Merge remote branch 'origin/master' into 965-ttm

Conflicts:

	src/mesa/drivers/dri/common/dri_bufmgr.h

58 files changed, 945 insertions, 673 deletions

diff --git a/src/glw/glw.pc.in b/src/glw/glw.pc.in
index 951e2dc2af..9c2682fbf3 100644
--- a/src/glw/glw.pc.in
+++ b/src/glw/glw.pc.in
@@ -7,5 +7,5 @@ Name: glw
 Description: Mesa OpenGL widget library
 Requires: gl
 Version: @VERSION@
-Libs: -L${libdir} -lGLU
+Libs: -L${libdir} -lGLw
 Cflags: -I${includedir}
diff --git a/src/glx/x11/dri_glx.c b/src/glx/x11/dri_glx.c
index c02f105611..dab454e8e3 100644
--- a/src/glx/x11/dri_glx.c
+++ b/src/glx/x11/dri_glx.c
@@ -194,7 +194,8 @@ static __DRIdriver *OpenDriver(const char *driverName)
    /* First, search Drivers list to see if we've already opened this driver */
    for (driver = Drivers; driver; driver = driver->next) {
       if (strcmp(driver->name, driverName) == 0) {
-         /* found it */
+         /* found it, increment library refcount & return */
+         dlopen(driver->libpath, RTLD_NOW | RTLD_GLOBAL);
          return driver;
       }
    }
@@ -238,7 +239,12 @@ static __DRIdriver *OpenDriver(const char *driverName)
             break; /* out of memory! */
          /* init the struct */
          driver->name = __glXstrdup(driverName);
-         if (!driver->name) {
+         driver->libpath = __glXstrdup(realDriverName);
+         if (!driver->name || !driver->libpath) {
+            if (driver->name)
+               Xfree(driver->name);
+            if (driver->libpath)
+               Xfree(driver->libpath);
             Xfree(driver);
             driver = NULL;
             break; /* out of memory! */
@@ -401,6 +407,7 @@ static void driDestroyDisplay(Display *dpy, void *private)
 		       Drivers = driver->next;
 
 		    Xfree(driver->name);
+		    Xfree(driver->libpath);
 		    Xfree(driver);
 		    break;
 		 }
diff --git a/src/glx/x11/glxclient.h b/src/glx/x11/glxclient.h
index 05354073c4..7054f3cc31 100644
--- a/src/glx/x11/glxclient.h
+++ b/src/glx/x11/glxclient.h
@@ -117,6 +117,7 @@ struct __DRIdisplayRec {
 */
 struct __DRIdriverRec {
    const char *name;
+   const char *libpath;
    void *handle;
    PFNCREATENEWSCREENFUNC createNewScreenFunc;
    struct __DRIdriverRec *next;
diff --git a/src/glx/x11/glxcmds.c b/src/glx/x11/glxcmds.c
index 2d217517af..7ff2bf700c 100644
--- a/src/glx/x11/glxcmds.c
+++ b/src/glx/x11/glxcmds.c
@@ -101,6 +101,7 @@ static void GarbageCollectDRIDrawables(Display *dpy, __GLXscreenConfigs *sc)
 		   longer exists in the Xserver */
 		(*pdraw->driDrawable.destroyDrawable)(&pdraw->driDrawable);
 		XF86DRIDestroyDrawable(dpy, sc->scr, draw);
+                __glxHashDelete(sc->drawHash, draw);
 		Xfree(pdraw);
 	    }
 	} while (__glxHashNext(sc->drawHash, &draw, (void *)&pdraw) == 1);
diff --git a/src/mesa/Makefile b/src/mesa/Makefile
index 77bb7be480..2c6d65d8a1 100644
--- a/src/mesa/Makefile
+++ b/src/mesa/Makefile
@@ -65,7 +65,7 @@ linux-solo: depend subdirs libmesa.a
 # Stand-alone Mesa libGL, no built-in drivers (DirectFB)
 
 libgl-core: $(CORE_OBJECTS)
-	@ $(TOP)/bin/mklib -o $(GL_LIB) -linker '$(CC)' -ldflags '$(LDFLAGS) \
+	@ $(TOP)/bin/mklib -o $(GL_LIB) -linker '$(CC)' -ldflags '$(LDFLAGS)' \
 		-major $(MESA_MAJOR) -minor $(MESA_MINOR) -patch $(MESA_TINY) \
 		-install $(TOP)/$(LIB_DIR) $(MKLIB_OPTIONS) $(CORE_OBJECTS) \
 		$(GL_LIB_DEPS)
diff --git a/src/mesa/descrip.mms b/src/mesa/descrip.mms
index a12e3fc1b7..dbfa336a91 100644
--- a/src/mesa/descrip.mms
+++ b/src/mesa/descrip.mms
@@ -1,5 +1,6 @@
 # Makefile for Mesa for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Date last revision : 7 March 2007
 
 all :
 	set default [.main]
@@ -22,5 +23,5 @@ all :
 	$(MMS)$(MMSQUALIFIERS)
 	set default [-.swrast_setup]
 	$(MMS)$(MMSQUALIFIERS)
-	set default [-.array_cache]
+	set default [-.vbo]
 	$(MMS)$(MMSQUALIFIERS)
diff --git a/src/mesa/drivers/common/descrip.mms b/src/mesa/drivers/common/descrip.mms
index 3cb856d12c..99a2ae6c37 100644
--- a/src/mesa/drivers/common/descrip.mms
+++ b/src/mesa/drivers/common/descrip.mms
@@ -1,12 +1,15 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 2 November 2005
+# Last revision : 3 October 2007
 
 .first
 	define gl [----.include.gl]
 	define math [--.math]
 	define tnl [--.tnl]
 	define swrast [--.swrast]
+	define glapi [--.glapi]
+	define shader [--.shader]
+	define main [--.main]
 
 .include [----]mms-config.
 
diff --git a/src/mesa/drivers/common/driverfuncs.c b/src/mesa/drivers/common/driverfuncs.c
index 1e7c279875..03fbab69e3 100644
--- a/src/mesa/drivers/common/driverfuncs.c
+++ b/src/mesa/drivers/common/driverfuncs.c
@@ -112,6 +112,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->DeleteTexture = _mesa_delete_texture_object;
    driver->NewTextureImage = _mesa_new_texture_image;
    driver->FreeTexImageData = _mesa_free_texture_image_data; 
+   driver->MapTexture = NULL;
+   driver->UnmapTexture = NULL;
    driver->TextureMemCpy = _mesa_memcpy; 
    driver->IsTextureResident = NULL;
    driver->PrioritizeTexture = NULL;
diff --git a/src/mesa/drivers/dri/common/dri_bufmgr.c b/src/mesa/drivers/dri/common/dri_bufmgr.c
index 72a4a17150..757a237494 100644
--- a/src/mesa/drivers/dri/common/dri_bufmgr.c
+++ b/src/mesa/drivers/dri/common/dri_bufmgr.c
@@ -35,7 +35,7 @@
 
 dri_bo *
 dri_bo_alloc(dri_bufmgr *bufmgr, const char *name, unsigned long size,
-	     unsigned int alignment, unsigned int location_mask)
+	     unsigned int alignment, uint64_t location_mask)
 {
    assert((location_mask & ~(DRM_BO_FLAG_MEM_LOCAL | DRM_BO_FLAG_MEM_TT |
 			     DRM_BO_FLAG_MEM_VRAM | DRM_BO_FLAG_MEM_PRIV0 |
@@ -48,7 +48,7 @@ dri_bo_alloc(dri_bufmgr *bufmgr, const char *name, unsigned long size,
 dri_bo *
 dri_bo_alloc_static(dri_bufmgr *bufmgr, const char *name, unsigned long offset,
 		    unsigned long size, void *virtual,
-		    unsigned int location_mask)
+		    uint64_t location_mask)
 {
    assert((location_mask & ~(DRM_BO_FLAG_MEM_LOCAL | DRM_BO_FLAG_MEM_TT |
 			     DRM_BO_FLAG_MEM_VRAM | DRM_BO_FLAG_MEM_PRIV0 |
@@ -139,9 +139,10 @@ dri_bufmgr_destroy(dri_bufmgr *bufmgr)
 }
 
 
-void dri_emit_reloc(dri_bo *batch_buf, GLuint flags, GLuint delta, GLuint offset, dri_bo *relocatee)
+void dri_emit_reloc(dri_bo *reloc_buf, uint64_t flags, GLuint delta,
+		    GLuint offset, dri_bo *target_buf)
 {
-   batch_buf->bufmgr->emit_reloc(batch_buf, flags, delta, offset, relocatee);
+   reloc_buf->bufmgr->emit_reloc(reloc_buf, flags, delta, offset, target_buf);
 }
 
 void *dri_process_relocs(dri_bo *batch_buf, GLuint *count)
diff --git a/src/mesa/drivers/dri/common/dri_bufmgr.h b/src/mesa/drivers/dri/common/dri_bufmgr.h
index 1290fa956b..d263ad279b 100644
--- a/src/mesa/drivers/dri/common/dri_bufmgr.h
+++ b/src/mesa/drivers/dri/common/dri_bufmgr.h
@@ -84,7 +84,7 @@ struct _dri_bufmgr {
     */
    dri_bo *(*bo_alloc)(dri_bufmgr *bufmgr_ctx, const char *name,
 		       unsigned long size, unsigned int alignment,
-		       unsigned int location_mask);
+		       uint64_t location_mask);
 
    /**
     * Allocates a buffer object for a static allocation.
@@ -94,7 +94,7 @@ struct _dri_bufmgr {
     */
    dri_bo *(*bo_alloc_static)(dri_bufmgr *bufmgr_ctx, const char *name,
 			      unsigned long offset, unsigned long size,
-			      void *virtual, unsigned int location_mask);
+			      void *virtual, uint64_t location_mask);
 
    /** Takes a reference on a buffer object */
    void (*bo_reference)(dri_bo *bo);
@@ -153,7 +153,7 @@ struct _dri_bufmgr {
     * \param target Buffer whose offset should be written into the relocation
     *	     entry.
     */
-   void (*emit_reloc)(dri_bo *reloc_buf, GLuint flags, GLuint delta,
+   void (*emit_reloc)(dri_bo *reloc_buf, uint64_t flags, GLuint delta,
 		      GLuint offset, dri_bo *target);
 
    /**
@@ -175,10 +175,10 @@ struct _dri_bufmgr {
 };
 
 dri_bo *dri_bo_alloc(dri_bufmgr *bufmgr, const char *name, unsigned long size,
-		     unsigned int alignment, unsigned int location_mask);
+		     unsigned int alignment, uint64_t location_mask);
 dri_bo *dri_bo_alloc_static(dri_bufmgr *bufmgr, const char *name,
 			    unsigned long offset, unsigned long size,
-			    void *virtual, unsigned int location_mask);
+			    void *virtual, uint64_t location_mask);
 void dri_bo_reference(dri_bo *bo);
 void dri_bo_unreference(dri_bo *bo);
 int dri_bo_map(dri_bo *buf, GLboolean write_enable);
@@ -202,6 +202,7 @@ dri_bufmgr *dri_bufmgr_fake_init(unsigned long low_offset, void *low_virtual,
 				 int (*fence_wait)(void *private,
 						   unsigned int cookie),
 				 void *driver_priv);
+void dri_bufmgr_fake_set_debug(dri_bufmgr *bufmgr, GLboolean enable_debug);
 void dri_bo_fake_disable_backing_store(dri_bo *bo,
 				       void (*invalidate_cb)(dri_bo *bo,
 							     void *ptr),
@@ -210,7 +211,8 @@ void dri_bufmgr_destroy(dri_bufmgr *bufmgr);
 dri_bo *dri_ttm_bo_create_from_handle(dri_bufmgr *bufmgr, const char *name,
 				      unsigned int handle);
 
-void dri_emit_reloc(dri_bo *batch_buf, GLuint flags, GLuint delta, GLuint offset, dri_bo *relocatee);
+void dri_emit_reloc(dri_bo *reloc_buf, uint64_t flags, GLuint delta,
+		    GLuint offset, dri_bo *target_buf);
 void *dri_process_relocs(dri_bo *batch_buf, uint32_t *count);
 void dri_post_process_relocs(dri_bo *batch_buf);
 void dri_post_submit(dri_bo *batch_buf, dri_fence **last_fence);
diff --git a/src/mesa/drivers/dri/common/dri_bufmgr_fake.c b/src/mesa/drivers/dri/common/dri_bufmgr_fake.c
index 9081f23e4a..4c5e0cac26 100644
--- a/src/mesa/drivers/dri/common/dri_bufmgr_fake.c
+++ b/src/mesa/drivers/dri/common/dri_bufmgr_fake.c
@@ -41,11 +41,10 @@
 #include "mm.h"
 #include "imports.h"
 
-#if 0
-#define DBG(...) _mesa_printf(__VA_ARGS__)
-#else
-#define DBG(...)
-#endif
+#define DBG(...) do {					\
+   if (bufmgr_fake->debug)				\
+      _mesa_printf(__VA_ARGS__);			\
+} while (0)
 
 /* Internal flags:
  */
@@ -67,7 +66,7 @@ struct fake_buffer_reloc
    dri_bo *target_buf;
    GLuint offset;
    GLuint delta;
-   GLuint validate_flags;
+   uint64_t validate_flags;
    GLboolean relocated;
 };
 
@@ -129,6 +128,7 @@ typedef struct _bufmgr_fake {
    /** Driver-supplied argument to driver callbacks */
    void *driver_priv;
 
+   GLboolean debug;
 
    /** fake relocation list */
    struct fake_buffer_reloc reloc[MAX_RELOCS];
@@ -161,7 +161,7 @@ typedef struct _dri_bo_fake {
     * DRM_BO_NO_BACKING_STORE and BM_NO_FENCE_SUBDATA, which are the first two
     * driver private flags.
     */
-   unsigned int flags;
+   uint64_t flags;
    unsigned int alignment;
    GLboolean is_static, validated;
    unsigned int map_count;
@@ -329,6 +329,7 @@ free_backing_store(dri_bo *bo)
 static void
 set_dirty(dri_bo *bo)
 {
+   dri_bufmgr_fake *bufmgr_fake = (dri_bufmgr_fake *)bo->bufmgr;
    dri_bo_fake *bo_fake = (dri_bo_fake *)bo;
 
    if (bo_fake->flags & BM_NO_BACKING_STORE && bo_fake->invalidate_cb != NULL)
@@ -575,7 +576,7 @@ dri_bufmgr_fake_contended_lock_take(dri_bufmgr *bufmgr)
 static dri_bo *
 dri_fake_bo_alloc(dri_bufmgr *bufmgr, const char *name,
 		  unsigned long size, unsigned int alignment,
-		  unsigned int location_mask)
+		  uint64_t location_mask)
 {
    dri_bufmgr_fake *bufmgr_fake;
    dri_bo_fake *bo_fake;
@@ -613,7 +614,7 @@ dri_fake_bo_alloc(dri_bufmgr *bufmgr, const char *name,
 static dri_bo *
 dri_fake_bo_alloc_static(dri_bufmgr *bufmgr, const char *name,
 			 unsigned long offset, unsigned long size,
-			 void *virtual, unsigned int location_mask)
+			 void *virtual, uint64_t location_mask)
 {
    dri_bufmgr_fake *bufmgr_fake;
    dri_bo_fake *bo_fake;
@@ -790,6 +791,7 @@ dri_fake_bo_unmap(dri_bo *bo)
       return 0;
 
    _glthread_LOCK_MUTEX(bufmgr_fake->mutex);
+   assert(bo_fake->map_count != 0);
    if (--bo_fake->map_count != 0) {
       _glthread_UNLOCK_MUTEX(bufmgr_fake->mutex);
       return 0;
@@ -806,7 +808,7 @@ dri_fake_bo_unmap(dri_bo *bo)
 }
 
 static int
-dri_fake_bo_validate(dri_bo *bo, unsigned int flags)
+dri_fake_bo_validate(dri_bo *bo, uint64_t flags)
 {
    dri_bufmgr_fake *bufmgr_fake;
    dri_bo_fake *bo_fake = (dri_bo_fake *)bo;
@@ -814,13 +816,19 @@ dri_fake_bo_validate(dri_bo *bo, unsigned int flags)
    /* XXX: Sanity-check whether we've already validated this one under
     * different flags.  See drmAddValidateItem().
     */
+   bufmgr_fake = (dri_bufmgr_fake *)bo->bufmgr;
 
    DBG("drm_bo_validate: (buf %d: %s, %d kb)\n", bo_fake->id, bo_fake->name,
        bo_fake->bo.size / 1024);
-   bufmgr_fake = (dri_bufmgr_fake *)bo->bufmgr;
 
    _glthread_LOCK_MUTEX(bufmgr_fake->mutex);
    {
+      /* Sanity check: Buffers should be unmapped before being validated.
+       * This is not so much of a problem for bufmgr_fake, but TTM refuses,
+       * and the problem is harder to debug there.
+       */
+      assert(bo_fake->map_count == 0);
+
       if (bo_fake->is_static) {
 	 /* Add it to the needs-fence list */
 	 bufmgr_fake->need_fence = 1;
@@ -950,7 +958,7 @@ dri_fake_destroy(dri_bufmgr *bufmgr)
 }
 
 static void
-dri_fake_emit_reloc(dri_bo *reloc_buf, GLuint flags, GLuint delta,
+dri_fake_emit_reloc(dri_bo *reloc_buf, uint64_t flags, GLuint delta,
 		    GLuint offset, dri_bo *target_buf)
 {
    dri_bufmgr_fake *bufmgr_fake = (dri_bufmgr_fake *)reloc_buf->bufmgr;
@@ -1133,6 +1141,14 @@ dri_fake_post_submit(dri_bo *batch_buf, dri_fence **last_fence)
    bufmgr_fake->nr_relocs = 0;
 }
 
+void
+dri_bufmgr_fake_set_debug(dri_bufmgr *bufmgr, GLboolean enable_debug)
+{
+   dri_bufmgr_fake *bufmgr_fake = (dri_bufmgr_fake *)bufmgr;
+
+   bufmgr_fake->debug = enable_debug;
+}
+
 dri_bufmgr *
 dri_bufmgr_fake_init(unsigned long low_offset, void *low_virtual,
 		     unsigned long size,
diff --git a/src/mesa/drivers/dri/common/dri_util.c b/src/mesa/drivers/dri/common/dri_util.c
index 2e2e64c4d1..2b378a531e 100644
--- a/src/mesa/drivers/dri/common/dri_util.c
+++ b/src/mesa/drivers/dri/common/dri_util.c
@@ -336,6 +336,9 @@ static void driSwapBuffers(__DRIdrawable *drawable)
     __DRIdrawablePrivate *dPriv = drawable->private;
     drm_clip_rect_t rect;
 
+    if (!dPriv->numClipRects)
+        return;
+
     dPriv->swapBuffers(dPriv);
 
     /* Check that we actually have the new damage report method */
diff --git a/src/mesa/drivers/dri/i915/i830_vtbl.c b/src/mesa/drivers/dri/i915/i830_vtbl.c
index eecff2729f..c23c02c060 100644
--- a/src/mesa/drivers/dri/i915/i830_vtbl.c
+++ b/src/mesa/drivers/dri/i915/i830_vtbl.c
@@ -383,9 +383,15 @@ do {							\
 } while (0)
 
 static GLuint
+get_dirty(struct i830_hw_state *state)
+{
+   return state->active & ~state->emitted;
+}
+
+static GLuint
 get_state_size(struct i830_hw_state *state)
 {
-   GLuint dirty = state->active & ~state->emitted;
+   GLuint dirty = get_dirty(state);
    GLuint sz = 0;
    GLuint i;
 
@@ -416,7 +422,7 @@ get_state_size(struct i830_hw_state *state)
 /* Push the state into the sarea and/or texture memory.
  */
 static void
-i830_emit_state(struct intel_context *intel)
+i830_do_emit_state(struct intel_context *intel)
 {
    struct i830_context *i830 = i830_context(&intel->ctx);
    struct i830_hw_state *state = i830->current;
@@ -433,10 +439,32 @@ i830_emit_state(struct intel_context *intel)
     */
    intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
 
+   /* Workaround.  There are cases I haven't been able to track down
+    * where we aren't emitting a full state at the start of a new
+    * batchbuffer.  This code spots that we are on a new batchbuffer
+    * and forces a full state emit no matter what.  
+    *
+    * In the normal case state->emitted is already zero, this code is
+    * another set of checks to make sure it really is.
+    */
+   if (intel->batch->id != intel->last_state_batch_id ||
+       intel->batch->map == intel->batch->ptr) 
+   {
+      state->emitted = 0;
+      intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
+   }
+
    /* Do this here as we may have flushed the batchbuffer above,
     * causing more state to be dirty!
     */
-   dirty = state->active & ~state->emitted;
+   dirty = get_dirty(state);
+   state->emitted |= dirty;
+   assert(get_dirty(state) == 0);
+
+   if (intel->batch->id != intel->last_state_batch_id) {
+      assert(dirty & I830_UPLOAD_CTX);
+      intel->last_state_batch_id = intel->batch->id;
+   }
 
    if (dirty & I830_UPLOAD_INVARIENT) {
       DBG("I830_UPLOAD_INVARIENT:\n");
@@ -515,7 +543,30 @@ i830_emit_state(struct intel_context *intel)
       }
    }
 
-   state->emitted |= dirty;
+   intel->batch->dirty_state &= ~dirty;
+   assert(get_dirty(state) == 0);
+}
+
+static void
+i830_emit_state(struct intel_context *intel)
+{
+   struct i830_context *i830 = i830_context(&intel->ctx);
+
+   i830_do_emit_state( intel );
+
+   /* Second chance - catch batchbuffer wrap in the middle of state
+    * emit.  This shouldn't happen but it has been observed in
+    * testing.
+    */
+   if (get_dirty( i830->current )) {
+      /* Force a full re-emit if this happens.
+       */
+      i830->current->emitted = 0;
+      i830_do_emit_state( intel );
+   }
+
+   assert(get_dirty(i830->current) == 0);
+   assert((intel->batch->dirty_state & (1<<1)) == 0);
 }
 
 static void
@@ -652,8 +703,7 @@ i830_assert_not_dirty( struct intel_context *intel )
 {
    struct i830_context *i830 = i830_context(&intel->ctx);
    struct i830_hw_state *state = i830->current;
-   GLuint dirty = state->active & ~state->emitted;
-   assert(!dirty);
+   assert(!get_dirty(state));
 }
 
 
diff --git a/src/mesa/drivers/dri/i915/i915_vtbl.c b/src/mesa/drivers/dri/i915/i915_vtbl.c
index 5decffa116..58f681f74b 100644
--- a/src/mesa/drivers/dri/i915/i915_vtbl.c
+++ b/src/mesa/drivers/dri/i915/i915_vtbl.c
@@ -442,14 +442,16 @@ i915_do_emit_state(struct intel_context *intel)
    }
 
    if (dirty & I915_UPLOAD_PROGRAM) {
-      if (INTEL_DEBUG & DEBUG_STATE)
-         fprintf(stderr, "I915_UPLOAD_PROGRAM:\n");
+      if (state->ProgramSize) {
+         if (INTEL_DEBUG & DEBUG_STATE)
+            fprintf(stderr, "I915_UPLOAD_PROGRAM:\n");
 
-      assert((state->Program[0] & 0x1ff) + 2 == state->ProgramSize);
+         assert((state->Program[0] & 0x1ff) + 2 == state->ProgramSize);
 
-      emit(intel, state->Program, state->ProgramSize * sizeof(GLuint));
-      if (INTEL_DEBUG & DEBUG_STATE)
-         i915_disassemble_program(state->Program, state->ProgramSize);
+         emit(intel, state->Program, state->ProgramSize * sizeof(GLuint));
+         if (INTEL_DEBUG & DEBUG_STATE)
+            i915_disassemble_program(state->Program, state->ProgramSize);
+      }
    }
 
    intel->batch->dirty_state &= ~dirty;
diff --git a/src/mesa/drivers/dri/i915/intel_context.c b/src/mesa/drivers/dri/i915/intel_context.c
index a26a424f86..326be9c76f 100644
--- a/src/mesa/drivers/dri/i915/intel_context.c
+++ b/src/mesa/drivers/dri/i915/intel_context.c
@@ -460,6 +460,8 @@ intelInitContext(struct intel_context *intel,
 
 #if DO_DEBUG
    INTEL_DEBUG = driParseDebugString(getenv("INTEL_DEBUG"), debug_control);
+   if (!intel->intelScreen->ttm && (INTEL_DEBUG & DEBUG_BUFMGR))
+      dri_bufmgr_fake_set_debug(intel->intelScreen->bufmgr, GL_TRUE);
 #endif
 
    if (getenv("INTEL_NO_RAST")) {
diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index bdfdfd1704..4654ab1ddf 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -70,12 +70,14 @@ static void brwUseProgram(GLcontext *ctx, GLuint program)
         if (sh_prog) {
 	    if (sh_prog->VertexProgram) {
 		brw->attribs.VertexProgram->Current = sh_prog->VertexProgram;
+		sh_prog->VertexProgram->Base.RefCount++;
 		ctx->VertexProgram.Enabled = GL_TRUE;
 	    }else
 		ctx->VertexProgram.Enabled = GL_FALSE;
 		
 	    if (sh_prog->FragmentProgram) {
 		brw->attribs.FragmentProgram->Current = sh_prog->FragmentProgram;
+		sh_prog->FragmentProgram->Base.RefCount++;
 		ctx->FragmentProgram.Enabled = GL_TRUE;
 	    } else
 		ctx->FragmentProgram.Enabled = GL_FALSE;
diff --git a/src/mesa/drivers/dri/i965/brw_curbe.c b/src/mesa/drivers/dri/i965/brw_curbe.c
index 43e66ff3fd..4007dbf9e9 100644
--- a/src/mesa/drivers/dri/i965/brw_curbe.c
+++ b/src/mesa/drivers/dri/i965/brw_curbe.c
@@ -184,6 +184,7 @@ static GLfloat fixed_plane[6][4] = {
  */
 static void upload_constant_buffer(struct brw_context *brw)
 {
+   struct intel_context *intel = &brw->intel;
    GLcontext *ctx = &brw->intel.ctx;
    struct brw_vertex_program *vp = (struct brw_vertex_program *)brw->vertex_program;
    struct brw_fragment_program *fp = (struct brw_fragment_program *)brw->fragment_program;
@@ -201,13 +202,10 @@ static void upload_constant_buffer(struct brw_context *brw)
    brw->curbe.tracked_state.dirty.mesa |= fp->param_state;
 
    if (sz == 0) {
-      struct brw_constant_buffer cb;
-      cb.header.opcode = CMD_CONST_BUFFER;
-      cb.header.length = sizeof(cb)/4 - 2;
-      cb.header.valid = 0;
-      cb.bits0.buffer_length = 0;
-      cb.bits0.buffer_address = 0;
-      BRW_BATCH_STRUCT(brw, &cb);
+      BEGIN_BATCH(2, INTEL_BATCH_NO_CLIPRECTS);
+      OUT_BATCH((CMD_CONST_BUFFER << 16) | (2 - 2));
+      OUT_BATCH(0);
+      ADVANCE_BATCH();
 
       if (brw->curbe.last_buf) {
 	 free(brw->curbe.last_buf);
@@ -317,39 +315,23 @@ static void upload_constant_buffer(struct brw_context *brw)
       dri_bo_subdata(pool->buffer, brw->curbe.gs_offset, bufsz, buf);
    }
 
-   /* TODO: only emit the constant_buffer packet when necessary, ie:
-      - contents have changed
-      - offset has changed
-      - hw requirements due to other packets emitted.
-   */
-   {
-      struct brw_constant_buffer cb;
-      
-      memset(&cb, 0, sizeof(cb));
-
-      cb.header.opcode = CMD_CONST_BUFFER;
-      cb.header.length = sizeof(cb)/4 - 2;
-      cb.header.valid = 1;
-      cb.bits0.buffer_length = sz - 1;
-      cb.bits0.buffer_address = brw->curbe.gs_offset >> 6;
-      
-      /* Because this provokes an action (ie copy the constants into the
-       * URB), it shouldn't be shortcircuited if identical to the
-       * previous time - because eg. the urb destination may have
-       * changed, or the urb contents different to last time.  
-       *
-       * Note that the data referred to is actually copied internally,
-       * not just used in place according to passed pointer.
-       *
-       * It appears that the CS unit takes care of using each available
-       * URB entry (Const URB Entry == CURBE) in turn, and issuing
-       * flushes as necessary when doublebuffering of CURBEs isn't
-       * possible.
-       */
-/*       intel_batchbuffer_align(brw->intel.batch, 64, sizeof(cb)); */
-      BRW_BATCH_STRUCT(brw, &cb);
-/*       intel_batchbuffer_align(brw->intel.batch, 64, 0); */
-   }
+   /* Because this provokes an action (ie copy the constants into the
+    * URB), it shouldn't be shortcircuited if identical to the
+    * previous time - because eg. the urb destination may have
+    * changed, or the urb contents different to last time.
+    *
+    * Note that the data referred to is actually copied internally,
+    * not just used in place according to passed pointer.
+    *
+    * It appears that the CS unit takes care of using each available
+    * URB entry (Const URB Entry == CURBE) in turn, and issuing
+    * flushes as necessary when doublebuffering of CURBEs isn't
+    * possible.
+    */
+   BEGIN_BATCH(2, INTEL_BATCH_NO_CLIPRECTS);
+   OUT_BATCH((CMD_CONST_BUFFER << 16) | (1 << 8) | (2 - 2));
+   OUT_BATCH(brw->curbe.gs_offset | (sz - 1));
+   ADVANCE_BATCH();
 }
 
 /* This tracked state is unique in that the state it monitors varies
diff --git a/src/mesa/drivers/dri/i965/brw_wm_emit.c b/src/mesa/drivers/dri/i965/brw_wm_emit.c
index 1f7158f7a7..6bafa44b82 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_emit.c
@@ -810,6 +810,7 @@ static void emit_kil( struct brw_wm_compile *c,
       brw_push_insn_state(p);
       brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_GE, arg0[i], brw_imm_f(0));   
       brw_set_predicate_control_flag_value(p, 0xff);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
       brw_AND(p, r0uw, brw_flag_reg(), r0uw);
       brw_pop_insn_state(p);
    }
diff --git a/src/mesa/drivers/dri/i965/brw_wm_glsl.c b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
index 4b273fefe9..e738086fef 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_glsl.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
@@ -324,6 +324,20 @@ static void emit_fb_write(struct brw_wm_compile *c,
 	nr += 8;
 	brw_pop_insn_state(p);
     }
+
+   if (c->key.source_depth_to_render_target)
+   {
+      if (c->key.computes_depth) {
+         src0 = get_src_reg(c, &inst->SrcReg[2], 2, 1);
+         brw_MOV(p, brw_message_reg(nr), src0);
+      } else {
+         src0 = get_src_reg(c, &inst->SrcReg[1], 1, 1);
+         brw_MOV(p, brw_message_reg(nr), src0);
+      }
+
+      nr += 2;
+   }
+
     fire_fb_write(c, 0, nr);
 }
 
diff --git a/src/mesa/drivers/dri/intel/intel_batchbuffer.c b/src/mesa/drivers/dri/intel/intel_batchbuffer.c
index 7d26bd0a16..21db0e7dcd 100644
--- a/src/mesa/drivers/dri/intel/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/intel/intel_batchbuffer.c
@@ -131,6 +131,7 @@ do_flush_locked(struct intel_batchbuffer *batch,
    void *start;
    GLuint count;
 
+   dri_bo_unmap(batch->buf);
    start = dri_process_relocs(batch->buf, &count);
 
    batch->map = NULL;
diff --git a/src/mesa/drivers/dri/intel/intel_bufmgr_ttm.c b/src/mesa/drivers/dri/intel/intel_bufmgr_ttm.c
index 32d9886091..3e0d818a0c 100644
--- a/src/mesa/drivers/dri/intel/intel_bufmgr_ttm.c
+++ b/src/mesa/drivers/dri/intel/intel_bufmgr_ttm.c
@@ -1,10 +1,10 @@
 /**************************************************************************
- * 
+ *
  * Copyright © 2007 Red Hat Inc.
  * Copyright © 2007 Intel Corporation
  * Copyright 2006 Tungsten Graphics, Inc., Bismarck, ND., USA
  * 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
@@ -12,20 +12,20 @@
  * 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 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 HOLDERS, AUTHORS 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 
+ * 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.
  *
  * The above copyright notice and this permission notice (including the
  * next paragraph) shall be included in all copies or substantial portions
  * of the Software.
- * 
- * 
+ *
+ *
  **************************************************************************/
 /*
  * Authors: Thomas Hellström <thomas-at-tungstengraphics-dot-com>
@@ -52,11 +52,11 @@
 
 struct intel_reloc_info
 {
-   GLuint type;
-   GLuint reloc;
-   GLuint delta;                /* not needed? */
-   GLuint index;
-   drm_handle_t handle;
+    GLuint type;
+    GLuint reloc;
+    GLuint delta;
+    GLuint index;
+    drm_handle_t handle;
 };
 
 struct intel_bo_node
@@ -64,8 +64,8 @@ struct intel_bo_node
     drmMMListHead head;
     drmBO *buf;
     struct drm_i915_op_arg bo_arg;
-    unsigned long arg0;
-    unsigned long arg1;
+    uint64_t flags;
+    uint64_t mask;
     void (*destroy)(void *);
     void *priv;
 };
@@ -92,39 +92,40 @@ struct intel_bo_list {
 };
 
 typedef struct _dri_bufmgr_ttm {
-   dri_bufmgr bufmgr;
+    dri_bufmgr bufmgr;
 
-   int fd;
-   _glthread_Mutex mutex;
-   unsigned int fence_type;
-   unsigned int fence_type_flush;
+    int fd;
+    _glthread_Mutex mutex;
+    unsigned int fence_type;
+    unsigned int fence_type_flush;
 
-   uint32_t max_relocs;
-   /** ttm relocation list */
-   struct intel_bo_list list;
-   struct intel_bo_list reloc_list;
+    uint32_t max_relocs;
+    /** ttm relocation list */
+    struct intel_bo_list list;
+    struct intel_bo_list reloc_list;
 
 } dri_bufmgr_ttm;
 
 typedef struct _dri_bo_ttm {
-   dri_bo bo;
+    dri_bo bo;
 
-   int refcount;		/* Protected by bufmgr->mutex */
-   drmBO drm_bo;
-   const char *name;
+    int refcount;		/* Protected by bufmgr->mutex */
+    drmBO drm_bo;
+    const char *name;
 } dri_bo_ttm;
 
 typedef struct _dri_fence_ttm
 {
-   dri_fence fence;
+    dri_fence fence;
 
-   int refcount;		/* Protected by bufmgr->mutex */
-   const char *name;
-   drmFence drm_fence;
+    int refcount;		/* Protected by bufmgr->mutex */
+    const char *name;
+    drmFence drm_fence;
 } dri_fence_ttm;
 
 
-static void intel_bo_free_list(struct intel_bo_list *list)
+static void
+intel_bo_free_list(struct intel_bo_list *list)
 {
     struct intel_bo_node *node;
     drmMMListHead *l;
@@ -139,12 +140,15 @@ static void intel_bo_free_list(struct intel_bo_list *list)
     }
 }
 
-static void generic_destroy(void *nodep)
+static void
+generic_destroy(void *nodep)
 {
     free(nodep);
 }
 
-static int intel_create_bo_list(int numTarget, struct intel_bo_list *list, void (*destroy)(void *))
+static int
+intel_create_bo_list(int numTarget, struct intel_bo_list *list,
+		     void (*destroy)(void *))
 {
     DRMINITLISTHEAD(&list->list);
     list->numCurrent = 0;
@@ -157,8 +161,10 @@ static int intel_create_bo_list(int numTarget, struct intel_bo_list *list, void
 
 
 static struct drm_i915_op_arg *
-intel_setup_validate_list(int fd, struct intel_bo_list *list, struct intel_bo_list *reloc_list, GLuint *count_p)
+intel_setup_validate_list(dri_bufmgr_ttm *bufmgr_ttm, GLuint *count_p)
 {
+    struct intel_bo_list *list = &bufmgr_ttm->list;
+    struct intel_bo_list *reloc_list = &bufmgr_ttm->reloc_list;
     struct intel_bo_node *node;
     struct intel_bo_reloc_node *rl_node;
     drmMMListHead *l, *rl;
@@ -185,13 +191,15 @@ intel_setup_validate_list(int fd, struct intel_bo_list *list, struct intel_bo_li
 	prevNext = &arg->next;
 	req->bo_req.handle = node->buf->handle;
 	req->op = drm_bo_validate;
-	req->bo_req.flags = node->arg0;
+	req->bo_req.flags = node->flags;
 	req->bo_req.hint = 0;
-	req->bo_req.mask = node->arg1;
+	req->bo_req.mask = node->mask;
 	req->bo_req.fence_class = 0; /* Backwards compat. */
 	arg->reloc_handle = 0;
 
-	for (rl = reloc_list->list.next; rl != &reloc_list->list; rl = rl->next) {
+	for (rl = reloc_list->list.next; rl != &reloc_list->list;
+	     rl = rl->next)
+	{
 	    rl_node = DRMLISTENTRY(struct intel_bo_reloc_node, rl, head);
 
 	    if (rl_node->handle == node->buf->handle) {
@@ -208,13 +216,15 @@ intel_setup_validate_list(int fd, struct intel_bo_list *list, struct intel_bo_li
     return first;
 }
 
-static void intel_free_validate_list(int fd, struct intel_bo_list *list)
+static void
+intel_free_validate_list(dri_bufmgr_ttm *bufmgr_ttm)
 {
+    struct intel_bo_list *list = &bufmgr_ttm->list;
     struct intel_bo_node *node;
     drmMMListHead *l;
 
     for (l = list->list.next; l != &list->list; l = l->next) {
-        node = DRMLISTENTRY(struct intel_bo_node, l, head);    
+        node = DRMLISTENTRY(struct intel_bo_node, l, head);
 
 	if (node->destroy)
 	    (*node->destroy)(node->priv);
@@ -222,30 +232,45 @@ static void intel_free_validate_list(int fd, struct intel_bo_list *list)
     }
 }
 
-static void intel_free_reloc_list(int fd, struct intel_bo_list *reloc_list)
+static void
+intel_free_reloc_list(dri_bufmgr_ttm *bufmgr_ttm)
 {
+    struct intel_bo_list *reloc_list = &bufmgr_ttm->reloc_list;
     struct intel_bo_reloc_node *reloc_node;
     drmMMListHead *rl, *tmp;
 
-    for (rl = reloc_list->list.next, tmp = rl->next; rl != &reloc_list->list; rl = tmp, tmp = rl->next) {
+    for (rl = reloc_list->list.next, tmp = rl->next; rl != &reloc_list->list;
+	 rl = tmp, tmp = rl->next)
+    {
 	reloc_node = DRMLISTENTRY(struct intel_bo_reloc_node, rl, head);
 
 	DRMLISTDEL(rl);
 
 	if (reloc_node->nr_reloc_types > 1) {
-	    
 	    /* TODO */
 	}
 
-	drmBOUnmap(fd, &reloc_node->type_list.buf);
-	drmBOUnreference(fd, &reloc_node->type_list.buf);
+	drmBOUnmap(bufmgr_ttm->fd, &reloc_node->type_list.buf);
+	drmBOUnreference(bufmgr_ttm->fd, &reloc_node->type_list.buf);
 	free(reloc_node);
     }
 }
 
-static int intel_add_validate_buffer(struct intel_bo_list *list, dri_bo *buf, unsigned flags,
-				     unsigned mask, int *itemLoc, void (*destroy_cb)(void *))
+/**
+ * Adds the given buffer to the list of buffers to be validated (moved into the
+ * appropriate memory type) with the next batch submission.
+ *
+ * If a buffer is validated multiple times in a batch submission, it ends up
+ * with the intersection of the memory type flags and the union of the
+ * remaining flags.
+ */
+static int
+intel_add_validate_buffer(dri_bufmgr_ttm *bufmgr_ttm,
+			  dri_bo *buf,
+			  uint64_t flags, uint64_t mask,
+			  int *itemLoc, void (*destroy_cb)(void *))
 {
+    struct intel_bo_list *list = &bufmgr_ttm->list;
     struct intel_bo_node *node, *cur;
     drmMMListHead *l;
     int count = 0;
@@ -253,6 +278,7 @@ static int intel_add_validate_buffer(struct intel_bo_list *list, dri_bo *buf, un
     drmBO *buf_bo = &((dri_bo_ttm *)buf)->drm_bo;
     cur = NULL;
 
+    /* Find the buffer in the validation list if it's already there. */
     for (l = list->list.next; l != &list->list; l = l->next) {
 	node = DRMLISTENTRY(struct intel_bo_node, l, head);
 	if (node->buf->handle == buf_bo->handle) {
@@ -261,7 +287,7 @@ static int intel_add_validate_buffer(struct intel_bo_list *list, dri_bo *buf, un
 	}
 	count++;
     }
-    
+
     if (!cur) {
 	cur = drmMalloc(sizeof(*cur));
 	if (!cur) {
@@ -269,55 +295,81 @@ static int intel_add_validate_buffer(struct intel_bo_list *list, dri_bo *buf, un
 	}
 	cur->buf = buf_bo;
 	cur->priv = buf;
-	cur->arg0 = flags;
-	cur->arg1 = mask;
+	cur->flags = flags;
+	cur->mask = mask;
 	cur->destroy = destroy_cb;
 	ret = 1;
 
 	DRMLISTADDTAIL(&cur->head, &list->list);
-
     } else {
-	unsigned memMask = (cur->arg1 | mask) & DRM_BO_MASK_MEM;
-	unsigned memFlags = cur->arg0 & flags & memMask;
-	
+	uint64_t memMask = (cur->mask | mask) & DRM_BO_MASK_MEM;
+	uint64_t memFlags = cur->flags & flags & memMask;
+
 	if (!memFlags) {
+	    fprintf(stderr,
+		    "%s: No shared memory types between "
+		    "0x%16llx and 0x%16llx\n",
+		    __FUNCTION__, cur->flags, flags);
 	    return -EINVAL;
 	}
-	if (mask & cur->arg1 & ~DRM_BO_MASK_MEM  & (cur->arg0 ^ flags)) {
+	if (mask & cur->mask & ~DRM_BO_MASK_MEM  & (cur->flags ^ flags)) {
+	    fprintf(stderr,
+		    "%s: Incompatible flags between 0x%16llx and 0x%16llx "
+		    "(0x%16llx, 0x%16llx masks)\n",
+		    __FUNCTION__, cur->flags, flags, cur->mask, mask);
 	    return -EINVAL;
 	}
-	cur->arg1 |= mask;
-	cur->arg0 = memFlags | ((cur->arg0 | flags) & 
-				cur->arg1 & ~DRM_BO_MASK_MEM);
+	cur->mask |= mask;
+	cur->flags = memFlags | ((cur->flags | flags) &
+				cur->mask & ~DRM_BO_MASK_MEM);
     }
     *itemLoc = count;
     return ret;
 }
 
 
-#define RELOC_BUF_SIZE(x) ((I915_RELOC_HEADER + x * I915_RELOC0_STRIDE) * sizeof(uint32_t))
+#define RELOC_BUF_SIZE(x) ((I915_RELOC_HEADER + x * I915_RELOC0_STRIDE) * \
+	sizeof(uint32_t))
 
-static int intel_create_new_reloc_type_list(int fd, struct intel_bo_reloc_list *cur_type, int max_relocs)
+static int
+intel_create_new_reloc_type_list(dri_bufmgr_ttm *bufmgr_ttm,
+				 struct intel_bo_reloc_list *cur_type)
 {
     int ret;
-    
+
     /* should allocate a drmBO here */
-    ret = drmBOCreate(fd, RELOC_BUF_SIZE(max_relocs), 0,
+    ret = drmBOCreate(bufmgr_ttm->fd, RELOC_BUF_SIZE(bufmgr_ttm->max_relocs), 0,
 		      NULL,
-		      DRM_BO_FLAG_MEM_LOCAL | DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE | DRM_BO_FLAG_MAPPABLE | DRM_BO_FLAG_CACHED,
+		      DRM_BO_FLAG_MEM_LOCAL |
+		      DRM_BO_FLAG_READ |
+		      DRM_BO_FLAG_WRITE |
+		      DRM_BO_FLAG_MAPPABLE |
+		      DRM_BO_FLAG_CACHED,
 		      0, &cur_type->buf);
-    if (ret)
+    if (ret) {
+	fprintf(stderr, "Failed to create relocation BO: %s\n",
+		strerror(-ret));
 	return ret;
-    
-    ret = drmBOMap(fd, &cur_type->buf, DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE, 0, (void **)&cur_type->relocs);
-    if (ret)
+    }
+
+    ret = drmBOMap(bufmgr_ttm->fd, &cur_type->buf,
+		   DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE,
+		   0, (void **)&cur_type->relocs);
+    if (ret) {
+	fprintf(stderr, "Failed to map relocation BO: %s\n", strerror(-ret));
 	return ret;
+    }
     return 0;
 }
 
-
-static int intel_add_validate_reloc(int fd, struct intel_bo_list *reloc_list, struct intel_reloc_info *reloc_info, uint32_t max_relocs)
+/**
+ * Adds the relocation @reloc_info to the relocation list.
+ */
+static int
+intel_add_validate_reloc(dri_bufmgr_ttm *bufmgr_ttm,
+			 struct intel_reloc_info *reloc_info)
 {
+    struct intel_bo_list *reloc_list = &bufmgr_ttm->reloc_list;
     struct intel_bo_reloc_node *rl_node, *cur;
     drmMMListHead *rl, *l;
     int ret = 0;
@@ -346,7 +398,7 @@ static int intel_add_validate_reloc(int fd, struct intel_bo_list *reloc_list, st
 	cur_type = &cur->type_list;
 
 	DRMINITLISTHEAD(&cur->type_list.head);
-	ret = intel_create_new_reloc_type_list(fd, cur_type, max_relocs);
+	ret = intel_create_new_reloc_type_list(bufmgr_ttm, cur_type);
 	if (ret) {
 	    return -1;
 	}
@@ -361,22 +413,24 @@ static int intel_add_validate_reloc(int fd, struct intel_bo_list *reloc_list, st
 		cur_type = &cur->type_list;
 		found = 1;
 	} else {
-	    for (l = cur->type_list.head.next; l != &cur->type_list.head; l = l->next) {
+	    for (l = cur->type_list.head.next; l != &cur->type_list.head;
+		 l = l->next)
+	    {
 	        cur_type = DRMLISTENTRY(struct intel_bo_reloc_list, l, head);
 	        if (((cur_type->relocs[0] >> 16) & 0xffff) == reloc_info->type)
 	    	    found = 1;
 		break;
 	    }
         }
-	
+
 	/* didn't find the relocation type */
 	if (!found) {
 	    cur_type = malloc(sizeof(*cur_type));
 	    if (!cur_type) {
 		return -ENOMEM;
 	    }
-	    
-	    ret = intel_create_new_reloc_type_list(fd, cur_type, max_relocs);
+
+	    ret = intel_create_new_reloc_type_list(bufmgr_ttm, cur_type);
 	    DRMLISTADDTAIL(&cur_type->head, &cur->type_list.head);
 
 	    cur_type->relocs[0] = (reloc_info->type << 16);
@@ -390,11 +444,14 @@ static int intel_add_validate_reloc(int fd, struct intel_bo_list *reloc_list, st
 
     num_relocs = (reloc_start[0] & 0xffff);
 
-    reloc_start[num_relocs*I915_RELOC0_STRIDE + I915_RELOC_HEADER] = reloc_info->reloc;
-    reloc_start[num_relocs*I915_RELOC0_STRIDE + I915_RELOC_HEADER+1] = reloc_info->delta;
-    reloc_start[num_relocs*I915_RELOC0_STRIDE + I915_RELOC_HEADER+2] = reloc_info->index;
+    reloc_start[num_relocs * I915_RELOC0_STRIDE + I915_RELOC_HEADER] =
+       reloc_info->reloc;
+    reloc_start[num_relocs * I915_RELOC0_STRIDE + I915_RELOC_HEADER + 1] =
+       reloc_info->delta;
+    reloc_start[num_relocs * I915_RELOC0_STRIDE + I915_RELOC_HEADER + 2] =
+       reloc_info->index;
     reloc_start[0]++;
-    if (((reloc_start[0] & 0xffff)) > (max_relocs)) {
+    if (((reloc_start[0] & 0xffff)) > (bufmgr_ttm->max_relocs)) {
 	return -ENOMEM;
     }
     return 0;
@@ -405,64 +462,64 @@ static int intel_add_validate_reloc(int fd, struct intel_bo_list *reloc_list, st
 int
 driFenceSignaled(DriFenceObject * fence, unsigned type)
 {
-   int signaled;
-   int ret;
+    int signaled;
+    int ret;
 
-   if (fence == NULL)
-      return GL_TRUE;
+    if (fence == NULL)
+	return GL_TRUE;
 
-   _glthread_LOCK_MUTEX(fence->mutex);
-   ret = drmFenceSignaled(bufmgr_ttm->fd, &fence->fence, type, &signaled);
-   _glthread_UNLOCK_MUTEX(fence->mutex);
-   BM_CKFATAL(ret);
-   return signaled;
+    _glthread_LOCK_MUTEX(fence->mutex);
+    ret = drmFenceSignaled(bufmgr_ttm->fd, &fence->fence, type, &signaled);
+    _glthread_UNLOCK_MUTEX(fence->mutex);
+    BM_CKFATAL(ret);
+    return signaled;
 }
 #endif
 
 static dri_bo *
 dri_ttm_alloc(dri_bufmgr *bufmgr, const char *name,
 	      unsigned long size, unsigned int alignment,
-	      unsigned int location_mask)
-{
-   dri_bufmgr_ttm *ttm_bufmgr;
-   dri_bo_ttm *ttm_buf;
-   unsigned int pageSize = getpagesize();
-   int ret;
-   unsigned int flags, hint;
-
-   ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
-
-   ttm_buf = malloc(sizeof(*ttm_buf));
-   if (!ttm_buf)
-      return NULL;
-
-   /* The mask argument doesn't do anything for us that we want other than
-    * determine which pool (TTM or local) the buffer is allocated into, so just
-    * pass all of the allocation class flags.
-    */
-   flags = location_mask | DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE |
-      DRM_BO_FLAG_EXE;
-   /* No hints we want to use. */
-   hint = 0;
-
-   ret = drmBOCreate(ttm_bufmgr->fd, size, alignment / pageSize,
-		     NULL, flags, hint, &ttm_buf->drm_bo);
-   if (ret != 0) {
-      free(ttm_buf);
-      return NULL;
-   }
-   ttm_buf->bo.size = ttm_buf->drm_bo.size;
-   ttm_buf->bo.offset = ttm_buf->drm_bo.offset;
-   ttm_buf->bo.virtual = NULL;
-   ttm_buf->bo.bufmgr = bufmgr;
-   ttm_buf->name = name;
-   ttm_buf->refcount = 1;
+	      uint64_t location_mask)
+{
+    dri_bufmgr_ttm *ttm_bufmgr;
+    dri_bo_ttm *ttm_buf;
+    unsigned int pageSize = getpagesize();
+    int ret;
+    unsigned int flags, hint;
+
+    ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
+
+    ttm_buf = malloc(sizeof(*ttm_buf));
+    if (!ttm_buf)
+	return NULL;
+
+    /* The mask argument doesn't do anything for us that we want other than
+     * determine which pool (TTM or local) the buffer is allocated into, so
+     * just pass all of the allocation class flags.
+     */
+    flags = location_mask | DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE |
+	DRM_BO_FLAG_EXE;
+    /* No hints we want to use. */
+    hint = 0;
+
+    ret = drmBOCreate(ttm_bufmgr->fd, size, alignment / pageSize,
+		      NULL, flags, hint, &ttm_buf->drm_bo);
+    if (ret != 0) {
+	free(ttm_buf);
+	return NULL;
+    }
+    ttm_buf->bo.size = ttm_buf->drm_bo.size;
+    ttm_buf->bo.offset = ttm_buf->drm_bo.offset;
+    ttm_buf->bo.virtual = NULL;
+    ttm_buf->bo.bufmgr = bufmgr;
+    ttm_buf->name = name;
+    ttm_buf->refcount = 1;
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "bo_create: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
+    fprintf(stderr, "bo_create: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
 #endif
 
-   return &ttm_buf->bo;
+    return &ttm_buf->bo;
 }
 
 /* Our TTM backend doesn't allow creation of static buffers, as that requires
@@ -472,12 +529,13 @@ dri_ttm_alloc(dri_bufmgr *bufmgr, const char *name,
 static dri_bo *
 dri_ttm_alloc_static(dri_bufmgr *bufmgr, const char *name,
 		     unsigned long offset, unsigned long size, void *virtual,
-		     unsigned int location_mask)
+		     uint64_t location_mask)
 {
-   return NULL;
+    return NULL;
 }
 
-/** Returns a dri_bo wrapping the given buffer object handle.
+/**
+ * Returns a dri_bo wrapping the given buffer object handle.
  *
  * This can be used when one application needs to pass a buffer object
  * to another.
@@ -486,121 +544,122 @@ dri_bo *
 intel_ttm_bo_create_from_handle(dri_bufmgr *bufmgr, const char *name,
 			      unsigned int handle)
 {
-   dri_bufmgr_ttm *ttm_bufmgr;
-   dri_bo_ttm *ttm_buf;
-   int ret;
+    dri_bufmgr_ttm *ttm_bufmgr;
+    dri_bo_ttm *ttm_buf;
+    int ret;
 
-   ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
+    ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
 
-   ttm_buf = malloc(sizeof(*ttm_buf));
-   if (!ttm_buf)
-      return NULL;
+    ttm_buf = malloc(sizeof(*ttm_buf));
+    if (!ttm_buf)
+	return NULL;
 
-   ret = drmBOReference(ttm_bufmgr->fd, handle, &ttm_buf->drm_bo);
-   if (ret != 0) {
-      free(ttm_buf);
-      return NULL;
-   }
-   ttm_buf->bo.size = ttm_buf->drm_bo.size;
-   ttm_buf->bo.offset = ttm_buf->drm_bo.offset;
-   ttm_buf->bo.virtual = NULL;
-   ttm_buf->bo.bufmgr = bufmgr;
-   ttm_buf->name = name;
-   ttm_buf->refcount = 1;
+    ret = drmBOReference(ttm_bufmgr->fd, handle, &ttm_buf->drm_bo);
+    if (ret != 0) {
+	free(ttm_buf);
+	return NULL;
+    }
+    ttm_buf->bo.size = ttm_buf->drm_bo.size;
+    ttm_buf->bo.offset = ttm_buf->drm_bo.offset;
+    ttm_buf->bo.virtual = NULL;
+    ttm_buf->bo.bufmgr = bufmgr;
+    ttm_buf->name = name;
+    ttm_buf->refcount = 1;
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "bo_create_from_handle: %p %08x (%s)\n", &ttm_buf->bo, handle,
-	   ttm_buf->name);
+    fprintf(stderr, "bo_create_from_handle: %p %08x (%s)\n",
+	    &ttm_buf->bo, handle, ttm_buf->name);
 #endif
 
-   return &ttm_buf->bo;
+    return &ttm_buf->bo;
 }
 
 static void
 dri_ttm_bo_reference(dri_bo *buf)
 {
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
-   dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
+    dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
 
-   _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
-   ttm_buf->refcount++;
-   _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+    _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
+    ttm_buf->refcount++;
+    _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
 }
 
 static void
 dri_ttm_bo_unreference(dri_bo *buf)
 {
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
-   dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
+    dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
 
-   if (!buf)
-      return;
+    if (!buf)
+	return;
 
-   _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
-   if (--ttm_buf->refcount == 0) {
-      int ret;
+    _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
+    if (--ttm_buf->refcount == 0) {
+	int ret;
 
-      ret = drmBOUnreference(bufmgr_ttm->fd, &ttm_buf->drm_bo);
-      if (ret != 0) {
-	 fprintf(stderr, "drmBOUnreference failed (%s): %s\n", ttm_buf->name,
-		 strerror(-ret));
-      }
+	ret = drmBOUnreference(bufmgr_ttm->fd, &ttm_buf->drm_bo);
+	if (ret != 0) {
+	    fprintf(stderr, "drmBOUnreference failed (%s): %s\n",
+		    ttm_buf->name, strerror(-ret));
+	}
 #if BUFMGR_DEBUG
-      fprintf(stderr, "bo_unreference final: %p (%s)\n",
-	      &ttm_buf->bo, ttm_buf->name);
+	fprintf(stderr, "bo_unreference final: %p (%s)\n",
+		&ttm_buf->bo, ttm_buf->name);
 #endif
-      _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
-      free(buf);
-      return;
-   }
-   _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+	_glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+	free(buf);
+	return;
+    }
+    _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
 }
 
 static int
 dri_ttm_bo_map(dri_bo *buf, GLboolean write_enable)
 {
-   dri_bufmgr_ttm *bufmgr_ttm;
-   dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
-   unsigned int flags;
+    dri_bufmgr_ttm *bufmgr_ttm;
+    dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
+    unsigned int flags;
 
-   bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
+    bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
 
-   flags = DRM_BO_FLAG_READ;
-   if (write_enable)
-       flags |= DRM_BO_FLAG_WRITE;
+    flags = DRM_BO_FLAG_READ;
+    if (write_enable)
+	flags |= DRM_BO_FLAG_WRITE;
 
-   assert(buf->virtual == NULL);
+    assert(buf->virtual == NULL);
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "bo_map: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
+    fprintf(stderr, "bo_map: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
 #endif
 
-   return drmBOMap(bufmgr_ttm->fd, &ttm_buf->drm_bo, flags, 0, &buf->virtual);
+    return drmBOMap(bufmgr_ttm->fd, &ttm_buf->drm_bo, flags, 0, &buf->virtual);
 }
 
 static int
 dri_ttm_bo_unmap(dri_bo *buf)
 {
-   dri_bufmgr_ttm *bufmgr_ttm;
-   dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
+    dri_bufmgr_ttm *bufmgr_ttm;
+    dri_bo_ttm *ttm_buf = (dri_bo_ttm *)buf;
 
-   if (buf == NULL)
-      return 0;
+    if (buf == NULL)
+	return 0;
 
-   bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
+    bufmgr_ttm = (dri_bufmgr_ttm *)buf->bufmgr;
 
-   assert(buf->virtual != NULL);
+    assert(buf->virtual != NULL);
 
-   buf->virtual = NULL;
+    buf->virtual = NULL;
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "bo_unmap: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
+    fprintf(stderr, "bo_unmap: %p (%s)\n", &ttm_buf->bo, ttm_buf->name);
 #endif
 
-   return drmBOUnmap(bufmgr_ttm->fd, &ttm_buf->drm_bo);
+    return drmBOUnmap(bufmgr_ttm->fd, &ttm_buf->drm_bo);
 }
 
-/* Returns a dri_bo wrapping the given buffer object handle.
+/**
+ * Returns a dri_bo wrapping the given buffer object handle.
  *
  * This can be used when one application needs to pass a buffer object
  * to another.
@@ -609,183 +668,185 @@ dri_fence *
 intel_ttm_fence_create_from_arg(dri_bufmgr *bufmgr, const char *name,
 				drm_fence_arg_t *arg)
 {
-   dri_bufmgr_ttm *ttm_bufmgr;
-   dri_fence_ttm *ttm_fence;
+    dri_bufmgr_ttm *ttm_bufmgr;
+    dri_fence_ttm *ttm_fence;
 
-   ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
+    ttm_bufmgr = (dri_bufmgr_ttm *)bufmgr;
 
-   ttm_fence = malloc(sizeof(*ttm_fence));
-   if (!ttm_fence)
-      return NULL;
+    ttm_fence = malloc(sizeof(*ttm_fence));
+    if (!ttm_fence)
+	return NULL;
 
-   ttm_fence->drm_fence.handle = arg->handle;
-   ttm_fence->drm_fence.fence_class = arg->fence_class;
-   ttm_fence->drm_fence.type = arg->type;
-   ttm_fence->drm_fence.flags = arg->flags;
-   ttm_fence->drm_fence.signaled = 0;
-   ttm_fence->drm_fence.sequence = arg->sequence;
+    ttm_fence->drm_fence.handle = arg->handle;
+    ttm_fence->drm_fence.fence_class = arg->fence_class;
+    ttm_fence->drm_fence.type = arg->type;
+    ttm_fence->drm_fence.flags = arg->flags;
+    ttm_fence->drm_fence.signaled = 0;
+    ttm_fence->drm_fence.sequence = arg->sequence;
 
-   ttm_fence->fence.bufmgr = bufmgr;
-   ttm_fence->name = name;
-   ttm_fence->refcount = 1;
+    ttm_fence->fence.bufmgr = bufmgr;
+    ttm_fence->name = name;
+    ttm_fence->refcount = 1;
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "fence_create_from_handle: %p (%s)\n", &ttm_fence->fence,
-	   ttm_fence->name);
+    fprintf(stderr, "fence_create_from_handle: %p (%s)\n", &ttm_fence->fence,
+	    ttm_fence->name);
 #endif
 
-   return &ttm_fence->fence;
+    return &ttm_fence->fence;
 }
 
 
 static void
 dri_ttm_fence_reference(dri_fence *fence)
 {
-   dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
+    dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
 
-   _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
-   ++fence_ttm->refcount;
-   _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+    _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
+    ++fence_ttm->refcount;
+    _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
 #if BUFMGR_DEBUG
-   fprintf(stderr, "fence_reference: %p (%s)\n", &fence_ttm->fence,
-	   fence_ttm->name);
+    fprintf(stderr, "fence_reference: %p (%s)\n", &fence_ttm->fence,
+	    fence_ttm->name);
 #endif
 }
 
 static void
 dri_ttm_fence_unreference(dri_fence *fence)
 {
-   dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
+    dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
 
-   if (!fence)
-      return;
+    if (!fence)
+	return;
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "fence_unreference: %p (%s)\n", &fence_ttm->fence,
-	   fence_ttm->name);
+    fprintf(stderr, "fence_unreference: %p (%s)\n", &fence_ttm->fence,
+	    fence_ttm->name);
 #endif
-   _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
-   if (--fence_ttm->refcount == 0) {
-      int ret;
-
-      ret = drmFenceUnreference(bufmgr_ttm->fd, &fence_ttm->drm_fence);
-      if (ret != 0) {
-	 fprintf(stderr, "drmFenceUnreference failed (%s): %s\n",
-		 fence_ttm->name, strerror(-ret));
-      }
-
-      _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
-      free(fence);
-      return;
-   }
-   _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+    _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
+    if (--fence_ttm->refcount == 0) {
+	int ret;
+
+	ret = drmFenceUnreference(bufmgr_ttm->fd, &fence_ttm->drm_fence);
+	if (ret != 0) {
+	    fprintf(stderr, "drmFenceUnreference failed (%s): %s\n",
+		    fence_ttm->name, strerror(-ret));
+	}
+
+	_glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+	free(fence);
+	return;
+    }
+    _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
 }
 
 static void
 dri_ttm_fence_wait(dri_fence *fence)
 {
-   dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
-   int ret;
+    dri_fence_ttm *fence_ttm = (dri_fence_ttm *)fence;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)fence->bufmgr;
+    int ret;
 
-   _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
-   ret = drmFenceWait(bufmgr_ttm->fd, 0, &fence_ttm->drm_fence, 0);
-   _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
-   if (ret != 0) {
-      _mesa_printf("%s:%d: Error %d waiting for fence %s.\n",
-		   __FILE__, __LINE__, ret, fence_ttm->name);
-      abort();
-   }
+    _glthread_LOCK_MUTEX(bufmgr_ttm->mutex);
+    ret = drmFenceWait(bufmgr_ttm->fd, 0, &fence_ttm->drm_fence, 0);
+    _glthread_UNLOCK_MUTEX(bufmgr_ttm->mutex);
+    if (ret != 0) {
+	_mesa_printf("%s:%d: Error %d waiting for fence %s.\n",
+		     __FILE__, __LINE__, ret, fence_ttm->name);
+	abort();
+    }
 
 #if BUFMGR_DEBUG
-   fprintf(stderr, "fence_wait: %p (%s)\n", &fence_ttm->fence,
-	   fence_ttm->name);
+    fprintf(stderr, "fence_wait: %p (%s)\n", &fence_ttm->fence,
+	    fence_ttm->name);
 #endif
 }
 
 static void
 dri_bufmgr_ttm_destroy(dri_bufmgr *bufmgr)
 {
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)bufmgr;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)bufmgr;
 
-   intel_bo_free_list(&bufmgr_ttm->list);
-   intel_bo_free_list(&bufmgr_ttm->reloc_list);
+    intel_bo_free_list(&bufmgr_ttm->list);
+    intel_bo_free_list(&bufmgr_ttm->reloc_list);
 
-   _glthread_DESTROY_MUTEX(bufmgr_ttm->mutex);
-   free(bufmgr);
+    _glthread_DESTROY_MUTEX(bufmgr_ttm->mutex);
+    free(bufmgr);
 }
 
 
-static void intel_dribo_destroy_callback(void *priv)
+static void
+intel_dribo_destroy_callback(void *priv)
 {
-   dri_bo *dribo = priv;
-   
-   if (dribo) {
-     dri_bo_unreference(dribo);
-   }
+    dri_bo *dribo = priv;
+
+    if (dribo)
+	dri_bo_unreference(dribo);
 }
 
 static void
-dri_ttm_emit_reloc(dri_bo *batch_buf, GLuint flags, GLuint delta, GLuint offset,
-		    dri_bo *relocatee)
-{
-   dri_bo_ttm *ttm_buf = (dri_bo_ttm *)batch_buf;
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)batch_buf->bufmgr;
-   int newItem;
-   struct intel_reloc_info reloc;
-   int mask;
-   int ret;
-
-   mask = DRM_BO_MASK_MEM;
-   mask |= flags & (DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE | DRM_BO_FLAG_EXE);
-
-   ret = intel_add_validate_buffer(&bufmgr_ttm->list, relocatee, flags, mask, &newItem, intel_dribo_destroy_callback);
-   if (ret < 0)
-     return;
-
-   if (ret == 1) {
-      dri_bo_reference(relocatee);
-   }
-
-   reloc.type = I915_RELOC_TYPE_0;
-   reloc.reloc = offset;
-   reloc.delta = delta;
-   reloc.index = newItem;
-   reloc.handle = ttm_buf->drm_bo.handle;
-
-   intel_add_validate_reloc(bufmgr_ttm->fd, &bufmgr_ttm->reloc_list, &reloc, bufmgr_ttm->max_relocs);
-   return;
+dri_ttm_emit_reloc(dri_bo *reloc_buf, uint64_t flags, GLuint delta,
+		   GLuint offset, dri_bo *target_buf)
+{
+    dri_bo_ttm *ttm_buf = (dri_bo_ttm *)reloc_buf;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)reloc_buf->bufmgr;
+    int newItem;
+    struct intel_reloc_info reloc;
+    int mask;
+    int ret;
+
+    mask = DRM_BO_MASK_MEM;
+    mask |= flags & (DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE | DRM_BO_FLAG_EXE);
+
+    ret = intel_add_validate_buffer(bufmgr_ttm, target_buf, flags, mask,
+				    &newItem, intel_dribo_destroy_callback);
+    if (ret < 0)
+	return;
+
+    if (ret == 1)
+	dri_bo_reference(target_buf);
+
+    reloc.type = I915_RELOC_TYPE_0;
+    reloc.reloc = offset;
+    reloc.delta = delta;
+    reloc.index = newItem;
+    reloc.handle = ttm_buf->drm_bo.handle;
+
+    intel_add_validate_reloc(bufmgr_ttm, &reloc);
 }
 
 
 static void *
 dri_ttm_process_reloc(dri_bo *batch_buf, GLuint *count)
 {
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)batch_buf->bufmgr;
-   void *ptr;
-   int itemLoc;
-
-   dri_bo_unmap(batch_buf);
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)batch_buf->bufmgr;
+    void *ptr;
+    int itemLoc;
 
-   intel_add_validate_buffer(&bufmgr_ttm->list, batch_buf, DRM_BO_FLAG_MEM_TT | DRM_BO_FLAG_EXE,
-			     DRM_BO_MASK_MEM | DRM_BO_FLAG_EXE, &itemLoc, NULL);
+    /* Add the batch buffer to the validation list.  There are no relocations
+     * pointing to it.
+     */
+    intel_add_validate_buffer(bufmgr_ttm, batch_buf,
+			      DRM_BO_FLAG_MEM_TT | DRM_BO_FLAG_EXE,
+			      DRM_BO_MASK_MEM | DRM_BO_FLAG_EXE,
+			      &itemLoc, NULL);
 
-   ptr = intel_setup_validate_list(bufmgr_ttm->fd, &bufmgr_ttm->list, &bufmgr_ttm->reloc_list, count);
+    ptr = intel_setup_validate_list(bufmgr_ttm, count);
 
-   return ptr;
+    return ptr;
 }
 
 static void
 dri_ttm_post_submit(dri_bo *batch_buf, dri_fence **last_fence)
 {
-   dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)batch_buf->bufmgr;
+    dri_bufmgr_ttm *bufmgr_ttm = (dri_bufmgr_ttm *)batch_buf->bufmgr;
 
-   intel_free_validate_list(bufmgr_ttm->fd, &bufmgr_ttm->list);
-   intel_free_reloc_list(bufmgr_ttm->fd, &bufmgr_ttm->reloc_list);
+    intel_free_validate_list(bufmgr_ttm);
+    intel_free_reloc_list(bufmgr_ttm);
 
-   intel_bo_free_list(&bufmgr_ttm->list);
+    intel_bo_free_list(&bufmgr_ttm->list);
 }
 
 /**
@@ -801,33 +862,34 @@ dri_bufmgr *
 intel_bufmgr_ttm_init(int fd, unsigned int fence_type,
 		      unsigned int fence_type_flush, int batch_size)
 {
-   dri_bufmgr_ttm *bufmgr_ttm;
-
-   bufmgr_ttm = malloc(sizeof(*bufmgr_ttm));
-   bufmgr_ttm->fd = fd;
-   bufmgr_ttm->fence_type = fence_type;
-   bufmgr_ttm->fence_type_flush = fence_type_flush;
-   _glthread_INIT_MUTEX(bufmgr_ttm->mutex);
-
-   /* lets go with one relocation per every four dwords - purely heuristic */
-   bufmgr_ttm->max_relocs = batch_size / sizeof(uint32_t) / 4;
-
-   intel_create_bo_list(10, &bufmgr_ttm->list, NULL);
-   intel_create_bo_list(1, &bufmgr_ttm->reloc_list, NULL);
-
-   bufmgr_ttm->bufmgr.bo_alloc = dri_ttm_alloc;
-   bufmgr_ttm->bufmgr.bo_alloc_static = dri_ttm_alloc_static;
-   bufmgr_ttm->bufmgr.bo_reference = dri_ttm_bo_reference;
-   bufmgr_ttm->bufmgr.bo_unreference = dri_ttm_bo_unreference;
-   bufmgr_ttm->bufmgr.bo_map = dri_ttm_bo_map;
-   bufmgr_ttm->bufmgr.bo_unmap = dri_ttm_bo_unmap;
-   bufmgr_ttm->bufmgr.fence_reference = dri_ttm_fence_reference;
-   bufmgr_ttm->bufmgr.fence_unreference = dri_ttm_fence_unreference;
-   bufmgr_ttm->bufmgr.fence_wait = dri_ttm_fence_wait;
-   bufmgr_ttm->bufmgr.destroy = dri_bufmgr_ttm_destroy;
-   bufmgr_ttm->bufmgr.emit_reloc = dri_ttm_emit_reloc;
-   bufmgr_ttm->bufmgr.process_relocs = dri_ttm_process_reloc;
-   bufmgr_ttm->bufmgr.post_submit = dri_ttm_post_submit;
-   return &bufmgr_ttm->bufmgr;
+    dri_bufmgr_ttm *bufmgr_ttm;
+
+    bufmgr_ttm = malloc(sizeof(*bufmgr_ttm));
+    bufmgr_ttm->fd = fd;
+    bufmgr_ttm->fence_type = fence_type;
+    bufmgr_ttm->fence_type_flush = fence_type_flush;
+    _glthread_INIT_MUTEX(bufmgr_ttm->mutex);
+
+    /* lets go with one relocation per every four dwords - purely heuristic */
+    bufmgr_ttm->max_relocs = batch_size / sizeof(uint32_t) / 4;
+
+    intel_create_bo_list(10, &bufmgr_ttm->list, NULL);
+    intel_create_bo_list(1, &bufmgr_ttm->reloc_list, NULL);
+
+    bufmgr_ttm->bufmgr.bo_alloc = dri_ttm_alloc;
+    bufmgr_ttm->bufmgr.bo_alloc_static = dri_ttm_alloc_static;
+    bufmgr_ttm->bufmgr.bo_reference = dri_ttm_bo_reference;
+    bufmgr_ttm->bufmgr.bo_unreference = dri_ttm_bo_unreference;
+    bufmgr_ttm->bufmgr.bo_map = dri_ttm_bo_map;
+    bufmgr_ttm->bufmgr.bo_unmap = dri_ttm_bo_unmap;
+    bufmgr_ttm->bufmgr.fence_reference = dri_ttm_fence_reference;
+    bufmgr_ttm->bufmgr.fence_unreference = dri_ttm_fence_unreference;
+    bufmgr_ttm->bufmgr.fence_wait = dri_ttm_fence_wait;
+    bufmgr_ttm->bufmgr.destroy = dri_bufmgr_ttm_destroy;
+    bufmgr_ttm->bufmgr.emit_reloc = dri_ttm_emit_reloc;
+    bufmgr_ttm->bufmgr.process_relocs = dri_ttm_process_reloc;
+    bufmgr_ttm->bufmgr.post_submit = dri_ttm_post_submit;
+
+    return &bufmgr_ttm->bufmgr;
 }
 
diff --git a/src/mesa/drivers/dri/r200/r200_texstate.c b/src/mesa/drivers/dri/r200/r200_texstate.c
index 755035b07a..94701ef1ca 100644
--- a/src/mesa/drivers/dri/r200/r200_texstate.c
+++ b/src/mesa/drivers/dri/r200/r200_texstate.c
@@ -71,7 +71,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define _INVALID(f) \
     [ MESA_FORMAT_ ## f ] = { 0xffffffff, 0 }
 #define VALID_FORMAT(f) ( ((f) <= MESA_FORMAT_RGBA_DXT5) \
-			     && (tx_table_le[f].format != 0xffffffff) )
+			     && (tx_table_be[f].format != 0xffffffff) )
 
 struct tx_table {
    GLuint format, filter;
@@ -164,12 +164,15 @@ static void r200SetTexImages( r200ContextPtr rmesa,
     */
    if ( !t->image_override ) {
       if ( VALID_FORMAT( baseImage->TexFormat->MesaFormat ) ) {
+	 const struct tx_table *table = _mesa_little_endian() ? tx_table_le :
+								tx_table_be;
+
          t->pp_txformat &= ~(R200_TXFORMAT_FORMAT_MASK |
                              R200_TXFORMAT_ALPHA_IN_MAP);
          t->pp_txfilter &= ~R200_YUV_TO_RGB;
 
-	 t->pp_txformat |= tx_table_le[ baseImage->TexFormat->MesaFormat ].format;
-	 t->pp_txfilter |= tx_table_le[ baseImage->TexFormat->MesaFormat ].filter;
+	 t->pp_txformat |= table[ baseImage->TexFormat->MesaFormat ].format;
+	 t->pp_txfilter |= table[ baseImage->TexFormat->MesaFormat ].filter;
       }
       else {
          _mesa_problem(NULL, "unexpected texture format in %s", __FUNCTION__);
diff --git a/src/mesa/drivers/osmesa/descrip.mms b/src/mesa/drivers/osmesa/descrip.mms
index 4035b24e4e..6c2f3226f6 100644
--- a/src/mesa/drivers/osmesa/descrip.mms
+++ b/src/mesa/drivers/osmesa/descrip.mms
@@ -1,15 +1,19 @@
 # Makefile for core library for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
-# Last revision : 16 June 2003
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Last revision : 3 October 2007
 
 .first
 	define gl [----.include.gl]
 	define math [--.math]
 	define tnl [--.tnl]
+	define vbo [--.vbo]
 	define swrast [--.swrast]
 	define swrast_setup [--.swrast_setup]
 	define array_cache [--.array_cache]
 	define drivers [-]
+	define glapi [--.glapi]
+	define main [--.main]
+	define shader [--.shader]
 
 .include [----]mms-config.
 
@@ -17,7 +21,7 @@
 
 VPATH = RCS
 
-INCDIR = [----.include],[--.main],[--.glapi]
+INCDIR = [----.include],[--.main],[--.glapi],[--.shader]
 LIBDIR = [----.lib]
 CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ieee/ieee=denorm
 
diff --git a/src/mesa/drivers/windows/gdi/wgl.c b/src/mesa/drivers/windows/gdi/wgl.c
index 17eed1dc32..f7028d01a3 100644
--- a/src/mesa/drivers/windows/gdi/wgl.c
+++ b/src/mesa/drivers/windows/gdi/wgl.c
@@ -55,7 +55,7 @@
 #include <windows.h>
 
 #endif
-
+#include "config.h"
 #include "glapi.h"
 #include "GL/wmesa.h"   /* protos for wmesa* functions */
 
@@ -70,10 +70,12 @@ struct __pixelformat__
     GLboolean doubleBuffered;
 };
 
+
+
 /* These are the PFD's supported by this driver. */
 struct __pixelformat__	pfd[] =
 {
-#if 0
+#if 0 
     /* Double Buffer, alpha */
     {	
 	{	
@@ -87,7 +89,7 @@ struct __pixelformat__	pfd[] =
 	    8, 16,	
 	    8, 24,
 	    0, 0, 0, 0, 0,	
-	    16,	8,	
+	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
 	    0, 0, 0,	
 	    0, 0, 0 
 	},
@@ -106,13 +108,13 @@ struct __pixelformat__	pfd[] =
 	    8, 16,	
 	    8, 24,
 	    0, 0, 0, 0,	0,	
-	    16,	8,
+	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
 	    0, 0, 0,	
 	    0, 0, 0
 	},
         GL_FALSE
     },
-#endif
+#endif 
     /* Double Buffer, no alpha */
     {	
 	{	
@@ -126,7 +128,7 @@ struct __pixelformat__	pfd[] =
 	    8, 16,
 	    0, 0,
 	    0, 0, 0, 0,	0,
-	    16,	8,
+	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
 	    0, 0, 0, 
 	    0, 0, 0 
 	},
@@ -145,7 +147,7 @@ struct __pixelformat__	pfd[] =
 	    8, 16,
 	    0, 0,
 	    0, 0, 0, 0,	0,
-	    16,	8,
+	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
 	    0, 0, 0,
 	    0, 0, 0 
 	},
@@ -601,8 +603,8 @@ WINGDIAPI HGLRC GLAPIENTRY wglCreateLayerContext(HDC hdc,
 WINGDIAPI BOOL GLAPIENTRY wglShareLists(HGLRC hglrc1,
 					HGLRC hglrc2)
 {
-	WMesaShareLists(hglrc1, hglrc2);
-	return(TRUE);
+    WMesaShareLists(hglrc1, hglrc2);
+    return(TRUE);
 }
 
 
diff --git a/src/mesa/drivers/x11/descrip.mms b/src/mesa/drivers/x11/descrip.mms
index 6c6184b2c3..648f683dfa 100644
--- a/src/mesa/drivers/x11/descrip.mms
+++ b/src/mesa/drivers/x11/descrip.mms
@@ -1,15 +1,19 @@
 # Makefile for core library for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
-# Last revision : 16 June 2003
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Last revision : 3 October 2007
 
 .first
 	define gl [----.include.gl]
 	define math [--.math]
 	define tnl [--.tnl]
+	define vbo [--.vbo]
 	define swrast [--.swrast]
 	define swrast_setup [--.swrast_setup]
 	define array_cache [--.array_cache]
 	define drivers [-]
+	define glapi [--.glapi]
+	define main [--.main]
+	define shader [--.shader]
 
 .include [----]mms-config.
 
@@ -17,7 +21,7 @@
 
 VPATH = RCS
 
-INCDIR = [----.include],[--.main],[--.glapi]
+INCDIR = [----.include],[--.main],[--.glapi],[--.shader]
 LIBDIR = [----.lib]
 CFLAGS =/include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ieee/ieee=denorm
 
diff --git a/src/mesa/main/config.h b/src/mesa/main/config.h
index 8eb02beb4c..ab0f035b45 100644
--- a/src/mesa/main/config.h
+++ b/src/mesa/main/config.h
@@ -213,7 +213,7 @@
 /** For GL_ARB_vertex_shader */
 /*@{*/
 #define MAX_VERTEX_ATTRIBS 16
-#define MAX_VERTEX_TEXTURE_IMAGE_UNITS 0
+#define MAX_VERTEX_TEXTURE_IMAGE_UNITS MAX_TEXTURE_UNITS
 #define MAX_COMBINED_TEXTURE_IMAGE_UNITS (MAX_TEXTURE_IMAGE_UNITS + MAX_VERTEX_TEXTURE_IMAGE_UNITS)
 /*@}*/
 
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 08db12b1b6..2ae0e1221d 100644
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -992,7 +992,6 @@ init_attrib_groups(GLcontext *ctx)
    /* Miscellaneous */
    ctx->NewState = _NEW_ALL;
    ctx->ErrorValue = (GLenum) GL_NO_ERROR;
-   ctx->_Facing = 0;
 
    return GL_TRUE;
 }
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index caa50dd682..77d27bb03d 100644
--- a/src/mesa/main/dd.h
+++ b/src/mesa/main/dd.h
@@ -493,6 +493,11 @@ struct dd_function_table {
     */
    void (*FreeTexImageData)( GLcontext *ctx, struct gl_texture_image *tImage );
 
+   /** Map texture image data into user space */
+   void (*MapTexture)( GLcontext *ctx, struct gl_texture_object *tObj );
+   /** Unmap texture images from user space */
+   void (*UnmapTexture)( GLcontext *ctx, struct gl_texture_object *tObj );
+
    /**
     * Note: no context argument.  This function doesn't initially look
     * like it belongs here, except that the driver is the only entity
diff --git a/src/mesa/main/descrip.mms b/src/mesa/main/descrip.mms
index d09c57b321..3ef215f47f 100644
--- a/src/mesa/main/descrip.mms
+++ b/src/mesa/main/descrip.mms
@@ -1,11 +1,13 @@
 # Makefile for core library for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
-# Last revision : 10 May 2005
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Last revision : 2 October 2007
 
 .first
 	define gl [---.include.gl]
 	define math [-.math]
 	define shader [-.shader]
+	define glapi [-.glapi]
+	define main [-.main]
 
 .include [---]mms-config.
 
@@ -58,13 +60,13 @@ SOURCES =accum.c \
 	matrix.c \
 	mipmap.c \
 	mm.c \
-	occlude.c \
 	pixel.c \
 	points.c \
 	polygon.c \
 	rastpos.c \
 	rbadaptors.c \
 	renderbuffer.c \
+	shaders.c \
 	state.c \
 	stencil.c \
 	texcompress.c \
@@ -78,7 +80,9 @@ SOURCES =accum.c \
 	texstate.c \
 	texstore.c \
 	varray.c \
-	vtxfmt.c
+	vtxfmt.c \
+	queryobj.c \
+	rbadaptors.c
 
 OBJECTS=accum.obj,\
 api_arrayelt.obj,\
@@ -121,12 +125,12 @@ lines.obj,\
 matrix.obj,\
 mipmap.obj,\
 mm.obj,\
-occlude.obj,\
 pixel.obj,\
 points.obj,\
 polygon.obj,\
 rastpos.obj,\
 renderbuffer.obj,\
+shaders.obj,\
 state.obj,\
 stencil.obj,\
 texcompress.obj,\
@@ -140,7 +144,9 @@ texrender.obj,\
 texstate.obj,\
 texstore.obj,\
 varray.obj,\
-vtxfmt.obj
+vtxfmt.obj,\
+queryobj.obj,\
+rbadaptors.obj
 
 ##### RULES #####
 
@@ -149,8 +155,7 @@ VERSION=Mesa V3.4
 ##### TARGETS #####
 # Make the library
 $(LIBDIR)$(GL_LIB) : $(OBJECTS)
-  @ $(MAKELIB) $(GL_LIB) $(OBJECTS)
-  @ rename $(GL_LIB)* $(LIBDIR)
+  @ $(MAKELIB) $(LIBDIR)$(GL_LIB) $(OBJECTS)
 
 clean :
 	purge
@@ -197,7 +202,6 @@ lines.obj : lines.c
 matrix.obj : matrix.c
 mipmap.obj : mipmap.c
 mm.obj : mm.c
-occlude.obj : occlude.c
 pixel.obj : pixel.c
 points.obj : points.c
 polygon.obj : polygon.c
@@ -219,3 +223,6 @@ texstate.obj : texstate.c
 texstore.obj : texstore.c
 varray.obj : varray.c
 vtxfmt.obj : vtxfmt.c
+shaders.obj : shaders.c
+queryobj.obj : queryobj.c
+rbadaptors.obj : rbadaptors.c
diff --git a/src/mesa/main/enable.c b/src/mesa/main/enable.c
index ba4972b167..9fe979a6b2 100644
--- a/src/mesa/main/enable.c
+++ b/src/mesa/main/enable.c
@@ -5,9 +5,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.1
+ * Version:  7.0.3
  *
- * 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"),
@@ -364,12 +364,12 @@ _mesa_set_enable(GLcontext *ctx, GLenum cap, GLboolean state)
       case GL_LIGHTING:
          if (ctx->Light.Enabled == state)
             return;
+         FLUSH_VERTICES(ctx, _NEW_LIGHT);
+         ctx->Light.Enabled = state;
          if (ctx->Light.Enabled && ctx->Light.Model.TwoSide)
             ctx->_TriangleCaps |= DD_TRI_LIGHT_TWOSIDE;
          else
             ctx->_TriangleCaps &= ~DD_TRI_LIGHT_TWOSIDE;
-         FLUSH_VERTICES(ctx, _NEW_LIGHT);
-         ctx->Light.Enabled = state;
          break;
       case GL_LINE_SMOOTH:
          if (ctx->Line.SmoothFlag == state)
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 1fbfeb7447..94b7094388 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -3059,12 +3059,6 @@ struct __GLcontextRec
 
    struct gl_list_extensions ListExt; /**< driver dlist extensions */
 
-
-   GLuint _Facing; /**< This is a hack for 2-sided stencil test.
-		    *
-		    * We don't have a better way to communicate this value from
-		    * swrast_setup to swrast. */
-
    /** \name For debugging/development only */
    /*@{*/
    GLboolean FirstTimeCurrent;
diff --git a/src/mesa/main/texstore.c b/src/mesa/main/texstore.c
index 3b5151ed17..a8e639be6d 100644
--- a/src/mesa/main/texstore.c
+++ b/src/mesa/main/texstore.c
@@ -1443,7 +1443,6 @@ _mesa_texstore_argb8888(TEXSTORE_PARAMS)
 	    (baseInternalFormat == GL_RGBA ||
 	     baseInternalFormat == GL_RGB) &&
             srcType == GL_UNSIGNED_BYTE) {
-
       int img, row, col;
       for (img = 0; img < srcDepth; img++) {
          const GLint srcRowStride = _mesa_image_row_stride(srcPacking,
@@ -1455,11 +1454,12 @@ _mesa_texstore_argb8888(TEXSTORE_PARAMS)
             + dstYoffset * dstRowStride
             + dstXoffset * dstFormat->TexelBytes;
          for (row = 0; row < srcHeight; row++) {
+            GLuint *d4 = (GLuint *) dstRow;
             for (col = 0; col < srcWidth; col++) {
-               dstRow[col * 4 + 0] = srcRow[col * 3 + BCOMP];
-               dstRow[col * 4 + 1] = srcRow[col * 3 + GCOMP];
-               dstRow[col * 4 + 2] = srcRow[col * 3 + RCOMP];
-               dstRow[col * 4 + 3] = 0xff;
+               d4[col] = ((0xff                    << 24) |
+                          (srcRow[col * 3 + RCOMP] << 16) |
+                          (srcRow[col * 3 + GCOMP] <<  8) |
+                          (srcRow[col * 3 + BCOMP] <<  0));
             }
             dstRow += dstRowStride;
             srcRow += srcRowStride;
@@ -1471,7 +1471,9 @@ _mesa_texstore_argb8888(TEXSTORE_PARAMS)
 	    dstFormat == &_mesa_texformat_argb8888 &&
             srcFormat == GL_RGBA &&
 	    baseInternalFormat == GL_RGBA &&
-            (srcType == GL_UNSIGNED_BYTE && littleEndian)) {
+            srcType == GL_UNSIGNED_BYTE &&
+            littleEndian) {
+      /* same as above case, but src data has alpha too */
       GLint img, row, col;
       /* For some reason, streaming copies to write-combined regions
        * are extremely sensitive to the characteristics of how the
@@ -1488,13 +1490,13 @@ _mesa_texstore_argb8888(TEXSTORE_PARAMS)
             + dstImageOffsets[dstZoffset + img] * dstFormat->TexelBytes
             + dstYoffset * dstRowStride
             + dstXoffset * dstFormat->TexelBytes;
-
          for (row = 0; row < srcHeight; row++) {
+            GLuint *d4 = (GLuint *) dstRow;
             for (col = 0; col < srcWidth; col++) {
-               *(GLuint *)(dstRow + col * 4)  = (srcRow[col * 4 + RCOMP] << 16 |
-						 srcRow[col * 4 + GCOMP] << 8 |
-						 srcRow[col * 4 + BCOMP] << 0 |
-						 srcRow[col * 4 + ACOMP] << 24);
+               d4[col] = ((srcRow[col * 4 + ACOMP] << 24) |
+                          (srcRow[col * 4 + RCOMP] << 16) |
+                          (srcRow[col * 4 + GCOMP] <<  8) |
+                          (srcRow[col * 4 + BCOMP] <<  0));
             }
             dstRow += dstRowStride;
             srcRow += srcRowStride;
diff --git a/src/mesa/math/descrip.mms b/src/mesa/math/descrip.mms
index a3f20c2f25..3aaa6eb8b4 100644
--- a/src/mesa/math/descrip.mms
+++ b/src/mesa/math/descrip.mms
@@ -1,10 +1,12 @@
 # Makefile for core library for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
-# Last revision : 16 June 2003
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Last revision : 3 October 2007
 
 .first
 	define gl [---.include.gl]
 	define math [-.math]
+	define glapi [-.glapi]
+	define main [-.main]
 
 .include [---]mms-config.
 
diff --git a/src/mesa/shader/descrip.mms b/src/mesa/shader/descrip.mms
index d70cec3830..157c193c79 100644
--- a/src/mesa/shader/descrip.mms
+++ b/src/mesa/shader/descrip.mms
@@ -1,11 +1,14 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 20 November 2006
+# Last revision : 3 October 2007
 .first
 	define gl [---.include.gl]
 	define math [-.math]
 	define swrast [-.swrast]
 	define array_cache [-.array_cache]
+	define glapi [-.glapi]
+	define main [-.main]
+	define shader [-.shader]
 
 .include [---]mms-config.
 
@@ -23,11 +26,16 @@ SOURCES = \
 	arbprogram.c \
 	nvfragparse.c \
 	nvprogram.c \
-	nvvertexec.c \
 	nvvertparse.c \
 	program.c \
-	shaderobjects.c \
-	shaderobjects_3dlabs.c
+	programopt.c \
+	prog_debug.c \
+	prog_execute.c \
+	prog_instruction.c \
+	prog_parameter.c \
+	prog_print.c \
+	prog_statevars.c \
+	shader_api.c
 
 OBJECTS = \
 	atifragshader.obj,\
@@ -35,12 +43,16 @@ OBJECTS = \
 	arbprogram.obj,\
 	nvfragparse.obj,\
 	nvprogram.obj,\
-	nvvertexec.obj,\
 	nvvertparse.obj,\
 	program.obj,\
-	shaderobjects.obj,\
-	shaderobjects_3dlabs.obj
-
+	programopt.obj,\
+	prog_debug.obj,\
+	prog_execute.obj,\
+	prog_instruction.obj,\
+	prog_parameter.obj,\
+	prog_print.obj,\
+	prog_statevars.obj,\
+	shader_api.obj
 
 ##### RULES #####
 
@@ -68,9 +80,13 @@ arbprogparse.obj : arbprogparse.c
 arbprogram.obj : arbprogram.c
 nvfragparse.obj : nvfragparse.c
 nvprogram.obj : nvprogram.c
-nvvertexec.obj : nvvertexec.c
 nvvertparse.obj : nvvertparse.c
 program.obj : program.c
-shaderobjects.obj : shaderobjects.c
-	cc$(CFLAGS)/nowarn shaderobjects.c
-shaderobjects_3dlabs.obj : shaderobjects_3dlabs.c
+programopt. obj : programopt.c
+prog_debug.obj : prog_debug.c
+prog_execute.obj : prog_execute.c
+prog_instruction.obj : prog_instruction.c
+prog_parameter.obj : prog_parameter.c
+prog_print.obj : prog_print.c
+prog_statevars.obj : prog_statevars.c
+shader_api.obj : shader_api.c
diff --git a/src/mesa/shader/prog_statevars.c b/src/mesa/shader/prog_statevars.c
index 57a9d2dae9..c5dda01861 100644
--- a/src/mesa/shader/prog_statevars.c
+++ b/src/mesa/shader/prog_statevars.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.
  *
@@ -824,3 +824,94 @@ _mesa_load_state_parameters(GLcontext *ctx,
    }
 }
 
+
+/**
+ * Copy the 16 elements of a matrix into four consecutive program
+ * registers starting at 'pos'.
+ */
+static void
+load_matrix(GLfloat registers[][4], GLuint pos, const GLfloat mat[16])
+{
+   GLuint i;
+   for (i = 0; i < 4; i++) {
+      registers[pos + i][0] = mat[0 + i];
+      registers[pos + i][1] = mat[4 + i];
+      registers[pos + i][2] = mat[8 + i];
+      registers[pos + i][3] = mat[12 + i];
+   }
+}
+
+
+/**
+ * As above, but transpose the matrix.
+ */
+static void
+load_transpose_matrix(GLfloat registers[][4], GLuint pos,
+                      const GLfloat mat[16])
+{
+   MEMCPY(registers[pos], mat, 16 * sizeof(GLfloat));
+}
+
+
+/**
+ * Load current vertex program's parameter registers with tracked
+ * matrices (if NV program).  This only needs to be done per
+ * glBegin/glEnd, not per-vertex.
+ */
+void
+_mesa_load_tracked_matrices(GLcontext *ctx)
+{
+   GLuint i;
+
+   for (i = 0; i < MAX_NV_VERTEX_PROGRAM_PARAMS / 4; i++) {
+      /* point 'mat' at source matrix */
+      GLmatrix *mat;
+      if (ctx->VertexProgram.TrackMatrix[i] == GL_MODELVIEW) {
+         mat = ctx->ModelviewMatrixStack.Top;
+      }
+      else if (ctx->VertexProgram.TrackMatrix[i] == GL_PROJECTION) {
+         mat = ctx->ProjectionMatrixStack.Top;
+      }
+      else if (ctx->VertexProgram.TrackMatrix[i] == GL_TEXTURE) {
+         mat = ctx->TextureMatrixStack[ctx->Texture.CurrentUnit].Top;
+      }
+      else if (ctx->VertexProgram.TrackMatrix[i] == GL_COLOR) {
+         mat = ctx->ColorMatrixStack.Top;
+      }
+      else if (ctx->VertexProgram.TrackMatrix[i]==GL_MODELVIEW_PROJECTION_NV) {
+         /* XXX verify the combined matrix is up to date */
+         mat = &ctx->_ModelProjectMatrix;
+      }
+      else if (ctx->VertexProgram.TrackMatrix[i] >= GL_MATRIX0_NV &&
+               ctx->VertexProgram.TrackMatrix[i] <= GL_MATRIX7_NV) {
+         GLuint n = ctx->VertexProgram.TrackMatrix[i] - GL_MATRIX0_NV;
+         ASSERT(n < MAX_PROGRAM_MATRICES);
+         mat = ctx->ProgramMatrixStack[n].Top;
+      }
+      else {
+         /* no matrix is tracked, but we leave the register values as-is */
+         assert(ctx->VertexProgram.TrackMatrix[i] == GL_NONE);
+         continue;
+      }
+
+      /* load the matrix values into sequential registers */
+      if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_IDENTITY_NV) {
+         load_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
+      }
+      else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_INVERSE_NV) {
+         _math_matrix_analyse(mat); /* update the inverse */
+         ASSERT(!_math_matrix_is_dirty(mat));
+         load_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+      }
+      else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_TRANSPOSE_NV) {
+         load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
+      }
+      else {
+         assert(ctx->VertexProgram.TrackMatrixTransform[i]
+                == GL_INVERSE_TRANSPOSE_NV);
+         _math_matrix_analyse(mat); /* update the inverse */
+         ASSERT(!_math_matrix_is_dirty(mat));
+         load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+      }
+   }
+}
diff --git a/src/mesa/shader/prog_statevars.h b/src/mesa/shader/prog_statevars.h
index 3281a4a2a0..22bb8e07ad 100644
--- a/src/mesa/shader/prog_statevars.h
+++ b/src/mesa/shader/prog_statevars.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"),
@@ -126,4 +126,8 @@ extern const char *
 _mesa_program_state_string(const gl_state_index state[STATE_LENGTH]);
 
 
+extern void
+_mesa_load_tracked_matrices(GLcontext *ctx);
+
+
 #endif /* PROG_STATEVARS_H */
diff --git a/src/mesa/shader/slang/descrip.mms b/src/mesa/shader/slang/descrip.mms
index c86763718a..6eefbcf5bd 100644
--- a/src/mesa/shader/slang/descrip.mms
+++ b/src/mesa/shader/slang/descrip.mms
@@ -1,12 +1,15 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 17 March 2006
+# Last revision : 3 October 2007
 
 .first
 	define gl [----.include.gl]
 	define math [--.math]
 	define swrast [--.swrast]
 	define array_cache [--.array_cache]
+	define main [--.main]
+	define glapi [--.glapi]
+	define shader [--.shader]
 
 .include [----]mms-config.
 
@@ -21,15 +24,13 @@ CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ie
 SOURCES = \
 	slang_compile.c,slang_preprocess.c
 
-OBJECTS = \
-	slang_compile.obj,slang_preprocess.obj,slang_utility.obj,\
-	slang_execute.obj,slang_assemble.obj,slang_assemble_conditional.obj,\
-	slang_assemble_constructor.obj,slang_assemble_typeinfo.obj,\
-	slang_storage.obj,slang_assemble_assignment.obj,\
-	slang_compile_function.obj,slang_compile_struct.obj,\
-	slang_compile_variable.obj,slang_compile_operation.obj,\
-	slang_library_noise.obj,slang_link.obj,slang_export.obj,\
-	slang_analyse.obj,slang_library_texsample.obj
+OBJECTS = slang_builtin.obj,slang_codegen.obj,slang_compile.obj,\
+	slang_compile_function.obj,slang_compile_operation.obj,\
+	slang_compile_struct.obj,slang_compile_variable.obj,slang_emit.obj,\
+	slang_ir.obj,slang_label.obj,slang_library_noise.obj,slang_link.obj,\
+	slang_log.obj,slang_mem.obj,slang_preprocess.obj,slang_print.obj,\
+	slang_simplify.obj,slang_storage.obj,slang_typeinfo.obj,\
+	slang_utility.obj,slang_vartable.obj
 
 ##### RULES #####
 
@@ -44,22 +45,24 @@ clean :
 	purge
 	delete *.obj;*
 
+slang_builtin.obj : slang_builtin.c
+slang_codegen.obj : slang_codegen.c
 slang_compile.obj : slang_compile.c
-slang_preprocess.obj : slang_preprocess.c
-slang_utility.obj : slang_utility.c
-slang_execute.obj : slang_execute.c
-slang_assemble.obj : slang_assemble.c
-slang_assemble_conditional.obj : slang_assemble_conditional.c
-slang_assemble_constructor.obj : slang_assemble_constructor.c
-slang_assemble_typeinfo.obj : slang_assemble_typeinfo.c
-slang_storage.obj : slang_storage.c
-slang_assemble_assignment.obj : slang_assemble_assignment.c
 slang_compile_function.obj : slang_compile_function.c
+slang_compile_operation.obj : slang_compile_operation.c
 slang_compile_struct.obj : slang_compile_struct.c
 slang_compile_variable.obj : slang_compile_variable.c
-slang_compile_operation.obj : slang_compile_operation.c
+slang_emit.obj : slang_emit.c
+slang_ir.obj : slang_ir.c
+slang_label.obj : slang_label.c
 slang_library_noise.obj : slang_library_noise.c
 slang_link.obj : slang_link.c
-slang_export.obj : slang_export.c
-slang_analyse.obj : slang_analyse.c
-slang_library_texsample.obj : slang_library_texsample.c
+slang_log.obj : slang_log.c
+slang_mem.obj : slang_mem.c
+slang_preprocess.obj : slang_preprocess.c
+slang_print.obj : slang_print.c
+slang_simplify.obj : slang_simplify.c
+slang_storage.obj : slang_storage.c
+slang_typeinfo.obj : slang_typeinfo.c
+slang_utility.obj : slang_utility.c
+slang_vartable.obj : slang_vartable.c
diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c
index 5191e0c222..f6ea57b3f2 100644
--- a/src/mesa/shader/slang/slang_codegen.c
+++ b/src/mesa/shader/slang/slang_codegen.c
@@ -2352,7 +2352,8 @@ _slang_gen_field(slang_assemble_ctx * A, slang_operation *oper)
       return n;
    }
    else if (   ti.spec.type == SLANG_SPEC_FLOAT
-            || ti.spec.type == SLANG_SPEC_INT) {
+            || ti.spec.type == SLANG_SPEC_INT
+            || ti.spec.type == SLANG_SPEC_BOOL) {
       const GLuint rows = 1;
       slang_swizzle swz;
       slang_ir_node *n;
diff --git a/src/mesa/swrast/descrip.mms b/src/mesa/swrast/descrip.mms
index 4d49673b5d..0b23deb3c2 100644
--- a/src/mesa/swrast/descrip.mms
+++ b/src/mesa/swrast/descrip.mms
@@ -1,12 +1,15 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 21 February 2006
+# Last revision : 3 October 2007
 
 .first
 	define gl [---.include.gl]
 	define math [-.math]
 	define swrast [-.swrast]
 	define array_cache [-.array_cache]
+	define glapi [-.glapi]
+	define main [-.main]
+	define shader [-.shader]
 
 .include [---]mms-config.
 
@@ -20,17 +23,17 @@ CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ie
 
 SOURCES = s_aaline.c s_aatriangle.c s_accum.c s_alpha.c \
 	s_bitmap.c s_blend.c s_blit.c s_buffers.c s_context.c \
-	s_copypix.c s_depth.c \
+	s_copypix.c s_depth.c s_fragprog.c \
         s_drawpix.c s_feedback.c s_fog.c s_imaging.c s_lines.c s_logic.c \
-	s_masking.c s_nvfragprog.c s_points.c s_readpix.c \
+	s_masking.c s_points.c s_readpix.c \
 	s_span.c s_stencil.c s_texstore.c s_texcombine.c s_texfilter.c \
-	s_triangle.c s_zoom.c s_atifragshader.c s_arbshader.c
+	s_triangle.c s_zoom.c s_atifragshader.c
  
 OBJECTS = s_aaline.obj,s_aatriangle.obj,s_accum.obj,s_alpha.obj,\
-	s_bitmap.obj,s_blend.obj,s_blit.obj,s_arbshader.obj,\
+	s_bitmap.obj,s_blend.obj,s_blit.obj,s_fragprog.obj,\
 	s_buffers.obj,s_context.obj,s_atifragshader.obj,\
 	s_copypix.obj,s_depth.obj,s_drawpix.obj,s_feedback.obj,s_fog.obj,\
-	s_imaging.obj,s_lines.obj,s_logic.obj,s_masking.obj,s_nvfragprog.obj,\
+	s_imaging.obj,s_lines.obj,s_logic.obj,s_masking.obj,\
 	s_points.obj,s_readpix.obj,s_span.obj,s_stencil.obj,\
 	s_texstore.obj,s_texcombine.obj,s_texfilter.obj,s_triangle.obj,\
 	s_zoom.obj
@@ -67,7 +70,6 @@ s_imaging.obj : s_imaging.c
 s_lines.obj : s_lines.c
 s_logic.obj : s_logic.c
 s_masking.obj : s_masking.c
-s_nvfragprog.obj : s_nvfragprog.c
 s_points.obj : s_points.c
 s_readpix.obj : s_readpix.c
 s_span.obj : s_span.c
@@ -77,4 +79,4 @@ s_texcombine.obj : s_texcombine.c
 s_texfilter.obj : s_texfilter.c
 s_triangle.obj : s_triangle.c
 s_zoom.obj : s_zoom.c
-s_arbshader.obj : s_arbshader.c
+s_fragprog.obj : s_fragprog.c
diff --git a/src/mesa/swrast/s_aalinetemp.h b/src/mesa/swrast/s_aalinetemp.h
index e7911fec3b..ca08203d83 100644
--- a/src/mesa/swrast/s_aalinetemp.h
+++ b/src/mesa/swrast/s_aalinetemp.h
@@ -141,6 +141,7 @@ NAME(line)(GLcontext *ctx, const SWvertex *v0, const SWvertex *v1)
 
    INIT_SPAN(line.span, GL_LINE);
    line.span.arrayMask = SPAN_XY | SPAN_COVERAGE;
+   line.span.facing = swrast->PointLineFacing;
    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 29609add17..0827b3db9e 100644
--- a/src/mesa/swrast/s_aatritemp.h
+++ b/src/mesa/swrast/s_aatritemp.h
@@ -65,7 +65,7 @@
    GLfloat attrPlane[FRAG_ATTRIB_MAX][4][4];
    GLfloat wPlane[4];  /* win[3] */
 #endif
-   GLfloat bf = SWRAST_CONTEXT(ctx)->_BackfaceSign;
+   GLfloat bf = SWRAST_CONTEXT(ctx)->_BackfaceCullSign;
    
    (void) swrast;
 
@@ -104,6 +104,7 @@
    majDx = vMax->attrib[FRAG_ATTRIB_WPOS][0] - vMin->attrib[FRAG_ATTRIB_WPOS][0];
    majDy = vMax->attrib[FRAG_ATTRIB_WPOS][1] - vMin->attrib[FRAG_ATTRIB_WPOS][1];
 
+   /* front/back-face determination and cullling */
    {
       const GLfloat botDx = vMid->attrib[FRAG_ATTRIB_WPOS][0] - vMin->attrib[FRAG_ATTRIB_WPOS][0];
       const GLfloat botDy = vMid->attrib[FRAG_ATTRIB_WPOS][1] - vMin->attrib[FRAG_ATTRIB_WPOS][1];
@@ -112,6 +113,8 @@
       if (area * bf < 0 || area == 0 || IS_INF_OR_NAN(area))
 	 return;
       ltor = (GLboolean) (area < 0.0F);
+
+      span.facing = area * swrast->_BackfaceSign > 0.0F;
    }
 
    /* Plane equation setup:
diff --git a/src/mesa/swrast/s_context.c b/src/mesa/swrast/s_context.c
index 3956925651..3bf9804bef 100644
--- a/src/mesa/swrast/s_context.c
+++ b/src/mesa/swrast/s_context.c
@@ -117,8 +117,8 @@ _swrast_update_rasterflags( GLcontext *ctx )
 
 
 /**
- * Examine polycon culls tate to compute the _BackfaceSign field.
- * _BackfaceSign will be 0 if no culling, -1 if culling back-faces,
+ * Examine polycon culls tate to compute the _BackfaceCullSign field.
+ * _BackfaceCullSign will be 0 if no culling, -1 if culling back-faces,
  * and 1 if culling front-faces.  The Polygon FrontFace state also
  * factors in.
  */
@@ -149,10 +149,15 @@ _swrast_update_polygon( GLcontext *ctx )
       backface_sign = 0.0;
    }
 
-   SWRAST_CONTEXT(ctx)->_BackfaceSign = backface_sign;
+   SWRAST_CONTEXT(ctx)->_BackfaceCullSign = backface_sign;
+
+   /* This is for front/back-face determination, but not for culling */
+   SWRAST_CONTEXT(ctx)->_BackfaceSign
+      = (ctx->Polygon.FrontFace == GL_CW) ? -1.0 : 1.0;
 }
 
 
+
 /**
  * Update the _PreferPixelFog field to indicate if we need to compute
  * fog blend factors (from the fog coords) per-fragment.
@@ -764,6 +769,12 @@ _swrast_ResetLineStipple( GLcontext *ctx )
 }
 
 void
+_swrast_SetFacing(GLcontext *ctx, GLuint facing)
+{
+   SWRAST_CONTEXT(ctx)->PointLineFacing = facing;
+}
+
+void
 _swrast_allow_vertex_fog( GLcontext *ctx, GLboolean value )
 {
    if (SWRAST_DEBUG) {
diff --git a/src/mesa/swrast/s_context.h b/src/mesa/swrast/s_context.h
index daa07e1578..aebc80c208 100644
--- a/src/mesa/swrast/s_context.h
+++ b/src/mesa/swrast/s_context.h
@@ -128,7 +128,8 @@ typedef struct
     * _swrast_validate_derived():
     */
    GLbitfield _RasterMask;
-   GLfloat _BackfaceSign;
+   GLfloat _BackfaceSign;      /** +1 or -1 */
+   GLfloat _BackfaceCullSign;  /** +1, 0, or -1 */
    GLboolean _PreferPixelFog;    /* Compute fog blend factor per fragment? */
    GLboolean _AnyTextureCombine;
    GLboolean _FogEnabled;
@@ -156,6 +157,7 @@ typedef struct
    /* Working values:
     */
    GLuint StippleCounter;    /**< Line stipple counter */
+   GLuint PointLineFacing;
    GLbitfield NewState;
    GLuint StateChanges;
    GLenum Primitive;    /* current primitive being drawn (ala glBegin) */
diff --git a/src/mesa/swrast/s_fragprog.c b/src/mesa/swrast/s_fragprog.c
index 4067fd6886..76c79eb0d8 100644
--- a/src/mesa/swrast/s_fragprog.c
+++ b/src/mesa/swrast/s_fragprog.c
@@ -120,7 +120,7 @@ init_machine(GLcontext *ctx, struct gl_program_machine *machine,
    /* if running a GLSL program (not ARB_fragment_program) */
    if (ctx->Shader.CurrentProgram) {
       /* Store front/back facing value in register FOGC.Y */
-      machine->Attribs[FRAG_ATTRIB_FOGC][col][1] = (GLfloat) ctx->_Facing;
+      machine->Attribs[FRAG_ATTRIB_FOGC][col][1] = 1.0 - span->facing;
    }
 
    machine->CurElement = col;
diff --git a/src/mesa/swrast/s_linetemp.h b/src/mesa/swrast/s_linetemp.h
index 1accfc67e2..1abf8d6c7f 100644
--- a/src/mesa/swrast/s_linetemp.h
+++ b/src/mesa/swrast/s_linetemp.h
@@ -308,6 +308,9 @@ NAME( GLcontext *ctx, const SWvertex *vert0, const SWvertex *vert1 )
    span.interpMask = interpFlags;
    span.arrayMask = SPAN_XY;
 
+   span.facing = swrast->PointLineFacing;
+
+
    /*
     * Draw
     */
diff --git a/src/mesa/swrast/s_points.c b/src/mesa/swrast/s_points.c
index dd664b980e..d60e175baa 100644
--- a/src/mesa/swrast/s_points.c
+++ b/src/mesa/swrast/s_points.c
@@ -106,6 +106,8 @@ sprite_point(GLcontext *ctx, const SWvertex *vert)
    INIT_SPAN(span, GL_POINT);
    span.interpMask = SPAN_Z | SPAN_RGBA;
 
+   span.facing = swrast->PointLineFacing;
+
    span.red   = ChanToFixed(vert->color[0]);
    span.green = ChanToFixed(vert->color[1]);
    span.blue  = ChanToFixed(vert->color[2]);
@@ -280,6 +282,8 @@ smooth_point(GLcontext *ctx, const SWvertex *vert)
    span.interpMask = SPAN_Z | SPAN_RGBA;
    span.arrayMask = SPAN_COVERAGE | SPAN_MASK;
 
+   span.facing = swrast->PointLineFacing;
+
    span.red   = ChanToFixed(vert->color[0]);
    span.green = ChanToFixed(vert->color[1]);
    span.blue  = ChanToFixed(vert->color[2]);
@@ -386,6 +390,7 @@ large_point(GLcontext *ctx, const SWvertex *vert)
    /* span init */
    INIT_SPAN(span, GL_POINT);
    span.arrayMask = SPAN_XY;
+   span.facing = swrast->PointLineFacing;
 
    if (ciMode) {
       span.interpMask = SPAN_Z | SPAN_INDEX;
@@ -492,7 +497,8 @@ pixel_point(GLcontext *ctx, const SWvertex *vert)
 
    /* check if we need to flush */
    if (span->end >= MAX_WIDTH ||
-       (swrast->_RasterMask & (BLEND_BIT | LOGIC_OP_BIT | MASKING_BIT))) {
+       (swrast->_RasterMask & (BLEND_BIT | LOGIC_OP_BIT | MASKING_BIT)) ||
+       span->facing != swrast->PointLineFacing) {
       if (ciMode)
          _swrast_write_index_span(ctx, span);
       else
@@ -502,6 +508,8 @@ pixel_point(GLcontext *ctx, const SWvertex *vert)
 
    count = span->end;
 
+   span->facing = swrast->PointLineFacing;
+
    /* fragment attributes */
    if (ciMode) {
       span->array->index[count] = (GLuint) vert->attrib[FRAG_ATTRIB_CI][0];
diff --git a/src/mesa/swrast/s_triangle.c b/src/mesa/swrast/s_triangle.c
index c255545217..938cdefa56 100644
--- a/src/mesa/swrast/s_triangle.c
+++ b/src/mesa/swrast/s_triangle.c
@@ -58,7 +58,7 @@ _swrast_culltriangle( GLcontext *ctx,
    GLfloat fy = v2->attrib[FRAG_ATTRIB_WPOS][1] - v0->attrib[FRAG_ATTRIB_WPOS][1];
    GLfloat c = ex*fy-ey*fx;
 
-   if (c * SWRAST_CONTEXT(ctx)->_BackfaceSign > 0)
+   if (c * SWRAST_CONTEXT(ctx)->_BackfaceCullSign > 0)
       return 0;
 
    return 1;
diff --git a/src/mesa/swrast/s_tritemp.h b/src/mesa/swrast/s_tritemp.h
index cded4a6c1c..8e3c5b5eeb 100644
--- a/src/mesa/swrast/s_tritemp.h
+++ b/src/mesa/swrast/s_tritemp.h
@@ -234,18 +234,18 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
    /* compute area, oneOverArea and perform backface culling */
    {
       const GLfloat area = eMaj.dx * eBot.dy - eBot.dx * eMaj.dy;
-      /* Do backface culling */
-      if (area * bf < 0.0)
-         return;
 
       if (IS_INF_OR_NAN(area) || area == 0.0F)
          return;
 
-      oneOverArea = 1.0F / area;
-   }
+      if (area * bf * swrast->_BackfaceCullSign < 0.0)
+         return;
 
+      oneOverArea = 1.0F / area;
 
-   span.facing = ctx->_Facing; /* for 2-sided stencil test */
+      /* 0 = front, 1 = back */
+      span.facing = oneOverArea * bf > 0.0F;
+   }
 
    /* Edge setup.  For a triangle strip these could be reused... */
    {
diff --git a/src/mesa/swrast/swrast.h b/src/mesa/swrast/swrast.h
index 85a27fd55b..047f7991e6 100644
--- a/src/mesa/swrast/swrast.h
+++ b/src/mesa/swrast/swrast.h
@@ -143,6 +143,13 @@ _swrast_Accum(GLcontext *ctx, GLenum op, GLfloat value);
 extern void
 _swrast_ResetLineStipple( GLcontext *ctx );
 
+/**
+ * Indicates front/back facing for subsequent points/lines when drawing
+ * unfilled polygons.  Needed for two-side stencil.
+ */
+extern void
+_swrast_SetFacing(GLcontext *ctx, GLuint facing);
+
 /* These will always render the correct point/line/triangle for the
  * current state.
  *
diff --git a/src/mesa/swrast_setup/descrip.mms b/src/mesa/swrast_setup/descrip.mms
index e5e48afb3d..32ffd80732 100644
--- a/src/mesa/swrast_setup/descrip.mms
+++ b/src/mesa/swrast_setup/descrip.mms
@@ -1,13 +1,16 @@
 # Makefile for core library for VMS
-# contributed by Jouk Jansen  joukj@hrem.stm.tudelft.nl
-# Last revision : 16 June 2003
+# contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
+# Last revision : 3 October 2007
 
 .first
 	define gl [---.include.gl]
 	define math [-.math]
 	define tnl [-.tnl]
+	define vbo [-.vbo]
 	define swrast [-.swrast]
 	define array_cache [-.array_cache]
+	define glapi [-.glapi]
+	define main [-.main]
 
 .include [---]mms-config.
 
diff --git a/src/mesa/swrast_setup/ss_context.c b/src/mesa/swrast_setup/ss_context.c
index a9c7d941e5..a4f949ddfe 100644
--- a/src/mesa/swrast_setup/ss_context.c
+++ b/src/mesa/swrast_setup/ss_context.c
@@ -201,6 +201,9 @@ _swsetup_RenderStart( GLcontext *ctx )
 
    swsetup->NewState = 0;
 
+   /* This will change if drawing unfilled tris */
+   _swrast_SetFacing(ctx, 0);
+
    _swrast_render_start(ctx);
 
    /* Important */
diff --git a/src/mesa/swrast_setup/ss_triangle.c b/src/mesa/swrast_setup/ss_triangle.c
index b4207f2c64..4817d23932 100644
--- a/src/mesa/swrast_setup/ss_triangle.c
+++ b/src/mesa/swrast_setup/ss_triangle.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"),
@@ -68,6 +68,8 @@ static void _swsetup_render_line_tri( GLcontext *ctx,
          return;
    }
 
+   _swrast_SetFacing(ctx, facing);
+
    if (ctx->Light.ShadeModel == GL_FLAT) {
       COPY_CHAN4(c[0], v0->color);
       COPY_CHAN4(c[1], v1->color);
@@ -127,6 +129,8 @@ static void _swsetup_render_point_tri( GLcontext *ctx,
          return;
    }
 
+   _swrast_SetFacing(ctx, facing);
+
    if (ctx->Light.ShadeModel == GL_FLAT) {
       /* save colors/indexes for v0, v1 vertices */
       COPY_CHAN4(c[0], v0->color);
@@ -311,6 +315,4 @@ void _swsetup_choose_trifuncs( GLcontext *ctx )
    tnl->Driver.Render.Quad = quad_tab[ind];
    tnl->Driver.Render.Line = swsetup_line;
    tnl->Driver.Render.Points = swsetup_points;
-
-   ctx->_Facing = 0;
 }
diff --git a/src/mesa/swrast_setup/ss_tritmp.h b/src/mesa/swrast_setup/ss_tritmp.h
index 7d1bc23465..97d2f4a16b 100644
--- a/src/mesa/swrast_setup/ss_tritmp.h
+++ b/src/mesa/swrast_setup/ss_tritmp.h
@@ -49,7 +49,6 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
    v[1] = &verts[e1];
    v[2] = &verts[e2];
 
-
    if (IND & (SS_TWOSIDE_BIT | SS_OFFSET_BIT | SS_UNFILLED_BIT))
    {
       GLfloat ex = v[0]->attrib[FRAG_ATTRIB_WPOS][0] - v[2]->attrib[FRAG_ATTRIB_WPOS][0];
@@ -61,7 +60,6 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
       if (IND & (SS_TWOSIDE_BIT | SS_UNFILLED_BIT))
       {
 	 facing = (cc < 0.0) ^ ctx->Polygon._FrontBit;
-         ctx->_Facing = facing;
 
 	 if (IND & SS_UNFILLED_BIT)
 	    mode = facing ? ctx->Polygon.BackMode : ctx->Polygon.FrontMode;
diff --git a/src/mesa/tnl/descrip.mms b/src/mesa/tnl/descrip.mms
index 91e32bf978..f77f672dc8 100644
--- a/src/mesa/tnl/descrip.mms
+++ b/src/mesa/tnl/descrip.mms
@@ -1,12 +1,17 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 21 February 2006
+# Last revision : 30 November 2007
 
 .first
 	define gl [---.include.gl]
 	define math [-.math]
+	define vbo [-.vbo]
 	define shader [-.shader]
+	define swrast [-.swrast]
 	define array_cache [-.array_cache]
+	define main [-.main]
+	define glapi [-.glapi]
+	define tnl [-.tnl]
 
 .include [---]mms-config.
 
@@ -18,21 +23,19 @@ INCDIR = [---.include],[-.main],[-.glapi],[-.shader],[-.shader.slang]
 LIBDIR = [---.lib]
 CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ieee/ieee=denorm
 
-SOURCES = t_array_api.c t_array_import.c t_context.c \
-	t_pipeline.c t_vb_fog.c t_save_api.c t_vtx_api.c \
+SOURCES = t_context.c t_draw.c \
+	t_pipeline.c t_vb_fog.c \
 	t_vb_light.c t_vb_normals.c t_vb_points.c t_vb_program.c \
 	t_vb_render.c t_vb_texgen.c t_vb_texmat.c t_vb_vertex.c \
-	t_vtx_eval.c t_vtx_exec.c t_save_playback.c t_save_loopback.c \
-	t_vertex.c t_vtx_generic.c t_vtx_x86.c t_vertex_generic.c \
-	t_vb_arbprogram.c t_vp_build.c t_vb_arbshader.c
+	t_vertex.c \
+	t_vertex_generic.c t_vp_build.c
 
-OBJECTS = t_array_api.obj,t_array_import.obj,t_context.obj,\
+OBJECTS = t_context.obj,t_draw.obj,\
 	t_pipeline.obj,t_vb_fog.obj,t_vb_light.obj,t_vb_normals.obj,\
 	t_vb_points.obj,t_vb_program.obj,t_vb_render.obj,t_vb_texgen.obj,\
-	t_vb_texmat.obj,t_vb_vertex.obj,t_save_api.obj,t_vtx_api.obj,\
-	t_vtx_eval.obj,t_vtx_exec.obj,t_save_playback.obj,t_save_loopback.obj,\
-	t_vertex.obj,t_vtx_generic.obj,t_vtx_x86.obj,t_vertex_generic.obj,\
-	t_vb_arbprogram.obj,t_vp_build.obj,t_vb_arbshader.obj
+	t_vb_texmat.obj,t_vb_vertex.obj,\
+	t_vertex.obj,t_vertex_generic.obj,\
+	t_vp_build.obj
 
 ##### RULES #####
 
@@ -47,9 +50,8 @@ clean :
 	purge
 	delete *.obj;*
 
-t_array_api.obj : t_array_api.c
-t_array_import.obj : t_array_import.c
 t_context.obj : t_context.c
+t_draw.obj : t_draw.c
 t_pipeline.obj : t_pipeline.c
 t_vb_fog.obj : t_vb_fog.c
 t_vb_light.obj : t_vb_light.c
@@ -60,16 +62,6 @@ t_vb_render.obj : t_vb_render.c
 t_vb_texgen.obj : t_vb_texgen.c
 t_vb_texmat.obj : t_vb_texmat.c
 t_vb_vertex.obj : t_vb_vertex.c
-t_save_api.obj : t_save_api.c
-t_vtx_api.obj : t_vtx_api.c
-t_vtx_eval.obj : t_vtx_eval.c
-t_vtx_exec.obj : t_vtx_exec.c
-t_save_playback.obj : t_save_playback.c
-t_save_loopback.obj : t_save_loopback.c
 t_vertex.obj : t_vertex.c
-t_vtx_x86.obj : t_vtx_x86.c
-t_vtx_generic.obj : t_vtx_generic.c
 t_vertex_generic.obj : t_vertex_generic.c
-t_vb_arbprogram.obj : t_vb_arbprogram.c
 t_vp_build.obj : t_vp_build.c
-t_vb_arbshader.obj : t_vb_arbshader.c
diff --git a/src/mesa/tnl/t_vb_program.c b/src/mesa/tnl/t_vb_program.c
index f15f9d8619..addaf76127 100644
--- a/src/mesa/tnl/t_vb_program.c
+++ b/src/mesa/tnl/t_vb_program.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.
  *
@@ -30,20 +30,20 @@
  */
 
 
-#include "glheader.h"
-#include "colormac.h"
-#include "context.h"
-#include "macros.h"
-#include "imports.h"
+#include "main/glheader.h"
+#include "main/colormac.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "main/imports.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 "tnl/tnl.h"
+#include "tnl/t_context.h"
+#include "tnl/t_pipeline.h"
 
 
 
@@ -63,18 +63,12 @@ struct vp_stage_data {
 #define VP_STAGE_DATA(stage) ((struct vp_stage_data *)(stage->privatePtr))
 
 
-/**
- * XXX the texture sampling code in this module is a bit of a hack.
- * The texture sampling code is in swrast, though it doesn't have any
- * real dependencies on the rest of swrast.  It should probably be
- * moved into main/ someday.
- */
-
-static void userclip( GLcontext *ctx,
-		      GLvector4f *clip,
-		      GLubyte *clipmask,
-		      GLubyte *clipormask,
-		      GLubyte *clipandmask )
+static void
+userclip( GLcontext *ctx,
+          GLvector4f *clip,
+          GLubyte *clipmask,
+          GLubyte *clipormask,
+          GLubyte *clipandmask )
 {
    GLuint p;
 
@@ -172,6 +166,12 @@ do_ndc_cliptest(GLcontext *ctx, struct vp_stage_data *store)
 }
 
 
+/**
+ * XXX the texture sampling code in this module is a bit of a hack.
+ * The texture sampling code is in swrast, though it doesn't have any
+ * real dependencies on the rest of swrast.  It should probably be
+ * moved into main/ someday.
+ */
 static void
 vp_fetch_texel(GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
                GLuint unit, GLfloat color[4])
@@ -246,92 +246,44 @@ init_machine(GLcontext *ctx, struct gl_program_machine *machine)
 
 
 /**
- * Copy the 16 elements of a matrix into four consecutive program
- * registers starting at 'pos'.
+ * Map the texture images which the vertex program will access (if any).
  */
 static void
-load_matrix(GLfloat registers[][4], GLuint pos, const GLfloat mat[16])
+map_textures(GLcontext *ctx, const struct gl_vertex_program *vp)
 {
-   GLuint i;
-   for (i = 0; i < 4; i++) {
-      registers[pos + i][0] = mat[0 + i];
-      registers[pos + i][1] = mat[4 + i];
-      registers[pos + i][2] = mat[8 + i];
-      registers[pos + i][3] = mat[12 + i];
-   }
-}
+   GLuint u;
 
+   if (!ctx->Driver.MapTexture)
+      return;
 
-/**
- * As above, but transpose the matrix.
- */
-static void
-load_transpose_matrix(GLfloat registers[][4], GLuint pos,
-                      const GLfloat mat[16])
-{
-   MEMCPY(registers[pos], mat, 16 * sizeof(GLfloat));
+   for (u = 0; u < ctx->Const.MaxVertexTextureImageUnits; u++) {
+      if (vp->Base.TexturesUsed[u]) {
+         /* Note: _Current *should* correspond to the target indicated
+          * in TexturesUsed[u].
+          */
+         ctx->Driver.MapTexture(ctx, ctx->Texture.Unit[u]._Current);
+      }
+   }
 }
 
 
 /**
- * Load current vertex program's parameter registers with tracked
- * matrices (if NV program).  This only needs to be done per
- * glBegin/glEnd, not per-vertex.
+ * Unmap the texture images which were used by the vertex program (if any).
  */
-void
-_mesa_load_tracked_matrices(GLcontext *ctx)
+static void
+unmap_textures(GLcontext *ctx, const struct gl_vertex_program *vp)
 {
-   GLuint i;
+   GLuint u;
 
-   for (i = 0; i < MAX_NV_VERTEX_PROGRAM_PARAMS / 4; i++) {
-      /* point 'mat' at source matrix */
-      GLmatrix *mat;
-      if (ctx->VertexProgram.TrackMatrix[i] == GL_MODELVIEW) {
-         mat = ctx->ModelviewMatrixStack.Top;
-      }
-      else if (ctx->VertexProgram.TrackMatrix[i] == GL_PROJECTION) {
-         mat = ctx->ProjectionMatrixStack.Top;
-      }
-      else if (ctx->VertexProgram.TrackMatrix[i] == GL_TEXTURE) {
-         mat = ctx->TextureMatrixStack[ctx->Texture.CurrentUnit].Top;
-      }
-      else if (ctx->VertexProgram.TrackMatrix[i] == GL_COLOR) {
-         mat = ctx->ColorMatrixStack.Top;
-      }
-      else if (ctx->VertexProgram.TrackMatrix[i]==GL_MODELVIEW_PROJECTION_NV) {
-         /* XXX verify the combined matrix is up to date */
-         mat = &ctx->_ModelProjectMatrix;
-      }
-      else if (ctx->VertexProgram.TrackMatrix[i] >= GL_MATRIX0_NV &&
-               ctx->VertexProgram.TrackMatrix[i] <= GL_MATRIX7_NV) {
-         GLuint n = ctx->VertexProgram.TrackMatrix[i] - GL_MATRIX0_NV;
-         ASSERT(n < MAX_PROGRAM_MATRICES);
-         mat = ctx->ProgramMatrixStack[n].Top;
-      }
-      else {
-         /* no matrix is tracked, but we leave the register values as-is */
-         assert(ctx->VertexProgram.TrackMatrix[i] == GL_NONE);
-         continue;
-      }
+   if (!ctx->Driver.MapTexture)
+      return;
 
-      /* load the matrix values into sequential registers */
-      if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_IDENTITY_NV) {
-         load_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
-      }
-      else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_INVERSE_NV) {
-         _math_matrix_analyse(mat); /* update the inverse */
-         ASSERT(!_math_matrix_is_dirty(mat));
-         load_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
-      }
-      else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_TRANSPOSE_NV) {
-         load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
-      }
-      else {
-         assert(ctx->VertexProgram.TrackMatrixTransform[i]
-                == GL_INVERSE_TRANSPOSE_NV);
-         _math_matrix_analyse(mat); /* update the inverse */
-         ASSERT(!_math_matrix_is_dirty(mat));
-         load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+   for (u = 0; u < ctx->Const.MaxVertexTextureImageUnits; u++) {
+      if (vp->Base.TexturesUsed[u]) {
+         /* Note: _Current *should* correspond to the target indicated
+          * in TexturesUsed[u].
+          */
+         ctx->Driver.UnmapTexture(ctx, ctx->Texture.Unit[u]._Current);
       }
    }
 }
@@ -362,6 +314,7 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
       _mesa_load_state_parameters(ctx, program->Base.Parameters);
    }
 
+   /* make list of outputs to save some time below */
    numOutputs = 0;
    for (i = 0; i < VERT_RESULT_MAX; i++) {
       if (program->Base.OutputsWritten & (1 << i)) {
@@ -369,6 +322,8 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
       }
    }
 
+   map_textures(ctx, program);
+
    for (i = 0; i < VB->Count; i++) {
       GLuint attr;
 
@@ -420,6 +375,8 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
 #endif
    }
 
+   unmap_textures(ctx, program);
+
    /* Fixup fog and point size results if needed */
    if (program->IsNVProgram) {
       if (ctx->Fog.Enabled &&
@@ -439,8 +396,9 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
 
    if (program->IsPositionInvariant) {
       /* We need the exact same transform as in the fixed function path here
-         to guarantee invariance, depending on compiler optimization flags results
-         could be different otherwise */
+       * to guarantee invariance, depending on compiler optimization flags
+       * results could be different otherwise.
+       */
       VB->ClipPtr = TransformRaw( &store->results[0],
 				  &ctx->_ModelProjectMatrix,
 				  VB->AttribPtr[0] );
@@ -460,16 +418,15 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
 	 break;
       }
    }
-
-
-   /* Setup the VB pointers so that the next pipeline stages get
-    * their data from the right place (the program output arrays).
-    */
    else {
+      /* Setup the VB pointers so that the next pipeline stages get
+       * their data from the right place (the program output arrays).
+       */
       VB->ClipPtr = &store->results[VERT_RESULT_HPOS];
       VB->ClipPtr->size = 4;
       VB->ClipPtr->count = VB->Count;
    }
+
    VB->ColorPtr[0] = &store->results[VERT_RESULT_COL0];
    VB->ColorPtr[1] = &store->results[VERT_RESULT_BFC0];
    VB->SecondaryColorPtr[0] = &store->results[VERT_RESULT_COL1];
diff --git a/src/mesa/tnl/tnl.h b/src/mesa/tnl/tnl.h
index 047b764dcb..c2b1f71caa 100644
--- a/src/mesa/tnl/tnl.h
+++ b/src/mesa/tnl/tnl.h
@@ -1,9 +1,8 @@
-
 /*
  * Mesa 3-D graphics library
- * Version:  3.5
+ * Version:  7.1
  *
- * Copyright (C) 1999-2001  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 +81,4 @@ _tnl_draw_prims( GLcontext *ctx,
 		 GLuint min_index,
 		 GLuint max_index);
 
-extern void
-_mesa_load_tracked_matrices(GLcontext *ctx);
-
 #endif
diff --git a/src/mesa/tnl_dd/t_dd_tritmp.h b/src/mesa/tnl_dd/t_dd_tritmp.h
index 0bf32ff344..6c2378f5e8 100644
--- a/src/mesa/tnl_dd/t_dd_tritmp.h
+++ b/src/mesa/tnl_dd/t_dd_tritmp.h
@@ -155,10 +155,6 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
       {
 	 facing = AREA_IS_CCW( cc ) ^ ctx->Polygon._FrontBit;
 
-         if (DO_TWOSTENCIL && ctx->Stencil.TestTwoSide) {
-            ctx->_Facing = facing; /* mixed mode rendering: for 2-sided stencil test */
-         }
-
 	 if (DO_UNFILLED) {
 	    if (facing) {
 	       mode = ctx->Polygon.BackMode;
@@ -421,10 +417,6 @@ static void TAG(quad)( GLcontext *ctx,
       {
 	 facing = AREA_IS_CCW( cc ) ^ ctx->Polygon._FrontBit;
 
-         if (DO_TWOSTENCIL && ctx->Stencil.TestTwoSide) {
-            ctx->_Facing = facing; /* mixed mode rendering: for 2-sided stencil test */
-         }
-
 	 if (DO_UNFILLED) {
 	    if (facing) {
 	       mode = ctx->Polygon.BackMode;
diff --git a/src/mesa/vbo/descrip.mms b/src/mesa/vbo/descrip.mms
index 4ab22e4005..c2a16a04f5 100644
--- a/src/mesa/vbo/descrip.mms
+++ b/src/mesa/vbo/descrip.mms
@@ -1,6 +1,6 @@
 # Makefile for core library for VMS
 # contributed by Jouk Jansen  joukj@hrem.nano.tudelft.nl
-# Last revision : 7 March 2007
+# Last revision : 3 October 2007
 
 .first
 	define gl [---.include.gl]
@@ -10,6 +10,8 @@
 	define shader [-.shader]
 	define swrast [-.swrast]
 	define swrast_setup [-.swrast_setup]
+	define main [-.main]
+	define glapi [-.glapi]
 
 .include [---]mms-config.