,0,2,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,0,1,9,18,95,95,114,
+101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,98,50,
+0,20,0,0,1,0,2,1,1,1,0,1,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,98,0,59,120,
+120,0,20,0,0,1,0,2,1,1,1,0,9,102,0,0,0,1,3,2,1,10,1,122,101,114,111,0,2,58,118,101,99,50,0,17,48,0,
+48,0,0,0,17,48,0,48,0,0,0,0,0,0,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,
+59,120,121,0,0,18,102,0,59,120,120,0,0,18,122,101,114,111,0,0,0,0,1,0,2,1,1,1,0,5,105,0,0,0,1,3,2,
+1,6,1,122,101,114,111,0,2,58,105,118,101,99,50,0,16,8,48,0,0,16,8,48,0,0,0,0,0,4,118,101,99,52,95,
+115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,105,0,59,120,120,0,0,18,122,101,
+114,111,0,0,0,0,1,0,3,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,1,1,0,1,98,51,0,0,0,1,9,18,95,95,114,
+101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,98,50,
+0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,98,51,0,20,0,0,1,0,3,1,1,1,0,1,98,0,0,0,1,9,
+18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,98,0,59,120,120,120,0,20,0,0,1,0,3,1,1,1,0,9,
+102,0,0,0,1,3,2,1,11,1,122,101,114,111,0,2,58,118,101,99,51,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,
+48,0,48,0,0,0,0,0,0,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,
+122,0,0,18,102,0,59,120,120,120,0,0,18,122,101,114,111,0,0,0,0,1,0,3,1,1,1,0,5,105,0,0,0,1,3,2,1,7,
+1,122,101,114,111,0,2,58,105,118,101,99,51,0,16,8,48,0,0,16,8,48,0,0,16,8,48,0,0,0,0,0,4,118,101,
+99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,105,0,59,120,120,120,
+0,0,18,122,101,114,111,0,0,0,0,1,0,4,1,1,1,0,1,98,49,0,0,1,1,0,1,98,50,0,0,1,1,0,1,98,51,0,0,1,1,0,
+1,98,52,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,
 86,97,108,0,59,121,0,18,98,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,98,51,0,20,0,9,
 18,95,95,114,101,116,86,97,108,0,59,119,0,18,98,52,0,20,0,0,1,0,4,1,1,1,0,1,98,0,0,0,1,9,18,95,95,
 114,101,116,86,97,108,0,59,120,121,122,119,0,18,98,0,59,120,120,120,120,0,20,0,0,1,0,4,1,1,1,0,9,
@@ -646,7 +647,7 @@
 1,0,12,2,21,1,1,0,15,109,0,0,1,1,0,12,118,0,0,0,1,2,3,2,0,12,1,114,48,0,0,0,9,18,114,48,0,59,120,0,
 18,109,0,16,8,48,0,57,59,120,0,20,0,9,18,114,48,0,59,121,0,18,109,0,16,10,49,0,57,59,120,0,20,0,9,
 18,114,48,0,59,122,0,18,109,0,16,10,50,0,57,59,120,0,20,0,9,18,114,48,0,59,119,0,18,109,0,16,10,51,
-0,57,59,120,0,20,0,4,118,101,99,52,95,100,111,116,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,
+0,57,59,120,0,20,0,4,118,101,99,52,95,100,111,116,0,18,95,95,114,101,116,86,97,108,0,59,115,0,0,18,
 114,48,0,0,18,118,0,0,0,0,2,3,2,0,12,1,114,49,0,0,0,9,18,114,49,0,59,120,0,18,109,0,16,8,48,0,57,
 59,121,0,20,0,9,18,114,49,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,20,0,9,18,114,49,0,59,122,0,
 18,109,0,16,10,50,0,57,59,121,0,20,0,9,18,114,49,0,59,119,0,18,109,0,16,10,51,0,57,59,121,0,20,0,4,
diff --git a/src/mesa/shader/slang/slang_builtin.c b/src/mesa/shader/slang/slang_builtin.c
index 6ee0fd33b6..1081d8ff8d 100644
--- a/src/mesa/shader/slang/slang_builtin.c
+++ b/src/mesa/shader/slang/slang_builtin.c
@@ -28,14 +28,14 @@
  * \author Brian Paul
  */
 
-#include "imports.h"
-#include "slang_builtin.h"
-#include "slang_ir.h"
-#include "mtypes.h"
-#include "program.h"
-#include "prog_instruction.h"
-#include "prog_parameter.h"
-#include "prog_statevars.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
+#include "shader/program.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_statevars.h"
+#include "shader/slang/slang_ir.h"
+#include "shader/slang/slang_builtin.h"
 
 
 /**
diff --git a/src/mesa/shader/slang/slang_builtin.h b/src/mesa/shader/slang/slang_builtin.h
index ae20c844d5..58629f4f7f 100644
--- a/src/mesa/shader/slang/slang_builtin.h
+++ b/src/mesa/shader/slang/slang_builtin.h
@@ -26,7 +26,7 @@
 #ifndef SLANG_BUILTIN_H
 #define SLANG_BUILTIN_H
 
-#include "prog_parameter.h"
+#include "shader/prog_parameter.h"
 #include "slang_utility.h"
 #include "slang_ir.h"
 
diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c
index 02260d3422..35e80e0452 100644
--- a/src/mesa/shader/slang/slang_codegen.c
+++ b/src/mesa/shader/slang/slang_codegen.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.3
+ * Version:  7.1
  *
  * Copyright (C) 2005-2007  Brian Paul   All Rights Reserved.
  *
@@ -37,13 +37,13 @@
 
 
 
-#include "imports.h"
-#include "macros.h"
-#include "mtypes.h"
-#include "program.h"
-#include "prog_instruction.h"
-#include "prog_parameter.h"
-#include "prog_statevars.h"
+#include "main/imports.h"
+#include "main/macros.h"
+#include "main/mtypes.h"
+#include "shader/program.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_statevars.h"
 #include "slang_typeinfo.h"
 #include "slang_codegen.h"
 #include "slang_compile.h"
@@ -1261,15 +1261,23 @@ make_writemask(const char *field)
    while (*field) {
       switch (*field) {
       case 'x':
+      case 's':
+      case 'r':
          mask |= WRITEMASK_X;
          break;
       case 'y':
+      case 't':
+      case 'g':
          mask |= WRITEMASK_Y;
          break;
       case 'z':
+      case 'p':
+      case 'b':
          mask |= WRITEMASK_Z;
          break;
       case 'w':
+      case 'q':
+      case 'a':
          mask |= WRITEMASK_W;
          break;
       default:
@@ -2320,7 +2328,8 @@ _slang_gen_field(slang_assemble_ctx * A, slang_operation *oper)
          n = _slang_gen_swizzle(n, swizzle);
       return n;
    }
-   else if (ti.spec.type == SLANG_SPEC_FLOAT) {
+   else if (   ti.spec.type == SLANG_SPEC_FLOAT
+            || ti.spec.type == SLANG_SPEC_INT) {
       const GLuint rows = 1;
       slang_swizzle swz;
       slang_ir_node *n;
diff --git a/src/mesa/shader/slang/slang_compile.c b/src/mesa/shader/slang/slang_compile.c
index a4dd5b8b4a..4e29e8dcc9 100644
--- a/src/mesa/shader/slang/slang_compile.c
+++ b/src/mesa/shader/slang/slang_compile.c
@@ -28,11 +28,11 @@
  * \author Michal Krol
  */
 
-#include "imports.h"
-#include "context.h"
-#include "program.h"
-#include "prog_parameter.h"
-#include "grammar_mesa.h"
+#include "main/imports.h"
+#include "main/context.h"
+#include "shader/program.h"
+#include "shader/prog_parameter.h"
+#include "shader/grammar/grammar_mesa.h"
 #include "slang_codegen.h"
 #include "slang_compile.h"
 #include "slang_preprocess.h"
@@ -2135,7 +2135,7 @@ _slang_compile(GLcontext *ctx, struct gl_shader *shader)
          progTarget = GL_FRAGMENT_PROGRAM_ARB;
       shader->Programs
          = (struct gl_program **) malloc(sizeof(struct gl_program*));
-      shader->Programs[0] = _mesa_new_program(ctx, progTarget, 1);
+      shader->Programs[0] = ctx->Driver.NewProgram(ctx, progTarget, 1);
       shader->NumPrograms = 1;
 
       shader->Programs[0]->Parameters = _mesa_new_parameter_list();
diff --git a/src/mesa/shader/slang/slang_emit.c b/src/mesa/shader/slang/slang_emit.c
index 7804e19236..02c74095a9 100644
--- a/src/mesa/shader/slang/slang_emit.c
+++ b/src/mesa/shader/slang/slang_emit.c
@@ -36,13 +36,13 @@
  ***/
 
 
-#include "imports.h"
-#include "context.h"
-#include "macros.h"
-#include "program.h"
-#include "prog_instruction.h"
-#include "prog_parameter.h"
-#include "prog_print.h"
+#include "main/imports.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "shader/program.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_print.h"
 #include "slang_builtin.h"
 #include "slang_emit.h"
 #include "slang_mem.h"
diff --git a/src/mesa/shader/slang/slang_ir.c b/src/mesa/shader/slang/slang_ir.c
index a6903cc8b6..a29f302687 100644
--- a/src/mesa/shader/slang/slang_ir.c
+++ b/src/mesa/shader/slang/slang_ir.c
@@ -27,7 +27,7 @@
 #include "context.h"
 #include "slang_ir.h"
 #include "slang_mem.h"
-#include "prog_print.h"
+#include "shader/prog_print.h"
 
 
 static const slang_ir_info IrInfo[] = {
diff --git a/src/mesa/shader/slang/slang_label.h b/src/mesa/shader/slang/slang_label.h
index 0f1a45b30f..87068ae7a7 100644
--- a/src/mesa/shader/slang/slang_label.h
+++ b/src/mesa/shader/slang/slang_label.h
@@ -1,9 +1,9 @@
 #ifndef SLANG_LABEL_H
 #define SLANG_LABEL_H 1
 
-#include "imports.h"
-#include "mtypes.h"
-#include "prog_instruction.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
+#include "shader/prog_instruction.h"
 
 
 struct slang_label_
diff --git a/src/mesa/shader/slang/slang_link.c b/src/mesa/shader/slang/slang_link.c
index d6d1c7523e..eaa29ba094 100644
--- a/src/mesa/shader/slang/slang_link.c
+++ b/src/mesa/shader/slang/slang_link.c
@@ -28,16 +28,16 @@
  * \author Brian Paul
  */
 
-#include "imports.h"
-#include "context.h"
-#include "hash.h"
-#include "macros.h"
-#include "program.h"
-#include "prog_instruction.h"
-#include "prog_parameter.h"
-#include "prog_print.h"
-#include "prog_statevars.h"
-#include "shader_api.h"
+#include "main/imports.h"
+#include "main/context.h"
+#include "main/hash.h"
+#include "main/macros.h"
+#include "shader/program.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_print.h"
+#include "shader/prog_statevars.h"
+#include "shader/shader_api.h"
 #include "slang_link.h"
 
 
diff --git a/src/mesa/shader/slang/slang_preprocess.c b/src/mesa/shader/slang/slang_preprocess.c
index 72281eda57..076e982f8f 100644
--- a/src/mesa/shader/slang/slang_preprocess.c
+++ b/src/mesa/shader/slang/slang_preprocess.c
@@ -29,7 +29,7 @@
  */
 
 #include "imports.h"
-#include "grammar_mesa.h"
+#include "shader/grammar/grammar_mesa.h"
 #include "slang_preprocess.h"
 
 LONGSTRING static const char *slang_pp_directives_syn =
diff --git a/src/mesa/shader/slang/slang_typeinfo.c b/src/mesa/shader/slang/slang_typeinfo.c
index da0b32bc44..8a1c3abf48 100644
--- a/src/mesa/shader/slang/slang_typeinfo.c
+++ b/src/mesa/shader/slang/slang_typeinfo.c
@@ -28,12 +28,12 @@
  * \author Michal Krol
  */
 
-#include "imports.h"
+#include "main/imports.h"
+#include "shader/prog_instruction.h"
 #include "slang_typeinfo.h"
 #include "slang_compile.h"
 #include "slang_log.h"
 #include "slang_mem.h"
-#include "prog_instruction.h"
 
 
 /**
diff --git a/src/mesa/shader/slang/slang_vartable.c b/src/mesa/shader/slang/slang_vartable.c
index 8a3c299d19..1d817000c6 100644
--- a/src/mesa/shader/slang/slang_vartable.c
+++ b/src/mesa/shader/slang/slang_vartable.c
@@ -1,11 +1,11 @@
 
-#include "imports.h"
+#include "main/imports.h"
+#include "shader/prog_instruction.h"
 #include "slang_compile.h"
 #include "slang_compile_variable.h"
 #include "slang_mem.h"
 #include "slang_vartable.h"
 #include "slang_ir.h"
-#include "prog_instruction.h"
 
 
 static int dbg = 0;
diff --git a/src/mesa/sparc/glapi_sparc.S b/src/mesa/sparc/glapi_sparc.S
index d4ea12870f..8725c7ee4d 100644
--- a/src/mesa/sparc/glapi_sparc.S
+++ b/src/mesa/sparc/glapi_sparc.S
@@ -1733,12 +1733,17 @@ _mesa_sparc_glapi_end:
 	.globl glVertexAttrib4Nubv ; .type glVertexAttrib4Nubv,#function ; glVertexAttrib4Nubv = glVertexAttrib4NubvARB
 	.globl glVertexAttrib4Nuiv ; .type glVertexAttrib4Nuiv,#function ; glVertexAttrib4Nuiv = glVertexAttrib4NuivARB
 	.globl glVertexAttrib4Nusv ; .type glVertexAttrib4Nusv,#function ; glVertexAttrib4Nusv = glVertexAttrib4NusvARB
+	.globl glVertexAttrib4bv ; .type glVertexAttrib4bv,#function ; glVertexAttrib4bv = glVertexAttrib4bvARB
 	.globl glVertexAttrib4d ; .type glVertexAttrib4d,#function ; glVertexAttrib4d = glVertexAttrib4dARB
 	.globl glVertexAttrib4dv ; .type glVertexAttrib4dv,#function ; glVertexAttrib4dv = glVertexAttrib4dvARB
 	.globl glVertexAttrib4f ; .type glVertexAttrib4f,#function ; glVertexAttrib4f = glVertexAttrib4fARB
 	.globl glVertexAttrib4fv ; .type glVertexAttrib4fv,#function ; glVertexAttrib4fv = glVertexAttrib4fvARB
+	.globl glVertexAttrib4iv ; .type glVertexAttrib4iv,#function ; glVertexAttrib4iv = glVertexAttrib4ivARB
 	.globl glVertexAttrib4s ; .type glVertexAttrib4s,#function ; glVertexAttrib4s = glVertexAttrib4sARB
 	.globl glVertexAttrib4sv ; .type glVertexAttrib4sv,#function ; glVertexAttrib4sv = glVertexAttrib4svARB
+	.globl glVertexAttrib4ubv ; .type glVertexAttrib4ubv,#function ; glVertexAttrib4ubv = glVertexAttrib4ubvARB
+	.globl glVertexAttrib4uiv ; .type glVertexAttrib4uiv,#function ; glVertexAttrib4uiv = glVertexAttrib4uivARB
+	.globl glVertexAttrib4usv ; .type glVertexAttrib4usv,#function ; glVertexAttrib4usv = glVertexAttrib4usvARB
 	.globl glVertexAttribPointer ; .type glVertexAttribPointer,#function ; glVertexAttribPointer = glVertexAttribPointerARB
 	.globl glBindBuffer ; .type glBindBuffer,#function ; glBindBuffer = glBindBufferARB
 	.globl glBufferData ; .type glBufferData,#function ; glBufferData = glBufferDataARB
diff --git a/src/mesa/swrast/s_aalinetemp.h b/src/mesa/swrast/s_aalinetemp.h
index 3d3511823d..e7911fec3b 100644
--- a/src/mesa/swrast/s_aalinetemp.h
+++ b/src/mesa/swrast/s_aalinetemp.h
@@ -132,13 +132,15 @@ NAME(line)(GLcontext *ctx, const SWvertex *v0, const SWvertex *v1)
    line.dx = line.x1 - line.x0;
    line.dy = line.y1 - line.y0;
    line.len = SQRTF(line.dx * line.dx + line.dy * line.dy);
-   line.halfWidth = 0.5F * ctx->Line._Width;
+   line.halfWidth = 0.5F * CLAMP(ctx->Line.Width,
+                                 ctx->Const.MinLineWidthAA,
+                                 ctx->Const.MaxLineWidthAA);
 
    if (line.len == 0.0 || IS_INF_OR_NAN(line.len))
       return;
 
-   INIT_SPAN(line.span, GL_LINE, 0, 0, SPAN_XY | SPAN_COVERAGE);
-
+   INIT_SPAN(line.span, GL_LINE);
+   line.span.arrayMask = SPAN_XY | SPAN_COVERAGE;
    line.xAdj = line.dx / line.len * line.halfWidth;
    line.yAdj = line.dy / line.len * line.halfWidth;
 
diff --git a/src/mesa/swrast/s_aatritemp.h b/src/mesa/swrast/s_aatritemp.h
index 34a2305b39..42d74a1632 100644
--- a/src/mesa/swrast/s_aatritemp.h
+++ b/src/mesa/swrast/s_aatritemp.h
@@ -69,7 +69,8 @@
    
    (void) swrast;
 
-   INIT_SPAN(span, GL_POLYGON, 0, 0, SPAN_COVERAGE);
+   INIT_SPAN(span, GL_POLYGON);
+   span.arrayMask = SPAN_COVERAGE;
 
    /* determine bottom to top order of vertices */
    {
diff --git a/src/mesa/swrast/s_accum.c b/src/mesa/swrast/s_accum.c
index f53e7f52c5..3c45dee399 100644
--- a/src/mesa/swrast/s_accum.c
+++ b/src/mesa/swrast/s_accum.c
@@ -474,7 +474,9 @@ accum_return(GLcontext *ctx, GLfloat value,
          SWspan span;
 
          /* init color span */
-         INIT_SPAN(span, GL_BITMAP, width, 0, SPAN_RGBA);
+         INIT_SPAN(span, GL_BITMAP);
+         span.end = width;
+         span.arrayMask = SPAN_RGBA;
          span.x = xpos;
          span.y = ypos + i;
 
diff --git a/src/mesa/swrast/s_atifragshader.c b/src/mesa/swrast/s_atifragshader.c
index 947054faa3..55ec757ee0 100644
--- a/src/mesa/swrast/s_atifragshader.c
+++ b/src/mesa/swrast/s_atifragshader.c
@@ -23,11 +23,10 @@
 #include "glheader.h"
 #include "colormac.h"
 #include "context.h"
-#include "atifragshader.h"
 #include "macros.h"
-#include "program.h"
-
-#include "s_atifragshader.h"
+#include "shader/program.h"
+#include "shader/atifragshader.h"
+#include "swrast/s_atifragshader.h"
 
 
 /**
diff --git a/src/mesa/swrast/s_bitmap.c b/src/mesa/swrast/s_bitmap.c
index 563b5fe602..1e7f6c18e6 100644
--- a/src/mesa/swrast/s_bitmap.c
+++ b/src/mesa/swrast/s_bitmap.c
@@ -82,7 +82,9 @@ _swrast_Bitmap( GLcontext *ctx, GLint px, GLint py,
    if (SWRAST_CONTEXT(ctx)->NewState)
       _swrast_validate_derived( ctx );
 
-   INIT_SPAN(span, GL_BITMAP, width, 0, SPAN_XY);
+   INIT_SPAN(span, GL_BITMAP);
+   span.end = width;
+   span.arrayMask = SPAN_XY;
    _swrast_span_default_attribs(ctx, &span);
 
    for (row = 0; row < height; row++) {
@@ -180,7 +182,9 @@ _swrast_Bitmap( GLcontext *ctx, GLint px, GLint py,
    if (SWRAST_CONTEXT(ctx)->NewState)
       _swrast_validate_derived( ctx );
 
-   INIT_SPAN(span, GL_BITMAP, width, 0, SPAN_MASK);
+   INIT_SPAN(span, GL_BITMAP);
+   span.end = width;
+   span.arrayMask = SPAN_MASK;
    _swrast_span_default_attribs(ctx, &span);
 
    /*span.arrayMask |= SPAN_MASK;*/  /* we'll init span.mask[] */
diff --git a/src/mesa/swrast/s_buffers.c b/src/mesa/swrast/s_buffers.c
index 35f2dd6490..90a56284c5 100644
--- a/src/mesa/swrast/s_buffers.c
+++ b/src/mesa/swrast/s_buffers.c
@@ -55,7 +55,9 @@ clear_rgba_buffer_with_masking(GLcontext *ctx, struct gl_renderbuffer *rb)
 
    /* Initialize color span with clear color */
    /* XXX optimize for clearcolor == black/zero (bzero) */
-   INIT_SPAN(span, GL_BITMAP, width, 0, SPAN_RGBA);
+   INIT_SPAN(span, GL_BITMAP);
+   span.end = width;
+   span.arrayMask = SPAN_RGBA;
    span.array->ChanType = rb->DataType;
    if (span.array->ChanType == GL_UNSIGNED_BYTE) {
       GLubyte clearColor[4];
@@ -119,7 +121,9 @@ clear_ci_buffer_with_masking(GLcontext *ctx, struct gl_renderbuffer *rb)
    ASSERT(rb->DataType == GL_UNSIGNED_INT);
 
    /* Initialize index span with clear index */
-   INIT_SPAN(span, GL_BITMAP, width, 0, SPAN_RGBA);
+   INIT_SPAN(span, GL_BITMAP);
+   span.end = width;
+   span.arrayMask = SPAN_INDEX;
    for (i = 0; i < width;i++) {
       span.array->index[i] = ctx->Color.ClearIndex;
    }
diff --git a/src/mesa/swrast/s_context.c b/src/mesa/swrast/s_context.c
index cb3bc756a4..3956925651 100644
--- a/src/mesa/swrast/s_context.c
+++ b/src/mesa/swrast/s_context.c
@@ -31,9 +31,9 @@
 #include "context.h"
 #include "colormac.h"
 #include "mtypes.h"
-#include "prog_statevars.h"
 #include "teximage.h"
 #include "swrast.h"
+#include "shader/prog_statevars.h"
 #include "s_blend.h"
 #include "s_context.h"
 #include "s_lines.h"
@@ -535,10 +535,11 @@ _swrast_update_texture_samplers(GLcontext *ctx)
 
 
 /**
- * Update swrast->_ActiveAttribs, swrast->_NumActiveAttribs, swrast->_ActiveAtttribMask.
+ * Update swrast->_ActiveAttribs, swrast->_NumActiveAttribs,
+ * swrast->_ActiveAtttribMask.
  */
 static void
-_swrast_update_fragment_attribs(GLcontext *ctx)
+_swrast_update_active_attribs(GLcontext *ctx)
 {
    SWcontext *swrast = SWRAST_CONTEXT(ctx);
    GLuint attribsMask;
@@ -679,7 +680,7 @@ _swrast_validate_derived( GLcontext *ctx )
                               _NEW_LIGHT |
                               _NEW_PROGRAM |
                               _NEW_TEXTURE))
-         _swrast_update_fragment_attribs(ctx);
+         _swrast_update_active_attribs(ctx);
 
       if (swrast->NewState & (_NEW_PROGRAM | _NEW_BUFFERS))
          _swrast_update_color_outputs(ctx);
diff --git a/src/mesa/swrast/s_context.h b/src/mesa/swrast/s_context.h
index f118eb92ca..daa07e1578 100644
--- a/src/mesa/swrast/s_context.h
+++ b/src/mesa/swrast/s_context.h
@@ -43,10 +43,10 @@
 #ifndef S_CONTEXT_H
 #define S_CONTEXT_H
 
-#include "mtypes.h"
+#include "main/mtypes.h"
+#include "shader/prog_execute.h"
 #include "swrast.h"
 #include "s_span.h"
-#include "prog_execute.h"
 
 
 typedef void (*texture_sample_func)(GLcontext *ctx,
diff --git a/src/mesa/swrast/s_copypix.c b/src/mesa/swrast/s_copypix.c
index 53e584b3b6..bbe1081860 100644
--- a/src/mesa/swrast/s_copypix.c
+++ b/src/mesa/swrast/s_copypix.c
@@ -102,8 +102,10 @@ copy_conv_rgba_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
    GLfloat *dest, *tmpImage, *convImage;
    SWspan span;
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_RGBA);
+   INIT_SPAN(span, GL_BITMAP);
    _swrast_span_default_attribs(ctx, &span);
+   span.arrayMask = SPAN_RGBA;
+   span.arrayAttribs = FRAG_BIT_COL0;
 
    /* allocate space for GLfloat image */
    tmpImage = (GLfloat *) _mesa_malloc(width * height * 4 * sizeof(GLfloat));
@@ -156,7 +158,7 @@ copy_conv_rgba_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
       /* write the new image */
       for (row = 0; row < height; row++) {
          const GLfloat *src = convImage + row * width * 4;
-         GLvoid *rgba = (GLvoid *) span.array->attribs[FRAG_ATTRIB_COL0];
+         GLfloat *rgba = (GLfloat *) span.array->attribs[FRAG_ATTRIB_COL0];
 
          /* copy convolved colors into span array */
          _mesa_memcpy(rgba, src, width * 4 * sizeof(GLfloat));
@@ -232,8 +234,10 @@ copy_rgba_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
       stepy = 1;
    }
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_RGBA);
+   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 *) _mesa_malloc(width * height * sizeof(GLfloat) * 4);
@@ -314,8 +318,9 @@ copy_ci_pixels( GLcontext *ctx, GLint srcx, GLint srcy,
       return;
    }
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_INDEX);
+   INIT_SPAN(span, GL_BITMAP);
    _swrast_span_default_attribs(ctx, &span);
+   span.arrayMask = SPAN_INDEX;
 
    if (ctx->DrawBuffer == ctx->ReadBuffer) {
       overlapping = regions_overlap(srcx, srcy, destx, desty, width, height,
@@ -448,8 +453,9 @@ copy_depth_pixels( GLcontext *ctx, GLint srcx, GLint srcy,
       return;
    }
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_Z);
+   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,
diff --git a/src/mesa/swrast/s_drawpix.c b/src/mesa/swrast/s_drawpix.c
index d971d90fb9..0cf425e1c6 100644
--- a/src/mesa/swrast/s_drawpix.c
+++ b/src/mesa/swrast/s_drawpix.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.2
+ * Version:  7.1
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -70,7 +70,9 @@ fast_draw_rgba_pixels(GLcontext *ctx, GLint x, GLint y,
       return GL_FALSE;
    }
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_RGBA);
+   INIT_SPAN(span, GL_BITMAP);
+   span.arrayMask = SPAN_RGBA;
+   span.arrayAttribs = FRAG_BIT_COL0;
    _swrast_span_default_attribs(ctx, &span);
 
    /* copy input params since clipping may change them */
@@ -332,7 +334,8 @@ draw_index_pixels( GLcontext *ctx, GLint x, GLint y,
    GLint row, skipPixels;
    SWspan span;
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_INDEX);
+   INIT_SPAN(span, GL_BITMAP);
+   span.arrayMask = SPAN_INDEX;
    _swrast_span_default_attribs(ctx, &span);
 
    /*
@@ -395,10 +398,9 @@ draw_stencil_pixels( GLcontext *ctx, GLint x, GLint y,
                                                       width, height,
                                                       GL_COLOR_INDEX, type,
                                                       row, skipPixels);
-         _mesa_unpack_index_span(ctx, spanWidth, destType, values,
-                                 type, source, unpack,
-                                 ctx->_ImageTransferState);
-         _mesa_apply_stencil_transfer_ops(ctx, spanWidth, values);
+         _mesa_unpack_stencil_span(ctx, spanWidth, destType, values,
+                                   type, source, unpack,
+                                   ctx->_ImageTransferState);
          if (zoom) {
             _swrast_write_zoomed_stencil_span(ctx, x, y, spanWidth,
                                               spanX, spanY, values);
@@ -427,7 +429,8 @@ draw_depth_pixels( GLcontext *ctx, GLint x, GLint y,
    const GLboolean zoom = ctx->Pixel.ZoomX != 1.0 || ctx->Pixel.ZoomY != 1.0;
    SWspan span;
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0, SPAN_Z);
+   INIT_SPAN(span, GL_BITMAP);
+   span.arrayMask = SPAN_Z;
    _swrast_span_default_attribs(ctx, &span);
 
    if (type == GL_UNSIGNED_SHORT
@@ -481,7 +484,7 @@ draw_depth_pixels( GLcontext *ctx, GLint x, GLint y,
    }
    else {
       /* General case */
-      const GLfloat depthMax = ctx->DrawBuffer->_DepthMaxF;
+      const GLuint depthMax = ctx->DrawBuffer->_DepthMax;
       GLint skipPixels = 0;
 
       /* in case width > MAX_WIDTH do the copy in chunks */
@@ -540,11 +543,14 @@ draw_rgba_pixels( GLcontext *ctx, GLint x, GLint y,
 
    /* Try an optimized glDrawPixels first */
    if (fast_draw_rgba_pixels(ctx, x, y, width, height, format, type,
-                             unpack, pixels))
+                             unpack, pixels)) {
       return;
+   }
 
-   INIT_SPAN(span, GL_BITMAP, 0, 0x0, SPAN_RGBA);
+   INIT_SPAN(span, GL_BITMAP);
    _swrast_span_default_attribs(ctx, &span);
+   span.arrayMask = SPAN_RGBA;
+   span.arrayAttribs = FRAG_BIT_COL0; /* we're fill in COL0 attrib values */
 
    if (ctx->Pixel.Convolution2DEnabled || ctx->Pixel.Separable2DEnabled) {
       /* Convolution has to be handled specially.  We'll create an
@@ -689,7 +695,7 @@ draw_depth_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
    const GLint imgX = x, imgY = y;
    const GLboolean scaleOrBias
       = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
-   const GLfloat depthScale = ctx->DrawBuffer->_DepthMaxF;
+   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;
@@ -777,7 +783,7 @@ draw_depth_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
                /* general case */
                GLuint zValues[MAX_WIDTH];  /* 16 or 32-bit Z value storage */
                _mesa_unpack_depth_span(ctx, width,
-                                       depthRb->DataType, zValues, depthScale,
+                                       depthRb->DataType, zValues, depthMax,
                                        type, depthStencilSrc, &clippedUnpack);
                if (zoom) {
                   _swrast_write_zoomed_z_span(ctx, imgX, imgY, width, x,
diff --git a/src/mesa/swrast/s_feedback.c b/src/mesa/swrast/s_feedback.c
index 606afc63ba..07b7409ab5 100644
--- a/src/mesa/swrast/s_feedback.c
+++ b/src/mesa/swrast/s_feedback.c
@@ -39,19 +39,14 @@ static void
 feedback_vertex(GLcontext * ctx, const SWvertex * v, const SWvertex * pv)
 {
    GLfloat win[4];
-   GLfloat color[4];
    const GLfloat *vtc = v->attrib[FRAG_ATTRIB_TEX0];
+   const GLfloat *color = v->attrib[FRAG_ATTRIB_COL0];
 
    win[0] = v->attrib[FRAG_ATTRIB_WPOS][0];
    win[1] = v->attrib[FRAG_ATTRIB_WPOS][1];
    win[2] = v->attrib[FRAG_ATTRIB_WPOS][2] / ctx->DrawBuffer->_DepthMaxF;
    win[3] = 1.0F / v->attrib[FRAG_ATTRIB_WPOS][3];
 
-   color[0] = CHAN_TO_FLOAT(pv->color[0]);
-   color[1] = CHAN_TO_FLOAT(pv->color[1]);
-   color[2] = CHAN_TO_FLOAT(pv->color[2]);
-   color[3] = CHAN_TO_FLOAT(pv->color[3]);
-
    _mesa_feedback_vertex(ctx, win, color, v->attrib[FRAG_ATTRIB_CI][0], vtc);
 }
 
diff --git a/src/mesa/swrast/s_fragprog.c b/src/mesa/swrast/s_fragprog.c
index a36c1ba491..14c9868c18 100644
--- a/src/mesa/swrast/s_fragprog.c
+++ b/src/mesa/swrast/s_fragprog.c
@@ -22,11 +22,11 @@
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include "glheader.h"
-#include "colormac.h"
-#include "context.h"
-#include "prog_instruction.h"
-#include "texstate.h"
+#include "main/glheader.h"
+#include "main/colormac.h"
+#include "main/context.h"
+#include "main/texstate.h"
+#include "shader/prog_instruction.h"
 
 #include "s_fragprog.h"
 #include "s_span.h"
@@ -200,7 +200,6 @@ void
 _swrast_exec_fragment_program( GLcontext *ctx, SWspan *span )
 {
    const struct gl_fragment_program *program = ctx->FragmentProgram._Current;
-   GLuint i;
 
    /* incoming colors should be floats */
    if (program->Base.InputsRead & FRAG_BIT_COL0) {
@@ -209,23 +208,8 @@ _swrast_exec_fragment_program( GLcontext *ctx, SWspan *span )
 
    ctx->_CurrentProgram = GL_FRAGMENT_PROGRAM_ARB; /* or NV, doesn't matter */
 
-   for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
-      if (ctx->Texture.Unit[i]._Current != NULL) {
-         const GLboolean enable_shadow = ((1 << i) & program->Base.ShadowSamplers);
-         _mesa_update_texture_compare_function(ctx->Texture.Unit[i]._Current,
-                                               !enable_shadow);
-      }
-   }
-
    run_program(ctx, span, 0, span->end);
 
-   for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
-      if (ctx->Texture.Unit[i]._Current != NULL) {
-         _mesa_update_texture_compare_function(ctx->Texture.Unit[i]._Current,
-                                               GL_FALSE);
-      }
-   }
-
    if (program->Base.OutputsWritten & (1 << FRAG_RESULT_COLR)) {
       span->interpMask &= ~SPAN_RGBA;
       span->arrayMask |= SPAN_RGBA;
diff --git a/src/mesa/swrast/s_lines.c b/src/mesa/swrast/s_lines.c
index 781146e67f..15ef6233ed 100644
--- a/src/mesa/swrast/s_lines.c
+++ b/src/mesa/swrast/s_lines.c
@@ -63,12 +63,13 @@ compute_stipple_mask( GLcontext *ctx, GLuint len, GLubyte mask[] )
 static void
 draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor )
 {
-   GLint width, start;
+   const GLint width = (GLint) CLAMP(ctx->Line.Width,
+                                     ctx->Const.MinLineWidth,
+                                     ctx->Const.MaxLineWidth);
+   GLint start;
 
    ASSERT(span->end < MAX_WIDTH);
 
-   width = (GLint) CLAMP( ctx->Line._Width, MIN_LINE_WIDTH, MAX_LINE_WIDTH );
-
    if (width & 1)
       start = width / 2;
    else
@@ -143,7 +144,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor )
       span.arrayMask |= SPAN_MASK;				\
       compute_stipple_mask(ctx, span.end, span.array->mask);    \
    }								\
-   if (ctx->Line._Width > 1.0) {				\
+   if (ctx->Line.Width > 1.0) {					\
       draw_wide_line(ctx, &span, (GLboolean)(dx > dy));		\
    }								\
    else {							\
@@ -161,7 +162,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor )
       span.arrayMask |= SPAN_MASK;				\
       compute_stipple_mask(ctx, span.end, span.array->mask);	\
    }								\
-   if (ctx->Line._Width > 1.0) {				\
+   if (ctx->Line.Width > 1.0) {					\
       draw_wide_line(ctx, &span, (GLboolean)(dx > dy));		\
    }								\
    else {							\
@@ -180,7 +181,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor )
       span.arrayMask |= SPAN_MASK;				\
       compute_stipple_mask(ctx, span.end, span.array->mask);	\
    }								\
-   if (ctx->Line._Width > 1.0) {				\
+   if (ctx->Line.Width > 1.0) {					\
       draw_wide_line(ctx, &span, (GLboolean)(dx > dy));		\
    }								\
    else {							\
@@ -274,7 +275,7 @@ _swrast_choose_line( GLcontext *ctx )
          USE(general_line);
       }
       else if (ctx->Depth.Test
-               || ctx->Line._Width != 1.0
+               || ctx->Line.Width != 1.0
                || ctx->Line.StippleFlag) {
          /* no texture, but Z, fog, width>1, stipple, etc. */
          if (rgbmode)
@@ -284,6 +285,7 @@ _swrast_choose_line( GLcontext *ctx )
       }
       else {
          ASSERT(!ctx->Depth.Test);
+         ASSERT(ctx->Line.Width == 1.0);
          /* simple lines */
          if (rgbmode)
             USE(simple_no_z_rgba_line);
diff --git a/src/mesa/swrast/s_linetemp.h b/src/mesa/swrast/s_linetemp.h
index 55548f27b0..1accfc67e2 100644
--- a/src/mesa/swrast/s_linetemp.h
+++ b/src/mesa/swrast/s_linetemp.h
@@ -303,7 +303,10 @@ NAME( GLcontext *ctx, const SWvertex *vert0, const SWvertex *vert1 )
    }
 #endif
 
-   INIT_SPAN(span, GL_LINE, numPixels, interpFlags, SPAN_XY);
+   INIT_SPAN(span, GL_LINE);
+   span.end = numPixels;
+   span.interpMask = interpFlags;
+   span.arrayMask = SPAN_XY;
 
    /*
     * Draw
diff --git a/src/mesa/swrast/s_points.c b/src/mesa/swrast/s_points.c
index 02c9d9b425..4768fbea97 100644
--- a/src/mesa/swrast/s_points.c
+++ b/src/mesa/swrast/s_points.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.1
+ * Version:  7.1
  *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -34,134 +34,493 @@
 #include "s_span.h"
 
 
-#define RGBA       0x1
-#define INDEX      0x2
-#define SMOOTH     0x4
-#define ATTRIBS    0x8
-#define SPECULAR  0x10
-#define LARGE     0x20
-#define ATTENUATE 0x40
-#define SPRITE    0x80
-
-
-/*
- * CI points with size == 1.0
- */
-#define FLAGS (INDEX)
-#define NAME size1_ci_point
-#include "s_pointtemp.h"
-
-
-/*
- * General CI points.
+/**
+ * Used to cull points with invalid coords
  */
-#define FLAGS (INDEX | LARGE)
-#define NAME general_ci_point
-#include "s_pointtemp.h"
+#define CULL_INVALID(V)                              \
+   do {                                              \
+      float tmp = (V)->attrib[FRAG_ATTRIB_WPOS][0]   \
+                + (V)->attrib[FRAG_ATTRIB_WPOS][1];  \
+      if (IS_INF_OR_NAN(tmp))                        \
+	 return;                                     \
+   } while(0)
 
 
-/*
- * Antialiased CI points.
+/**
+ * Draw a point sprite
  */
-#define FLAGS (INDEX | SMOOTH)
-#define NAME antialiased_ci_point
-#include "s_pointtemp.h"
+static void
+sprite_point(GLcontext *ctx, const SWvertex *vert)
+{
+   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   SWspan span;
+   GLfloat size;
+   GLuint tCoords[MAX_TEXTURE_COORD_UNITS];
+   GLuint numTcoords = 0;
+   GLfloat t0, dtdy;
+
+   CULL_INVALID(vert);
+
+   /* z coord */
+   if (ctx->DrawBuffer->Visual.depthBits <= 16)
+      span.z = FloatToFixed(vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   else
+      span.z = (GLuint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   span.zStep = 0;
+
+   /* compute size */
+   if (ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) {
+      /* use vertex's point size */
+      /* first, clamp attenuated size to the user-specifed range */
+      size = CLAMP(vert->pointSize, ctx->Point.MinSize, ctx->Point.MaxSize);
+   }
+   else {
+      /* use constant point size */
+      size = ctx->Point.Size;
+   }
+   /* clamp to non-AA implementation limits */
+   size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize);
+
+   /* span init */
+   INIT_SPAN(span, GL_POINT);
+   span.interpMask = SPAN_Z | SPAN_RGBA;
+
+   span.red   = ChanToFixed(vert->color[0]);
+   span.green = ChanToFixed(vert->color[1]);
+   span.blue  = ChanToFixed(vert->color[2]);
+   span.alpha = ChanToFixed(vert->color[3]);
+   span.redStep = 0;
+   span.greenStep = 0;
+   span.blueStep = 0;
+   span.alphaStep = 0;
+
+   /* need these for fragment programs */
+   span.attrStart[FRAG_ATTRIB_WPOS][3] = 1.0F;
+   span.attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
+   span.attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
+
+   ATTRIB_LOOP_BEGIN
+      if (attr >= FRAG_ATTRIB_TEX0 && attr < FRAG_ATTRIB_VAR0) {
+         const GLuint u = attr - FRAG_ATTRIB_TEX0;
+         /* a texcoord */
+         if (ctx->Point.CoordReplace[u]) {
+            GLfloat s, r, dsdx;
+
+            s = 0.0;
+            dsdx = 1.0 / size;
+
+            if (ctx->Point.SpriteOrigin == GL_LOWER_LEFT) {
+               t0 = 0.0;
+               dtdy = 1.0 / size;
+            }
+            else {
+               /* GL_UPPER_LEFT */
+               t0 = 1.0;
+               dtdy = -1.0 / size;
+            }
+            tCoords[numTcoords++] = attr;
+
+            if (ctx->Point.SpriteRMode == GL_ZERO)
+               r = 0.0F;
+            else if (ctx->Point.SpriteRMode == GL_S)
+               r = vert->attrib[attr][0];
+            else /* GL_R */
+               r = vert->attrib[attr][2];
+
+            span.attrStart[attr][0] = s;
+            span.attrStart[attr][1] = 0.0; /* overwritten below */
+            span.attrStart[attr][2] = r;
+            span.attrStart[attr][3] = 1.0;
+
+            span.attrStepX[attr][0] = dsdx;
+            span.attrStepX[attr][1] = 0.0;
+            span.attrStepX[attr][2] = 0.0;
+            span.attrStepX[attr][3] = 0.0;
+
+            span.attrStepY[attr][0] = 0.0;
+            span.attrStepY[attr][1] = dtdy;
+            span.attrStepY[attr][2] = 0.0;
+            span.attrStepY[attr][3] = 0.0;
+
+            continue;
+         }
+      }
+      /* use vertex's texcoord/attrib */
+      COPY_4V(span.attrStart[attr], vert->attrib[attr]);
+      ASSIGN_4V(span.attrStepX[attr], 0, 0, 0, 0);
+      ASSIGN_4V(span.attrStepY[attr], 0, 0, 0, 0);
+   ATTRIB_LOOP_END
+
+   /* compute pos, bounds and render */
+   {
+      const GLfloat x = vert->attrib[FRAG_ATTRIB_WPOS][0];
+      const GLfloat y = vert->attrib[FRAG_ATTRIB_WPOS][1];
+      GLint iSize = (GLint) (size + 0.5F);
+      GLint xmin, xmax, ymin, ymax, iy;
+      GLint iRadius;
+      GLfloat tcoord = t0;
+
+      iSize = MAX2(1, iSize);
+      iRadius = iSize / 2;
+
+      if (iSize & 1) {
+         /* odd size */
+         xmin = (GLint) (x - iRadius);
+         xmax = (GLint) (x + iRadius);
+         ymin = (GLint) (y - iRadius);
+         ymax = (GLint) (y + iRadius);
+      }
+      else {
+         /* even size */
+         xmin = (GLint) x - iRadius + 1;
+         xmax = xmin + iSize - 1;
+         ymin = (GLint) y - iRadius + 1;
+         ymax = ymin + iSize - 1;
+      }
 
+      /* render spans */
+      for (iy = ymin; iy <= ymax; iy++) {
+         GLuint i;
+         /* setup texcoord T for this row */
+         for (i = 0; i < numTcoords; i++) {
+            span.attrStart[tCoords[i]][1] = tcoord;
+         }
 
-/*
- * Distance attenuated, general CI points.
- */
-#define FLAGS (INDEX | ATTENUATE)
-#define NAME atten_general_ci_point
-#include "s_pointtemp.h"
+         /* these might get changed by span clipping */
+         span.x = xmin;
+         span.y = iy;
+         span.end = xmax - xmin + 1;
 
+         _swrast_write_rgba_span(ctx, &span);
 
-/*
- * RGBA points with size == 1.0
- */
-#define FLAGS (RGBA)
-#define NAME size1_rgba_point
-#include "s_pointtemp.h"
+         tcoord += dtdy;
+      }
+   }
+}
 
 
-/*
- * General RGBA points.
+/**
+ * Draw smooth/antialiased point.  RGB or CI mode.
  */
-#define FLAGS (RGBA | LARGE)
-#define NAME general_rgba_point
-#include "s_pointtemp.h"
-
+static void
+smooth_point(GLcontext *ctx, const SWvertex *vert)
+{
+   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   const GLboolean ciMode = !ctx->Visual.rgbMode;
+   SWspan span;
+   GLfloat size, alphaAtten;
+
+   CULL_INVALID(vert);
+
+   /* z coord */
+   if (ctx->DrawBuffer->Visual.depthBits <= 16)
+      span.z = FloatToFixed(vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   else
+      span.z = (GLuint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   span.zStep = 0;
+
+   /* compute size */
+   if (ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) {
+      /* use vertex's point size */
+      /* first, clamp attenuated size to the user-specifed range */
+      size = CLAMP(vert->pointSize, ctx->Point.MinSize, ctx->Point.MaxSize);
+   }
+   else {
+      /* use constant point size */
+      size = ctx->Point.Size;
+   }
+   /* clamp to AA implementation limits */
+   size = CLAMP(size, ctx->Const.MinPointSizeAA, ctx->Const.MaxPointSizeAA);
 
-/*
- * Antialiased RGBA points.
- */
-#define FLAGS (RGBA | SMOOTH)
-#define NAME antialiased_rgba_point
-#include "s_pointtemp.h"
+   /* alpha attenuation / fade factor */
+   if (ctx->Multisample.Enabled) {
+      if (vert->pointSize >= ctx->Point.Threshold) {
+         alphaAtten = 1.0F;
+      }
+      else {
+         GLfloat dsize = vert->pointSize / ctx->Point.Threshold;
+         alphaAtten = dsize * dsize;
+      }
+   }
+   else {
+      alphaAtten = 1.0;
+   }
+   (void) alphaAtten; /* not used */
+
+   /* span init */
+   INIT_SPAN(span, GL_POINT);
+   span.interpMask = SPAN_Z | SPAN_RGBA;
+   span.arrayMask = SPAN_COVERAGE | SPAN_MASK;
+
+   span.red   = ChanToFixed(vert->color[0]);
+   span.green = ChanToFixed(vert->color[1]);
+   span.blue  = ChanToFixed(vert->color[2]);
+   span.alpha = ChanToFixed(vert->color[3]);
+   span.redStep = 0;
+   span.greenStep = 0;
+   span.blueStep = 0;
+   span.alphaStep = 0;
+
+   /* need these for fragment programs */
+   span.attrStart[FRAG_ATTRIB_WPOS][3] = 1.0F;
+   span.attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
+   span.attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
+
+   ATTRIB_LOOP_BEGIN
+      COPY_4V(span.attrStart[attr], vert->attrib[attr]);
+      ASSIGN_4V(span.attrStepX[attr], 0, 0, 0, 0);
+      ASSIGN_4V(span.attrStepY[attr], 0, 0, 0, 0);
+   ATTRIB_LOOP_END
+
+   /* compute pos, bounds and render */
+   {
+      const GLfloat x = vert->attrib[FRAG_ATTRIB_WPOS][0];
+      const GLfloat y = vert->attrib[FRAG_ATTRIB_WPOS][1];
+      const GLfloat radius = 0.5F * size;
+      const GLfloat rmin = radius - 0.7071F;  /* 0.7071 = sqrt(2)/2 */
+      const GLfloat rmax = radius + 0.7071F;
+      const GLfloat rmin2 = MAX2(0.0F, rmin * rmin);
+      const GLfloat rmax2 = rmax * rmax;
+      const GLfloat cscale = 1.0F / (rmax2 - rmin2);
+      const GLint xmin = (GLint) (x - radius);
+      const GLint xmax = (GLint) (x + radius);
+      const GLint ymin = (GLint) (y - radius);
+      const GLint ymax = (GLint) (y + radius);
+      GLint ix, iy;
+
+      for (iy = ymin; iy <= ymax; iy++) {
+
+         /* these might get changed by span clipping */
+         span.x = xmin;
+         span.y = iy;
+         span.end = xmax - xmin + 1;
+
+         /* compute coverage for each pixel in span */
+         for (ix = xmin; ix <= xmax; ix++) {
+            const GLfloat dx = ix - x + 0.5F;
+            const GLfloat dy = iy - y + 0.5F;
+            const GLfloat dist2 = dx * dx + dy * dy;
+            GLfloat coverage;
+
+            if (dist2 < rmax2) {
+               if (dist2 >= rmin2) {
+                  /* compute partial coverage */
+                  coverage = 1.0F - (dist2 - rmin2) * cscale;
+                  if (ciMode) {
+                     /* coverage in [0,15] */
+                     coverage *= 15.0;
+                  }
+               }
+               else {
+                  /* full coverage */
+                  coverage = 1.0F;
+               }
+               span.array->mask[ix - xmin] = 1;
+            }
+            else {
+               /* zero coverage - fragment outside the radius */
+               coverage = 0.0;
+               span.array->mask[ix - xmin] = 0;
+            }
+            span.array->coverage[ix - xmin] = coverage;
+         }
 
+         /* render span */
+         _swrast_write_rgba_span(ctx, &span);
 
-/*
- * Textured RGBA points.
- */
-#define FLAGS (RGBA | LARGE | ATTRIBS | SPECULAR)
-#define NAME textured_rgba_point
-#include "s_pointtemp.h"
+      }
+   }
+}
 
 
-/*
- * Antialiased points with texture mapping.
+/**
+ * Draw large (size >= 1) non-AA point.  RGB or CI mode.
  */
-#define FLAGS (RGBA | SMOOTH | ATTRIBS | SPECULAR)
-#define NAME antialiased_tex_rgba_point
-#include "s_pointtemp.h"
+static void
+large_point(GLcontext *ctx, const SWvertex *vert)
+{
+   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   const GLboolean ciMode = !ctx->Visual.rgbMode;
+   SWspan span;
+   GLfloat size;
+
+   CULL_INVALID(vert);
+
+   /* z coord */
+   if (ctx->DrawBuffer->Visual.depthBits <= 16)
+      span.z = FloatToFixed(vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   else
+      span.z = (GLuint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
+   span.zStep = 0;
+
+   /* compute size */
+   if (ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) {
+      /* use vertex's point size */
+      /* first, clamp attenuated size to the user-specifed range */
+      size = CLAMP(vert->pointSize, ctx->Point.MinSize, ctx->Point.MaxSize);
+   }
+   else {
+      /* use constant point size */
+      size = ctx->Point.Size;
+   }
+   /* clamp to non-AA implementation limits */
+   size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize);
 
+   /* span init */
+   INIT_SPAN(span, GL_POINT);
+   span.arrayMask = SPAN_XY;
 
-/*
- * Distance attenuated, general RGBA points.
- */
-#define FLAGS (RGBA | ATTENUATE)
-#define NAME atten_general_rgba_point
-#include "s_pointtemp.h"
+   if (ciMode) {
+      span.interpMask = SPAN_Z | SPAN_INDEX;
+      span.index = FloatToFixed(vert->attrib[FRAG_ATTRIB_CI][0]);
+      span.indexStep = 0;
+   }
+   else {
+      span.interpMask = SPAN_Z | SPAN_RGBA;
+      span.red   = ChanToFixed(vert->color[0]);
+      span.green = ChanToFixed(vert->color[1]);
+      span.blue  = ChanToFixed(vert->color[2]);
+      span.alpha = ChanToFixed(vert->color[3]);
+      span.redStep = 0;
+      span.greenStep = 0;
+      span.blueStep = 0;
+      span.alphaStep = 0;
+   }
 
+   /* need these for fragment programs */
+   span.attrStart[FRAG_ATTRIB_WPOS][3] = 1.0F;
+   span.attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
+   span.attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
+
+   ATTRIB_LOOP_BEGIN
+      COPY_4V(span.attrStart[attr], vert->attrib[attr]);
+      ASSIGN_4V(span.attrStepX[attr], 0, 0, 0, 0);
+      ASSIGN_4V(span.attrStepY[attr], 0, 0, 0, 0);
+   ATTRIB_LOOP_END
+
+   /* compute pos, bounds and render */
+   {
+      const GLfloat x = vert->attrib[FRAG_ATTRIB_WPOS][0];
+      const GLfloat y = vert->attrib[FRAG_ATTRIB_WPOS][1];
+      GLint iSize = (GLint) (size + 0.5F);
+      GLint xmin, xmax, ymin, ymax, ix, iy;
+      GLint iRadius;
+
+      iSize = MAX2(1, iSize);
+      iRadius = iSize / 2;
+
+      if (iSize & 1) {
+         /* odd size */
+         xmin = (GLint) (x - iRadius);
+         xmax = (GLint) (x + iRadius);
+         ymin = (GLint) (y - iRadius);
+         ymax = (GLint) (y + iRadius);
+      }
+      else {
+         /* even size */
+         xmin = (GLint) x - iRadius + 1;
+         xmax = xmin + iSize - 1;
+         ymin = (GLint) y - iRadius + 1;
+         ymax = ymin + iSize - 1;
+      }
 
-/*
- * Distance attenuated, textured RGBA points.
- */
-#define FLAGS (RGBA | ATTENUATE | ATTRIBS | SPECULAR)
-#define NAME atten_textured_rgba_point
-#include "s_pointtemp.h"
+      /* generate fragments */
+      span.end = 0;
+      for (iy = ymin; iy <= ymax; iy++) {
+         for (ix = xmin; ix <= xmax; ix++) {
+            span.array->x[span.end] = ix;
+            span.array->y[span.end] = iy;
+            span.end++;
+         }
+      }
+      assert(span.end <= MAX_WIDTH);
+      _swrast_write_rgba_span(ctx, &span);
+   }
+}
 
 
-/*
- * Distance attenuated, antialiased points with or without texture mapping.
+/**
+ * Draw size=1, single-pixel point
  */
-#define FLAGS (RGBA | ATTENUATE | ATTRIBS | SMOOTH)
-#define NAME atten_antialiased_rgba_point
-#include "s_pointtemp.h"
+static void
+pixel_point(GLcontext *ctx, const SWvertex *vert)
+{
+   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   const GLboolean ciMode = !ctx->Visual.rgbMode;
+   /*
+    * Note that unlike the other functions, we put single-pixel points
+    * into a special span array in order to render as many points as
+    * possible with a single _swrast_write_rgba_span() call.
+    */
+   SWspan *span = &(swrast->PointSpan);
+   GLuint count;
+
+   CULL_INVALID(vert);
+
+   /* Span init */
+   span->interpMask = 0;
+   span->arrayMask = SPAN_XY | SPAN_Z;
+   if (ciMode)
+      span->arrayMask |= SPAN_INDEX;
+   else
+      span->arrayMask |= SPAN_RGBA;
+   /*span->arrayMask |= SPAN_LAMBDA;*/
+   span->arrayAttribs = swrast->_ActiveAttribMask; /* we'll produce these vals */
+
+   /* need these for fragment programs */
+   span->attrStart[FRAG_ATTRIB_WPOS][3] = 1.0F;
+   span->attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
+   span->attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
+
+   /* check if we need to flush */
+   if (span->end >= MAX_WIDTH ||
+       (swrast->_RasterMask & (BLEND_BIT | LOGIC_OP_BIT | MASKING_BIT))) {
+      if (ciMode)
+         _swrast_write_index_span(ctx, span);
+      else
+         _swrast_write_rgba_span(ctx, span);
+      span->end = 0;
+   }
 
+   count = span->end;
 
-/*
- * Sprite (textured point)
- */
-#define FLAGS (RGBA | SPRITE | SPECULAR)
-#define NAME sprite_point
-#include "s_pointtemp.h"
+   /* fragment attributes */
+   if (ciMode) {
+      span->array->index[count] = (GLuint) vert->attrib[FRAG_ATTRIB_CI][0];
+   }
+   else {
+      span->array->rgba[count][RCOMP] = vert->color[0];
+      span->array->rgba[count][GCOMP] = vert->color[1];
+      span->array->rgba[count][BCOMP] = vert->color[2];
+      span->array->rgba[count][ACOMP] = vert->color[3];
+   }
+   ATTRIB_LOOP_BEGIN
+      COPY_4V(span->array->attribs[attr][count], vert->attrib[attr]);
+   ATTRIB_LOOP_END
 
+   /* fragment position */
+   span->array->x[count] = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][0];
+   span->array->y[count] = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][1];
+   span->array->z[count] = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
 
-#define FLAGS (RGBA | SPRITE | SPECULAR | ATTENUATE)
-#define NAME atten_sprite_point
-#include "s_pointtemp.h"
+   span->end = count + 1;
+   ASSERT(span->end <= MAX_WIDTH);
+}
 
 
+/**
+ * Add specular color to primary color, draw point, restore original
+ * primary color.
+ */
 void
 _swrast_add_spec_terms_point(GLcontext *ctx, const SWvertex *v0)
 {
-   SWvertex *ncv0 = (SWvertex *) v0;
+   SWvertex *ncv0 = (SWvertex *) v0; /* cast away const */
    GLfloat rSum, gSum, bSum;
    GLchan cSave[4];
 
    /* save */
-   COPY_CHAN4( cSave, ncv0->color );
+   COPY_CHAN4(cSave, ncv0->color);
    /* sum */
    rSum = CHAN_TO_FLOAT(ncv0->color[0]) + ncv0->attrib[FRAG_ATTRIB_COL1][0];
    gSum = CHAN_TO_FLOAT(ncv0->color[1]) + ncv0->attrib[FRAG_ATTRIB_COL1][1];
@@ -176,115 +535,35 @@ _swrast_add_spec_terms_point(GLcontext *ctx, const SWvertex *v0)
 }
 
 
-
-/* record the current point function name */
-#ifdef DEBUG
-
-static const char *pntFuncName = NULL;
-
-#define USE(pntFunc)                   \
-do {                                   \
-    pntFuncName = #pntFunc;            \
-    /*printf("%s\n", pntFuncName);*/   \
-    swrast->Point = pntFunc;           \
-} while (0)
-
-#else
-
-#define USE(pntFunc)  swrast->Point = pntFunc
-
-#endif
-
-
-/*
- * Examine the current context to determine which point drawing function
- * should be used.
+/**
+ * Examine current state to determine which point drawing function to use.
  */
 void
-_swrast_choose_point( GLcontext *ctx )
+_swrast_choose_point(GLcontext *ctx)
 {
    SWcontext *swrast = SWRAST_CONTEXT(ctx);
-   GLboolean rgbMode = ctx->Visual.rgbMode;
-   GLboolean specular = (ctx->Fog.ColorSumEnabled ||
-                         (ctx->Light.Enabled &&
-                          ctx->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR));
 
-   if (ctx->RenderMode==GL_RENDER) {
+   if (ctx->RenderMode == GL_RENDER) {
       if (ctx->Point.PointSprite) {
-         /* GL_ARB_point_sprite / GL_NV_point_sprite */
-         /* XXX this might not be good enough */
-         if (ctx->Point._Attenuated)
-            USE(atten_sprite_point);
-         else
-            USE(sprite_point);
+         swrast->Point = sprite_point;
       }
       else if (ctx->Point.SmoothFlag) {
-         /* Smooth points */
-         if (rgbMode) {
-            if (ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) {
-               USE(atten_antialiased_rgba_point);
-            }
-            else if (ctx->Texture._EnabledCoordUnits) {
-               USE(antialiased_tex_rgba_point);
-            }
-            else {
-               USE(antialiased_rgba_point);
-            }
-         }
-         else {
-            USE(antialiased_ci_point);
-         }
+         swrast->Point = smooth_point;
       }
-      else if (ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) {
-         if (rgbMode) {
-            if (ctx->Texture._EnabledCoordUnits) {
-               if (ctx->Point.SmoothFlag) {
-                  USE(atten_antialiased_rgba_point);
-               }
-               else {
-                  USE(atten_textured_rgba_point);
-               }
-            }
-            else {
-               USE(atten_general_rgba_point);
-            }
-         }
-         else {
-            /* ci, atten */
-            USE(atten_general_ci_point);
-         }
-      }
-      else if ((ctx->Texture._EnabledCoordUnits
-                || specular
-                || swrast->_FogEnabled) && rgbMode) {
-         /* textured, fogged */
-         USE(textured_rgba_point);
-      }
-      else if (ctx->Point._Size != 1.0) {
-         /* large points */
-         if (rgbMode) {
-            USE(general_rgba_point);
-         }
-         else {
-            USE(general_ci_point);
-         }
+      else if (ctx->Point.Size > 1.0 ||
+               ctx->Point._Attenuated ||
+               ctx->VertexProgram.PointSizeEnabled) {
+         swrast->Point = large_point;
       }
       else {
-         /* single pixel points */
-         if (rgbMode) {
-            assert((swrast->_ActiveAttribMask & FRAG_BIT_COL1) == 0);
-            USE(size1_rgba_point);
-         }
-         else {
-            USE(size1_ci_point);
-         }
+         swrast->Point = pixel_point;
       }
    }
-   else if (ctx->RenderMode==GL_FEEDBACK) {
-      USE(_swrast_feedback_point);
+   else if (ctx->RenderMode == GL_FEEDBACK) {
+      swrast->Point = _swrast_feedback_point;
    }
    else {
       /* GL_SELECT mode */
-      USE(_swrast_select_point);
+      swrast->Point = _swrast_select_point;
    }
 }
diff --git a/src/mesa/swrast/s_pointtemp.h b/src/mesa/swrast/s_pointtemp.h
deleted file mode 100644
index 206085b5b8..0000000000
--- a/src/mesa/swrast/s_pointtemp.h
+++ /dev/null
@@ -1,406 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.5.3
- *
- * 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.
- */
-
-/*
- * Regarding GL_NV_point_sprite:
- *
- * Portions of this software may use or implement intellectual
- * property owned and licensed by NVIDIA Corporation. NVIDIA disclaims
- * any and all warranties with respect to such intellectual property,
- * including any use thereof or modifications thereto.
- */
-
-
-/*
- * Point rendering template code.
- *
- * Set FLAGS = bitwise-OR of the following tokens:
- *
- *   RGBA = do rgba instead of color index
- *   SMOOTH = do antialiasing
- *   ATTRIBS = general attributes (texcoords, etc)
- *   LARGE = do points with diameter > 1 pixel
- *   ATTENUATE = compute point size attenuation
- *   SPRITE = GL_ARB_point_sprite / GL_NV_point_sprite
- *
- * Notes: LARGE and ATTENUATE are exclusive of each other.
- *        ATTRIBS requires RGBA
- */
-
-
-/*
- * NOTES on antialiased point rasterization:
- *
- * Let d = distance of fragment center from vertex.
- * if d < rmin2 then
- *    fragment has 100% coverage
- * else if d > rmax2 then
- *    fragment has 0% coverage
- * else
- *    fragment has % coverage = (d - rmin2) / (rmax2 - rmin2)
- */
-
-
-static void
-NAME ( GLcontext *ctx, const SWvertex *vert )
-{
-#if FLAGS & (ATTENUATE | LARGE | SMOOTH | SPRITE)
-   GLfloat size;
-#endif
-#if FLAGS & RGBA
-#if (FLAGS & ATTENUATE) && (FLAGS & SMOOTH)
-   GLfloat alphaAtten;
-#endif
-   const GLchan red   = vert->color[0];
-   const GLchan green = vert->color[1];
-   const GLchan blue  = vert->color[2];
-   const GLchan alpha = vert->color[3];
-#endif
-#if FLAGS & INDEX
-   const GLuint colorIndex = (GLuint) vert->attrib[FRAG_ATTRIB_CI][0]; /* XXX round? */
-#endif
-#if FLAGS & ATTRIBS
-   GLfloat attrib[FRAG_ATTRIB_MAX][4]; /* texture & varying */
-#endif
-   SWcontext *swrast = SWRAST_CONTEXT(ctx);
-   SWspan *span = &(swrast->PointSpan);
-
-   /*
-   printf("%s  %g %g %g %g\n", __FUNCTION__,
-          vert->attrib[FRAG_ATTRIB_COL1][0],
-          vert->attrib[FRAG_ATTRIB_COL1][1],
-          vert->attrib[FRAG_ATTRIB_COL1][2],
-          vert->attrib[FRAG_ATTRIB_COL1][3]);
-   if ( vert->attrib[FRAG_ATTRIB_COL1][0] == 0.0 &&
-        vert->attrib[FRAG_ATTRIB_COL1][1] == 1.0 &&
-        vert->attrib[FRAG_ATTRIB_COL1][2] == 0.0)
-      foo();
-   */
-
-   /* Cull primitives with malformed coordinates.
-    */
-   {
-      float tmp = vert->attrib[FRAG_ATTRIB_WPOS][0] + vert->attrib[FRAG_ATTRIB_WPOS][1];
-      if (IS_INF_OR_NAN(tmp))
-	 return;
-   }
-
-   /*
-    * Span init
-    */
-   span->interpMask = 0;
-   span->arrayMask = SPAN_XY | SPAN_Z;
-#if FLAGS & RGBA
-   span->arrayMask |= SPAN_RGBA;
-#endif
-#if FLAGS & INDEX
-   span->arrayMask |= SPAN_INDEX;
-#endif
-#if FLAGS & ATTRIBS
-   span->arrayMask |= SPAN_LAMBDA;
-
-   /* we're filling in the attrib arrays: */
-   span->arrayAttribs = swrast->_ActiveAttribMask;
-
-   ATTRIB_LOOP_BEGIN
-      COPY_4V(attrib[attr], vert->attrib[attr]);
-   ATTRIB_LOOP_END
-
-   /* need these for fragment programs */
-   span->attrStart[FRAG_ATTRIB_WPOS][3] = 1.0F;
-   span->attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
-   span->attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
-#else
-   assert((swrast->_ActiveAttribMask & FRAG_BIT_COL1) == 0);
-#endif
-
-#if FLAGS & SMOOTH
-   span->arrayMask |= SPAN_COVERAGE;
-#endif
-#if FLAGS & SPRITE
-   span->arrayMask |= SPAN_LAMBDA;
-#endif
-
-   /* Compute point size if not known to be one */
-#if FLAGS & ATTENUATE
-   /* first, clamp attenuated size to the user-specifed range */
-   size = CLAMP(vert->pointSize, ctx->Point.MinSize, ctx->Point.MaxSize);
-#if (FLAGS & RGBA) && (FLAGS & SMOOTH)
-   /* only if multisampling, compute the fade factor */
-   if (ctx->Multisample.Enabled) {
-      if (vert->pointSize >= ctx->Point.Threshold) {
-         alphaAtten = 1.0F;
-      }
-      else {
-         GLfloat dsize = vert->pointSize / ctx->Point.Threshold;
-         alphaAtten = dsize * dsize;
-      }
-   }
-   else {
-      alphaAtten = 1.0;
-   }
-#endif
-#elif FLAGS & (LARGE | SMOOTH | SPRITE)
-   /* constant, non-attenuated size */
-   size = ctx->Point._Size; /* this is already clamped */
-#endif
-
-
-#if FLAGS & (ATTENUATE | LARGE | SMOOTH | SPRITE)
-   /***
-    *** Multi-pixel points
-    ***/
-
-   /* do final clamping now */
-   if (ctx->Point.SmoothFlag) {
-      size = CLAMP(size, ctx->Const.MinPointSizeAA, ctx->Const.MaxPointSizeAA);
-   }
-   else {
-      size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize);
-   }
-
-   {{
-      GLint x, y;
-      const GLfloat radius = 0.5F * size;
-      const GLuint z = (GLuint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
-      GLuint count;
-#if FLAGS & SMOOTH
-      const GLfloat rmin = radius - 0.7071F;  /* 0.7071 = sqrt(2)/2 */
-      const GLfloat rmax = radius + 0.7071F;
-      const GLfloat rmin2 = MAX2(0.0F, rmin * rmin);
-      const GLfloat rmax2 = rmax * rmax;
-      const GLfloat cscale = 1.0F / (rmax2 - rmin2);
-      const GLint xmin = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][0] - radius);
-      const GLint xmax = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][0] + radius);
-      const GLint ymin = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][1] - radius);
-      const GLint ymax = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][1] + radius);
-#else
-      /* non-smooth */
-      GLint xmin, xmax, ymin, ymax;
-      GLint iSize = (GLint) (size + 0.5F);
-      GLint iRadius;
-      iSize = MAX2(1, iSize);
-      iRadius = iSize / 2;
-      if (iSize & 1) {
-         /* odd size */
-         xmin = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][0] - iRadius);
-         xmax = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][0] + iRadius);
-         ymin = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][1] - iRadius);
-         ymax = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][1] + iRadius);
-      }
-      else {
-         /* even size */
-         xmin = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][0] - iRadius + 1;
-         xmax = xmin + iSize - 1;
-         ymin = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][1] - iRadius + 1;
-         ymax = ymin + iSize - 1;
-      }
-#endif /*SMOOTH*/
-
-      /* check if we need to flush */
-      if (span->end + (xmax-xmin+1) * (ymax-ymin+1) >= MAX_WIDTH ||
-          (swrast->_RasterMask & (BLEND_BIT | LOGIC_OP_BIT | MASKING_BIT))) {
-         if (span->end > 0) {
-#if FLAGS & RGBA
-            _swrast_write_rgba_span(ctx, span);
-#else
-            _swrast_write_index_span(ctx, span);
-#endif
-            span->end = 0;
-         }
-      }
-
-      /*
-       * OK, generate fragments
-       */
-      count = span->end;
-      (void) radius;
-      for (y = ymin; y <= ymax; y++) {
-         /* check if we need to flush */
-         if (count + (xmax-xmin+1) >= MAX_WIDTH) {
-	     span->end = count;
-#if FLAGS & RGBA
-            _swrast_write_rgba_span(ctx, span);
-#else
-            _swrast_write_index_span(ctx, span);
-#endif
-            count = span->end = 0;
-         }
-         for (x = xmin; x <= xmax; x++) {
-#if FLAGS & SPRITE
-            GLuint u;
-#endif
-
-#if FLAGS & RGBA
-            span->array->rgba[count][RCOMP] = red;
-            span->array->rgba[count][GCOMP] = green;
-            span->array->rgba[count][BCOMP] = blue;
-            span->array->rgba[count][ACOMP] = alpha;
-#endif
-#if FLAGS & INDEX
-            span->array->index[count] = colorIndex;
-#endif
-#if FLAGS & ATTRIBS
-            ATTRIB_LOOP_BEGIN
-               COPY_4V(span->array->attribs[attr][count], attrib[attr]);
-            /**
-               if (attr < FRAG_ATTRIB_VAR0) {
-                  const GLuint u = attr - FRAG_ATTRIB_TEX0;
-                  span->array->lambda[u][count] = 0.0;
-               }
-            **/
-            ATTRIB_LOOP_END
-#endif
-
-#if FLAGS & SMOOTH
-            /* compute coverage */
-            {
-               const GLfloat dx = x - vert->attrib[FRAG_ATTRIB_WPOS][0] + 0.5F;
-               const GLfloat dy = y - vert->attrib[FRAG_ATTRIB_WPOS][1] + 0.5F;
-               const GLfloat dist2 = dx * dx + dy * dy;
-               if (dist2 < rmax2) {
-                  if (dist2 >= rmin2) {
-                     /* compute partial coverage */
-                     span->array->coverage[count] = 1.0F - (dist2 - rmin2) * cscale;
-#if FLAGS & INDEX
-                     /* coverage in [0,15] */
-                     span->array->coverage[count] *= 15.0;
-#endif
-                  }
-                  else {
-                     /* full coverage */
-                     span->array->coverage[count] = 1.0F;
-                  }
-
-                  span->array->x[count] = x;
-                  span->array->y[count] = y;
-                  span->array->z[count] = z;
-
-#if (FLAGS & ATTENUATE) && (FLAGS & RGBA)
-                  span->array->rgba[count][ACOMP] = (GLchan) (alpha * alphaAtten);
-#elif FLAGS & RGBA
-                  span->array->rgba[count][ACOMP] = alpha;
-#endif /*ATTENUATE*/
-                  count++;
-               } /*if*/
-            }
-
-#else /*SMOOTH*/
-
-            /* not smooth (square points) */
-            span->array->x[count] = x;
-            span->array->y[count] = y;
-            span->array->z[count] = z;
-
-#if FLAGS & SPRITE
-            for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
-               GLuint attr = FRAG_ATTRIB_TEX0 + u;
-               if (ctx->Texture.Unit[u]._ReallyEnabled) {
-                  if (ctx->Point.CoordReplace[u]) {
-                     GLfloat s = 0.5F + (x + 0.5F - vert->attrib[FRAG_ATTRIB_WPOS][0]) / size;
-                     GLfloat t, r;
-                     if (ctx->Point.SpriteOrigin == GL_LOWER_LEFT)
-                        t = 0.5F + (y + 0.5F - vert->attrib[FRAG_ATTRIB_WPOS][1]) / size;
-                     else /* GL_UPPER_LEFT */
-                        t = 0.5F - (y + 0.5F - vert->attrib[FRAG_ATTRIB_WPOS][1]) / size;
-                     if (ctx->Point.SpriteRMode == GL_ZERO)
-                        r = 0.0F;
-                     else if (ctx->Point.SpriteRMode == GL_S)
-                        r = vert->attrib[attr][0];
-                     else /* GL_R */
-                        r = vert->attrib[attr][2];
-                     span->array->attribs[attr][count][0] = s;
-                     span->array->attribs[attr][count][1] = t;
-                     span->array->attribs[attr][count][2] = r;
-                     span->array->attribs[attr][count][3] = 1.0F;
-                     span->array->lambda[u][count] = 0.0; /* XXX fix? */
-                  }
-                  else {
-                     COPY_4V(span->array->attribs[attr][count],
-                             vert->attrib[attr]);
-                  }
-               }
-            }
-#endif /*SPRITE*/
-
-            count++;  /* square point */
-
-#endif /*SMOOTH*/
-
-	 } /*for x*/
-      } /*for y*/
-      span->end = count;
-   }}
-
-#else /* LARGE | ATTENUATE | SMOOTH | SPRITE */
-
-   /***
-    *** Single-pixel points
-    ***/
-   {{
-      GLuint count;
-
-      /* check if we need to flush */
-      if (span->end >= MAX_WIDTH ||
-          (swrast->_RasterMask & (BLEND_BIT | LOGIC_OP_BIT | MASKING_BIT))) {
-#if FLAGS & RGBA
-         _swrast_write_rgba_span(ctx, span);
-#else
-         _swrast_write_index_span(ctx, span);
-#endif
-         span->end = 0;
-      }
-
-      count = span->end;
-
-#if FLAGS & RGBA
-      span->array->rgba[count][RCOMP] = red;
-      span->array->rgba[count][GCOMP] = green;
-      span->array->rgba[count][BCOMP] = blue;
-      span->array->rgba[count][ACOMP] = alpha;
-#endif
-#if FLAGS & INDEX
-      span->array->index[count] = colorIndex;
-#endif
-#if FLAGS & ATTRIBS
-      ATTRIB_LOOP_BEGIN
-         COPY_4V(span->array->attribs[attr][count], attribs[attr]);
-      ATTRIB_LOOP_END
-#endif
-
-      span->array->x[count] = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][0];
-      span->array->y[count] = (GLint) vert->attrib[FRAG_ATTRIB_WPOS][1];
-      span->array->z[count] = (GLint) (vert->attrib[FRAG_ATTRIB_WPOS][2] + 0.5F);
-
-      span->end = count + 1;
-   }}
-
-#endif /* LARGE || ATTENUATE || SMOOTH */
-
-   ASSERT(span->end <= MAX_WIDTH);
-}
-
-
-#undef FLAGS
-#undef NAME
diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c
index 90a3d5545f..f23272c2be 100644
--- a/src/mesa/swrast/s_span.c
+++ b/src/mesa/swrast/s_span.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.3
+ * Version:  7.1
  *
  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
@@ -171,6 +171,12 @@ interpolate_active_attribs(GLcontext *ctx, SWspan *span, GLbitfield attrMask)
 {
    const SWcontext *swrast = SWRAST_CONTEXT(ctx);
 
+   /*
+    * Don't overwrite existing array values, such as colors that may have
+    * been produced by glDraw/CopyPixels.
+    */
+   attrMask &= ~span->arrayAttribs;
+
    ATTRIB_LOOP_BEGIN
       if (attrMask & (1 << attr)) {
          const GLfloat dwdx = span->attrStepX[FRAG_ATTRIB_WPOS][3];
@@ -196,6 +202,7 @@ interpolate_active_attribs(GLcontext *ctx, SWspan *span, GLbitfield attrMask)
             v3 += dv3dx;
             w += dwdx;
          }
+         ASSERT((span->arrayAttribs & (1 << attr)) == 0);
          span->arrayAttribs |= (1 << attr);
       }
    ATTRIB_LOOP_END
@@ -1145,6 +1152,7 @@ convert_color_type(SWspan *span, GLenum newType, GLuint output)
                         span->end, span->array->mask);
 
    span->array->ChanType = newType;
+   span->array->rgba = dst;
 }
 
 
@@ -1169,10 +1177,15 @@ shade_texture_span(GLcontext *ctx, SWspan *span)
    if (ctx->FragmentProgram._Current ||
        ctx->ATIFragmentShader._Enabled) {
       /* programmable shading */
+      if (span->primitive == GL_BITMAP && span->array->ChanType != GL_FLOAT) {
+         convert_color_type(span, GL_FLOAT, 0);
+      }
+      if (span->primitive != GL_POINT) {
+         /* for points, we populated the arrays already */
+         interpolate_active_attribs(ctx, span, ~0);
+      }
       span->array->ChanType = GL_FLOAT;
 
-      interpolate_active_attribs(ctx, span, ~0);
-
       if (!(span->arrayMask & SPAN_Z))
          _swrast_span_interpolate_z (ctx, span);
 
@@ -1227,7 +1240,8 @@ _swrast_write_rgba_span( GLcontext *ctx, SWspan *span)
    const GLbitfield origInterpMask = span->interpMask;
    const GLbitfield origArrayMask = span->arrayMask;
    const GLbitfield origArrayAttribs = span->arrayAttribs;
-   const GLenum chanType = span->array->ChanType;
+   const GLenum origChanType = span->array->ChanType;
+   void * const origRgba = span->array->rgba;
    const GLboolean shader = (ctx->FragmentProgram._Current
                              || ctx->ATIFragmentShader._Enabled);
    const GLboolean shaderOrTexture = shader || ctx->Texture._EnabledUnits;
@@ -1449,7 +1463,8 @@ end:
    span->interpMask = origInterpMask;
    span->arrayMask = origArrayMask;
    span->arrayAttribs = origArrayAttribs;
-   span->array->ChanType = chanType;
+   span->array->ChanType = origChanType;
+   span->array->rgba = origRgba;
 }
 
 
diff --git a/src/mesa/swrast/s_span.h b/src/mesa/swrast/s_span.h
index 585cce91ee..512134db0f 100644
--- a/src/mesa/swrast/s_span.h
+++ b/src/mesa/swrast/s_span.h
@@ -159,15 +159,15 @@ typedef struct sw_span
 
 
 
-#define INIT_SPAN(S, PRIMITIVE, END, INTERP_MASK, ARRAY_MASK)	\
-do {								\
-   (S).primitive = (PRIMITIVE);					\
-   (S).interpMask = (INTERP_MASK);				\
-   (S).arrayMask = (ARRAY_MASK);				\
-   (S).arrayAttribs = 0x0;					\
-   (S).end = (END);						\
-   (S).facing = 0;						\
-   (S).array = SWRAST_CONTEXT(ctx)->SpanArrays;			\
+#define INIT_SPAN(S, PRIMITIVE)			\
+do {						\
+   (S).primitive = (PRIMITIVE);			\
+   (S).interpMask = 0x0;			\
+   (S).arrayMask = 0x0;				\
+   (S).arrayAttribs = 0x0;			\
+   (S).end = 0;					\
+   (S).facing = 0;				\
+   (S).array = SWRAST_CONTEXT(ctx)->SpanArrays;	\
 } while (0)
 
 
diff --git a/src/mesa/swrast/s_stencil.c b/src/mesa/swrast/s_stencil.c
index 89991fad02..d0cbdd6917 100644
--- a/src/mesa/swrast/s_stencil.c
+++ b/src/mesa/swrast/s_stencil.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5
+ * Version:  7.1
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -392,6 +392,23 @@ do_stencil_test( GLcontext *ctx, GLuint face, GLuint n, GLstencil stencil[],
 }
 
 
+/**
+ * Compute the zpass/zfail masks by comparing the pre- and post-depth test
+ * masks.
+ */
+static INLINE void
+compute_pass_fail_masks(GLuint n, const GLubyte origMask[],
+                        const GLubyte newMask[],
+                        GLubyte passMask[], GLubyte failMask[])
+{
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      ASSERT(newMask[i] == 0 || newMask[i] == 1);
+      passMask[i] = origMask[i] & newMask[i];
+      failMask[i] = origMask[i] & (newMask[i] ^ 1);
+   }
+}
+
 
 /**
  * Apply stencil and depth testing to the span of pixels.
@@ -460,39 +477,24 @@ stencil_and_ztest_span(GLcontext *ctx, SWspan *span, GLuint face)
       /*
        * Perform depth buffering, then apply zpass or zfail stencil function.
        */
-      GLubyte passmask[MAX_WIDTH], failmask[MAX_WIDTH], oldmask[MAX_WIDTH];
-      GLuint i;
+      GLubyte passMask[MAX_WIDTH], failMask[MAX_WIDTH], origMask[MAX_WIDTH];
 
       /* save the current mask bits */
-      _mesa_memcpy(oldmask, mask, n * sizeof(GLubyte));
+      _mesa_memcpy(origMask, mask, n * sizeof(GLubyte));
 
       /* apply the depth test */
       _swrast_depth_test_span(ctx, span);
 
-      /* Set the stencil pass/fail flags according to result of depth testing.
-       * if oldmask[i] == 0 then
-       *    Don't touch the stencil value
-       * else if oldmask[i] and newmask[i] then
-       *    Depth test passed
-       * else
-       *    assert(oldmask[i] && !newmask[i])
-       *    Depth test failed
-       * endif
-       */
-      for (i=0;i<n;i++) {
-         ASSERT(mask[i] == 0 || mask[i] == 1);
-         passmask[i] = oldmask[i] & mask[i];
-         failmask[i] = oldmask[i] & (mask[i] ^ 1);
-      }
+      compute_pass_fail_masks(n, origMask, mask, passMask, failMask);
 
       /* apply the pass and fail operations */
       if (ctx->Stencil.ZFailFunc[face] != GL_KEEP) {
          apply_stencil_op( ctx, ctx->Stencil.ZFailFunc[face], face,
-                           n, stencil, failmask );
+                           n, stencil, failMask );
       }
       if (ctx->Stencil.ZPassFunc[face] != GL_KEEP) {
          apply_stencil_op( ctx, ctx->Stencil.ZPassFunc[face], face,
-                           n, stencil, passmask );
+                           n, stencil, passMask );
       }
    }
 
@@ -902,6 +904,7 @@ stencil_test_pixels( GLcontext *ctx, GLuint face, GLuint n,
 static GLboolean
 stencil_and_ztest_pixels( GLcontext *ctx, SWspan *span, GLuint face )
 {
+   GLubyte passMask[MAX_WIDTH], failMask[MAX_WIDTH], origMask[MAX_WIDTH];
    struct gl_framebuffer *fb = ctx->DrawBuffer;
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
    const GLuint n = span->end;
@@ -916,13 +919,10 @@ stencil_and_ztest_pixels( GLcontext *ctx, SWspan *span, GLuint face )
    if (!rb->GetPointer(ctx, rb, 0, 0)) {
       /* No direct access */
       GLstencil stencil[MAX_WIDTH];
-      GLubyte origMask[MAX_WIDTH];
 
       ASSERT(rb->DataType == GL_UNSIGNED_BYTE);
       _swrast_get_values(ctx, rb, n, x, y, stencil, sizeof(GLubyte));
 
-      _mesa_memcpy(origMask, mask, n * sizeof(GLubyte));
-
       (void) do_stencil_test(ctx, face, n, stencil, mask);
 
       if (ctx->Depth.Test == GL_FALSE) {
@@ -930,27 +930,19 @@ stencil_and_ztest_pixels( GLcontext *ctx, SWspan *span, GLuint face )
                           n, stencil, mask);
       }
       else {
+         _mesa_memcpy(origMask, mask, n * sizeof(GLubyte));
+
          _swrast_depth_test_span(ctx, span);
 
+         compute_pass_fail_masks(n, origMask, mask, passMask, failMask);
+
          if (ctx->Stencil.ZFailFunc[face] != GL_KEEP) {
-            GLubyte failmask[MAX_WIDTH];
-            GLuint i;
-            for (i = 0; i < n; i++) {
-               ASSERT(mask[i] == 0 || mask[i] == 1);
-               failmask[i] = origMask[i] & (mask[i] ^ 1);
-            }
             apply_stencil_op(ctx, ctx->Stencil.ZFailFunc[face], face,
-                             n, stencil, failmask);
+                             n, stencil, failMask);
          }
          if (ctx->Stencil.ZPassFunc[face] != GL_KEEP) {
-            GLubyte passmask[MAX_WIDTH];
-            GLuint i;
-            for (i = 0; i < n; i++) {
-               ASSERT(mask[i] == 0 || mask[i] == 1);
-               passmask[i] = origMask[i] & mask[i];
-            }
             apply_stencil_op(ctx, ctx->Stencil.ZPassFunc[face], face,
-                             n, stencil, passmask);
+                             n, stencil, passMask);
          }
       }
 
@@ -972,28 +964,21 @@ stencil_and_ztest_pixels( GLcontext *ctx, SWspan *span, GLuint face )
                                     ctx->Stencil.ZPassFunc[face], face, mask);
       }
       else {
-         GLubyte passmask[MAX_WIDTH], failmask[MAX_WIDTH], oldmask[MAX_WIDTH];
-         GLuint i;
-
-         _mesa_memcpy(oldmask, mask, n * sizeof(GLubyte));
+         _mesa_memcpy(origMask, mask, n * sizeof(GLubyte));
 
          _swrast_depth_test_span(ctx, span);
 
-         for (i=0;i<n;i++) {
-            ASSERT(mask[i] == 0 || mask[i] == 1);
-            passmask[i] = oldmask[i] & mask[i];
-            failmask[i] = oldmask[i] & (mask[i] ^ 1);
-         }
+         compute_pass_fail_masks(n, origMask, mask, passMask, failMask);
 
          if (ctx->Stencil.ZFailFunc[face] != GL_KEEP) {
             apply_stencil_op_to_pixels(ctx, n, x, y,
                                        ctx->Stencil.ZFailFunc[face],
-                                       face, failmask);
+                                       face, failMask);
          }
          if (ctx->Stencil.ZPassFunc[face] != GL_KEEP) {
             apply_stencil_op_to_pixels(ctx, n, x, y,
                                        ctx->Stencil.ZPassFunc[face],
-                                       face, passmask);
+                                       face, passMask);
          }
       }
 
diff --git a/src/mesa/swrast/s_texfilter.c b/src/mesa/swrast/s_texfilter.c
index d4516f6faa..c2a7512388 100644
--- a/src/mesa/swrast/s_texfilter.c
+++ b/src/mesa/swrast/s_texfilter.c
@@ -3347,7 +3347,7 @@ texture_sample_func
 _swrast_choose_texture_sample_func( GLcontext *ctx,
 				    const struct gl_texture_object *t )
 {
-   if (!t || !t->Complete) {
+   if (!t || !t->_Complete) {
       return &null_sample_func;
    }
    else {
diff --git a/src/mesa/swrast/s_tritemp.h b/src/mesa/swrast/s_tritemp.h
index 2a90ffd85f..cded4a6c1c 100644
--- a/src/mesa/swrast/s_tritemp.h
+++ b/src/mesa/swrast/s_tritemp.h
@@ -144,7 +144,7 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
 
    (void) swrast;
 
-   INIT_SPAN(span, GL_POLYGON, 0, 0, 0);
+   INIT_SPAN(span, GL_POLYGON);
    span.y = 0; /* silence warnings */
 
 #ifdef INTERP_Z
@@ -393,7 +393,7 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
 	 span.greenStep = 0;
 	 span.blueStep  = 0;
 #  ifdef INTERP_ALPHA
-         span.attrStepX[FRAG_ATTRIB_COL0][3] = span.attrStepX[FRAG_ATTRIB_COL0][3] = 0.0F;
+         span.attrStepX[FRAG_ATTRIB_COL0][3] = span.attrStepY[FRAG_ATTRIB_COL0][3] = 0.0F;
 	 span.alphaStep = 0;
 #  endif
       }
diff --git a/src/mesa/swrast/s_zoom.c b/src/mesa/swrast/s_zoom.c
index 78fa137d3f..9f1a4c6f0a 100644
--- a/src/mesa/swrast/s_zoom.c
+++ b/src/mesa/swrast/s_zoom.c
@@ -148,26 +148,25 @@ zoom_span( GLcontext *ctx, GLint imgX, GLint imgY, const SWspan *span,
    ASSERT((span->arrayMask & SPAN_XY) == 0);
    ASSERT(span->primitive == GL_BITMAP);
 
-   INIT_SPAN(zoomed, GL_BITMAP, 0, 0, 0);
+   INIT_SPAN(zoomed, GL_BITMAP);
    zoomed.x = x0;
    zoomed.end = zoomedWidth;
    zoomed.array = &zoomed_arrays;
    zoomed_arrays.ChanType = span->array->ChanType;
-   /* XXX temporary */
-#if CHAN_TYPE == GL_UNSIGNED_BYTE
-   zoomed_arrays.rgba = zoomed_arrays.rgba8;
-#elif CHAN_TYPE == GL_UNSIGNED_SHORT
-   zoomed_arrays.rgba = zoomed_arrays.rgba16;
-#else
-   zoomed_arrays.rgba = zoomed_arrays.attribs[FRAG_ATTRIB_COL0];
-#endif
+   if (zoomed_arrays.ChanType == GL_UNSIGNED_BYTE)
+      zoomed_arrays.rgba = (GLchan (*)[4]) zoomed_arrays.rgba8;
+   else if (zoomed_arrays.ChanType == GL_UNSIGNED_SHORT)
+      zoomed_arrays.rgba = (GLchan (*)[4]) zoomed_arrays.rgba16;
+   else
+      zoomed_arrays.rgba = (GLchan (*)[4]) zoomed_arrays.attribs[FRAG_ATTRIB_COL0];
 
+   COPY_4V(zoomed.attrStart[FRAG_ATTRIB_WPOS], span->attrStart[FRAG_ATTRIB_WPOS]);
+   COPY_4V(zoomed.attrStepX[FRAG_ATTRIB_WPOS], span->attrStepX[FRAG_ATTRIB_WPOS]);
+   COPY_4V(zoomed.attrStepY[FRAG_ATTRIB_WPOS], span->attrStepY[FRAG_ATTRIB_WPOS]);
 
-   /* copy fog interp info */
    zoomed.attrStart[FRAG_ATTRIB_FOGC][0] = span->attrStart[FRAG_ATTRIB_FOGC][0];
    zoomed.attrStepX[FRAG_ATTRIB_FOGC][0] = span->attrStepX[FRAG_ATTRIB_FOGC][0];
    zoomed.attrStepY[FRAG_ATTRIB_FOGC][0] = span->attrStepY[FRAG_ATTRIB_FOGC][0];
-   /* XXX copy texcoord info? */
 
    if (format == GL_RGBA || format == GL_RGB) {
       /* copy Z info */
@@ -176,6 +175,7 @@ zoom_span( GLcontext *ctx, GLint imgX, GLint imgY, const SWspan *span,
       /* we'll generate an array of colorss */
       zoomed.interpMask = span->interpMask & ~SPAN_RGBA;
       zoomed.arrayMask |= SPAN_RGBA;
+      zoomed.arrayAttribs |= FRAG_BIT_COL0;  /* we'll produce these values */
       ASSERT(span->arrayMask & SPAN_RGBA);
    }
    else if (format == GL_COLOR_INDEX) {
diff --git a/src/mesa/swrast/swrast.h b/src/mesa/swrast/swrast.h
index d101a9e2ae..85a27fd55b 100644
--- a/src/mesa/swrast/swrast.h
+++ b/src/mesa/swrast/swrast.h
@@ -32,7 +32,7 @@
 #ifndef SWRAST_H
 #define SWRAST_H
 
-#include "mtypes.h"
+#include "main/mtypes.h"
 
 /**
  * \struct SWvertex
diff --git a/src/mesa/swrast_setup/ss_context.c b/src/mesa/swrast_setup/ss_context.c
index cd4ac57d37..a9c7d941e5 100644
--- a/src/mesa/swrast_setup/ss_context.c
+++ b/src/mesa/swrast_setup/ss_context.c
@@ -125,6 +125,7 @@ setup_vertex_format(GLcontext *ctx)
       if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR0 )) {
          swsetup->intColors = !ctx->FragmentProgram._Current
                            && !ctx->ATIFragmentShader._Enabled
+                           && ctx->RenderMode == GL_RENDER
                            && CHAN_TYPE == GL_UNSIGNED_BYTE;
          if (swsetup->intColors)
             EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4CHAN_4F_RGBA, color );
@@ -279,28 +280,29 @@ _swsetup_Translate( GLcontext *ctx, const void *vertex, SWvertex *dest )
 
    /** XXX try to limit these loops someday */
    for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++)
-      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_TEX0+i,
+      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_TEX0 + i,
                      dest->attrib[FRAG_ATTRIB_TEX0 + i] );
 
    for (i = 0 ; i < ctx->Const.MaxVarying ; i++)
-      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_GENERIC0+i,
+      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_GENERIC0 + i,
                      dest->attrib[FRAG_ATTRIB_VAR0 + i] );
 
-   _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR0, tmp );
-   UNCLAMPED_FLOAT_TO_RGBA_CHAN( dest->color, tmp );
+   if (ctx->Visual.rgbMode) {
+      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR0,
+                     dest->attrib[FRAG_ATTRIB_COL0] );
+      UNCLAMPED_FLOAT_TO_RGBA_CHAN( dest->color, tmp );
 
-   _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR1, tmp );
-   COPY_4V(dest->attrib[FRAG_ATTRIB_COL1], tmp);
-   /*
-   UNCLAMPED_FLOAT_TO_RGBA_CHAN( dest->specular, tmp );
-   */
+      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR1,
+                     dest->attrib[FRAG_ATTRIB_COL1]);
+   }
+   else {
+      _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR_INDEX, tmp );
+      dest->attrib[FRAG_ATTRIB_CI][0] = tmp[0];
+   }
 
    _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_FOG, tmp );
    dest->attrib[FRAG_ATTRIB_FOGC][0] = tmp[0];
 
-   _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_COLOR_INDEX, tmp );
-   dest->attrib[FRAG_ATTRIB_CI][0] = tmp[0];
-
    /* XXX See _tnl_get_attr about pointsize ... */
    _tnl_get_attr( ctx, vertex, _TNL_ATTRIB_POINTSIZE, tmp );
    dest->pointSize = tmp[0];
diff --git a/src/mesa/swrast_setup/ss_tritmp.h b/src/mesa/swrast_setup/ss_tritmp.h
index c14468e951..f6b738d60d 100644
--- a/src/mesa/swrast_setup/ss_tritmp.h
+++ b/src/mesa/swrast_setup/ss_tritmp.h
@@ -37,7 +37,7 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
    SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts;
    SWvertex *v[3];
    GLfloat z[3];
-   GLfloat offset;
+   GLfloat offset, oz0, oz1, oz2;
    GLenum mode = GL_FILL;
    GLuint facing = 0;
    GLchan saved_color[3][4];
@@ -142,12 +142,16 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 	 }
       }
 
-      if (IND & SS_OFFSET_BIT)
-      {
-	 offset = ctx->Polygon.OffsetUnits * ctx->DrawBuffer->_MRD;
+      if (IND & SS_OFFSET_BIT) {
+         const GLfloat max = ctx->DrawBuffer->_DepthMaxF;
+         /* save original Z values (restored later) */
 	 z[0] = v[0]->attrib[FRAG_ATTRIB_WPOS][2];
 	 z[1] = v[1]->attrib[FRAG_ATTRIB_WPOS][2];
 	 z[2] = v[2]->attrib[FRAG_ATTRIB_WPOS][2];
+         /* Note that Z values are already scaled to [0,65535] (for example)
+          * so no MRD value is used here.
+          */
+	 offset = ctx->Polygon.OffsetUnits;
 	 if (cc * cc > 1e-16) {
 	    const GLfloat ez = z[0] - z[2];
 	    const GLfloat fz = z[1] - z[2];
@@ -155,35 +159,33 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 	    const GLfloat dzdx = FABSF((ey * fz - ez * fy) * oneOverArea);
 	    const GLfloat dzdy = FABSF((ez * fx - ex * fz) * oneOverArea);
 	    offset += MAX2(dzdx, dzdy) * ctx->Polygon.OffsetFactor;
-            /* Unfortunately, we need to clamp to prevent negative Zs below.
-             * Technically, we should do the clamping per-fragment.
-             */
-            offset = MAX2(offset, -v[0]->attrib[FRAG_ATTRIB_WPOS][2]);
-            offset = MAX2(offset, -v[1]->attrib[FRAG_ATTRIB_WPOS][2]);
-            offset = MAX2(offset, -v[2]->attrib[FRAG_ATTRIB_WPOS][2]);
 	 }
+         /* new Z values */
+         oz0 = CLAMP(v[0]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0, max);
+         oz1 = CLAMP(v[1]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0, max);
+         oz2 = CLAMP(v[2]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0, max);
       }
    }
 
    if (mode == GL_POINT) {
       if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetPoint) {
-	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
+	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0;
+	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1;
+	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2;
       }
       _swsetup_render_point_tri( ctx, e0, e1, e2, facing );
    } else if (mode == GL_LINE) {
       if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetLine) {
-	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
+	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0;
+	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1;
+	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2;
       }
       _swsetup_render_line_tri( ctx, e0, e1, e2, facing );
    } else {
       if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetFill) {
-	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
-	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] += offset;
+	 v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0;
+	 v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1;
+	 v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2;
       }
       _swrast_Triangle( ctx, v[0], v[1], v[2] );
    }
diff --git a/src/mesa/tnl/t_context.c b/src/mesa/tnl/t_context.c
index 3017c73cf1..3b8dd18bbb 100644
--- a/src/mesa/tnl/t_context.c
+++ b/src/mesa/tnl/t_context.c
@@ -26,12 +26,12 @@
  */
 
 
-#include "glheader.h"
-#include "imports.h"
-#include "context.h"
-#include "macros.h"
-#include "mtypes.h"
-#include "light.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "main/mtypes.h"
+#include "main/light.h"
 
 #include "tnl.h"
 #include "t_context.h"
diff --git a/src/mesa/tnl/t_context.h b/src/mesa/tnl/t_context.h
index 31b89aca41..baf283ef0f 100644
--- a/src/mesa/tnl/t_context.h
+++ b/src/mesa/tnl/t_context.h
@@ -49,8 +49,8 @@
 #ifndef _T_CONTEXT_H
 #define _T_CONTEXT_H
 
-#include "glheader.h"
-#include "mtypes.h"
+#include "main/glheader.h"
+#include "main/mtypes.h"
 
 #include "math/m_matrix.h"
 #include "math/m_vector.h"
diff --git a/src/mesa/tnl/t_pipeline.c b/src/mesa/tnl/t_pipeline.c
index c7188da34a..2a0ed8852a 100644
--- a/src/mesa/tnl/t_pipeline.c
+++ b/src/mesa/tnl/t_pipeline.c
@@ -25,11 +25,11 @@
  *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "glheader.h"
-#include "context.h"
-#include "imports.h"
-#include "state.h"
-#include "mtypes.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/imports.h"
+#include "main/state.h"
+#include "main/mtypes.h"
 
 #include "t_context.h"
 #include "t_pipeline.h"
diff --git a/src/mesa/tnl/t_save_api.c b/src/mesa/tnl/t_save_api.c
deleted file mode 100644
index b08f05374e..0000000000
--- a/src/mesa/tnl/t_save_api.c
+++ /dev/null
@@ -1,1834 +0,0 @@
-/**************************************************************************
-
-Copyright 2002 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
-on 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 THEIR 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.
-
-**************************************************************************/
-
-/*
- * Authors:
- *   Keith Whitwell <keith@tungstengraphics.com>
- */
-
-
-
-/**
- * The display list compiler attempts to store lists of vertices with the
- * same vertex layout.  Additionally it attempts to minimize the need
- * for execute-time fixup of these vertex lists, allowing them to be
- * cached on hardware.
- *
- * There are still some circumstances where this can be thwarted, for
- * example by building a list that consists of one very long primitive
- * (eg Begin(Triangles), 1000 vertices, End), and calling that list
- * from inside a different begin/end object (Begin(Lines), CallList,
- * End).  
- *
- * In that case the code will have to replay the list as individual
- * commands through the Exec dispatch table, or fix up the copied
- * vertices at execute-time.
- *
- * The other case where fixup is required is when a vertex attribute
- * is introduced in the middle of a primitive.  Eg:
- *  Begin(Lines)
- *  TexCoord1f()           Vertex2f()
- *  TexCoord1f() Color3f() Vertex2f()
- *  End()
- *
- *  If the current value of Color isn't known at compile-time, this
- *  primitive will require fixup.
- *
- *
- * The list compiler currently doesn't attempt to compile lists
- * containing EvalCoord or EvalPoint commands.  On encountering one of
- * these, compilation falls back to opcodes.  
- *
- * This could be improved to fallback only when a mix of EvalCoord and
- * Vertex commands are issued within a single primitive.
- */
-
-
-#include "glheader.h"
-#include "context.h"
-#include "dlist.h"
-#include "enums.h"
-#include "macros.h"
-#include "api_validate.h"
-#include "api_arrayelt.h"
-#include "vtxfmt.h"
-#include "t_save_api.h"
-#include "dispatch.h"
-
-/*
- * NOTE: Old 'parity' issue is gone, but copying can still be
- * wrong-footed on replay.
- */
-static GLuint _save_copy_vertices( GLcontext *ctx, 
-				   const struct tnl_vertex_list *node )
-{
-   TNLcontext *tnl = TNL_CONTEXT( ctx );
-   const struct tnl_prim *prim = &node->prim[node->prim_count-1];
-   GLuint nr = prim->count;
-   GLuint sz = tnl->save.vertex_size;
-   const GLfloat *src = node->buffer + prim->start * sz;
-   GLfloat *dst = tnl->save.copied.buffer;
-   GLuint ovf, i;
-
-   if (prim->mode & PRIM_END)
-      return 0;
-	 
-   switch( prim->mode & PRIM_MODE_MASK )
-   {
-   case GL_POINTS:
-      return 0;
-   case GL_LINES:
-      ovf = nr&1;
-      for (i = 0 ; i < ovf ; i++)
-	 _mesa_memcpy( dst+i*sz, src+(nr-ovf+i)*sz, sz*sizeof(GLfloat) );
-      return i;
-   case GL_TRIANGLES:
-      ovf = nr%3;
-      for (i = 0 ; i < ovf ; i++)
-	 _mesa_memcpy( dst+i*sz, src+(nr-ovf+i)*sz, sz*sizeof(GLfloat) );
-      return i;
-   case GL_QUADS:
-      ovf = nr&3;
-      for (i = 0 ; i < ovf ; i++)
-	 _mesa_memcpy( dst+i*sz, src+(nr-ovf+i)*sz, sz*sizeof(GLfloat) );
-      return i;
-   case GL_LINE_STRIP:
-      if (nr == 0) 
-	 return 0;
-      else {
-	 _mesa_memcpy( dst, src+(nr-1)*sz, sz*sizeof(GLfloat) );
-	 return 1;
-      }
-   case GL_LINE_LOOP:
-   case GL_TRIANGLE_FAN:
-   case GL_POLYGON:
-      if (nr == 0) 
-	 return 0;
-      else if (nr == 1) {
-	 _mesa_memcpy( dst, src+0, sz*sizeof(GLfloat) );
-	 return 1;
-      } else {
-	 _mesa_memcpy( dst, src+0, sz*sizeof(GLfloat) );
-	 _mesa_memcpy( dst+sz, src+(nr-1)*sz, sz*sizeof(GLfloat) );
-	 return 2;
-      }
-   case GL_TRIANGLE_STRIP:
-   case GL_QUAD_STRIP:
-      switch (nr) {
-      case 0: ovf = 0; break;
-      case 1: ovf = 1; break;
-      default: ovf = 2 + (nr&1); break;
-      }
-      for (i = 0 ; i < ovf ; i++)
-	 _mesa_memcpy( dst+i*sz, src+(nr-ovf+i)*sz, sz*sizeof(GLfloat) );
-      return i;
-   default:
-      assert(0);
-      return 0;
-   }
-}
-
-
-static void
-build_normal_lengths( struct tnl_vertex_list *node )
-{
-   GLuint i;
-   GLfloat *len;
-   GLfloat *n = node->buffer;
-   GLuint stride = node->vertex_size;
-   GLuint count = node->count;
-
-   len = node->normal_lengths = (GLfloat *) MALLOC( count * sizeof(GLfloat) );
-   if (!len)
-      return;
-
-   /* Find the normal of the first vertex:
-    */
-   for (i = 0 ; i < _TNL_ATTRIB_NORMAL ; i++) 
-      n += node->attrsz[i];
-
-   for (i = 0 ; i < count ; i++, n += stride) {
-      len[i] = LEN_3FV( n );
-      if (len[i] > 0.0F) len[i] = 1.0F / len[i];
-   } 
-}
-
-static struct tnl_vertex_store *alloc_vertex_store( GLcontext *ctx )
-{
-   struct tnl_vertex_store *store = MALLOC_STRUCT(tnl_vertex_store);
-   (void) ctx;
-   store->used = 0;
-   store->refcount = 1;
-   return store;
-}
-
-static struct tnl_primitive_store *alloc_prim_store( GLcontext *ctx )
-{
-   struct tnl_primitive_store *store = MALLOC_STRUCT(tnl_primitive_store);
-   (void) ctx;
-   store->used = 0;
-   store->refcount = 1;
-   return store;
-}
-
-static void _save_reset_counters( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   tnl->save.prim = tnl->save.prim_store->buffer + tnl->save.prim_store->used;
-   tnl->save.buffer = (tnl->save.vertex_store->buffer + 
-		       tnl->save.vertex_store->used);
-
-   if (tnl->save.vertex_size)
-      tnl->save.initial_counter = ((SAVE_BUFFER_SIZE - 
-				    tnl->save.vertex_store->used) / 
-				   tnl->save.vertex_size);
-   else
-      tnl->save.initial_counter = 0;
-
-   if (tnl->save.initial_counter > ctx->Const.MaxArrayLockSize )
-      tnl->save.initial_counter = ctx->Const.MaxArrayLockSize;
-
-   tnl->save.counter = tnl->save.initial_counter;
-   tnl->save.prim_count = 0;
-   tnl->save.prim_max = SAVE_PRIM_SIZE - tnl->save.prim_store->used;
-   tnl->save.copied.nr = 0;
-   tnl->save.dangling_attr_ref = 0;
-}
-
-
-/* Insert the active immediate struct onto the display list currently
- * being built.
- */
-static void _save_compile_vertex_list( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   struct tnl_vertex_list *node;
-
-   /* Allocate space for this structure in the display list currently
-    * being compiled.
-    */
-   node = (struct tnl_vertex_list *)
-      _mesa_alloc_instruction(ctx, tnl->save.opcode_vertex_list, sizeof(*node));
-
-   if (!node)
-      return;
-
-   /* Duplicate our template, increment refcounts to the storage structs:
-    */
-   _mesa_memcpy(node->attrsz, tnl->save.attrsz, sizeof(node->attrsz));
-   node->vertex_size = tnl->save.vertex_size;
-   node->buffer = tnl->save.buffer;
-   node->count = tnl->save.initial_counter - tnl->save.counter;
-   node->wrap_count = tnl->save.copied.nr;
-   node->have_materials = tnl->save.have_materials;
-   node->dangling_attr_ref = tnl->save.dangling_attr_ref;
-   node->normal_lengths = NULL;
-   node->prim = tnl->save.prim;
-   node->prim_count = tnl->save.prim_count;
-   node->vertex_store = tnl->save.vertex_store;
-   node->prim_store = tnl->save.prim_store;
-
-   node->vertex_store->refcount++;
-   node->prim_store->refcount++;
-
-   assert(node->attrsz[_TNL_ATTRIB_POS] != 0 ||
-	  node->count == 0);
-
-   if (tnl->save.dangling_attr_ref)
-      ctx->ListState.CurrentList->flags |= MESA_DLIST_DANGLING_REFS;
-
-   /* Maybe calculate normal lengths:
-    */
-   if (tnl->CalcDListNormalLengths && 
-       node->attrsz[_TNL_ATTRIB_NORMAL] == 3 &&
-       !(ctx->ListState.CurrentList->flags & MESA_DLIST_DANGLING_REFS))
-      build_normal_lengths( node );
-
-
-   tnl->save.vertex_store->used += tnl->save.vertex_size * node->count;
-   tnl->save.prim_store->used += node->prim_count;
-
-   /* Decide whether the storage structs are full, or can be used for
-    * the next vertex lists as well.
-    */
-   if (tnl->save.vertex_store->used > 
-       SAVE_BUFFER_SIZE - 16 * (tnl->save.vertex_size + 4)) {
-
-      tnl->save.vertex_store->refcount--; 
-      assert(tnl->save.vertex_store->refcount != 0);
-      tnl->save.vertex_store = alloc_vertex_store( ctx );
-      tnl->save.vbptr = tnl->save.vertex_store->buffer;
-   } 
-
-   if (tnl->save.prim_store->used > SAVE_PRIM_SIZE - 6) {
-      tnl->save.prim_store->refcount--; 
-      assert(tnl->save.prim_store->refcount != 0);
-      tnl->save.prim_store = alloc_prim_store( ctx );
-   } 
-
-   /* Reset our structures for the next run of vertices:
-    */
-   _save_reset_counters( ctx );
-
-   /* Copy duplicated vertices 
-    */
-   tnl->save.copied.nr = _save_copy_vertices( ctx, node );
-
-
-   /* Deal with GL_COMPILE_AND_EXECUTE:
-    */
-   if (ctx->ExecuteFlag) {
-      _tnl_playback_vertex_list( ctx, (void *) node );
-   }
-}
-
-
-/* TODO -- If no new vertices have been stored, don't bother saving
- * it.
- */
-static void _save_wrap_buffers( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLint i = tnl->save.prim_count - 1;
-   GLenum mode;
-
-   assert(i < (GLint) tnl->save.prim_max);
-   assert(i >= 0);
-
-   /* Close off in-progress primitive.
-    */
-   tnl->save.prim[i].count = ((tnl->save.initial_counter - tnl->save.counter) - 
-			     tnl->save.prim[i].start);
-   mode = tnl->save.prim[i].mode & ~(PRIM_BEGIN|PRIM_END);
-   
-   /* store the copied vertices, and allocate a new list.
-    */
-   _save_compile_vertex_list( ctx );
-
-   /* Restart interrupted primitive
-    */
-   tnl->save.prim[0].mode = mode;
-   tnl->save.prim[0].start = 0;
-   tnl->save.prim[0].count = 0;
-   tnl->save.prim_count = 1;
-}
-
-
-
-/* Called only when buffers are wrapped as the result of filling the
- * vertex_store struct.  
- */
-static void _save_wrap_filled_vertex( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLfloat *data = tnl->save.copied.buffer;
-   GLuint i;
-
-   /* Emit a glEnd to close off the last vertex list.
-    */
-   _save_wrap_buffers( ctx );
-   
-    /* Copy stored stored vertices to start of new list.
-    */
-   assert(tnl->save.counter > tnl->save.copied.nr);
-
-   for (i = 0 ; i < tnl->save.copied.nr ; i++) {
-      _mesa_memcpy( tnl->save.vbptr, data, tnl->save.vertex_size * sizeof(GLfloat));
-      data += tnl->save.vertex_size;
-      tnl->save.vbptr += tnl->save.vertex_size;
-      tnl->save.counter--;
-   }
-}
-
-
-static void _save_copy_to_current( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 
-   GLuint i;
-
-   /* XXX Use _TNL_FIRST_* and _TNL_LAST_* values instead? */
-   for (i = _TNL_ATTRIB_POS+1 ; i <= _TNL_ATTRIB_EDGEFLAG ; i++) {
-      if (tnl->save.attrsz[i]) {
-	 tnl->save.currentsz[i][0] = tnl->save.attrsz[i];
-	 COPY_CLEAN_4V(tnl->save.current[i], 
-		    tnl->save.attrsz[i], 
-		    tnl->save.attrptr[i]);
-      }
-   }
-
-   /* Edgeflag requires special treatment: 
-    *
-    * TODO: change edgeflag to GLfloat in Mesa.
-    */
-   if (tnl->save.attrsz[_TNL_ATTRIB_EDGEFLAG]) {
-      ctx->ListState.ActiveEdgeFlag = 1;
-      tnl->save.CurrentFloatEdgeFlag = 
-	 tnl->save.attrptr[_TNL_ATTRIB_EDGEFLAG][0];
-      ctx->ListState.CurrentEdgeFlag = 
-	 (tnl->save.CurrentFloatEdgeFlag == 1.0);
-   }
-}
-
-
-static void _save_copy_from_current( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 
-   GLint i;
-
-   for (i = _TNL_ATTRIB_POS+1 ; i <= _TNL_ATTRIB_EDGEFLAG ; i++) 
-      switch (tnl->save.attrsz[i]) {
-      case 4: tnl->save.attrptr[i][3] = tnl->save.current[i][3];
-      case 3: tnl->save.attrptr[i][2] = tnl->save.current[i][2];
-      case 2: tnl->save.attrptr[i][1] = tnl->save.current[i][1];
-      case 1: tnl->save.attrptr[i][0] = tnl->save.current[i][0];
-      case 0: break;
-      }
-
-   /* Edgeflag requires special treatment:
-    */
-   if (tnl->save.attrsz[_TNL_ATTRIB_EDGEFLAG]) {
-      tnl->save.CurrentFloatEdgeFlag = (GLfloat)ctx->ListState.CurrentEdgeFlag;
-      tnl->save.attrptr[_TNL_ATTRIB_EDGEFLAG][0] = tnl->save.CurrentFloatEdgeFlag;
-   }
-}
-
-
-
-
-/* Flush existing data, set new attrib size, replay copied vertices.
- */ 
-static void _save_upgrade_vertex( GLcontext *ctx, 
-				 GLuint attr,
-				 GLuint newsz )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLuint oldsz;
-   GLuint i;
-   GLfloat *tmp;
-
-   /* Store the current run of vertices, and emit a GL_END.  Emit a
-    * BEGIN in the new buffer.
-    */
-   if (tnl->save.initial_counter != tnl->save.counter) 
-      _save_wrap_buffers( ctx );
-   else
-      assert( tnl->save.copied.nr == 0 );
-
-   /* Do a COPY_TO_CURRENT to ensure back-copying works for the case
-    * when the attribute already exists in the vertex and is having
-    * its size increased.  
-    */
-   _save_copy_to_current( ctx );
-
-   /* Fix up sizes:
-    */
-   oldsz = tnl->save.attrsz[attr];
-   tnl->save.attrsz[attr] = newsz;
-
-   tnl->save.vertex_size += newsz - oldsz;
-   tnl->save.counter = ((SAVE_BUFFER_SIZE - tnl->save.vertex_store->used) / 
-			tnl->save.vertex_size);
-   if (tnl->save.counter > ctx->Const.MaxArrayLockSize )
-      tnl->save.counter = ctx->Const.MaxArrayLockSize;
-   tnl->save.initial_counter = tnl->save.counter;
-
-   /* Recalculate all the attrptr[] values:
-    */
-   for (i = 0, tmp = tnl->save.vertex ; i < _TNL_ATTRIB_MAX ; i++) {
-      if (tnl->save.attrsz[i]) {
-	 tnl->save.attrptr[i] = tmp;
-	 tmp += tnl->save.attrsz[i];
-      }
-      else 
-	 tnl->save.attrptr[i] = NULL; /* will not be dereferenced. */
-   }
-
-   /* Copy from current to repopulate the vertex with correct values.
-    */
-   _save_copy_from_current( ctx );
-
-   /* Replay stored vertices to translate them to new format here.
-    *
-    * If there are copied vertices and the new (upgraded) attribute
-    * has not been defined before, this list is somewhat degenerate,
-    * and will need fixup at runtime.
-    */
-   if (tnl->save.copied.nr)
-   {
-      GLfloat *data = tnl->save.copied.buffer;
-      GLfloat *dest = tnl->save.buffer;
-      GLuint j;
-
-      /* Need to note this and fix up at runtime (or loopback):
-       */
-      if (tnl->save.currentsz[attr][0] == 0) {
-	 assert(oldsz == 0);
-	 tnl->save.dangling_attr_ref = GL_TRUE;
-
-/* 	 _mesa_debug(NULL, "_save_upgrade_vertex: dangling reference attr %d\n",  */
-/* 		     attr);  */
-
-#if 0
-	 /* The current strategy is to punt these degenerate cases
-	  * through _tnl_loopback_vertex_list(), a lower-performance
-	  * option.  To minimize the impact of this, artificially
-	  * reduce the size of this vertex_list.
-	  */
-	 if (t->save.counter > 10) {
-	    t->save.initial_counter = 10;
-	    t->save.counter = 10;
-	 }
-#endif
-      }
-
-      for (i = 0 ; i < tnl->save.copied.nr ; i++) {
-	 for (j = 0 ; j < _TNL_ATTRIB_MAX ; j++) {
-	    if (tnl->save.attrsz[j]) {
-	       if (j == attr) {
-		  if (oldsz) {
-		     COPY_CLEAN_4V( dest, oldsz, data );
-		     data += oldsz;
-		     dest += newsz;
-		  }
-		  else {
-		     COPY_SZ_4V( dest, newsz, tnl->save.current[attr] );
-		     dest += newsz;
-		  }
-	       }
-	       else {
-		  GLint sz = tnl->save.attrsz[j];
-		  COPY_SZ_4V( dest, sz, data );
-		  data += sz;
-		  dest += sz;
-	       }
-	    }
-	 }
-      }
-
-      tnl->save.vbptr = dest;
-      tnl->save.counter -= tnl->save.copied.nr;
-   }
-}
-
-
-
-
-/* Helper function for 'CHOOSE' macro.  Do what's necessary when an
- * entrypoint is called for the first time.
- */
-static void do_choose( GLuint attr, GLuint sz, 
-		       void (*attr_func)( const GLfloat *),
-		       void (*choose1)( const GLfloat *),
-		       void (*choose2)( const GLfloat *),
-		       void (*choose3)( const GLfloat *),
-		       void (*choose4)( const GLfloat *),
-		       const GLfloat *v )
-{ 
-   GET_CURRENT_CONTEXT( ctx ); 
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 
-   static GLfloat id[4] = { 0, 0, 0, 1 };
-   int i;
-
-   if (tnl->save.attrsz[attr] < sz) {
-      /* New size is larger.  Need to flush existing vertices and get
-       * an enlarged vertex format.
-       */
-      _save_upgrade_vertex( ctx, attr, sz );
-   }
-   else {
-      /* New size is equal or smaller - just need to fill in some
-       * zeros.
-       */
-      for (i = sz ; i <= tnl->save.attrsz[attr] ; i++)
-	 tnl->save.attrptr[attr][i-1] = id[i-1];
-   }
-
-   /* Reset any active pointers for this attribute 
-    */
-   tnl->save.tabfv[attr][0] = choose1;
-   tnl->save.tabfv[attr][1] = choose2;
-   tnl->save.tabfv[attr][2] = choose3;
-   tnl->save.tabfv[attr][3] = choose4;
-
-   /* Update the secondary dispatch table with the new function
-    */
-   tnl->save.tabfv[attr][sz-1] = attr_func;
-
-   (*attr_func)(v);
-}
-
-
-
-/* Only one size for each attribute may be active at once.  Eg. if
- * Color3f is installed/active, then Color4f may not be, even if the
- * vertex actually contains 4 color coordinates.  This is because the
- * 3f version won't otherwise set color[3] to 1.0 -- this is the job
- * of the chooser function when switching between Color4f and Color3f.
- */
-#define ATTRFV( ATTR, N )					\
-static void save_choose_##ATTR##_##N( const GLfloat *v );	\
-								\
-static void save_attrib_##ATTR##_##N( const GLfloat *v )	\
-{								\
-   GET_CURRENT_CONTEXT( ctx );					\
-   TNLcontext *tnl = TNL_CONTEXT(ctx);				\
-								\
-   if ((ATTR) == 0) {						\
-      GLuint i;							\
-								\
-      if (N>0) tnl->save.vbptr[0] = v[0];			\
-      if (N>1) tnl->save.vbptr[1] = v[1];			\
-      if (N>2) tnl->save.vbptr[2] = v[2];			\
-      if (N>3) tnl->save.vbptr[3] = v[3];			\
-								\
-      for (i = N; i < tnl->save.vertex_size; i++)		\
-	 tnl->save.vbptr[i] = tnl->save.vertex[i];		\
-								\
-      tnl->save.vbptr += tnl->save.vertex_size;			\
-								\
-      if (--tnl->save.counter == 0)				\
-	 _save_wrap_filled_vertex( ctx );			\
-   }								\
-   else {							\
-      GLfloat *dest = tnl->save.attrptr[ATTR];			\
-      if (N>0) dest[0] = v[0];					\
-      if (N>1) dest[1] = v[1];					\
-      if (N>2) dest[2] = v[2];					\
-      if (N>3) dest[3] = v[3];					\
-   }								\
-}
-
-#define CHOOSE( ATTR, N )					\
-static void save_choose_##ATTR##_##N( const GLfloat *v )	\
-{								\
-   do_choose(ATTR, N,						\
-	     save_attrib_##ATTR##_##N,				\
-	     save_choose_##ATTR##_1,				\
-	     save_choose_##ATTR##_2,				\
-	     save_choose_##ATTR##_3,				\
-	     save_choose_##ATTR##_4,				\
-	     v );						\
-}
-
-#define INIT(ATTR)					\
-static void save_init_##ATTR( TNLcontext *tnl )		\
-{							\
-   tnl->save.tabfv[ATTR][0] = save_choose_##ATTR##_1;	\
-   tnl->save.tabfv[ATTR][1] = save_choose_##ATTR##_2;	\
-   tnl->save.tabfv[ATTR][2] = save_choose_##ATTR##_3;	\
-   tnl->save.tabfv[ATTR][3] = save_choose_##ATTR##_4;	\
-}
-   
-#define ATTRS( ATTRIB )				\
-   ATTRFV( ATTRIB, 1 )				\
-   ATTRFV( ATTRIB, 2 )				\
-   ATTRFV( ATTRIB, 3 )				\
-   ATTRFV( ATTRIB, 4 )				\
-   CHOOSE( ATTRIB, 1 )				\
-   CHOOSE( ATTRIB, 2 )				\
-   CHOOSE( ATTRIB, 3 )				\
-   CHOOSE( ATTRIB, 4 )				\
-   INIT( ATTRIB )				\
-
-
-/* Generate a lot of functions.  These are the actual worker
- * functions, which are equivalent to those generated via codegen
- * elsewhere.
- */
-ATTRS( 0 )
-ATTRS( 1 )
-ATTRS( 2 )
-ATTRS( 3 )
-ATTRS( 4 )
-ATTRS( 5 )
-ATTRS( 6 )
-ATTRS( 7 )
-ATTRS( 8 )
-ATTRS( 9 )
-ATTRS( 10 )
-ATTRS( 11 )
-ATTRS( 12 )
-ATTRS( 13 )
-ATTRS( 14 )
-ATTRS( 15 )
-
-ATTRS( 16 )
-ATTRS( 17 )
-ATTRS( 18 )
-ATTRS( 19 )
-ATTRS( 20 )
-ATTRS( 21 )
-ATTRS( 22 )
-ATTRS( 23 )
-ATTRS( 24 )
-ATTRS( 25 )
-ATTRS( 26 )
-ATTRS( 27 )
-ATTRS( 28 )
-ATTRS( 29 )
-ATTRS( 30 )
-ATTRS( 31 )
-
-
-static void _save_reset_vertex( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLuint i;
-
-   /* conventional attributes */
-   save_init_0( tnl );
-   save_init_1( tnl );
-   save_init_2( tnl );
-   save_init_3( tnl );
-   save_init_4( tnl );
-   save_init_5( tnl );
-   save_init_6( tnl );
-   save_init_7( tnl );
-   save_init_8( tnl );
-   save_init_9( tnl );
-   save_init_10( tnl );
-   save_init_11( tnl );
-   save_init_12( tnl );
-   save_init_13( tnl );
-   save_init_14( tnl );
-   save_init_15( tnl );
-
-   /* generic attributes */
-   save_init_16( tnl );
-   save_init_17( tnl );
-   save_init_18( tnl );
-   save_init_19( tnl );
-   save_init_20( tnl );
-   save_init_21( tnl );
-   save_init_22( tnl );
-   save_init_23( tnl );
-   save_init_24( tnl );
-   save_init_25( tnl );
-   save_init_26( tnl );
-   save_init_27( tnl );
-   save_init_28( tnl );
-   save_init_29( tnl );
-   save_init_30( tnl );
-   save_init_31( tnl );
-      
-   for (i = 0 ; i < _TNL_ATTRIB_MAX ; i++)
-      tnl->save.attrsz[i] = 0;
-      
-   tnl->save.vertex_size = 0;
-   tnl->save.have_materials = 0;
-
-   _save_reset_counters( ctx ); 
-}
-
-
-
-/* Cope with aliasing of classic Vertex, Normal, etc. and the fan-out
- * of glMultTexCoord and glProgramParamterNV by routing all these
- * through a second level dispatch table.
- */
-#define DISPATCH_ATTRFV( ATTR, COUNT, P )	\
-do {						\
-   GET_CURRENT_CONTEXT( ctx ); 			\
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 		\
-   tnl->save.tabfv[ATTR][COUNT-1]( P );		\
-} while (0)
-
-#define DISPATCH_ATTR1FV( ATTR, V ) DISPATCH_ATTRFV( ATTR, 1, V )
-#define DISPATCH_ATTR2FV( ATTR, V ) DISPATCH_ATTRFV( ATTR, 2, V )
-#define DISPATCH_ATTR3FV( ATTR, V ) DISPATCH_ATTRFV( ATTR, 3, V )
-#define DISPATCH_ATTR4FV( ATTR, V ) DISPATCH_ATTRFV( ATTR, 4, V )
-
-#define DISPATCH_ATTR1F( ATTR, S ) DISPATCH_ATTRFV( ATTR, 1, &(S) )
-
-#if defined(USE_X86_ASM) && 0 /* will break register calling convention */
-/* Naughty cheat:
- */
-#define DISPATCH_ATTR2F( ATTR, S,T ) DISPATCH_ATTRFV( ATTR, 2, &(S) )
-#define DISPATCH_ATTR3F( ATTR, S,T,R ) DISPATCH_ATTRFV( ATTR, 3, &(S) )
-#define DISPATCH_ATTR4F( ATTR, S,T,R,Q ) DISPATCH_ATTRFV( ATTR, 4, &(S) )
-#else
-/* Safe:
- */
-#define DISPATCH_ATTR2F( ATTR, S,T ) 		\
-do { 						\
-   GLfloat v[2]; 				\
-   v[0] = S; v[1] = T;				\
-   DISPATCH_ATTR2FV( ATTR, v );			\
-} while (0)
-#define DISPATCH_ATTR3F( ATTR, S,T,R ) 		\
-do { 						\
-   GLfloat v[3]; 				\
-   v[0] = S; v[1] = T; v[2] = R;		\
-   DISPATCH_ATTR3FV( ATTR, v );			\
-} while (0)
-#define DISPATCH_ATTR4F( ATTR, S,T,R,Q )	\
-do { 						\
-   GLfloat v[4]; 				\
-   v[0] = S; v[1] = T; v[2] = R; v[3] = Q;	\
-   DISPATCH_ATTR4FV( ATTR, v );			\
-} while (0)
-#endif
-
-
-static void enum_error( void )
-{
-   GET_CURRENT_CONTEXT( ctx );
-   _mesa_compile_error( ctx, GL_INVALID_ENUM, "glVertexAttrib" );
-}
-
-static void GLAPIENTRY _save_Vertex2f( GLfloat x, GLfloat y )
-{
-   DISPATCH_ATTR2F( _TNL_ATTRIB_POS, x, y );
-}
-
-static void GLAPIENTRY _save_Vertex2fv( const GLfloat *v )
-{
-   DISPATCH_ATTR2FV( _TNL_ATTRIB_POS, v );
-}
-
-static void GLAPIENTRY _save_Vertex3f( GLfloat x, GLfloat y, GLfloat z )
-{
-   DISPATCH_ATTR3F( _TNL_ATTRIB_POS, x, y, z );
-}
-
-static void GLAPIENTRY _save_Vertex3fv( const GLfloat *v )
-{
-   DISPATCH_ATTR3FV( _TNL_ATTRIB_POS, v );
-}
-
-static void GLAPIENTRY _save_Vertex4f( GLfloat x, GLfloat y, GLfloat z, GLfloat w )
-{
-   DISPATCH_ATTR4F( _TNL_ATTRIB_POS, x, y, z, w );
-}
-
-static void GLAPIENTRY _save_Vertex4fv( const GLfloat *v )
-{
-   DISPATCH_ATTR4FV( _TNL_ATTRIB_POS, v );
-}
-
-static void GLAPIENTRY _save_TexCoord1f( GLfloat x )
-{
-   DISPATCH_ATTR1F( _TNL_ATTRIB_TEX0, x );
-}
-
-static void GLAPIENTRY _save_TexCoord1fv( const GLfloat *v )
-{
-   DISPATCH_ATTR1FV( _TNL_ATTRIB_TEX0, v );
-}
-
-static void GLAPIENTRY _save_TexCoord2f( GLfloat x, GLfloat y )
-{
-   DISPATCH_ATTR2F( _TNL_ATTRIB_TEX0, x, y );
-}
-
-static void GLAPIENTRY _save_TexCoord2fv( const GLfloat *v )
-{
-   DISPATCH_ATTR2FV( _TNL_ATTRIB_TEX0, v );
-}
-
-static void GLAPIENTRY _save_TexCoord3f( GLfloat x, GLfloat y, GLfloat z )
-{
-   DISPATCH_ATTR3F( _TNL_ATTRIB_TEX0, x, y, z );
-}
-
-static void GLAPIENTRY _save_TexCoord3fv( const GLfloat *v )
-{
-   DISPATCH_ATTR3FV( _TNL_ATTRIB_TEX0, v );
-}
-
-static void GLAPIENTRY _save_TexCoord4f( GLfloat x, GLfloat y, GLfloat z, GLfloat w )
-{
-   DISPATCH_ATTR4F( _TNL_ATTRIB_TEX0, x, y, z, w );
-}
-
-static void GLAPIENTRY _save_TexCoord4fv( const GLfloat *v )
-{
-   DISPATCH_ATTR4FV( _TNL_ATTRIB_TEX0, v );
-}
-
-static void GLAPIENTRY _save_Normal3f( GLfloat x, GLfloat y, GLfloat z )
-{
-   DISPATCH_ATTR3F( _TNL_ATTRIB_NORMAL, x, y, z );
-}
-
-static void GLAPIENTRY _save_Normal3fv( const GLfloat *v )
-{
-   DISPATCH_ATTR3FV( _TNL_ATTRIB_NORMAL, v );
-}
-
-static void GLAPIENTRY _save_FogCoordfEXT( GLfloat x )
-{
-   DISPATCH_ATTR1F( _TNL_ATTRIB_FOG, x );
-}
-
-static void GLAPIENTRY _save_FogCoordfvEXT( const GLfloat *v )
-{
-   DISPATCH_ATTR1FV( _TNL_ATTRIB_FOG, v );
-}
-
-static void GLAPIENTRY _save_Color3f( GLfloat x, GLfloat y, GLfloat z )
-{
-   DISPATCH_ATTR3F( _TNL_ATTRIB_COLOR0, x, y, z );
-}
-
-static void GLAPIENTRY _save_Color3fv( const GLfloat *v )
-{
-   DISPATCH_ATTR3FV( _TNL_ATTRIB_COLOR0, v );
-}
-
-static void GLAPIENTRY _save_Color4f( GLfloat x, GLfloat y, GLfloat z, GLfloat w )
-{
-   DISPATCH_ATTR4F( _TNL_ATTRIB_COLOR0, x, y, z, w );
-}
-
-static void GLAPIENTRY _save_Color4fv( const GLfloat *v )
-{
-   DISPATCH_ATTR4FV( _TNL_ATTRIB_COLOR0, v );
-}
-
-static void GLAPIENTRY _save_SecondaryColor3fEXT( GLfloat x, GLfloat y, GLfloat z )
-{
-   DISPATCH_ATTR3F( _TNL_ATTRIB_COLOR1, x, y, z );
-}
-
-static void GLAPIENTRY _save_SecondaryColor3fvEXT( const GLfloat *v )
-{
-   DISPATCH_ATTR3FV( _TNL_ATTRIB_COLOR1, v );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord1f( GLenum target, GLfloat x  )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR1F( attr, x );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord1fv( GLenum target, const GLfloat *v )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR1FV( attr, v );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord2f( GLenum target, GLfloat x, GLfloat y )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR2F( attr, x, y );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord2fv( GLenum target, const GLfloat *v )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR2FV( attr, v );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord3f( GLenum target, GLfloat x, GLfloat y,
-				    GLfloat z)
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR3F( attr, x, y, z );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord3fv( GLenum target, const GLfloat *v )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR3FV( attr, v );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord4f( GLenum target, GLfloat x, GLfloat y,
-				    GLfloat z, GLfloat w )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR4F( attr, x, y, z, w );
-}
-
-static void GLAPIENTRY _save_MultiTexCoord4fv( GLenum target, const GLfloat *v )
-{
-   GLuint attr = (target & 0x7) + _TNL_ATTRIB_TEX0;
-   DISPATCH_ATTR4FV( attr, v );
-}
-
-
-
-static void GLAPIENTRY
-_save_VertexAttrib1fNV(GLuint index, GLfloat x)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR1F( index, x );
-   }
-   else
-      enum_error(); 
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib1fvNV(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR1FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib2fNV(GLuint index, GLfloat x, GLfloat y)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR2F( index, x, y );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib2fvNV(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR2FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib3fNV(GLuint index, GLfloat x, GLfloat y, GLfloat z)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR3F( index, x, y, z );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib3fvNV(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR3FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib4fNV(GLuint index, GLfloat x, GLfloat y,
-                       GLfloat z, GLfloat w)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR4F( index, x, y, z, w );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib4fvNV(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR4FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib1fARB(GLuint index, GLfloat x)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR1F( index, x );
-   }
-   else
-      enum_error(); 
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib1fvARB(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR1FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib2fARB(GLuint index, GLfloat x, GLfloat y)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR2F( index, x, y );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib2fvARB(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR2FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib3fARB(GLuint index, GLfloat x, GLfloat y, GLfloat z)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR3F( index, x, y, z );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib3fvARB(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR3FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib4fARB(GLuint index, GLfloat x, GLfloat y,
-                        GLfloat z, GLfloat w)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR4F( index, x, y, z, w );
-   }
-   else
-      enum_error();
-}
-
-static void GLAPIENTRY
-_save_VertexAttrib4fvARB(GLuint index, const GLfloat *v)
-{
-   if (index < MAX_VERTEX_ATTRIBS) {
-      if (index > 0)
-         index += VERT_ATTRIB_GENERIC0;
-      DISPATCH_ATTR4FV( index, v );
-   }
-   else
-      enum_error();
-}
-
-
-/* Materials:  
- * 
- * These are treated as per-vertex attributes, at indices above where
- * the NV_vertex_program leaves off.  There are a lot of good things
- * about treating materials this way.  
- *
- * However: I don't want to double the number of generated functions
- * just to cope with this, so I unroll the 'C' varients of CHOOSE and
- * ATTRF into this function, and dispense with codegen and
- * second-level dispatch.
- *
- * There is no aliasing of material attributes with other entrypoints.
- */
-#define MAT_ATTR( A, N, params )			\
-do {							\
-   if (tnl->save.attrsz[A] < N) {			\
-      _save_upgrade_vertex( ctx, A, N );		\
-      tnl->save.have_materials = GL_TRUE;               \
-   }							\
-							\
-   {							\
-      GLfloat *dest = tnl->save.attrptr[A];	      	\
-      if (N>0) dest[0] = params[0];			\
-      if (N>1) dest[1] = params[1];			\
-      if (N>2) dest[2] = params[2];			\
-      if (N>3) dest[3] = params[3];			\
-   }							\
-} while (0)
-
-
-#define MAT( ATTR, N, face, params )			\
-do {							\
-   if (face != GL_BACK)					\
-      MAT_ATTR( ATTR, N, params ); /* front */		\
-   if (face != GL_FRONT)				\
-      MAT_ATTR( ATTR + 1, N, params ); /* back */	\
-} while (0)
-
-
-/* NOTE: Have to remove/deal-with colormaterial crossovers, probably
- * later on - in the meantime just store everything.  
- */
-static void GLAPIENTRY _save_Materialfv( GLenum face, GLenum pname, 
-			       const GLfloat *params )
-{
-   GET_CURRENT_CONTEXT( ctx ); 
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   switch (pname) {
-   case GL_EMISSION:
-      MAT( _TNL_ATTRIB_MAT_FRONT_EMISSION, 4, face, params );
-      break;
-   case GL_AMBIENT:
-      MAT( _TNL_ATTRIB_MAT_FRONT_AMBIENT, 4, face, params );
-      break;
-   case GL_DIFFUSE:
-      MAT( _TNL_ATTRIB_MAT_FRONT_DIFFUSE, 4, face, params );
-      break;
-   case GL_SPECULAR:
-      MAT( _TNL_ATTRIB_MAT_FRONT_SPECULAR, 4, face, params );
-      break;
-   case GL_SHININESS:
-      MAT( _TNL_ATTRIB_MAT_FRONT_SHININESS, 1, face, params );
-      break;
-   case GL_COLOR_INDEXES:
-      MAT( _TNL_ATTRIB_MAT_FRONT_INDEXES, 3, face, params );
-      break;
-   case GL_AMBIENT_AND_DIFFUSE:
-      MAT( _TNL_ATTRIB_MAT_FRONT_AMBIENT, 4, face, params );
-      MAT( _TNL_ATTRIB_MAT_FRONT_DIFFUSE, 4, face, params );
-      break;
-   default:
-      _mesa_compile_error( ctx, GL_INVALID_ENUM, "glMaterialfv" );
-      return;
-   }
-}
-
-
-#define IDX_ATTR( A, IDX )				\
-do {							\
-   GET_CURRENT_CONTEXT( ctx );				\
-   TNLcontext *tnl = TNL_CONTEXT(ctx);			\
-							\
-   if (tnl->save.attrsz[A] < 1) {			\
-      _save_upgrade_vertex( ctx, A, 1 );		\
-   }							\
-							\
-   {							\
-      GLfloat *dest = tnl->save.attrptr[A];		\
-      dest[0] = IDX;				\
-   }							\
-} while (0)
-
-
-static void GLAPIENTRY _save_EdgeFlag( GLboolean b )
-{
-   IDX_ATTR( _TNL_ATTRIB_EDGEFLAG, (GLfloat)b );
-}
-
-
-static void GLAPIENTRY _save_Indexf( GLfloat f )
-{
-   IDX_ATTR( _TNL_ATTRIB_COLOR_INDEX, f );
-}
-
-static void GLAPIENTRY _save_Indexfv( const GLfloat *f )
-{
-   IDX_ATTR( _TNL_ATTRIB_COLOR_INDEX, f[0] );
-}
-
-
-
-
-/* Cope with EvalCoord/CallList called within a begin/end object:
- *     -- Flush current buffer
- *     -- Fallback to opcodes for the rest of the begin/end object.
- */
-#define FALLBACK(ctx) 							\
-do {									\
-   TNLcontext *tnl = TNL_CONTEXT(ctx);					\
-									\
-   if (tnl->save.initial_counter != tnl->save.counter ||		\
-       tnl->save.prim_count) 						\
-      _save_compile_vertex_list( ctx );					\
-									\
-   _save_copy_to_current( ctx );					\
-   _save_reset_vertex( ctx );						\
-   _mesa_install_save_vtxfmt( ctx, &ctx->ListState.ListVtxfmt );	\
-   ctx->Driver.SaveNeedFlush = 0;					\
-} while (0)
-
-static void GLAPIENTRY _save_EvalCoord1f( GLfloat u )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalCoord1f(ctx->Save, ( u ));
-}
-
-static void GLAPIENTRY _save_EvalCoord1fv( const GLfloat *v )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalCoord1fv(ctx->Save, ( v ));
-}
-
-static void GLAPIENTRY _save_EvalCoord2f( GLfloat u, GLfloat v )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalCoord2f(ctx->Save, ( u, v ));
-}
-
-static void GLAPIENTRY _save_EvalCoord2fv( const GLfloat *v )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalCoord2fv(ctx->Save, ( v ));
-}
-
-static void GLAPIENTRY _save_EvalPoint1( GLint i )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalPoint1(ctx->Save, ( i ));
-}
-
-static void GLAPIENTRY _save_EvalPoint2( GLint i, GLint j )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_EvalPoint2(ctx->Save, ( i, j ));
-}
-
-static void GLAPIENTRY _save_CallList( GLuint l )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_CallList(ctx->Save, ( l ));
-}
-
-static void GLAPIENTRY _save_CallLists( GLsizei n, GLenum type, const GLvoid *v )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   FALLBACK(ctx);
-   CALL_CallLists(ctx->Save, ( n, type, v ));
-}
-
-
-
-
-/**
- * Called via ctx->Driver.NotifySaveBegin(ctx, mode) when we get a
- * glBegin() call while compiling a display list.
- * See save_Begin() in dlist.c
- *
- * This plugs in our special TNL-related display list functions.
- * All subsequent glBegin/glVertex/glEnd()s found while compiling a
- * display list will get routed to the functions in this file.
- *
- * Updating of ctx->Driver.CurrentSavePrimitive is already taken care of.
- */
-static GLboolean _save_NotifyBegin( GLcontext *ctx, GLenum mode )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 
-
-   if (1) {
-      GLuint i = tnl->save.prim_count++;
-
-      assert(i < tnl->save.prim_max);
-      tnl->save.prim[i].mode = mode | PRIM_BEGIN;
-      tnl->save.prim[i].start = tnl->save.initial_counter - tnl->save.counter;
-      tnl->save.prim[i].count = 0;   
-
-      _mesa_install_save_vtxfmt( ctx, &tnl->save_vtxfmt );      
-      ctx->Driver.SaveNeedFlush = 1;
-      return GL_TRUE;
-   }
-   else 
-      return GL_FALSE;
-}
-
-
-
-static void GLAPIENTRY _save_End( void )
-{
-   GET_CURRENT_CONTEXT( ctx ); 
-   TNLcontext *tnl = TNL_CONTEXT(ctx); 
-   GLint i = tnl->save.prim_count - 1;
-
-   ctx->Driver.CurrentSavePrimitive = PRIM_OUTSIDE_BEGIN_END;
-   if (ctx->ExecuteFlag)
-      ctx->Driver.CurrentExecPrimitive = PRIM_OUTSIDE_BEGIN_END;
-
-   tnl->save.prim[i].mode |= PRIM_END;
-   tnl->save.prim[i].count = ((tnl->save.initial_counter - tnl->save.counter) - 
-			      tnl->save.prim[i].start);
-
-   if (i == (GLint) tnl->save.prim_max - 1) {
-      _save_compile_vertex_list( ctx );
-      assert(tnl->save.copied.nr == 0);
-   }
-
-   /* Swap out this vertex format while outside begin/end.  Any color,
-    * etc. received between here and the next begin will be compiled
-    * as opcodes.
-    */   
-   _mesa_install_save_vtxfmt( ctx, &ctx->ListState.ListVtxfmt );
-}
-
-
-/* These are all errors as this vtxfmt is only installed inside
- * begin/end pairs.
- */
-static void GLAPIENTRY _save_DrawElements(GLenum mode, GLsizei count, GLenum type,
-			       const GLvoid *indices)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) mode; (void) count; (void) type; (void) indices;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glDrawElements" );
-}
-
-
-static void GLAPIENTRY _save_DrawRangeElements(GLenum mode,
-				    GLuint start, GLuint end,
-				    GLsizei count, GLenum type,
-				    const GLvoid *indices)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) mode; (void) start; (void) end; (void) count; (void) type; (void) indices;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glDrawRangeElements" );
-}
-
-static void GLAPIENTRY _save_DrawArrays(GLenum mode, GLint start, GLsizei count)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) mode; (void) start; (void) count;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glDrawArrays" );
-}
-
-static void GLAPIENTRY _save_Rectf( GLfloat x1, GLfloat y1, GLfloat x2, GLfloat y2 )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) x1; (void) y1; (void) x2; (void) y2;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glRectf" );
-}
-
-static void GLAPIENTRY _save_EvalMesh1( GLenum mode, GLint i1, GLint i2 )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) mode; (void) i1; (void) i2;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glEvalMesh1" );
-}
-
-static void GLAPIENTRY _save_EvalMesh2( GLenum mode, GLint i1, GLint i2,
-				  GLint j1, GLint j2 )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   (void) mode; (void) i1; (void) i2; (void) j1; (void) j2;
-   _mesa_compile_error( ctx, GL_INVALID_OPERATION, "glEvalMesh2" );
-}
-
-/**
- * This is only called if someone tries to compile nested glBegin()s 
- * in their display list.
- */
-static void GLAPIENTRY _save_Begin( GLenum mode )
-{
-   GET_CURRENT_CONTEXT( ctx );
-   (void) mode;
-   _mesa_compile_error(ctx, GL_INVALID_OPERATION,
-                       "glBegin(called inside glBegin/End)");
-}
-
-
-/* Unlike the functions above, these are to be hooked into the vtxfmt
- * maintained in ctx->ListState, active when the list is known or
- * suspected to be outside any begin/end primitive.
- */
-static void GLAPIENTRY _save_OBE_Rectf( GLfloat x1, GLfloat y1, GLfloat x2, GLfloat y2 )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   _save_NotifyBegin( ctx, GL_QUADS | PRIM_WEAK );
-   CALL_Vertex2f(GET_DISPATCH(), ( x1, y1 ));
-   CALL_Vertex2f(GET_DISPATCH(), ( x2, y1 ));
-   CALL_Vertex2f(GET_DISPATCH(), ( x2, y2 ));
-   CALL_Vertex2f(GET_DISPATCH(), ( x1, y2 ));
-   CALL_End(GET_DISPATCH(), ());
-}
-
-
-static void GLAPIENTRY _save_OBE_DrawArrays(GLenum mode, GLint start, GLsizei count)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   GLint i;
-
-   if (!_mesa_validate_DrawArrays( ctx, mode, start, count ))
-      return;
-
-   _ae_map_vbos( ctx );
-
-   _save_NotifyBegin( ctx, mode | PRIM_WEAK );
-   for (i = 0; i < count; i++)
-       CALL_ArrayElement(GET_DISPATCH(), (start + i));
-   CALL_End(GET_DISPATCH(), ());
-
-   _ae_unmap_vbos( ctx );
-}
-
-
-static void GLAPIENTRY _save_OBE_DrawElements(GLenum mode, GLsizei count, GLenum type,
-				   const GLvoid *indices)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   GLint i;
-
-   if (!_mesa_validate_DrawElements( ctx, mode, count, type, indices ))
-      return;
-
-   _ae_map_vbos( ctx );
-
-   _save_NotifyBegin( ctx, mode | PRIM_WEAK );
-
-   switch (type) {
-   case GL_UNSIGNED_BYTE:
-      for (i = 0 ; i < count ; i++)
-	  CALL_ArrayElement(GET_DISPATCH(), ( ((GLubyte *)indices)[i] ));
-      break;
-   case GL_UNSIGNED_SHORT:
-      for (i = 0 ; i < count ; i++)
-	  CALL_ArrayElement(GET_DISPATCH(), ( ((GLushort *)indices)[i] ));
-      break;
-   case GL_UNSIGNED_INT:
-      for (i = 0 ; i < count ; i++)
-	  CALL_ArrayElement(GET_DISPATCH(), ( ((GLuint *)indices)[i] ));
-      break;
-   default:
-      _mesa_error( ctx, GL_INVALID_ENUM, "glDrawElements(type)" );
-      break;
-   }
-
-   CALL_End(GET_DISPATCH(), ());
-
-   _ae_unmap_vbos( ctx );
-}
-
-static void GLAPIENTRY _save_OBE_DrawRangeElements(GLenum mode,
-					GLuint start, GLuint end,
-					GLsizei count, GLenum type,
-					const GLvoid *indices)
-{
-   GET_CURRENT_CONTEXT(ctx);
-   if (_mesa_validate_DrawRangeElements( ctx, mode,
-					 start, end,
-					 count, type, indices ))
-      _save_OBE_DrawElements( mode, count, type, indices );
-}
-
-
-
-
-
-static void _save_vtxfmt_init( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLvertexformat *vfmt = &tnl->save_vtxfmt;
-
-   vfmt->ArrayElement = _ae_loopback_array_elt;	        /* generic helper */
-   vfmt->Begin = _save_Begin;
-   vfmt->Color3f = _save_Color3f;
-   vfmt->Color3fv = _save_Color3fv;
-   vfmt->Color4f = _save_Color4f;
-   vfmt->Color4fv = _save_Color4fv;
-   vfmt->EdgeFlag = _save_EdgeFlag;
-   vfmt->End = _save_End;
-   vfmt->FogCoordfEXT = _save_FogCoordfEXT;
-   vfmt->FogCoordfvEXT = _save_FogCoordfvEXT;
-   vfmt->Indexf = _save_Indexf;
-   vfmt->Indexfv = _save_Indexfv;
-   vfmt->Materialfv = _save_Materialfv;
-   vfmt->MultiTexCoord1fARB = _save_MultiTexCoord1f;
-   vfmt->MultiTexCoord1fvARB = _save_MultiTexCoord1fv;
-   vfmt->MultiTexCoord2fARB = _save_MultiTexCoord2f;
-   vfmt->MultiTexCoord2fvARB = _save_MultiTexCoord2fv;
-   vfmt->MultiTexCoord3fARB = _save_MultiTexCoord3f;
-   vfmt->MultiTexCoord3fvARB = _save_MultiTexCoord3fv;
-   vfmt->MultiTexCoord4fARB = _save_MultiTexCoord4f;
-   vfmt->MultiTexCoord4fvARB = _save_MultiTexCoord4fv;
-   vfmt->Normal3f = _save_Normal3f;
-   vfmt->Normal3fv = _save_Normal3fv;
-   vfmt->SecondaryColor3fEXT = _save_SecondaryColor3fEXT;
-   vfmt->SecondaryColor3fvEXT = _save_SecondaryColor3fvEXT;
-   vfmt->TexCoord1f = _save_TexCoord1f;
-   vfmt->TexCoord1fv = _save_TexCoord1fv;
-   vfmt->TexCoord2f = _save_TexCoord2f;
-   vfmt->TexCoord2fv = _save_TexCoord2fv;
-   vfmt->TexCoord3f = _save_TexCoord3f;
-   vfmt->TexCoord3fv = _save_TexCoord3fv;
-   vfmt->TexCoord4f = _save_TexCoord4f;
-   vfmt->TexCoord4fv = _save_TexCoord4fv;
-   vfmt->Vertex2f = _save_Vertex2f;
-   vfmt->Vertex2fv = _save_Vertex2fv;
-   vfmt->Vertex3f = _save_Vertex3f;
-   vfmt->Vertex3fv = _save_Vertex3fv;
-   vfmt->Vertex4f = _save_Vertex4f;
-   vfmt->Vertex4fv = _save_Vertex4fv;
-   vfmt->VertexAttrib1fNV = _save_VertexAttrib1fNV;
-   vfmt->VertexAttrib1fvNV = _save_VertexAttrib1fvNV;
-   vfmt->VertexAttrib2fNV = _save_VertexAttrib2fNV;
-   vfmt->VertexAttrib2fvNV = _save_VertexAttrib2fvNV;
-   vfmt->VertexAttrib3fNV = _save_VertexAttrib3fNV;
-   vfmt->VertexAttrib3fvNV = _save_VertexAttrib3fvNV;
-   vfmt->VertexAttrib4fNV = _save_VertexAttrib4fNV;
-   vfmt->VertexAttrib4fvNV = _save_VertexAttrib4fvNV;
-   vfmt->VertexAttrib1fARB = _save_VertexAttrib1fARB;
-   vfmt->VertexAttrib1fvARB = _save_VertexAttrib1fvARB;
-   vfmt->VertexAttrib2fARB = _save_VertexAttrib2fARB;
-   vfmt->VertexAttrib2fvARB = _save_VertexAttrib2fvARB;
-   vfmt->VertexAttrib3fARB = _save_VertexAttrib3fARB;
-   vfmt->VertexAttrib3fvARB = _save_VertexAttrib3fvARB;
-   vfmt->VertexAttrib4fARB = _save_VertexAttrib4fARB;
-   vfmt->VertexAttrib4fvARB = _save_VertexAttrib4fvARB;
-   
-   /* This will all require us to fallback to saving the list as opcodes:
-    */ 
-   vfmt->CallList = _save_CallList; /* inside begin/end */
-   vfmt->CallLists = _save_CallLists; /* inside begin/end */
-   vfmt->EvalCoord1f = _save_EvalCoord1f;
-   vfmt->EvalCoord1fv = _save_EvalCoord1fv;
-   vfmt->EvalCoord2f = _save_EvalCoord2f;
-   vfmt->EvalCoord2fv = _save_EvalCoord2fv;
-   vfmt->EvalPoint1 = _save_EvalPoint1;
-   vfmt->EvalPoint2 = _save_EvalPoint2;
-
-   /* These are all errors as we at least know we are in some sort of
-    * begin/end pair:
-    */
-   vfmt->EvalMesh1 = _save_EvalMesh1;	
-   vfmt->EvalMesh2 = _save_EvalMesh2;
-   vfmt->Begin = _save_Begin;
-   vfmt->Rectf = _save_Rectf;
-   vfmt->DrawArrays = _save_DrawArrays;
-   vfmt->DrawElements = _save_DrawElements;
-   vfmt->DrawRangeElements = _save_DrawRangeElements;
-
-}
-
-
-void _tnl_SaveFlushVertices( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   /* Noop when we are actually active:
-    */
-   if (ctx->Driver.CurrentSavePrimitive == PRIM_INSIDE_UNKNOWN_PRIM ||
-       ctx->Driver.CurrentSavePrimitive <= GL_POLYGON)
-      return;
-
-   if (tnl->save.initial_counter != tnl->save.counter ||
-       tnl->save.prim_count) 
-      _save_compile_vertex_list( ctx );
-   
-   _save_copy_to_current( ctx );
-   _save_reset_vertex( ctx );
-   ctx->Driver.SaveNeedFlush = 0;
-}
-
-void _tnl_NewList( GLcontext *ctx, GLuint list, GLenum mode )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   (void) list; (void) mode;
-
-   if (!tnl->save.prim_store)
-      tnl->save.prim_store = alloc_prim_store( ctx );
-
-   if (!tnl->save.vertex_store) {
-      tnl->save.vertex_store = alloc_vertex_store( ctx );
-      tnl->save.vbptr = tnl->save.vertex_store->buffer;
-   }
-   
-   _save_reset_vertex( ctx );
-   ctx->Driver.SaveNeedFlush = 0;
-}
-
-void _tnl_EndList( GLcontext *ctx )
-{
-   (void) ctx;
-   assert(TNL_CONTEXT(ctx)->save.vertex_size == 0);
-}
- 
-void _tnl_BeginCallList( GLcontext *ctx, struct mesa_display_list *dlist )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   tnl->save.replay_flags |= dlist->flags;
-   tnl->save.replay_flags |= tnl->LoopbackDListCassettes;
-}
-
-void _tnl_EndCallList( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   
-   if (ctx->ListState.CallDepth == 1)
-      tnl->save.replay_flags = 0;
-}
-
-
-static void _tnl_destroy_vertex_list( GLcontext *ctx, void *data )
-{
-   struct tnl_vertex_list *node = (struct tnl_vertex_list *)data;
-   (void) ctx;
-
-   if ( --node->vertex_store->refcount == 0 )
-      FREE( node->vertex_store );
-
-   if ( --node->prim_store->refcount == 0 )
-      FREE( node->prim_store );
-
-   if ( node->normal_lengths )
-      FREE( node->normal_lengths );
-}
-
-
-static void _tnl_print_vertex_list( GLcontext *ctx, void *data )
-{
-   struct tnl_vertex_list *node = (struct tnl_vertex_list *)data;
-   GLuint i;
-   (void) ctx;
-
-   _mesa_debug(NULL, "TNL-VERTEX-LIST, %u vertices %d primitives, %d vertsize\n",
-               node->count,
-	       node->prim_count,
-	       node->vertex_size);
-
-   for (i = 0 ; i < node->prim_count ; i++) {
-      struct tnl_prim *prim = &node->prim[i];
-      _mesa_debug(NULL, "   prim %d: %s %d..%d %s %s\n",
-		  i, 
-		  _mesa_lookup_enum_by_nr(prim->mode & PRIM_MODE_MASK),
-		  prim->start, 
-		  prim->start + prim->count,
-		  (prim->mode & PRIM_BEGIN) ? "BEGIN" : "(wrap)",
-		  (prim->mode & PRIM_END) ? "END" : "(wrap)");
-   }
-}
-
-
-static void _save_current_init( GLcontext *ctx ) 
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLint i;
-
-   for (i = 0; i < _TNL_ATTRIB_MAT_FRONT_AMBIENT; i++) {
-      ASSERT(i < VERT_ATTRIB_MAX);
-      tnl->save.currentsz[i] = &ctx->ListState.ActiveAttribSize[i];
-      tnl->save.current[i] = ctx->ListState.CurrentAttrib[i];
-   }
-
-   for (i = _TNL_FIRST_MAT; i <= _TNL_LAST_MAT; i++) {
-      const GLuint j = i - _TNL_FIRST_MAT;
-      ASSERT(j < MAT_ATTRIB_MAX);
-      tnl->save.currentsz[i] = &ctx->ListState.ActiveMaterialSize[j];
-      tnl->save.current[i] = ctx->ListState.CurrentMaterial[j];
-   }
-
-   tnl->save.currentsz[_TNL_ATTRIB_EDGEFLAG] = &ctx->ListState.ActiveEdgeFlag;
-   tnl->save.current[_TNL_ATTRIB_EDGEFLAG] = &tnl->save.CurrentFloatEdgeFlag;
-}
-
-/**
- * Initialize the display list compiler
- */
-void _tnl_save_init( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-   struct tnl_vertex_arrays *tmp = &tnl->save_inputs;
-   GLuint i;
-
-
-   for (i = 0; i < _TNL_ATTRIB_MAX; i++)
-      _mesa_vector4f_init( &tmp->Attribs[i], 0, NULL);
-
-   tnl->save.opcode_vertex_list =
-      _mesa_alloc_opcode( ctx,
-			  sizeof(struct tnl_vertex_list),
-			  _tnl_playback_vertex_list,
-			  _tnl_destroy_vertex_list,
-			  _tnl_print_vertex_list );
-
-   ctx->Driver.NotifySaveBegin = _save_NotifyBegin;
-
-   _save_vtxfmt_init( ctx );
-   _save_current_init( ctx );
-
-   /* Hook our array functions into the outside-begin-end vtxfmt in 
-    * ctx->ListState.
-    */
-   ctx->ListState.ListVtxfmt.Rectf = _save_OBE_Rectf;
-   ctx->ListState.ListVtxfmt.DrawArrays = _save_OBE_DrawArrays;
-   ctx->ListState.ListVtxfmt.DrawElements = _save_OBE_DrawElements;
-   ctx->ListState.ListVtxfmt.DrawRangeElements = _save_OBE_DrawRangeElements;
-   _mesa_install_save_vtxfmt( ctx, &ctx->ListState.ListVtxfmt );
-}
-
-
-/**
- * Deallocate the immediate-mode buffer for the given context, if
- * its reference count goes to zero.
- */
-void _tnl_save_destroy( GLcontext *ctx )
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   /* Decrement the refcounts.  References may still be held by
-    * display lists yet to be destroyed, so it may not yet be time to
-    * free these items.
-    */
-   if (tnl->save.prim_store &&
-       --tnl->save.prim_store->refcount == 0 )
-      FREE( tnl->save.prim_store );
-
-   if (tnl->save.vertex_store &&
-       --tnl->save.vertex_store->refcount == 0 )
-      FREE( tnl->save.vertex_store );
-}
diff --git a/src/mesa/tnl/t_vb_program.c b/src/mesa/tnl/t_vb_program.c
index 9961af70ce..f8e561ac57 100644
--- a/src/mesa/tnl/t_vb_program.c
+++ b/src/mesa/tnl/t_vb_program.c
@@ -35,16 +35,16 @@
 #include "context.h"
 #include "macros.h"
 #include "imports.h"
-#include "prog_instruction.h"
-#include "prog_statevars.h"
-#include "prog_execute.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_statevars.h"
+#include "shader/prog_execute.h"
+#include "swrast/s_context.h"
+#include "swrast/s_texfilter.h"
 
 #include "tnl.h"
 #include "t_context.h"
 #include "t_pipeline.h"
 
-#include "swrast/s_context.h"
-#include "swrast/s_texfilter.h"
 
 /**
  * XXX the texture sampling code in this module is a bit of a hack.
diff --git a/src/mesa/tnl/t_vertex.c b/src/mesa/tnl/t_vertex.c
index 6aae602037..a6728c318f 100644
--- a/src/mesa/tnl/t_vertex.c
+++ b/src/mesa/tnl/t_vertex.c
@@ -233,7 +233,7 @@ void _tnl_get_attr( GLcontext *ctx, const void *vin,
       /* If the hardware vertex doesn't have point size then use size from
        * GLcontext.  XXX this will be wrong if drawing attenuated points!
        */
-      dest[0] = ctx->Point._Size;
+      dest[0] = ctx->Point.Size;
    }
    else {
       _mesa_memcpy( dest, ctx->Current.Attrib[attr], 4*sizeof(GLfloat));
diff --git a/src/mesa/tnl/t_vp_build.c b/src/mesa/tnl/t_vp_build.c
index 2a1cae77f2..ee1a2498b3 100644
--- a/src/mesa/tnl/t_vp_build.c
+++ b/src/mesa/tnl/t_vp_build.c
@@ -33,11 +33,11 @@
 #include "glheader.h"
 #include "macros.h"
 #include "enums.h"
-#include "program.h"
-#include "prog_instruction.h"
-#include "prog_parameter.h"
-#include "prog_print.h"
-#include "prog_statevars.h"
+#include "shader/program.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_parameter.h"
+#include "shader/prog_print.h"
+#include "shader/prog_statevars.h"
 #include "t_context.h" /* NOTE: very light dependency on this */
 #include "t_vp_build.h"
 
diff --git a/src/mesa/tnl/tnl.h b/src/mesa/tnl/tnl.h
index 20bed5546d..047b764dcb 100644
--- a/src/mesa/tnl/tnl.h
+++ b/src/mesa/tnl/tnl.h
@@ -29,7 +29,7 @@
 #ifndef _TNL_H
 #define _TNL_H
 
-#include "mtypes.h"
+#include "main/mtypes.h"
 
 
 
diff --git a/src/mesa/tnl_dd/t_dd_vb.c b/src/mesa/tnl_dd/t_dd_vb.c
index ab3bb37631..3cedd90119 100644
--- a/src/mesa/tnl_dd/t_dd_vb.c
+++ b/src/mesa/tnl_dd/t_dd_vb.c
@@ -184,7 +184,7 @@ void TAG(translate_vertex)(GLcontext *ctx,
       }
    }
 
-   dst->pointSize = ctx->Point._Size;
+   dst->pointSize = ctx->Point.Size;
 }
 
 
diff --git a/src/mesa/vbo/vbo.h b/src/mesa/vbo/vbo.h
index 874a5f9e0e..04c59c05b2 100644
--- a/src/mesa/vbo/vbo.h
+++ b/src/mesa/vbo/vbo.h
@@ -32,7 +32,7 @@
 #ifndef _VBO_H
 #define _VBO_H
 
-#include "mtypes.h"
+#include "main/mtypes.h"
 
 struct _mesa_prim {
    GLuint mode:8;
diff --git a/src/mesa/vbo/vbo_context.c b/src/mesa/vbo/vbo_context.c
index f64f59d11e..ad4556c500 100644
--- a/src/mesa/vbo/vbo_context.c
+++ b/src/mesa/vbo/vbo_context.c
@@ -25,11 +25,11 @@
  *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "mtypes.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
+#include "main/api_arrayelt.h"
 #include "vbo.h"
 #include "vbo_context.h"
-#include "imports.h"
-#include "api_arrayelt.h"
 
 /* Reach out and grab this to use as the default:
  */
diff --git a/src/mesa/vbo/vbo_exec.c b/src/mesa/vbo/vbo_exec.c
index 7d95873247..1efa74945d 100644
--- a/src/mesa/vbo/vbo_exec.c
+++ b/src/mesa/vbo/vbo_exec.c
@@ -26,14 +26,14 @@
  */
 
 
-#include "api_arrayelt.h"
-#include "glheader.h"
-#include "imports.h"
-#include "context.h"
-#include "macros.h"
-#include "mtypes.h"
-#include "dlist.h"
-#include "vtxfmt.h"
+#include "main/api_arrayelt.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "main/mtypes.h"
+#include "main/dlist.h"
+#include "main/vtxfmt.h"
 
 #include "vbo_context.h"
 
diff --git a/src/mesa/vbo/vbo_exec.h b/src/mesa/vbo/vbo_exec.h
index a9b01e08e6..b7e8c9fe79 100644
--- a/src/mesa/vbo/vbo_exec.h
+++ b/src/mesa/vbo/vbo_exec.h
@@ -34,7 +34,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #ifndef __VBO_EXEC_H__
 #define __VBO_EXEC_H__
 
-#include "mtypes.h"
+#include "main/mtypes.h"
 #include "vbo.h"
 #include "vbo_attrib.h"
 
diff --git a/src/mesa/vbo/vbo_exec_api.c b/src/mesa/vbo/vbo_exec_api.c
index 8d13ef9d08..7f56b3b629 100644
--- a/src/mesa/vbo/vbo_exec_api.c
+++ b/src/mesa/vbo/vbo_exec_api.c
@@ -30,19 +30,24 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
  *   Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "glheader.h"
-#include "context.h"
-#include "macros.h"
-#include "vtxfmt.h"
-#include "dlist.h"
-#include "state.h"
-#include "light.h"
-#include "api_arrayelt.h"
-#include "api_noop.h"
-#include "dispatch.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "main/vtxfmt.h"
+#include "main/dlist.h"
+#include "main/state.h"
+#include "main/light.h"
+#include "main/api_arrayelt.h"
+#include "main/api_noop.h"
+#include "glapi/dispatch.h"
 
 #include "vbo_context.h"
 
+#ifdef ERROR
+#undef ERROR
+#endif
+
+
 static void reset_attrfv( struct vbo_exec_context *exec );
 
 
diff --git a/src/mesa/vbo/vbo_exec_array.c b/src/mesa/vbo/vbo_exec_array.c
index 1e4c310203..77f3cf1455 100644
--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -25,12 +25,12 @@
  * 
  **************************************************************************/
 
-#include "glheader.h"
-#include "context.h"
-#include "state.h"
-#include "api_validate.h"
-#include "api_noop.h"
-#include "dispatch.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/state.h"
+#include "main/api_validate.h"
+#include "main/api_noop.h"
+#include "glapi/dispatch.h"
 
 #include "vbo_context.h"
 
diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index d8f167b357..0ef26cdfe3 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -25,11 +25,11 @@
  *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "glheader.h"
-#include "context.h"
-#include "enums.h"
-#include "state.h"
-#include "macros.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/enums.h"
+#include "main/state.h"
+#include "main/macros.h"
 
 #include "vbo_context.h"
 
diff --git a/src/mesa/vbo/vbo_exec_eval.c b/src/mesa/vbo/vbo_exec_eval.c
index fe533290bd..0ba5585d24 100644
--- a/src/mesa/vbo/vbo_exec_eval.c
+++ b/src/mesa/vbo/vbo_exec_eval.c
@@ -25,13 +25,13 @@
  *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "glheader.h"
-#include "api_eval.h"
-#include "context.h"
-#include "macros.h"
+#include "main/glheader.h"
+#include "main/api_eval.h"
+#include "main/context.h"
+#include "main/macros.h"
 #include "math/m_eval.h"
+#include "glapi/dispatch.h"
 #include "vbo_exec.h"
-#include "dispatch.h"
 
 
 static void clear_active_eval1( struct vbo_exec_context *exec, GLuint attr ) 
diff --git a/src/mesa/vbo/vbo_rebase.c b/src/mesa/vbo/vbo_rebase.c
index bc4211d852..dae778e741 100644
--- a/src/mesa/vbo/vbo_rebase.c
+++ b/src/mesa/vbo/vbo_rebase.c
@@ -46,9 +46,9 @@
  * of zero.
  */
 
-#include "glheader.h"
-#include "imports.h"
-#include "mtypes.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
 
 #include "vbo.h"
 
diff --git a/src/mesa/vbo/vbo_save.c b/src/mesa/vbo/vbo_save.c
index e7f4687963..87248e10f3 100644
--- a/src/mesa/vbo/vbo_save.c
+++ b/src/mesa/vbo/vbo_save.c
@@ -26,10 +26,10 @@
  */
 
 
-#include "mtypes.h"
-#include "dlist.h"
-#include "vtxfmt.h"
-#include "imports.h"
+#include "main/mtypes.h"
+#include "main/dlist.h"
+#include "main/vtxfmt.h"
+#include "main/imports.h"
 
 #include "vbo_context.h"
 
diff --git a/src/mesa/vbo/vbo_save.h b/src/mesa/vbo/vbo_save.h
index b81f275a60..b7e9baabf8 100644
--- a/src/mesa/vbo/vbo_save.h
+++ b/src/mesa/vbo/vbo_save.h
@@ -34,7 +34,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #ifndef VBO_SAVE_H
 #define VBO_SAVE_H
 
-#include "mtypes.h"
+#include "main/mtypes.h"
 #include "vbo.h"
 #include "vbo_attrib.h"
 
diff --git a/src/mesa/vbo/vbo_save_api.c b/src/mesa/vbo/vbo_save_api.c
index c08fd1fe55..aded738143 100644
--- a/src/mesa/vbo/vbo_save_api.c
+++ b/src/mesa/vbo/vbo_save_api.c
@@ -67,19 +67,23 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
 
-#include "glheader.h"
-#include "context.h"
-#include "dlist.h"
-#include "enums.h"
-#include "macros.h"
-#include "api_validate.h"
-#include "api_arrayelt.h"
-#include "vtxfmt.h"
-#include "dispatch.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/dlist.h"
+#include "main/enums.h"
+#include "main/macros.h"
+#include "main/api_validate.h"
+#include "main/api_arrayelt.h"
+#include "main/vtxfmt.h"
+#include "glapi/dispatch.h"
 
 #include "vbo_context.h"
 
 
+#ifdef ERROR
+#undef ERROR
+#endif
+
 
 /*
  * NOTE: Old 'parity' issue is gone, but copying can still be
diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
index 8940551d08..3c6f0fccd9 100644
--- a/src/mesa/vbo/vbo_save_draw.c
+++ b/src/mesa/vbo/vbo_save_draw.c
@@ -26,13 +26,13 @@
  *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
-#include "glheader.h"
-#include "context.h"
-#include "imports.h"
-#include "mtypes.h"
-#include "macros.h"
-#include "light.h"
-#include "state.h"
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
+#include "main/macros.h"
+#include "main/light.h"
+#include "main/state.h"
 
 #include "vbo_context.h"
 
diff --git a/src/mesa/vbo/vbo_save_loopback.c b/src/mesa/vbo/vbo_save_loopback.c
index 430333b84d..f2cef698fb 100644
--- a/src/mesa/vbo/vbo_save_loopback.c
+++ b/src/mesa/vbo/vbo_save_loopback.c
@@ -28,17 +28,17 @@
 #include "swrast_setup/swrast_setup.h"
 #include "swrast/swrast.h"
 #include "tnl/tnl.h"
-#include "context.h"
+#include "main/context.h"
+#include "main/glheader.h"
+#include "main/enums.h"
+#include "main/imports.h"
+#include "main/macros.h"
+#include "main/mtypes.h"
+#include "glapi/dispatch.h"
+#include "glapi/glapi.h"
 
 #include "vbo_context.h"
 
-#include "glheader.h"
-#include "enums.h"
-#include "glapi.h"
-#include "imports.h"
-#include "macros.h"
-#include "mtypes.h"
-#include "dispatch.h"
 
 
 typedef void (*attr_func)( GLcontext *ctx, GLint target, const GLfloat * );
diff --git a/src/mesa/vbo/vbo_split.c b/src/mesa/vbo/vbo_split.c
index ef205a3bb1..58e879628d 100644
--- a/src/mesa/vbo/vbo_split.c
+++ b/src/mesa/vbo/vbo_split.c
@@ -47,9 +47,9 @@
  * limitations on drivers which want to use it as a fallback path.
  */
 
-#include "glheader.h"
-#include "imports.h"
-#include "mtypes.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "main/mtypes.h"
 
 #include "vbo_split.h"
 #include "vbo.h"
diff --git a/src/mesa/vbo/vbo_split_copy.c b/src/mesa/vbo/vbo_split_copy.c
index e142dde680..e5c4429350 100644
--- a/src/mesa/vbo/vbo_split_copy.c
+++ b/src/mesa/vbo/vbo_split_copy.c
@@ -29,11 +29,11 @@
 /* Split indexed primitives with per-vertex copying.
  */
 
-#include "glheader.h"
-#include "imports.h"
-#include "macros.h"
-#include "enums.h"
-#include "mtypes.h"
+#include "main/glheader.h"
+#include "main/imports.h"
+#include "main/macros.h"
+#include "main/enums.h"
+#include "main/mtypes.h"
 
 #include "vbo_split.h"
 #include "vbo.h"
diff --git a/src/mesa/vbo/vbo_split_inplace.c b/src/mesa/vbo/vbo_split_inplace.c
index ea62866e7c..958afccd0c 100644
--- a/src/mesa/vbo/vbo_split_inplace.c
+++ b/src/mesa/vbo/vbo_split_inplace.c
@@ -27,9 +27,9 @@
  */
 
 
-#include "mtypes.h"
-#include "macros.h"
-#include "enums.h"
+#include "main/mtypes.h"
+#include "main/macros.h"
+#include "main/enums.h"
 #include "vbo_split.h"
 
 
diff --git a/src/mesa/x86-64/glapi_x86-64.S b/src/mesa/x86-64/glapi_x86-64.S
index e62dde8a2f..171e95d1b8 100644
--- a/src/mesa/x86-64/glapi_x86-64.S
+++ b/src/mesa/x86-64/glapi_x86-64.S
@@ -29398,12 +29398,17 @@ GL_PREFIX(_dispatch_stub_772):
 	.globl GL_PREFIX(VertexAttrib4Nubv) ; .set GL_PREFIX(VertexAttrib4Nubv), GL_PREFIX(VertexAttrib4NubvARB)
 	.globl GL_PREFIX(VertexAttrib4Nuiv) ; .set GL_PREFIX(VertexAttrib4Nuiv), GL_PREFIX(VertexAttrib4NuivARB)
 	.globl GL_PREFIX(VertexAttrib4Nusv) ; .set GL_PREFIX(VertexAttrib4Nusv), GL_PREFIX(VertexAttrib4NusvARB)
+	.globl GL_PREFIX(VertexAttrib4bv) ; .set GL_PREFIX(VertexAttrib4bv), GL_PREFIX(VertexAttrib4bvARB)
 	.globl GL_PREFIX(VertexAttrib4d) ; .set GL_PREFIX(VertexAttrib4d), GL_PREFIX(VertexAttrib4dARB)
 	.globl GL_PREFIX(VertexAttrib4dv) ; .set GL_PREFIX(VertexAttrib4dv), GL_PREFIX(VertexAttrib4dvARB)
 	.globl GL_PREFIX(VertexAttrib4f) ; .set GL_PREFIX(VertexAttrib4f), GL_PREFIX(VertexAttrib4fARB)
 	.globl GL_PREFIX(VertexAttrib4fv) ; .set GL_PREFIX(VertexAttrib4fv), GL_PREFIX(VertexAttrib4fvARB)
+	.globl GL_PREFIX(VertexAttrib4iv) ; .set GL_PREFIX(VertexAttrib4iv), GL_PREFIX(VertexAttrib4ivARB)
 	.globl GL_PREFIX(VertexAttrib4s) ; .set GL_PREFIX(VertexAttrib4s), GL_PREFIX(VertexAttrib4sARB)
 	.globl GL_PREFIX(VertexAttrib4sv) ; .set GL_PREFIX(VertexAttrib4sv), GL_PREFIX(VertexAttrib4svARB)
+	.globl GL_PREFIX(VertexAttrib4ubv) ; .set GL_PREFIX(VertexAttrib4ubv), GL_PREFIX(VertexAttrib4ubvARB)
+	.globl GL_PREFIX(VertexAttrib4uiv) ; .set GL_PREFIX(VertexAttrib4uiv), GL_PREFIX(VertexAttrib4uivARB)
+	.globl GL_PREFIX(VertexAttrib4usv) ; .set GL_PREFIX(VertexAttrib4usv), GL_PREFIX(VertexAttrib4usvARB)
 	.globl GL_PREFIX(VertexAttribPointer) ; .set GL_PREFIX(VertexAttribPointer), GL_PREFIX(VertexAttribPointerARB)
 	.globl GL_PREFIX(BindBuffer) ; .set GL_PREFIX(BindBuffer), GL_PREFIX(BindBufferARB)
 	.globl GL_PREFIX(BufferData) ; .set GL_PREFIX(BufferData), GL_PREFIX(BufferDataARB)
diff --git a/src/mesa/x86/glapi_x86.S b/src/mesa/x86/glapi_x86.S
index bdf42ac088..1d04779d14 100644
--- a/src/mesa/x86/glapi_x86.S
+++ b/src/mesa/x86/glapi_x86.S
@@ -58,12 +58,18 @@
 
 #ifdef GLX_USE_TLS
 
+#ifdef GLX_X86_READONLY_TEXT
+# define CTX_INSNS MOV_L(GS:(EAX), EAX)
+#else
+# define CTX_INSNS NOP /* Pad for init_glapi_relocs() */
+#endif
+
 #  define GL_STUB(fn,off,fn_alt)			\
 ALIGNTEXT16;						\
 GLOBL_FN(GL_PREFIX(fn, fn_alt));			\
 GL_PREFIX(fn, fn_alt):					\
 	CALL(_x86_get_dispatch) ;			\
-	NOP ;						\
+	CTX_INSNS ;					\
 	JMP(GL_OFFSET(off))
 
 #elif defined(PTHREADS)
@@ -114,7 +120,10 @@ SEG_TEXT
 	HIDDEN(GLNAME(_x86_get_dispatch))
 ALIGNTEXT16
 GLNAME(_x86_get_dispatch):
-	movl	%gs:_glapi_tls_Dispatch@NTPOFF, %eax
+	call	1f
+1:	popl	%eax
+	addl	$_GLOBAL_OFFSET_TABLE_+[.-1b], %eax
+	movl	_glapi_tls_Dispatch@GOTNTPOFF(%eax), %eax
 	ret
 
 #elif defined(PTHREADS)
@@ -133,7 +142,7 @@ GLNAME(_x86_get_dispatch):
 EXTERN GLNAME(_glapi_get_dispatch)
 #endif
 
-#if defined( GLX_USE_TLS )
+#if defined( GLX_USE_TLS ) && !defined( GLX_X86_READONLY_TEXT )
 		.section	wtext, "awx", @progbits
 #endif /* defined( GLX_USE_TLS ) */
 
@@ -1120,12 +1129,17 @@ GLNAME(gl_dispatch_functions_start):
 	GL_STUB_ALIAS(VertexAttrib4Nubv, _gloffset_VertexAttrib4NubvARB, VertexAttrib4Nubv@8, VertexAttrib4NubvARB, VertexAttrib4NubvARB@8)
 	GL_STUB_ALIAS(VertexAttrib4Nuiv, _gloffset_VertexAttrib4NuivARB, VertexAttrib4Nuiv@8, VertexAttrib4NuivARB, VertexAttrib4NuivARB@8)
 	GL_STUB_ALIAS(VertexAttrib4Nusv, _gloffset_VertexAttrib4NusvARB, VertexAttrib4Nusv@8, VertexAttrib4NusvARB, VertexAttrib4NusvARB@8)
+	GL_STUB_ALIAS(VertexAttrib4bv, _gloffset_VertexAttrib4bvARB, VertexAttrib4bv@8, VertexAttrib4bvARB, VertexAttrib4bvARB@8)
 	GL_STUB_ALIAS(VertexAttrib4d, _gloffset_VertexAttrib4dARB, VertexAttrib4d@36, VertexAttrib4dARB, VertexAttrib4dARB@36)
 	GL_STUB_ALIAS(VertexAttrib4dv, _gloffset_VertexAttrib4dvARB, VertexAttrib4dv@8, VertexAttrib4dvARB, VertexAttrib4dvARB@8)
 	GL_STUB_ALIAS(VertexAttrib4f, _gloffset_VertexAttrib4fARB, VertexAttrib4f@20, VertexAttrib4fARB, VertexAttrib4fARB@20)
 	GL_STUB_ALIAS(VertexAttrib4fv, _gloffset_VertexAttrib4fvARB, VertexAttrib4fv@8, VertexAttrib4fvARB, VertexAttrib4fvARB@8)
+	GL_STUB_ALIAS(VertexAttrib4iv, _gloffset_VertexAttrib4ivARB, VertexAttrib4iv@8, VertexAttrib4ivARB, VertexAttrib4ivARB@8)
 	GL_STUB_ALIAS(VertexAttrib4s, _gloffset_VertexAttrib4sARB, VertexAttrib4s@20, VertexAttrib4sARB, VertexAttrib4sARB@20)
 	GL_STUB_ALIAS(VertexAttrib4sv, _gloffset_VertexAttrib4svARB, VertexAttrib4sv@8, VertexAttrib4svARB, VertexAttrib4svARB@8)
+	GL_STUB_ALIAS(VertexAttrib4ubv, _gloffset_VertexAttrib4ubvARB, VertexAttrib4ubv@8, VertexAttrib4ubvARB, VertexAttrib4ubvARB@8)
+	GL_STUB_ALIAS(VertexAttrib4uiv, _gloffset_VertexAttrib4uivARB, VertexAttrib4uiv@8, VertexAttrib4uivARB, VertexAttrib4uivARB@8)
+	GL_STUB_ALIAS(VertexAttrib4usv, _gloffset_VertexAttrib4usvARB, VertexAttrib4usv@8, VertexAttrib4usvARB, VertexAttrib4usvARB@8)
 	GL_STUB_ALIAS(VertexAttribPointer, _gloffset_VertexAttribPointerARB, VertexAttribPointer@24, VertexAttribPointerARB, VertexAttribPointerARB@24)
 	GL_STUB_ALIAS(BindBuffer, _gloffset_BindBufferARB, BindBuffer@8, BindBufferARB, BindBufferARB@8)
 	GL_STUB_ALIAS(BufferData, _gloffset_BufferDataARB, BufferData@16, BufferDataARB, BufferDataARB@16)
author	Jouk <joukj@tarantella.(none)>	2007-07-25 10:16:11 +0200
committer	Jouk <joukj@tarantella.(none)>	2007-07-25 10:16:11 +0200
commit	eb9a5b6d5127858b01ec12672c999e7d25cd7aed (patch)
tree	91c0302c4d007ddcd0c42fbb8d8de71bf23d3914 /src
parent	55f8b7053065ce88296608071feed6f9540f6c0d (diff)
parent	03ec41ddc51e539c989a546f33d22daa2af69095 (diff)