Merge commit 'origin/master' into gallium-0.2

Conflicts:
	windows/VC8/mesa/osmesa/osmesa.vcproj
	windows/VC8/progs/demos/gears.vcproj
	windows/VC8/progs/progs.sln

37 files changed, 922 insertions, 446 deletions

diff --git a/src/glx/x11/glxcmds.c b/src/glx/x11/glxcmds.c
index 1d373ed29d..c68b6ac4ef 100644
--- a/src/glx/x11/glxcmds.c
+++ b/src/glx/x11/glxcmds.c
@@ -864,6 +864,12 @@ PUBLIC void glXSwapBuffers(Display *dpy, GLXDrawable drawable)
     GLXContext gc;
     GLXContextTag tag;
     CARD8 opcode;
+#ifdef USE_XCB
+    xcb_connection_t *c;
+#else
+    xGLXSwapBuffersReq *req;
+#endif
+
 #ifdef GLX_DIRECT_RENDERING
     __GLXDRIdrawable *pdraw = GetGLXDRIDrawable(dpy, drawable, NULL);
 
@@ -892,12 +898,10 @@ PUBLIC void glXSwapBuffers(Display *dpy, GLXDrawable drawable)
     }
 
 #ifdef USE_XCB
-    xcb_connection_t* c = XGetXCBConnection(dpy);
+    c = XGetXCBConnection(dpy);
     xcb_glx_swap_buffers(c, tag, drawable);
     xcb_flush(c);
 #else
-    xGLXSwapBuffersReq *req;
-
     /* Send the glXSwapBuffers request */
     LockDisplay(dpy);
     GetReq(GLXSwapBuffers,req);
diff --git a/src/glx/x11/glxcurrent.c b/src/glx/x11/glxcurrent.c
index 5af46cf0ba..4d0a7c65eb 100644
--- a/src/glx/x11/glxcurrent.c
+++ b/src/glx/x11/glxcurrent.c
@@ -312,13 +312,11 @@ SendMakeCurrentRequest(Display * dpy, CARD8 opcode,
 
 #ifdef GLX_DIRECT_RENDERING
 static __GLXDRIdrawable *
-FetchDRIDrawable(Display * dpy,
-                 GLXDrawable glxDrawable, GLXContext gc, Bool pre13)
+FetchDRIDrawable(Display * dpy, GLXDrawable glxDrawable, GLXContext gc)
 {
    __GLXdisplayPrivate *const priv = __glXInitialize(dpy);
    __GLXDRIdrawable *pdraw;
    __GLXscreenConfigs *psc;
-   XID drawable;
 
    if (priv == NULL)
       return NULL;
@@ -330,15 +328,7 @@ FetchDRIDrawable(Display * dpy,
    if (__glxHashLookup(psc->drawHash, glxDrawable, (void *) &pdraw) == 0)
       return pdraw;
 
-   /* If this is glXMakeCurrent (pre GLX 1.3) we allow creating the
-    * GLX drawable on the fly.  Otherwise we pass None as the X
-    * drawable */
-   if (pre13)
-      drawable = glxDrawable;
-   else
-      drawable = None;
-
-   pdraw = psc->driScreen->createDrawable(psc, drawable,
+   pdraw = psc->driScreen->createDrawable(psc, glxDrawable,
                                           glxDrawable, gc->mode);
    if (__glxHashInsert(psc->drawHash, glxDrawable, pdraw)) {
       (*pdraw->destroyDrawable) (pdraw);
@@ -357,7 +347,7 @@ FetchDRIDrawable(Display * dpy,
  */
 static Bool
 MakeContextCurrent(Display * dpy, GLXDrawable draw,
-                   GLXDrawable read, GLXContext gc, Bool pre13)
+                   GLXDrawable read, GLXContext gc)
 {
    xGLXMakeCurrentReply reply;
    const GLXContext oldGC = __glXGetCurrentContext();
@@ -384,8 +374,21 @@ MakeContextCurrent(Display * dpy, GLXDrawable draw,
 #ifdef GLX_DIRECT_RENDERING
    /* Bind the direct rendering context to the drawable */
    if (gc && gc->driContext) {
-      __GLXDRIdrawable *pdraw = FetchDRIDrawable(dpy, draw, gc, pre13);
-      __GLXDRIdrawable *pread = FetchDRIDrawable(dpy, read, gc, pre13);
+      __GLXDRIdrawable *pdraw = FetchDRIDrawable(dpy, draw, gc);
+      __GLXDRIdrawable *pread = FetchDRIDrawable(dpy, read, gc);
+
+      if ((pdraw == NULL) || (pread == NULL)) {
+         xError error;
+
+         error.errorCode = GLXBadDrawable;
+         error.resourceID = (pdraw == NULL) ? draw : read;
+         error.sequenceNumber = dpy->request;
+         error.type = X_Error;
+         error.majorCode = gc->majorOpcode;
+         error.minorCode = X_GLXMakeContextCurrent;
+         _XError(dpy, &error);
+         return False;
+      }
 
       bindReturnValue =
          (gc->driContext->bindContext) (gc->driContext, pdraw, pread);
@@ -516,15 +519,16 @@ MakeContextCurrent(Display * dpy, GLXDrawable draw,
 PUBLIC Bool
 glXMakeCurrent(Display * dpy, GLXDrawable draw, GLXContext gc)
 {
-   return MakeContextCurrent(dpy, draw, draw, gc, True);
+   return MakeContextCurrent(dpy, draw, draw, gc);
 }
 
 PUBLIC
 GLX_ALIAS(Bool, glXMakeCurrentReadSGI,
           (Display * dpy, GLXDrawable d, GLXDrawable r, GLXContext ctx),
-          (dpy, d, r, ctx, False), MakeContextCurrent)
+          (dpy, d, r, ctx), MakeContextCurrent)
 
-     PUBLIC GLX_ALIAS(Bool, glXMakeContextCurrent,
-                      (Display * dpy, GLXDrawable d, GLXDrawable r,
-                       GLXContext ctx), (dpy, d, r, ctx, False),
-                      MakeContextCurrent)
+PUBLIC
+GLX_ALIAS(Bool, glXMakeContextCurrent,
+	  (Display * dpy, GLXDrawable d, GLXDrawable r,
+	   GLXContext ctx), (dpy, d, r, ctx),
+	  MakeContextCurrent)
diff --git a/src/mesa/drivers/dri/common/dri_util.h b/src/mesa/drivers/dri/common/dri_util.h
index 0feb57b3c6..c95a5c8299 100644
--- a/src/mesa/drivers/dri/common/dri_util.h
+++ b/src/mesa/drivers/dri/common/dri_util.h
@@ -106,6 +106,28 @@ do {                                                                    \
     }                                                                   \
 } while (0)
 
+/**
+ * Same as above, but for two drawables simultaneously.
+ *
+ */
+
+#define DRI_VALIDATE_TWO_DRAWABLES_INFO(psp, pdp, prp)			\
+do {								\
+    while (*((pdp)->pStamp) != (pdp)->lastStamp ||			\
+	   *((prp)->pStamp) != (prp)->lastStamp) {			\
+        register unsigned int hwContext = (psp)->pSAREA->lock.lock &	\
+	    ~(DRM_LOCK_HELD | DRM_LOCK_CONT);				\
+	DRM_UNLOCK((psp)->fd, &(psp)->pSAREA->lock, hwContext);		\
+									\
+	DRM_SPINLOCK(&(psp)->pSAREA->drawable_lock, (psp)->drawLockID);	\
+	DRI_VALIDATE_DRAWABLE_INFO_ONCE(pdp);                           \
+	DRI_VALIDATE_DRAWABLE_INFO_ONCE(prp);				\
+	DRM_SPINUNLOCK(&(psp)->pSAREA->drawable_lock, (psp)->drawLockID); \
+									\
+	DRM_LIGHT_LOCK((psp)->fd, &(psp)->pSAREA->lock, hwContext);	\
+    }                                                                   \
+} while (0)
+
 
 /**
  * Driver callback functions.
diff --git a/src/mesa/drivers/dri/i915/i915_fragprog.c b/src/mesa/drivers/dri/i915/i915_fragprog.c
index 4760906a7e..f091d600c3 100644
--- a/src/mesa/drivers/dri/i915/i915_fragprog.c
+++ b/src/mesa/drivers/dri/i915/i915_fragprog.c
@@ -33,6 +33,7 @@
 #include "shader/prog_parameter.h"
 #include "shader/program.h"
 #include "shader/programopt.h"
+#include "shader/prog_print.h"
 
 #include "tnl/tnl.h"
 #include "tnl/t_context.h"
@@ -1048,6 +1049,9 @@ i915ProgramStringNotify(GLcontext * ctx,
          _mesa_append_fog_code(ctx, &p->FragProg);
          p->FragProg.FogOption = GL_NONE;
       }
+
+      if (INTEL_DEBUG & DEBUG_STATE)
+	 _mesa_print_program(prog);
    }
 
    _tnl_program_string(ctx, target, prog);
diff --git a/src/mesa/drivers/dri/i915/intel_tris.c b/src/mesa/drivers/dri/i915/intel_tris.c
index c4708dc7ab..e80996580c 100644
--- a/src/mesa/drivers/dri/i915/intel_tris.c
+++ b/src/mesa/drivers/dri/i915/intel_tris.c
@@ -92,7 +92,6 @@ static void intel_start_inline(struct intel_context *intel, uint32_t prim)
    BATCH_LOCALS;
    uint32_t batch_flags = LOOP_CLIPRECTS;
 
-   intel_wait_flips(intel);
    intel->vtbl.emit_state(intel);
 
    intel->no_batch_wrap = GL_TRUE;
@@ -214,8 +213,7 @@ void intel_flush_prim(struct intel_context *intel)
       return;
 
    /* Clear the current prims out of the context state so that a batch flush
-    * flush triggered by wait_flips or emit_state doesn't loop back to
-    * flush_prim again.
+    * flush triggered by emit_state doesn't loop back to flush_prim again.
     */
    vb_bo = intel->prim.vb_bo;
    dri_bo_reference(vb_bo);
@@ -227,8 +225,6 @@ void intel_flush_prim(struct intel_context *intel)
       intel->prim.start_offset = ALIGN(intel->prim.start_offset, 128);
    intel->prim.flush = NULL;
 
-   intel_wait_flips(intel);
-
    intel->vtbl.emit_state(intel);
 
    aper_array[0] = intel->batch->buf;
diff --git a/src/mesa/drivers/dri/i965/brw_sf_state.c b/src/mesa/drivers/dri/i965/brw_sf_state.c
index 741a7c53c4..242b7047a1 100644
--- a/src/mesa/drivers/dri/i965/brw_sf_state.c
+++ b/src/mesa/drivers/dri/i965/brw_sf_state.c
@@ -42,22 +42,15 @@ static void upload_sf_vp(struct brw_context *brw)
    GLcontext *ctx = &brw->intel.ctx;
    const GLfloat depth_scale = 1.0F / ctx->DrawBuffer->_DepthMaxF;
    struct brw_sf_viewport sfv;
-   struct intel_renderbuffer *irb =
-      intel_renderbuffer(ctx->DrawBuffer->_ColorDrawBuffers[0]);
    GLfloat y_scale, y_bias;
 
    memset(&sfv, 0, sizeof(sfv));
 
-   if (ctx->DrawBuffer->Name) {
-      /* User-created FBO */
-      if (irb && !irb->RenderToTexture) {
-	 y_scale = -1.0;
-	 y_bias = ctx->DrawBuffer->Height;
-      } else {
-	 y_scale = 1.0;
-	 y_bias = 0;
-      }
-   } else {
+   if (intel_rendering_to_texture(ctx)) {
+      y_scale = 1.0;
+      y_bias = 0;
+   }
+   else {
       y_scale = -1.0;
       y_bias = ctx->DrawBuffer->Height;
    }
@@ -120,6 +113,7 @@ struct brw_sf_unit_key {
    GLboolean scissor, line_smooth, point_sprite, point_attenuated;
    float line_width;
    float point_size;
+   GLboolean render_to_texture;
 };
 
 static void
@@ -150,6 +144,8 @@ sf_unit_populate_key(struct brw_context *brw, struct brw_sf_unit_key *key)
    key->point_sprite = brw->attribs.Point->PointSprite;
    key->point_size = brw->attribs.Point->Size;
    key->point_attenuated = brw->attribs.Point->_Attenuated;
+
+   key->render_to_texture = intel_rendering_to_texture(&brw->intel.ctx);
 }
 
 static dri_bo *
@@ -196,6 +192,11 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
    else
       sf.sf5.front_winding = BRW_FRONTWINDING_CW;
 
+   /* The viewport is inverted for rendering to texture, and that inverts
+    * polygon front/back orientation.
+    */
+   sf.sf5.front_winding ^= key->render_to_texture;
+
    switch (key->cull_face) {
    case GL_FRONT:
       sf.sf6.cull_mode = BRW_CULLMODE_FRONT;
diff --git a/src/mesa/drivers/dri/i965/brw_vs_emit.c b/src/mesa/drivers/dri/i965/brw_vs_emit.c
index 2084480e89..174331a765 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_emit.c
@@ -1006,7 +1006,7 @@ void brw_vs_emit(struct brw_vs_compile *c )
    GLuint file;
 
    if (INTEL_DEBUG & DEBUG_VS) {
-      _mesa_printf("\n\n\nvs-emit:\n");
+      _mesa_printf("vs-emit:\n");
       _mesa_print_program(&c->vp->program.Base); 
       _mesa_printf("\n");
    }
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 5b4ee20ecb..c50b0d2dd9 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -40,62 +40,20 @@
 GLuint brw_wm_nr_args( GLuint opcode )
 {
    switch (opcode) {
-
    case WM_PIXELXY:
-   case OPCODE_ABS:
-   case OPCODE_FLR:
-   case OPCODE_FRC:
-   case OPCODE_SWZ:
-   case OPCODE_MOV:
-   case OPCODE_COS:
-   case OPCODE_EX2:
-   case OPCODE_LG2:
-   case OPCODE_RCP:
-   case OPCODE_RSQ:
-   case OPCODE_SIN:
-   case OPCODE_SCS:
-   case OPCODE_TEX:
-   case OPCODE_TXB:
-   case OPCODE_TXP:	
-   case OPCODE_KIL:
-   case OPCODE_LIT: 
-   case OPCODE_NRM3:
-   case OPCODE_NRM4:
-   case WM_CINTERP: 
-   case WM_WPOSXY: 
+   case WM_CINTERP:
+   case WM_WPOSXY:
       return 1;
-
-   case OPCODE_POW:
-   case OPCODE_SUB:
-   case OPCODE_SGE:
-   case OPCODE_SGT:
-   case OPCODE_SLE:
-   case OPCODE_SLT:
-   case OPCODE_SEQ:
-   case OPCODE_SNE:
-   case OPCODE_ADD:
-   case OPCODE_MAX:
-   case OPCODE_MIN:
-   case OPCODE_MUL:
-   case OPCODE_XPD:
-   case OPCODE_DP3:	
-   case OPCODE_DP4:
-   case OPCODE_DPH:
-   case OPCODE_DST:
-   case WM_LINTERP: 
+   case WM_LINTERP:
    case WM_DELTAXY:
    case WM_PIXELW:
       return 2;
-
    case WM_FB_WRITE:
-   case WM_PINTERP: 
-   case OPCODE_MAD:	
-   case OPCODE_CMP:
-   case OPCODE_LRP:
+   case WM_PINTERP:
       return 3;
-      
    default:
-      return 0;
+      assert(opcode < MAX_OPCODE);
+      return _mesa_num_inst_src_regs(opcode);
    }
 }
 
@@ -178,6 +136,9 @@ static void do_wm_prog( struct brw_context *brw,
 	*/
        brw_wm_emit(c);
    }
+   if (INTEL_DEBUG & DEBUG_WM)
+      fprintf(stderr, "\n");
+
    /* get the program
     */
    program = brw_get_program(&c->func, &program_size);
diff --git a/src/mesa/drivers/dri/i965/brw_wm_debug.c b/src/mesa/drivers/dri/i965/brw_wm_debug.c
index f31d0974ec..8f07f89ebc 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_debug.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_debug.c
@@ -163,9 +163,9 @@ void brw_wm_print_program( struct brw_wm_compile *c,
 {
    GLuint insn;
 
-   _mesa_printf("\n\n\n%s:\n", stage);
+   _mesa_printf("%s:\n", stage);
    for (insn = 0; insn < c->nr_insns; insn++)
       brw_wm_print_insn(c, &c->instruction[insn]);
-   _mesa_printf("\n\n\n");
+   _mesa_printf("\n");
 }
 
diff --git a/src/mesa/drivers/dri/i965/brw_wm_fp.c b/src/mesa/drivers/dri/i965/brw_wm_fp.c
index 1a00b69825..6df2c95d80 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_fp.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_fp.c
@@ -913,7 +913,7 @@ void brw_wm_pass_fp( struct brw_wm_compile *c )
    GLuint insn;
 
    if (INTEL_DEBUG & DEBUG_WM) {
-      _mesa_printf("\n\n\npre-fp:\n");
+      _mesa_printf("pre-fp:\n");
       _mesa_print_program(&fp->program.Base); 
       _mesa_printf("\n");
    }
@@ -1020,7 +1020,7 @@ void brw_wm_pass_fp( struct brw_wm_compile *c )
    }
 
    if (INTEL_DEBUG & DEBUG_WM) {
-	   _mesa_printf("\n\n\npass_fp:\n");
+	   _mesa_printf("pass_fp:\n");
 	   print_insns( c->prog_instructions, c->nr_fp_insns );
 	   _mesa_printf("\n");
    }
diff --git a/src/mesa/drivers/dri/intel/intel_blit.c b/src/mesa/drivers/dri/intel/intel_blit.c
index 2f1639d4a4..208f90c0ab 100644
--- a/src/mesa/drivers/dri/intel/intel_blit.c
+++ b/src/mesa/drivers/dri/intel/intel_blit.c
@@ -544,7 +544,6 @@ intelClearWithBlit(GLcontext *ctx, GLbitfield mask)
                   _mesa_debug(ctx, "hardware blit clear buf %d rb id %d\n",
                   buf, irb->Base.Name);
                 */
-	       intel_wait_flips(intel);
 
                assert(b.x1 < b.x2);
                assert(b.y1 < b.y2);
diff --git a/src/mesa/drivers/dri/intel/intel_buffers.c b/src/mesa/drivers/dri/intel/intel_buffers.c
index 4d036dee42..0fd2f16a8f 100644
--- a/src/mesa/drivers/dri/intel/intel_buffers.c
+++ b/src/mesa/drivers/dri/intel/intel_buffers.c
@@ -43,22 +43,6 @@
 #include "vblank.h"
 #include "i915_drm.h"
 
-/* This block can be removed when libdrm >= 2.3.1 is required */
-
-#ifndef DRM_IOCTL_I915_FLIP
-
-#define DRM_VBLANK_FLIP 0x8000000
-
-typedef struct drm_i915_flip {
-   int pipes;
-} drm_i915_flip_t;
-
-#undef DRM_IOCTL_I915_FLIP
-#define DRM_IOCTL_I915_FLIP DRM_IOW(DRM_COMMAND_BASE + DRM_I915_FLIP, \
-				    drm_i915_flip_t)
-
-#endif
-
 #define FILE_DEBUG_FLAG DEBUG_BLIT
 
 /**
@@ -143,8 +127,7 @@ intel_get_cliprects(struct intel_context *intel,
       *num_cliprects = 1;
       *x_off = 0;
       *y_off = 0;
-   } else if (intel->front_cliprects ||
-       intel_fb->pf_active || dPriv->numBackClipRects == 0) {
+   } else if (intel->front_cliprects || dPriv->numBackClipRects == 0) {
       /* use the front clip rects */
       *cliprects = dPriv->pClipRects;
       *num_cliprects = dPriv->numClipRects;
@@ -160,76 +143,6 @@ intel_get_cliprects(struct intel_context *intel,
    }
 }
 
-static void
-intelUpdatePageFlipping(struct intel_context *intel,
-			GLint areaA, GLint areaB)
-{
-   __DRIdrawablePrivate *dPriv = intel->driDrawable;
-   struct intel_framebuffer *intel_fb = dPriv->driverPrivate;
-   GLboolean pf_active;
-   GLint pf_planes;
-
-   /* Update page flipping info */
-   pf_planes = 0;
-
-   if (areaA > 0)
-      pf_planes |= 1;
-
-   if (areaB > 0)
-      pf_planes |= 2;
-
-   intel_fb->pf_current_page = (intel->sarea->pf_current_page >>
-				(intel_fb->pf_planes & 0x2)) & 0x3;
-
-   intel_fb->pf_num_pages = 2;
-
-   pf_active = pf_planes && (pf_planes & intel->sarea->pf_active) == pf_planes;
-
-   if (INTEL_DEBUG & DEBUG_LOCK)
-      if (pf_active != intel_fb->pf_active)
-	 _mesa_printf("%s - Page flipping %sactive\n", __progname,
-		      pf_active ? "" : "in");
-
-   if (pf_active) {
-      /* Sync pages between planes if flipping on both at the same time */
-      if (pf_planes == 0x3 && pf_planes != intel_fb->pf_planes &&
-	  (intel->sarea->pf_current_page & 0x3) !=
-	  (((intel->sarea->pf_current_page) >> 2) & 0x3)) {
-	 drm_i915_flip_t flip;
-
-	 if (intel_fb->pf_current_page ==
-	     (intel->sarea->pf_current_page & 0x3)) {
-	    /* XXX: This is ugly, but emitting two flips 'in a row' can cause
-	     * lockups for unknown reasons.
-	     */
-	    intel->sarea->pf_current_page =
-	       intel->sarea->pf_current_page & 0x3;
-	    intel->sarea->pf_current_page |=
-	       ((intel_fb->pf_current_page + intel_fb->pf_num_pages - 1) %
-		intel_fb->pf_num_pages) << 2;
-
-	    flip.pipes = 0x2;
-	 } else {
-	    intel->sarea->pf_current_page =
-	       intel->sarea->pf_current_page & (0x3 << 2);
-	    intel->sarea->pf_current_page |=
-	       (intel_fb->pf_current_page + intel_fb->pf_num_pages - 1) %
-	       intel_fb->pf_num_pages;
-
-	    flip.pipes = 0x1;
-	 }
-
-	 drmCommandWrite(intel->driFd, DRM_I915_FLIP, &flip, sizeof(flip));
-      }
-
-      intel_fb->pf_planes = pf_planes;
-   }
-
-   intel_fb->pf_active = pf_active;
-   intel_flip_renderbuffers(intel_fb);
-   intel_draw_buffer(&intel->ctx, intel->ctx.DrawBuffer);
-}
-
 /**
  * This will be called whenever the currently bound window is moved/resized.
  * XXX: actually, it seems to NOT be called when the window is only moved (BP).
@@ -243,7 +156,7 @@ intelWindowMoved(struct intel_context *intel)
 
    if (!intel->intelScreen->driScrnPriv->dri2.enabled &&
        intel->intelScreen->driScrnPriv->ddx_version.minor >= 7) {
-      volatile struct drm_i915_sarea *sarea = intel->sarea;
+      volatile drm_i915_sarea_t *sarea = intel->sarea;
       drm_clip_rect_t drw_rect = { .x1 = dPriv->x, .x2 = dPriv->x + dPriv->w,
 				   .y1 = dPriv->y, .y2 = dPriv->y + dPriv->h };
       drm_clip_rect_t planeA_rect = { .x1 = sarea->planeA_x, .y1 = sarea->planeA_y,
@@ -256,8 +169,6 @@ intelWindowMoved(struct intel_context *intel)
       GLint areaB = driIntersectArea( drw_rect, planeB_rect );
       GLuint flags = dPriv->vblFlags;
 
-      intelUpdatePageFlipping(intel, areaA, areaB);
-
       /* Update vblank info
        */
       if (areaB > areaA || (areaA == areaB && areaB > 0)) {
@@ -282,7 +193,7 @@ intelWindowMoved(struct intel_context *intel)
 	    vbl.request.type |= DRM_VBLANK_SECONDARY;
 	 }
 
-	 for (i = 0; i < intel_fb->pf_num_pages; i++) {
+	 for (i = 0; i < 2; i++) {
 	    if (!intel_fb->color_rb[i] ||
 		(intel_fb->vbl_waited - intel_fb->color_rb[i]->vbl_pending) <=
 		(1<<23))
@@ -306,7 +217,7 @@ intelWindowMoved(struct intel_context *intel)
 
 	 intel_fb->vbl_waited = dPriv->vblSeq;
 
-	 for (i = 0; i < intel_fb->pf_num_pages; i++) {
+	 for (i = 0; i < 2; i++) {
 	    if (intel_fb->color_rb[i])
 	       intel_fb->color_rb[i]->vbl_pending = intel_fb->vbl_waited;
 	 }
@@ -545,176 +456,6 @@ intelClear(GLcontext *ctx, GLbitfield mask)
    }
 }
 
-
-/* Emit wait for pending flips */
-void
-intel_wait_flips(struct intel_context *intel)
-{
-   struct intel_framebuffer *intel_fb =
-      (struct intel_framebuffer *) intel->ctx.DrawBuffer;
-   struct intel_renderbuffer *intel_rb =
-      intel_get_renderbuffer(&intel_fb->Base,
-			     intel_fb->Base._ColorDrawBufferIndexes[0] ==
-			     BUFFER_FRONT_LEFT ? BUFFER_FRONT_LEFT :
-			     BUFFER_BACK_LEFT);
-
-   if (intel->intelScreen->driScrnPriv->dri2.enabled)
-      return;
-
-   if (intel_fb->Base.Name == 0 && intel_rb &&
-       intel_rb->pf_pending == intel_fb->pf_seq) {
-      GLint pf_planes = intel_fb->pf_planes;
-      BATCH_LOCALS;
-
-      /* Wait for pending flips to take effect */
-      BEGIN_BATCH(2, NO_LOOP_CLIPRECTS);
-      OUT_BATCH(pf_planes & 0x1 ? (MI_WAIT_FOR_EVENT | MI_WAIT_FOR_PLANE_A_FLIP)
-		: 0);
-      OUT_BATCH(pf_planes & 0x2 ? (MI_WAIT_FOR_EVENT | MI_WAIT_FOR_PLANE_B_FLIP)
-		: 0);
-      ADVANCE_BATCH();
-
-      intel_rb->pf_pending--;
-   }
-}
-
-
-/* Flip the front & back buffers
- */
-static GLboolean
-intelPageFlip(const __DRIdrawablePrivate * dPriv)
-{
-   struct intel_context *intel;
-   int ret;
-   struct intel_framebuffer *intel_fb = dPriv->driverPrivate;
-
-   if (INTEL_DEBUG & DEBUG_IOCTL)
-      fprintf(stderr, "%s\n", __FUNCTION__);
-
-   assert(dPriv);
-   assert(dPriv->driContextPriv);
-   assert(dPriv->driContextPriv->driverPrivate);
-
-   intel = (struct intel_context *) dPriv->driContextPriv->driverPrivate;
-
-   if (intel->intelScreen->drmMinor < 9)
-      return GL_FALSE;
-
-   intelFlush(&intel->ctx);
-
-   ret = 0;
-
-   LOCK_HARDWARE(intel);
-
-   if (dPriv->numClipRects && intel_fb->pf_active) {
-      drm_i915_flip_t flip;
-
-      flip.pipes = intel_fb->pf_planes;
-
-      ret = drmCommandWrite(intel->driFd, DRM_I915_FLIP, &flip, sizeof(flip));
-   }
-
-   UNLOCK_HARDWARE(intel);
-
-   if (ret || !intel_fb->pf_active)
-      return GL_FALSE;
-
-   if (!dPriv->numClipRects) {
-      usleep(10000);	/* throttle invisible client 10ms */
-   }
-
-   intel_fb->pf_current_page = (intel->sarea->pf_current_page >>
-				(intel_fb->pf_planes & 0x2)) & 0x3;
-
-   if (dPriv->numClipRects != 0) {
-      intel_get_renderbuffer(&intel_fb->Base, BUFFER_FRONT_LEFT)->pf_pending =
-      intel_get_renderbuffer(&intel_fb->Base, BUFFER_BACK_LEFT)->pf_pending =
-	 ++intel_fb->pf_seq;
-   }
-
-   intel_flip_renderbuffers(intel_fb);
-   intel_draw_buffer(&intel->ctx, &intel_fb->Base);
-
-   return GL_TRUE;
-}
-
-static GLboolean
-intelScheduleSwap(__DRIdrawablePrivate * dPriv, GLboolean *missed_target)
-{
-   struct intel_framebuffer *intel_fb = dPriv->driverPrivate;
-   unsigned int interval;
-   struct intel_context *intel =
-      intelScreenContext(dPriv->driScreenPriv->private);
-   const intelScreenPrivate *intelScreen = intel->intelScreen;
-   unsigned int target;
-   drm_i915_vblank_swap_t swap;
-   GLboolean ret;
-
-   if (!dPriv->vblFlags ||
-       (dPriv->vblFlags & VBLANK_FLAG_NO_IRQ) ||
-       intelScreen->drmMinor < (intel_fb->pf_active ? 9 : 6))
-      return GL_FALSE;
-
-   interval = driGetVBlankInterval(dPriv);
-
-   swap.seqtype = DRM_VBLANK_ABSOLUTE;
-
-   if (dPriv->vblFlags & VBLANK_FLAG_SYNC) {
-      swap.seqtype |= DRM_VBLANK_NEXTONMISS;
-   } else if (interval == 0)
-      return GL_FALSE;
-
-   swap.drawable = dPriv->hHWDrawable;
-   target = swap.sequence = dPriv->vblSeq + interval;
-
-   if ( dPriv->vblFlags & VBLANK_FLAG_SECONDARY ) {
-      swap.seqtype |= DRM_VBLANK_SECONDARY;
-   }
-
-   LOCK_HARDWARE(intel);
-
-   intel_batchbuffer_flush(intel->batch);
-
-   if ( intel_fb->pf_active ) {
-      swap.seqtype |= DRM_VBLANK_FLIP;
-
-      intel_fb->pf_current_page = (((intel->sarea->pf_current_page >>
-				     (intel_fb->pf_planes & 0x2)) & 0x3) + 1) %
-				  intel_fb->pf_num_pages;
-   }
-
-   if (!drmCommandWriteRead(intel->driFd, DRM_I915_VBLANK_SWAP, &swap,
-			    sizeof(swap))) {
-      dPriv->vblSeq = swap.sequence;
-      swap.sequence -= target;
-      *missed_target = swap.sequence > 0 && swap.sequence <= (1 << 23);
-
-      intel_get_renderbuffer(&intel_fb->Base, BUFFER_BACK_LEFT)->vbl_pending =
-	 intel_get_renderbuffer(&intel_fb->Base,
-				BUFFER_FRONT_LEFT)->vbl_pending =
-	 dPriv->vblSeq;
-
-      if (swap.seqtype & DRM_VBLANK_FLIP) {
-	 intel_flip_renderbuffers(intel_fb);
-	 intel_draw_buffer(&intel->ctx, intel->ctx.DrawBuffer);
-      }
-
-      ret = GL_TRUE;
-   } else {
-      if (swap.seqtype & DRM_VBLANK_FLIP) {
-	 intel_fb->pf_current_page = ((intel->sarea->pf_current_page >>
-					(intel_fb->pf_planes & 0x2)) & 0x3) %
-				     intel_fb->pf_num_pages;
-      }
-
-      ret = GL_FALSE;
-   }
-
-   UNLOCK_HARDWARE(intel);
-
-   return ret;
-}
-  
 void
 intelSwapBuffers(__DRIdrawablePrivate * dPriv)
 {
@@ -736,22 +477,22 @@ intelSwapBuffers(__DRIdrawablePrivate * dPriv)
          
 	 _mesa_notifySwapBuffers(ctx);  /* flush pending rendering comands */
 
-         if (!intelScheduleSwap(dPriv, &missed_target)) {
-	    driWaitForVBlank(dPriv, &missed_target);
+	/*
+	 * The old swapping ioctl was incredibly racy, just wait for vblank
+	 * and do the swap ourselves.
+	 */
+	 driWaitForVBlank(dPriv, &missed_target);
 
-	    /*
-	     * Update each buffer's vbl_pending so we don't get too out of
-	     * sync
-	     */
-	    intel_get_renderbuffer(&intel_fb->Base,
-				   BUFFER_BACK_LEFT)->vbl_pending = 
-		    intel_get_renderbuffer(&intel_fb->Base,
-					   BUFFER_FRONT_LEFT)->vbl_pending =
-		    dPriv->vblSeq;
-	    if (!intelPageFlip(dPriv)) {
-	       intelCopyBuffer(dPriv, NULL);
-	    }
-	 }
+	 /*
+	  * Update each buffer's vbl_pending so we don't get too out of
+	  * sync
+	  */
+	 intel_get_renderbuffer(&intel_fb->Base,
+		   		BUFFER_BACK_LEFT)->vbl_pending = dPriv->vblSeq;
+         intel_get_renderbuffer(&intel_fb->Base,
+		   		BUFFER_FRONT_LEFT)->vbl_pending = dPriv->vblSeq;
+
+	 intelCopyBuffer(dPriv, NULL);
 
 	 intel_fb->swap_count++;
 	 (*psp->systemTime->getUST) (&ust);
diff --git a/src/mesa/drivers/dri/intel/intel_buffers.h b/src/mesa/drivers/dri/intel/intel_buffers.h
index e5afb37dd1..0be1cee091 100644
--- a/src/mesa/drivers/dri/intel/intel_buffers.h
+++ b/src/mesa/drivers/dri/intel/intel_buffers.h
@@ -45,8 +45,6 @@ extern struct intel_region *intel_readbuf_region(struct intel_context *intel);
 
 extern struct intel_region *intel_drawbuf_region(struct intel_context *intel);
 
-extern void intel_wait_flips(struct intel_context *intel);
-
 extern void intelSwapBuffers(__DRIdrawablePrivate * dPriv);
 
 extern void intelWindowMoved(struct intel_context *intel);
diff --git a/src/mesa/drivers/dri/intel/intel_context.c b/src/mesa/drivers/dri/intel/intel_context.c
index 7b7f7d8c14..c4a24d7397 100644
--- a/src/mesa/drivers/dri/intel/intel_context.c
+++ b/src/mesa/drivers/dri/intel/intel_context.c
@@ -886,7 +886,7 @@ intelContendedLock(struct intel_context *intel, GLuint flags)
 {
    __DRIdrawablePrivate *dPriv = intel->driDrawable;
    __DRIscreenPrivate *sPriv = intel->driScreen;
-   volatile struct drm_i915_sarea *sarea = intel->sarea;
+   volatile drm_i915_sarea_t *sarea = intel->sarea;
    int me = intel->hHWContext;
 
    drmGetLock(intel->driFd, intel->hHWContext, flags);
diff --git a/src/mesa/drivers/dri/intel/intel_decode.c b/src/mesa/drivers/dri/intel/intel_decode.c
index 0b8a287f6f..5f90ca22ec 100644
--- a/src/mesa/drivers/dri/intel/intel_decode.c
+++ b/src/mesa/drivers/dri/intel/intel_decode.c
@@ -37,6 +37,7 @@
 
 #include <stdio.h>
 #include <stdarg.h>
+#include <string.h>
 #include <inttypes.h>
 
 #include "intel_decode.h"
@@ -324,6 +325,474 @@ decode_3d_1c(uint32_t *data, int count, uint32_t hw_offset, int *failures)
     return 1;
 }
 
+/** Sets the string dstname to describe the destination of the PS instruction */
+static void
+i915_get_instruction_dst(uint32_t *data, int i, char *dstname, int do_mask)
+{
+    uint32_t a0 = data[i];
+    int dst_nr = (a0 >> 14) & 0xf;
+    char dstmask[8];
+    char *sat;
+
+    if (do_mask) {
+	if (((a0 >> 10) & 0xf) == 0xf) {
+	    dstmask[0] = 0;
+	} else {
+	    int dstmask_index = 0;
+
+	    dstmask[dstmask_index++] = '.';
+	    if (a0 & (1 << 10))
+		dstmask[dstmask_index++] = 'x';
+	    if (a0 & (1 << 11))
+		dstmask[dstmask_index++] = 'y';
+	    if (a0 & (1 << 12))
+		dstmask[dstmask_index++] = 'z';
+	    if (a0 & (1 << 13))
+		dstmask[dstmask_index++] = 'w';
+	    dstmask[dstmask_index++] = 0;
+	}
+
+	if (a0 & (1 << 22))
+	    sat = ".sat";
+	else
+	    sat = "";
+    } else {
+	dstmask[0] = 0;
+	sat = "";
+    }
+
+    switch ((a0 >> 19) & 0x7) {
+    case 0:
+	if (dst_nr > 15)
+	    fprintf(out, "bad destination reg R%d\n", dst_nr);
+	sprintf(dstname, "R%d%s%s", dst_nr, dstmask, sat);
+	break;
+    case 4:
+	if (dst_nr > 0)
+	    fprintf(out, "bad destination reg oC%d\n", dst_nr);
+	sprintf(dstname, "oC%s%s", dstmask, sat);
+	break;
+    case 5:
+	if (dst_nr > 0)
+	    fprintf(out, "bad destination reg oD%d\n", dst_nr);
+	sprintf(dstname, "oD%s%s",  dstmask, sat);
+	break;
+    case 6:
+	if (dst_nr > 2)
+	    fprintf(out, "bad destination reg U%d\n", dst_nr);
+	sprintf(dstname, "U%d%s%s", dst_nr, dstmask, sat);
+	break;
+    default:
+	sprintf(dstname, "RESERVED");
+	break;
+    }
+}
+
+static char *
+i915_get_channel_swizzle(uint32_t select)
+{
+    switch (select & 0x7) {
+    case 0:
+	return (select & 8) ? "-x" : "x";
+    case 1:
+	return (select & 8) ? "-y" : "y";
+    case 2:
+	return (select & 8) ? "-z" : "z";
+    case 3:
+	return (select & 8) ? "-w" : "w";
+    case 4:
+	return (select & 8) ? "-0" : "0";
+    case 5:
+	return (select & 8) ? "-1" : "1";
+    default:
+	return (select & 8) ? "-bad" : "bad";
+    }
+}
+
+static void
+i915_get_instruction_src_name(uint32_t src_type, uint32_t src_nr, char *name)
+{
+    switch (src_type) {
+    case 0:
+	sprintf(name, "R%d", src_nr);
+	if (src_nr > 15)
+	    fprintf(out, "bad src reg %s\n", name);
+	break;
+    case 1:
+	if (src_nr < 8)
+	    sprintf(name, "T%d", src_nr);
+	else if (src_nr == 8)
+	    sprintf(name, "DIFFUSE");
+	else if (src_nr == 9)
+	    sprintf(name, "SPECULAR");
+	else if (src_nr == 10)
+	    sprintf(name, "FOG");
+	else {
+	    fprintf(out, "bad src reg T%d\n", src_nr);
+	    sprintf(name, "RESERVED");
+	}
+	break;
+    case 2:
+	sprintf(name, "C%d", src_nr);
+	if (src_nr > 31)
+	    fprintf(out, "bad src reg %s\n", name);
+	break;
+    case 4:
+	sprintf(name, "oC");
+	if (src_nr > 0)
+	    fprintf(out, "bad src reg oC%d\n", src_nr);
+	break;
+    case 5:
+	sprintf(name, "oD");
+	if (src_nr > 0)
+	    fprintf(out, "bad src reg oD%d\n", src_nr);
+	break;
+    case 6:
+	sprintf(name, "U%d", src_nr);
+	if (src_nr > 2)
+	    fprintf(out, "bad src reg %s\n", name);
+	break;
+    default:
+	fprintf(out, "bad src reg type %d\n", src_type);
+	sprintf(name, "RESERVED");
+	break;
+    }
+}
+
+static void
+i915_get_instruction_src0(uint32_t *data, int i, char *srcname)
+{
+    uint32_t a0 = data[i];
+    uint32_t a1 = data[i + 1];
+    int src_nr = (a0 >> 2) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a1 >> 28) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a1 >> 24) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a1 >> 20) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a1 >> 16) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a0 >> 7) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+	strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_src1(uint32_t *data, int i, char *srcname)
+{
+    uint32_t a1 = data[i + 1];
+    uint32_t a2 = data[i + 2];
+    int src_nr = (a1 >> 8) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a1 >> 4) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a1 >> 0) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a2 >> 28) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a2 >> 24) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a1 >> 13) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+	strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_src2(uint32_t *data, int i, char *srcname)
+{
+    uint32_t a2 = data[i + 2];
+    int src_nr = (a2 >> 16) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a2 >> 12) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a2 >> 8) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a2 >> 4) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a2 >> 0) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a2 >> 21) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+	strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_addr(uint32_t src_type, uint32_t src_nr, char *name)
+{
+    switch (src_type) {
+    case 0:
+	sprintf(name, "R%d", src_nr);
+	if (src_nr > 15)
+	    fprintf(out, "bad src reg %s\n", name);
+	break;
+    case 1:
+	if (src_nr < 8)
+	    sprintf(name, "T%d", src_nr);
+	else if (src_nr == 8)
+	    sprintf(name, "DIFFUSE");
+	else if (src_nr == 9)
+	    sprintf(name, "SPECULAR");
+	else if (src_nr == 10)
+	    sprintf(name, "FOG");
+	else {
+	    fprintf(out, "bad src reg T%d\n", src_nr);
+	    sprintf(name, "RESERVED");
+	}
+	break;
+    case 4:
+	sprintf(name, "oC");
+	if (src_nr > 0)
+	    fprintf(out, "bad src reg oC%d\n", src_nr);
+	break;
+    case 5:
+	sprintf(name, "oD");
+	if (src_nr > 0)
+	    fprintf(out, "bad src reg oD%d\n", src_nr);
+	break;
+    default:
+	fprintf(out, "bad src reg type %d\n", src_type);
+	sprintf(name, "RESERVED");
+	break;
+    }
+}
+
+static void
+i915_decode_alu1(uint32_t *data, uint32_t hw_offset,
+		 int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s\n", instr_prefix,
+	      op_name, dst, src0);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_alu2(uint32_t *data, uint32_t hw_offset,
+		 int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100], src1[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+    i915_get_instruction_src1(data, i, src1);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s, %s\n", instr_prefix,
+	      op_name, dst, src0, src1);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_alu3(uint32_t *data, uint32_t hw_offset,
+		 int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100], src1[100], src2[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+    i915_get_instruction_src1(data, i, src1);
+    i915_get_instruction_src2(data, i, src2);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s, %s, %s\n", instr_prefix,
+	      op_name, dst, src0, src1, src2);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_tex(uint32_t *data, uint32_t hw_offset, int i, char *instr_prefix,
+		char *tex_name)
+{
+    uint32_t t0 = data[i];
+    uint32_t t1 = data[i + 1];
+    char dst_name[100];
+    char addr_name[100];
+    int sampler_nr;
+
+    i915_get_instruction_dst(data, i, dst_name, 0);
+    i915_get_instruction_addr((t1 >> 24) & 0x7,
+			      (t1 >> 17) & 0xf,
+			      addr_name);
+    sampler_nr = t0 & 0xf;
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, S%d, %s\n", instr_prefix,
+	      tex_name, dst_name, sampler_nr, addr_name);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_dcl(uint32_t *data, uint32_t hw_offset, int i, char *instr_prefix)
+{
+    uint32_t d0 = data[i];
+    char *sampletype;
+    int dcl_nr = (d0 >> 14) & 0xf;
+    char *dcl_x = d0 & (1 << 10) ? "x" : "";
+    char *dcl_y = d0 & (1 << 11) ? "y" : "";
+    char *dcl_z = d0 & (1 << 12) ? "z" : "";
+    char *dcl_w = d0 & (1 << 13) ? "w" : "";
+    char dcl_mask[10];
+
+    switch ((d0 >> 19) & 0x3) {
+    case 1:
+	sprintf(dcl_mask, ".%s%s%s%s", dcl_x, dcl_y, dcl_z, dcl_w);
+	if (strcmp(dcl_mask, ".") == 0)
+	    fprintf(out, "bad (empty) dcl mask\n");
+
+	if (dcl_nr > 10)
+	    fprintf(out, "bad T%d dcl register number\n", dcl_nr);
+	if (dcl_nr < 8) {
+	    if (strcmp(dcl_mask, ".x") != 0 &&
+		strcmp(dcl_mask, ".xy") != 0 &&
+		strcmp(dcl_mask, ".xz") != 0 &&
+		strcmp(dcl_mask, ".w") != 0 &&
+		strcmp(dcl_mask, ".xyzw") != 0) {
+		fprintf(out, "bad T%d.%s dcl mask\n", dcl_nr, dcl_mask);
+	    }
+	    instr_out(data, hw_offset, i++, "%s: DCL T%d%s\n", instr_prefix,
+		      dcl_nr, dcl_mask);
+	} else {
+	    if (strcmp(dcl_mask, ".xz") == 0)
+		fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+	    else if (strcmp(dcl_mask, ".xw") == 0)
+		fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+	    else if (strcmp(dcl_mask, ".xzw") == 0)
+		fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+
+	    if (dcl_nr == 8) {
+		instr_out(data, hw_offset, i++, "%s: DCL DIFFUSE%s\n", instr_prefix,
+			  dcl_mask);
+	    } else if (dcl_nr == 9) {
+		instr_out(data, hw_offset, i++, "%s: DCL SPECULAR%s\n", instr_prefix,
+			  dcl_mask);
+	    } else if (dcl_nr == 10) {
+		instr_out(data, hw_offset, i++, "%s: DCL FOG%s\n", instr_prefix,
+			  dcl_mask);
+	    }
+	}
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	break;
+    case 3:
+	switch ((d0 >> 22) & 0x3) {
+	case 0:
+	    sampletype = "2D";
+	    break;
+	case 1:
+	    sampletype = "CUBE";
+	    break;
+	case 2:
+	    sampletype = "3D";
+	    break;
+	default:
+	    sampletype = "RESERVED";
+	    break;
+	}
+	if (dcl_nr > 15)
+	    fprintf(out, "bad S%d dcl register number\n", dcl_nr);
+	instr_out(data, hw_offset, i++, "%s: DCL S%d %s\n", instr_prefix,
+		  dcl_nr, sampletype);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	break;
+    default:
+	instr_out(data, hw_offset, i++, "%s: DCL RESERVED%d\n", instr_prefix, dcl_nr);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    }
+}
+
+static void
+i915_decode_instruction(uint32_t *data, uint32_t hw_offset,
+			int i, char *instr_prefix)
+{
+    switch ((data[i] >> 24) & 0x1f) {
+    case 0x0:
+	instr_out(data, hw_offset, i++, "%s: NOP\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	break;
+    case 0x01:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "ADD");
+	break;
+    case 0x02:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "MOV");
+	break;
+    case 0x03:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "MUL");
+	break;
+    case 0x04:
+	i915_decode_alu3(data, hw_offset, i, instr_prefix, "MAD");
+	break;
+    case 0x05:
+	i915_decode_alu3(data, hw_offset, i, instr_prefix, "DP2ADD");
+	break;
+    case 0x06:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "DP3");
+	break;
+    case 0x07:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "DP4");
+	break;
+    case 0x08:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "FRC");
+	break;
+    case 0x09:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "RCP");
+	break;
+    case 0x0a:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "RSQ");
+	break;
+    case 0x0b:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "EXP");
+	break;
+    case 0x0c:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "LOG");
+	break;
+    case 0x0d:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "CMP");
+	break;
+    case 0x0e:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "MIN");
+	break;
+    case 0x0f:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "MAX");
+	break;
+    case 0x10:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "FLR");
+	break;
+    case 0x11:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "MOD");
+	break;
+    case 0x12:
+	i915_decode_alu1(data, hw_offset, i, instr_prefix, "TRC");
+	break;
+    case 0x13:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "SGE");
+	break;
+    case 0x14:
+	i915_decode_alu2(data, hw_offset, i, instr_prefix, "SLT");
+	break;
+    case 0x15:
+	i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLD");
+	break;
+    case 0x16:
+	i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLDP");
+	break;
+    case 0x17:
+	i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLDB");
+	break;
+    case 0x19:
+	i915_decode_dcl(data, hw_offset, i, instr_prefix);
+	break;
+    default:
+	instr_out(data, hw_offset, i++, "%s: unknown\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+	break;
+    }
+}
+
 static int
 decode_3d_1d(uint32_t *data, int count, uint32_t hw_offset, int *failures, int i830)
 {
@@ -441,8 +910,9 @@ decode_3d_1d(uint32_t *data, int count, uint32_t hw_offset, int *failures, int i
     case 0x00:
 	instr_out(data, hw_offset, 0, "3DSTATE_MAP_STATE\n");
 	len = (data[0] & 0x0000003f) + 2;
+	instr_out(data, hw_offset, 1, "mask\n");
 
-	i = 1;
+	i = 2;
 	for (map = 0; map <= 15; map++) {
 	    if (data[1] & (1 << map)) {
 		if (i + 3 >= count)
@@ -495,19 +965,22 @@ decode_3d_1d(uint32_t *data, int count, uint32_t hw_offset, int *failures, int i
 	}
 	i = 1;
 	for (instr = 0; instr < (len - 1) / 3; instr++) {
+	    char instr_prefix[10];
+
 	    if (i + 3 >= count)
-		BUFFER_FAIL(count, len, "3DSTATE_MAP_STATE");
-	    instr_out(data, hw_offset, i++, "PS%03x\n", instr);
-	    instr_out(data, hw_offset, i++, "PS%03x\n", instr);
-	    instr_out(data, hw_offset, i++, "PS%03x\n", instr);
+		BUFFER_FAIL(count, len, "3DSTATE_PIXEL_SHADER_PROGRAM");
+	    sprintf(instr_prefix, "PS%03d", instr);
+	    i915_decode_instruction(data, hw_offset, i, instr_prefix);
+	    i += 3;
 	}
 	return len;
     case 0x01:
 	if (i830)
 	    break;
 	instr_out(data, hw_offset, 0, "3DSTATE_SAMPLER_STATE\n");
+	instr_out(data, hw_offset, 1, "mask\n");
 	len = (data[0] & 0x0000003f) + 2;
-	i = 1;
+	i = 2;
 	for (sampler = 0; sampler <= 15; sampler++) {
 	    if (data[1] & (1 << sampler)) {
 		if (i + 3 >= count)
diff --git a/src/mesa/drivers/dri/intel/intel_fbo.c b/src/mesa/drivers/dri/intel/intel_fbo.c
index 54f2fa5287..7cf12619d6 100644
--- a/src/mesa/drivers/dri/intel/intel_fbo.c
+++ b/src/mesa/drivers/dri/intel/intel_fbo.c
@@ -77,43 +77,6 @@ intel_get_renderbuffer(struct gl_framebuffer *fb, int attIndex)
       return NULL;
 }
 
-
-void
-intel_flip_renderbuffers(struct intel_framebuffer *intel_fb)
-{
-   int current_page = intel_fb->pf_current_page;
-   int next_page = (current_page + 1) % intel_fb->pf_num_pages;
-   struct gl_renderbuffer *tmp_rb;
-
-   /* Exchange renderbuffers if necessary but make sure their reference counts
-    * are preserved.
-    */
-   if (intel_fb->color_rb[current_page] &&
-       intel_fb->Base.Attachment[BUFFER_FRONT_LEFT].Renderbuffer !=
-       &intel_fb->color_rb[current_page]->Base) {
-      tmp_rb = NULL;
-      _mesa_reference_renderbuffer(&tmp_rb,
-	 intel_fb->Base.Attachment[BUFFER_FRONT_LEFT].Renderbuffer);
-      tmp_rb = &intel_fb->color_rb[current_page]->Base;
-      _mesa_reference_renderbuffer(
-	 &intel_fb->Base.Attachment[BUFFER_FRONT_LEFT].Renderbuffer, tmp_rb);
-      _mesa_reference_renderbuffer(&tmp_rb, NULL);
-   }
-
-   if (intel_fb->color_rb[next_page] &&
-       intel_fb->Base.Attachment[BUFFER_BACK_LEFT].Renderbuffer !=
-       &intel_fb->color_rb[next_page]->Base) {
-      tmp_rb = NULL;
-      _mesa_reference_renderbuffer(&tmp_rb,
-	 intel_fb->Base.Attachment[BUFFER_BACK_LEFT].Renderbuffer);
-      tmp_rb = &intel_fb->color_rb[next_page]->Base;
-      _mesa_reference_renderbuffer(
-	 &intel_fb->Base.Attachment[BUFFER_BACK_LEFT].Renderbuffer, tmp_rb);
-      _mesa_reference_renderbuffer(&tmp_rb, NULL);
-   }
-}
-
-
 struct intel_region *
 intel_get_rb_region(struct gl_framebuffer *fb, GLuint attIndex)
 {
@@ -351,7 +314,7 @@ intel_resize_buffers(GLcontext *ctx, struct gl_framebuffer *fb,
    }
 
    /* Make sure all window system renderbuffers are up to date */
-   for (i = 0; i < 3; i++) {
+   for (i = 0; i < 2; i++) {
       struct gl_renderbuffer *rb = &intel_fb->color_rb[i]->Base;
 
       /* only resize if size is changing */
diff --git a/src/mesa/drivers/dri/intel/intel_fbo.h b/src/mesa/drivers/dri/intel/intel_fbo.h
index 9d15582d78..b7e9280e8c 100644
--- a/src/mesa/drivers/dri/intel/intel_fbo.h
+++ b/src/mesa/drivers/dri/intel/intel_fbo.h
@@ -39,14 +39,7 @@ struct intel_framebuffer
 {
    struct gl_framebuffer Base;
 
-   struct intel_renderbuffer *color_rb[3];
-
-   /* Drawable page flipping state */
-   GLboolean pf_active;
-   GLuint pf_seq;
-   GLint pf_planes;
-   GLint pf_current_page;
-   GLint pf_num_pages;
+   struct intel_renderbuffer *color_rb[2];
 
    /* VBI
     */
@@ -76,8 +69,6 @@ struct intel_renderbuffer
    GLuint PairedDepth;   /**< only used if this is a depth renderbuffer */
    GLuint PairedStencil; /**< only used if this is a stencil renderbuffer */
 
-   GLuint pf_pending;  /**< sequence number of pending flip */
-
    GLuint vbl_pending;   /**< vblank sequence number of pending flip */
 
    uint8_t *span_cache;
@@ -111,5 +102,22 @@ extern struct intel_region *intel_get_rb_region(struct gl_framebuffer *fb,
 
 
 
+/**
+ * Are we currently rendering into a texture?
+ */
+static INLINE GLboolean
+intel_rendering_to_texture(const GLcontext *ctx)
+{
+   if (ctx->DrawBuffer->Name) {
+      /* User-created FBO */
+      const struct intel_renderbuffer *irb =
+         intel_renderbuffer(ctx->DrawBuffer->_ColorDrawBuffers[0]);
+      return irb && irb->RenderToTexture;
+   }
+   else {
+      return GL_FALSE;
+   }
+}
+
 
 #endif /* INTEL_FBO_H */
diff --git a/src/mesa/drivers/dri/intel/intel_screen.c b/src/mesa/drivers/dri/intel/intel_screen.c
index fc4e82b56c..7042c25564 100644
--- a/src/mesa/drivers/dri/intel/intel_screen.c
+++ b/src/mesa/drivers/dri/intel/intel_screen.c
@@ -134,7 +134,7 @@ intelPrintDRIInfo(intelScreenPrivate * intelScreen,
 
 
 static void
-intelPrintSAREA(const struct drm_i915_sarea * sarea)
+intelPrintSAREA(const drm_i915_sarea_t * sarea)
 {
    fprintf(stderr, "SAREA: sarea width %d  height %d\n", sarea->width,
            sarea->height);
@@ -161,7 +161,7 @@ intelPrintSAREA(const struct drm_i915_sarea * sarea)
  */
 void
 intelUpdateScreenFromSAREA(intelScreenPrivate * intelScreen,
-                           struct drm_i915_sarea * sarea)
+                           drm_i915_sarea_t * sarea)
 {
    intelScreen->width = sarea->width;
    intelScreen->height = sarea->height;
@@ -244,7 +244,7 @@ static GLboolean intelInitDriver(__DRIscreenPrivate *sPriv)
 {
    intelScreenPrivate *intelScreen;
    I830DRIPtr gDRIPriv = (I830DRIPtr) sPriv->pDevPriv;
-   struct drm_i915_sarea *sarea;
+   drm_i915_sarea_t *sarea;
 
    if (sPriv->devPrivSize != sizeof(I830DRIRec)) {
       fprintf(stderr,
@@ -264,7 +264,7 @@ static GLboolean intelInitDriver(__DRIscreenPrivate *sPriv)
 
    intelScreen->driScrnPriv = sPriv;
    sPriv->private = (void *) intelScreen;
-   sarea = (struct drm_i915_sarea *)
+   sarea = (drm_i915_sarea_t *)
       (((GLubyte *) sPriv->pSAREA) + gDRIPriv->sarea_priv_offset);
    intelScreen->sarea = sarea;
 
diff --git a/src/mesa/drivers/dri/intel/intel_screen.h b/src/mesa/drivers/dri/intel/intel_screen.h
index cf5359baae..fcd0d9c28c 100644
--- a/src/mesa/drivers/dri/intel/intel_screen.h
+++ b/src/mesa/drivers/dri/intel/intel_screen.h
@@ -68,7 +68,7 @@ typedef struct
 
    __DRIscreenPrivate *driScrnPriv;
 
-   volatile struct drm_i915_sarea *sarea;
+   volatile drm_i915_sarea_t *sarea;
 
    int drmMinor;
 
@@ -94,7 +94,7 @@ extern void intelUnmapScreenRegions(intelScreenPrivate * intelScreen);
 
 extern void
 intelUpdateScreenFromSAREA(intelScreenPrivate * intelScreen,
-                           struct drm_i915_sarea * sarea);
+                           drm_i915_sarea_t * sarea);
 
 extern void intelDestroyContext(__DRIcontextPrivate * driContextPriv);
 
diff --git a/src/mesa/drivers/glslcompiler/Makefile b/src/mesa/drivers/glslcompiler/Makefile
index dc4abd44d4..ffe4e0a234 100644
--- a/src/mesa/drivers/glslcompiler/Makefile
+++ b/src/mesa/drivers/glslcompiler/Makefile
@@ -33,7 +33,7 @@ default: $(PROGRAM)
 
 
 glslcompiler: $(OBJECTS)
-	$(CC) $(OBJECTS) -lm -lpthread -o $@
+	$(CC) $(OBJECTS) $(GL_LIB_DEPS) -o $@
 
 
 glslcompiler.o: glslcompiler.c
diff --git a/src/mesa/main/colormac.h b/src/mesa/main/colormac.h
index a34bd2ed38..74692e9a98 100644
--- a/src/mesa/main/colormac.h
+++ b/src/mesa/main/colormac.h
@@ -195,6 +195,10 @@ do {						\
 #define PACK_COLOR_565_REV( X, Y, Z ) \
    (((X) & 0xf8) | ((Y) & 0xe0) >> 5 | (((Y) & 0x1c) << 11) | (((Z) & 0xf8) << 5))
 
+#define PACK_COLOR_5551( R, G, B, A )					\
+   ((((R) & 0xf8) << 8) | (((G) & 0xf8) << 3) | (((B) & 0xf8) >> 2) |	\
+    ((A) ? 1 : 0))
+
 #define PACK_COLOR_1555( A, B, G, R )					\
    ((((B) & 0xf8) << 7) | (((G) & 0xf8) << 2) | (((R) & 0xf8) >> 3) |	\
     ((A) ? 0x8000 : 0))
diff --git a/src/mesa/main/enable.c b/src/mesa/main/enable.c
index 72ed50808c..2ef4b1b207 100644
--- a/src/mesa/main/enable.c
+++ b/src/mesa/main/enable.c
@@ -341,10 +341,6 @@ _mesa_set_enable(GLcontext *ctx, GLenum cap, GLboolean state)
          ctx->Transform.CullVertexFlag = state;
          break;
       case GL_DEPTH_TEST:
-         if (state && ctx->DrawBuffer->Visual.depthBits == 0) {
-            _mesa_warning(ctx,"glEnable(GL_DEPTH_TEST) but no depth buffer");
-            return;
-         }
          if (ctx->Depth.Test == state)
             return;
          FLUSH_VERTICES(ctx, _NEW_DEPTH);
diff --git a/src/mesa/main/texformat.c b/src/mesa/main/texformat.c
index 50e8d7a3b8..f49c916b33 100644
--- a/src/mesa/main/texformat.c
+++ b/src/mesa/main/texformat.c
@@ -896,6 +896,30 @@ const struct gl_texture_format _mesa_texformat_rgb565_rev = {
    store_texel_rgb565_rev		/* StoreTexel */
 };
 
+const struct gl_texture_format _mesa_texformat_rgba4444 = {
+   MESA_FORMAT_RGBA4444,		/* MesaFormat */
+   GL_RGBA,				/* BaseFormat */
+   GL_UNSIGNED_NORMALIZED_ARB,		/* DataType */
+   4,					/* RedBits */
+   4,					/* GreenBits */
+   4,					/* BlueBits */
+   4,					/* AlphaBits */
+   0,					/* LuminanceBits */
+   0,					/* IntensityBits */
+   0,					/* IndexBits */
+   0,					/* DepthBits */
+   0,					/* StencilBits */
+   2,					/* TexelBytes */
+   _mesa_texstore_rgba4444,		/* StoreTexImageFunc */
+   fetch_texel_1d_rgba4444,		/* FetchTexel1D */
+   fetch_texel_2d_rgba4444,		/* FetchTexel2D */
+   fetch_texel_3d_rgba4444,		/* FetchTexel3D */
+   NULL,				/* FetchTexel1Df */
+   NULL,				/* FetchTexel2Df */
+   NULL,				/* FetchTexel3Df */
+   store_texel_rgba4444			/* StoreTexel */
+};
+
 const struct gl_texture_format _mesa_texformat_argb4444 = {
    MESA_FORMAT_ARGB4444,		/* MesaFormat */
    GL_RGBA,				/* BaseFormat */
@@ -944,6 +968,30 @@ const struct gl_texture_format _mesa_texformat_argb4444_rev = {
    store_texel_argb4444_rev		/* StoreTexel */
 };
 
+const struct gl_texture_format _mesa_texformat_rgba5551 = {
+   MESA_FORMAT_RGBA5551,		/* MesaFormat */
+   GL_RGBA,				/* BaseFormat */
+   GL_UNSIGNED_NORMALIZED_ARB,		/* DataType */
+   5,					/* RedBits */
+   5,					/* GreenBits */
+   5,					/* BlueBits */
+   1,					/* AlphaBits */
+   0,					/* LuminanceBits */
+   0,					/* IntensityBits */
+   0,					/* IndexBits */
+   0,					/* DepthBits */
+   0,					/* StencilBits */
+   2,					/* TexelBytes */
+   _mesa_texstore_rgba5551,		/* StoreTexImageFunc */
+   fetch_texel_1d_rgba5551,		/* FetchTexel1D */
+   fetch_texel_2d_rgba5551,		/* FetchTexel2D */
+   fetch_texel_3d_rgba5551,		/* FetchTexel3D */
+   NULL,				/* FetchTexel1Df */
+   NULL,				/* FetchTexel2Df */
+   NULL,				/* FetchTexel3Df */
+   store_texel_rgba5551			/* StoreTexel */
+};
+
 const struct gl_texture_format _mesa_texformat_argb1555 = {
    MESA_FORMAT_ARGB1555,		/* MesaFormat */
    GL_RGBA,				/* BaseFormat */
diff --git a/src/mesa/main/texformat.h b/src/mesa/main/texformat.h
index c7a754b0b7..31364c36b1 100644
--- a/src/mesa/main/texformat.h
+++ b/src/mesa/main/texformat.h
@@ -71,8 +71,10 @@ enum _format {
    MESA_FORMAT_BGR888,		/*           BBBB BBBB GGGG GGGG RRRR RRRR */
    MESA_FORMAT_RGB565,		/*                     RRRR RGGG GGGB BBBB */
    MESA_FORMAT_RGB565_REV,	/*                     GGGB BBBB RRRR RGGG */
+   MESA_FORMAT_RGBA4444,        /*                     RRRR GGGG BBBB AAAA */
    MESA_FORMAT_ARGB4444,	/*                     AAAA RRRR GGGG BBBB */
    MESA_FORMAT_ARGB4444_REV,	/*                     GGGG BBBB AAAA RRRR */
+   MESA_FORMAT_RGBA5551,        /*                     RRRR RGGG GGBB BBBA */
    MESA_FORMAT_ARGB1555,	/*                     ARRR RRGG GGGB BBBB */
    MESA_FORMAT_ARGB1555_REV,	/*                     GGGB BBBB ARRR RRGG */
    MESA_FORMAT_AL88,		/*                     AAAA AAAA LLLL LLLL */
@@ -217,10 +219,12 @@ extern const struct gl_texture_format _mesa_texformat_rgb888;
 extern const struct gl_texture_format _mesa_texformat_bgr888;
 extern const struct gl_texture_format _mesa_texformat_rgb565;
 extern const struct gl_texture_format _mesa_texformat_rgb565_rev;
+extern const struct gl_texture_format _mesa_texformat_rgba4444;
 extern const struct gl_texture_format _mesa_texformat_argb4444;
 extern const struct gl_texture_format _mesa_texformat_argb4444_rev;
 extern const struct gl_texture_format _mesa_texformat_argb1555;
 extern const struct gl_texture_format _mesa_texformat_argb1555_rev;
+extern const struct gl_texture_format _mesa_texformat_rgba5551;
 extern const struct gl_texture_format _mesa_texformat_al88;
 extern const struct gl_texture_format _mesa_texformat_al88_rev;
 extern const struct gl_texture_format _mesa_texformat_rgb332;
diff --git a/src/mesa/main/texformat_tmp.h b/src/mesa/main/texformat_tmp.h
index b1031b0cfe..08b6d8f12f 100644
--- a/src/mesa/main/texformat_tmp.h
+++ b/src/mesa/main/texformat_tmp.h
@@ -695,7 +695,7 @@ static void store_texel_argb8888_rev(struct gl_texture_image *texImage,
 {
    const GLubyte *rgba = (const GLubyte *) texel;
    GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
-   *dst = PACK_COLOR_8888(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+   *dst = PACK_COLOR_8888(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP], rgba[ACOMP]);
 }
 #endif
 
@@ -804,6 +804,30 @@ static void store_texel_rgb565_rev(struct gl_texture_image *texImage,
 }
 #endif
 
+/* MESA_FORMAT_RGBA4444 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb444 texture, return 4 GLchans */
+static void FETCH(rgba4444)( const struct gl_texture_image *texImage,
+			     GLint i, GLint j, GLint k, GLchan *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = UBYTE_TO_CHAN( ((s >> 12) & 0xf) | ((s >> 8) & 0xf0) );
+   texel[GCOMP] = UBYTE_TO_CHAN( ((s >>  8) & 0xf) | ((s >> 4) & 0xf0) );
+   texel[BCOMP] = UBYTE_TO_CHAN( ((s >>  4) & 0xf) | ((s     ) & 0xf0) );
+   texel[ACOMP] = UBYTE_TO_CHAN( ((s      ) & 0xf) | ((s << 4) & 0xf0) );
+}
+
+#if DIM == 3
+static void store_texel_rgba4444(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_4444(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
 
 /* MESA_FORMAT_ARGB4444 ******************************************************/
 
@@ -825,7 +849,7 @@ static void store_texel_argb4444(struct gl_texture_image *texImage,
 {
    const GLubyte *rgba = (const GLubyte *) texel;
    GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
-   *dst = PACK_COLOR_4444(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+   *dst = PACK_COLOR_4444(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
 }
 #endif
 
@@ -853,6 +877,29 @@ static void store_texel_argb4444_rev(struct gl_texture_image *texImage,
 }
 #endif
 
+/* MESA_FORMAT_RGBA5551 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb1555 texture, return 4 GLchans */
+static void FETCH(rgba5551)( const struct gl_texture_image *texImage,
+			     GLint i, GLint j, GLint k, GLchan *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = UBYTE_TO_CHAN( ((s >>  8) & 0xf8) | ((s >> 13) & 0x7) );
+   texel[GCOMP] = UBYTE_TO_CHAN( ((s >>  3) & 0xf8) | ((s >>  8) & 0x7) );
+   texel[BCOMP] = UBYTE_TO_CHAN( ((s <<  2) & 0xf8) | ((s >>  3) & 0x7) );
+   texel[ACOMP] = UBYTE_TO_CHAN( ((s) & 0x01) ? 255 : 0);
+}
+
+#if DIM == 3
+static void store_texel_rgba5551(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_5551(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
 
 /* MESA_FORMAT_ARGB1555 ******************************************************/
 
diff --git a/src/mesa/main/texobj.c b/src/mesa/main/texobj.c
index d8e8b559f5..7848f0be35 100644
--- a/src/mesa/main/texobj.c
+++ b/src/mesa/main/texobj.c
@@ -762,24 +762,31 @@ unbind_texobj_from_texunits(GLcontext *ctx, struct gl_texture_object *texObj)
       struct gl_texture_unit *unit = &ctx->Texture.Unit[u];
       if (texObj == unit->Current1D) {
          _mesa_reference_texobj(&unit->Current1D, ctx->Shared->Default1D);
+         ASSERT(unit->Current1D);
       }
       else if (texObj == unit->Current2D) {
          _mesa_reference_texobj(&unit->Current2D, ctx->Shared->Default2D);
+         ASSERT(unit->Current2D);
       }
       else if (texObj == unit->Current3D) {
          _mesa_reference_texobj(&unit->Current3D, ctx->Shared->Default3D);
+         ASSERT(unit->Current3D);
       }
       else if (texObj == unit->CurrentCubeMap) {
          _mesa_reference_texobj(&unit->CurrentCubeMap, ctx->Shared->DefaultCubeMap);
+         ASSERT(unit->CurrentCubeMap);
       }
       else if (texObj == unit->CurrentRect) {
          _mesa_reference_texobj(&unit->CurrentRect, ctx->Shared->DefaultRect);
+         ASSERT(unit->CurrentRect);
       }
       else if (texObj == unit->Current1DArray) {
          _mesa_reference_texobj(&unit->Current1DArray, ctx->Shared->Default1DArray);
+         ASSERT(unit->Current1DArray);
       }
       else if (texObj == unit->Current2DArray) {
          _mesa_reference_texobj(&unit->Current2DArray, ctx->Shared->Default2DArray);
+         ASSERT(unit->Current2DArray);
       }
    }
 }
@@ -953,24 +960,31 @@ _mesa_BindTexture( GLenum target, GLuint texName )
    switch (target) {
       case GL_TEXTURE_1D:
          _mesa_reference_texobj(&texUnit->Current1D, newTexObj);
+         ASSERT(texUnit->Current1D);
          break;
       case GL_TEXTURE_2D:
          _mesa_reference_texobj(&texUnit->Current2D, newTexObj);
+         ASSERT(texUnit->Current2D);
          break;
       case GL_TEXTURE_3D:
          _mesa_reference_texobj(&texUnit->Current3D, newTexObj);
+         ASSERT(texUnit->Current3D);
          break;
       case GL_TEXTURE_CUBE_MAP_ARB:
          _mesa_reference_texobj(&texUnit->CurrentCubeMap, newTexObj);
+         ASSERT(texUnit->CurrentCubeMap);
          break;
       case GL_TEXTURE_RECTANGLE_NV:
          _mesa_reference_texobj(&texUnit->CurrentRect, newTexObj);
+         ASSERT(texUnit->CurrentRect);
          break;
       case GL_TEXTURE_1D_ARRAY_EXT:
          texUnit->Current1DArray = newTexObj;
+         ASSERT(texUnit->Current1DArray);
          break;
       case GL_TEXTURE_2D_ARRAY_EXT:
          texUnit->Current2DArray = newTexObj;
+         ASSERT(texUnit->Current2DArray);
          break;
       default:
          /* Bad target should be caught above */
diff --git a/src/mesa/main/texstate.c b/src/mesa/main/texstate.c
index 29955d76cb..9bfb7e0ec2 100644
--- a/src/mesa/main/texstate.c
+++ b/src/mesa/main/texstate.c
@@ -517,6 +517,14 @@ update_texture_state( GLcontext *ctx )
          enableBits = texUnit->Enabled;
       }
 
+      ASSERT(texUnit->Current1D);
+      ASSERT(texUnit->Current2D);
+      ASSERT(texUnit->Current3D);
+      ASSERT(texUnit->CurrentCubeMap);
+      ASSERT(texUnit->CurrentRect);
+      ASSERT(texUnit->Current1DArray);
+      ASSERT(texUnit->Current2DArray);
+
       /* Look for the highest-priority texture target that's enabled and
        * complete.  That's the one we'll use for texturing.  If we're using
        * a fragment program we're guaranteed that bitcount(enabledBits) <= 1.
diff --git a/src/mesa/main/texstore.c b/src/mesa/main/texstore.c
index 8afb947fa1..2b06796aba 100644
--- a/src/mesa/main/texstore.c
+++ b/src/mesa/main/texstore.c
@@ -1899,6 +1899,60 @@ _mesa_texstore_bgr888(TEXSTORE_PARAMS)
    return GL_TRUE;
 }
 
+GLboolean
+_mesa_texstore_rgba4444(TEXSTORE_PARAMS)
+{
+   ASSERT(dstFormat == &_mesa_texformat_rgba4444);
+   ASSERT(dstFormat->TexelBytes == 2);
+
+   if (!ctx->_ImageTransferState &&
+       !srcPacking->SwapBytes &&
+       dstFormat == &_mesa_texformat_rgba4444 &&
+       baseInternalFormat == GL_RGBA &&
+       srcFormat == GL_RGBA &&
+       srcType == GL_UNSIGNED_SHORT_4_4_4_4){
+      /* simple memcpy path */
+      memcpy_texture(ctx, dims,
+                     dstFormat, dstAddr, dstXoffset, dstYoffset, dstZoffset,
+                     dstRowStride,
+                     dstImageOffsets,
+                     srcWidth, srcHeight, srcDepth, srcFormat, srcType,
+                     srcAddr, srcPacking);
+   }
+   else {
+      /* general path */
+      const GLchan *tempImage = _mesa_make_temp_chan_image(ctx, dims,
+                                                 baseInternalFormat,
+                                                 dstFormat->BaseFormat,
+                                                 srcWidth, srcHeight, srcDepth,
+                                                 srcFormat, srcType, srcAddr,
+                                                 srcPacking);
+      const GLchan *src = tempImage;
+      GLint img, row, col;
+      if (!tempImage)
+         return GL_FALSE;
+      _mesa_adjust_image_for_convolution(ctx, dims, &srcWidth, &srcHeight);
+      for (img = 0; img < srcDepth; img++) {
+         GLubyte *dstRow = (GLubyte *) dstAddr
+            + dstImageOffsets[dstZoffset + img] * dstFormat->TexelBytes
+            + dstYoffset * dstRowStride
+            + dstXoffset * dstFormat->TexelBytes;
+         for (row = 0; row < srcHeight; row++) {
+            GLushort *dstUS = (GLushort *) dstRow;
+	    for (col = 0; col < srcWidth; col++) {
+	      dstUS[col] = PACK_COLOR_4444( CHAN_TO_UBYTE(src[RCOMP]),
+					    CHAN_TO_UBYTE(src[GCOMP]),
+					    CHAN_TO_UBYTE(src[BCOMP]),
+					    CHAN_TO_UBYTE(src[ACOMP]) );
+	      src += 4;
+            }
+            dstRow += dstRowStride;
+         }
+      }
+      _mesa_free((void *) tempImage);
+   }
+   return GL_TRUE;
+}
 
 GLboolean
 _mesa_texstore_argb4444(TEXSTORE_PARAMS)
@@ -1967,7 +2021,60 @@ _mesa_texstore_argb4444(TEXSTORE_PARAMS)
    return GL_TRUE;
 }
 
+GLboolean
+_mesa_texstore_rgba5551(TEXSTORE_PARAMS)
+{
+   ASSERT(dstFormat == &_mesa_texformat_rgba5551);
+   ASSERT(dstFormat->TexelBytes == 2);
 
+   if (!ctx->_ImageTransferState &&
+       !srcPacking->SwapBytes &&
+       dstFormat == &_mesa_texformat_rgba5551 &&
+       baseInternalFormat == GL_RGBA &&
+       srcFormat == GL_RGBA &&
+       srcType == GL_UNSIGNED_SHORT_5_5_5_1) {
+      /* simple memcpy path */
+      memcpy_texture(ctx, dims,
+                     dstFormat, dstAddr, dstXoffset, dstYoffset, dstZoffset,
+                     dstRowStride,
+                     dstImageOffsets,
+                     srcWidth, srcHeight, srcDepth, srcFormat, srcType,
+                     srcAddr, srcPacking);
+   }
+   else {
+      /* general path */
+      const GLchan *tempImage = _mesa_make_temp_chan_image(ctx, dims,
+                                                 baseInternalFormat,
+                                                 dstFormat->BaseFormat,
+                                                 srcWidth, srcHeight, srcDepth,
+                                                 srcFormat, srcType, srcAddr,
+                                                 srcPacking);
+      const GLchan *src =tempImage;
+      GLint img, row, col;
+      if (!tempImage)
+         return GL_FALSE;
+      _mesa_adjust_image_for_convolution(ctx, dims, &srcWidth, &srcHeight);
+      for (img = 0; img < srcDepth; img++) {
+         GLubyte *dstRow = (GLubyte *) dstAddr
+            + dstImageOffsets[dstZoffset + img] * dstFormat->TexelBytes
+            + dstYoffset * dstRowStride
+            + dstXoffset * dstFormat->TexelBytes;
+         for (row = 0; row < srcHeight; row++) {
+            GLushort *dstUS = (GLushort *) dstRow;
+	    for (col = 0; col < srcWidth; col++) {
+	       dstUS[col] = PACK_COLOR_5551( CHAN_TO_UBYTE(src[RCOMP]),
+					     CHAN_TO_UBYTE(src[GCOMP]),
+					     CHAN_TO_UBYTE(src[BCOMP]),
+					     CHAN_TO_UBYTE(src[ACOMP]) );
+	      src += 4;
+	    }
+            dstRow += dstRowStride;
+         }
+      }
+      _mesa_free((void *) tempImage);
+   }
+   return GL_TRUE;
+}
 
 GLboolean
 _mesa_texstore_argb1555(TEXSTORE_PARAMS)
diff --git a/src/mesa/main/texstore.h b/src/mesa/main/texstore.h
index 8dc1c963cd..b03386b2ac 100644
--- a/src/mesa/main/texstore.h
+++ b/src/mesa/main/texstore.h
@@ -47,8 +47,10 @@ extern GLboolean _mesa_texstore_rgb888(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_bgr888(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_rgb565(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_rgb565_rev(TEXSTORE_PARAMS);
+extern GLboolean _mesa_texstore_rgba4444(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_argb4444(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_argb4444_rev(TEXSTORE_PARAMS);
+extern GLboolean _mesa_texstore_rgba5551(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_argb1555(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_argb1555_rev(TEXSTORE_PARAMS);
 extern GLboolean _mesa_texstore_al88(TEXSTORE_PARAMS);
diff --git a/src/mesa/main/version.h b/src/mesa/main/version.h
index 24495d608a..3d874c8ba8 100644
--- a/src/mesa/main/version.h
+++ b/src/mesa/main/version.h
@@ -31,7 +31,7 @@
 #define MESA_MAJOR 7
 #define MESA_MINOR 3
 #define MESA_PATCH 0
-#define MESA_VERSION_STRING "7.3-rc2"
+#define MESA_VERSION_STRING "7.3-rc3"
 
 /* To make version comparison easy */
 #define MESA_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c))
diff --git a/src/mesa/shader/arbprogram.c b/src/mesa/shader/arbprogram.c
index 760dac2399..7c2b747c43 100644
--- a/src/mesa/shader/arbprogram.c
+++ b/src/mesa/shader/arbprogram.c
@@ -807,6 +807,7 @@ _mesa_GetProgramLocalParameterdvARB(GLenum target, GLuint index,
 {
    GET_CURRENT_CONTEXT(ctx);
    GLfloat floatParams[4];
+   ASSIGN_4V(floatParams, 0.0F, 0.0F, 0.0F, 0.0F);
    _mesa_GetProgramLocalParameterfvARB(target, index, floatParams);
    if (ctx->ErrorValue == GL_NO_ERROR) {
       COPY_4V(params, floatParams);
diff --git a/src/mesa/shader/slang/slang_mem.c b/src/mesa/shader/slang/slang_mem.c
index 21d6bfce88..9224578edb 100644
--- a/src/mesa/shader/slang/slang_mem.c
+++ b/src/mesa/shader/slang/slang_mem.c
@@ -184,6 +184,7 @@ _slang_realloc(void *oldBuffer, GLuint oldSize, GLuint newSize)
 #else
    GET_CURRENT_CONTEXT(ctx);
    slang_mempool *pool = (slang_mempool *) ctx->Shader.MemPool;
+   (void) pool;
 
    if (newSize < oldSize) {
       return oldBuffer;
@@ -235,6 +236,7 @@ _slang_free(void *addr)
    if (addr) {
       GET_CURRENT_CONTEXT(ctx);
       slang_mempool *pool = (slang_mempool *) ctx->Shader.MemPool;
+      (void) pool;
       ASSERT(is_valid_address(pool, addr));
    }
 #endif
diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c
index 214c2a1b6f..7f9cc64bb2 100644
--- a/src/mesa/swrast/s_span.c
+++ b/src/mesa/swrast/s_span.c
@@ -1032,6 +1032,7 @@ add_specular(GLcontext *ctx, SWspan *span)
    ASSERT(!ctx->FragmentProgram._Current);
    ASSERT(span->arrayMask & SPAN_RGBA);
    ASSERT(swrast->_ActiveAttribMask & FRAG_BIT_COL1);
+   (void) swrast; /* silence warning */
 
    if (span->array->ChanType == GL_FLOAT) {
       if ((span->arrayAttribs & FRAG_BIT_COL0) == 0) {
diff --git a/src/mesa/swrast/s_texfilter.c b/src/mesa/swrast/s_texfilter.c
index 6b1dfd5d50..a095b255ab 100644
--- a/src/mesa/swrast/s_texfilter.c
+++ b/src/mesa/swrast/s_texfilter.c
@@ -2064,6 +2064,7 @@ clamp_rect_coord_linear(GLenum wrapMode, GLfloat coord, GLint max,
    default:
       _mesa_problem(NULL, "bad wrapMode in clamp_rect_coord_linear");
       i0 = i1 = 0;
+      fcol = 0.0F;
    }
    *i0out = i0;
    *i1out = i1;
@@ -2872,6 +2873,8 @@ sample_depth_texture( GLcontext *ctx,
                                          texcoords[i][1]);
             slice = clamp_rect_coord_nearest(tObj->WrapR, texcoords[i][2], depth);
             break;
+         default:
+            col = row = slice = 0;
          }
 
          if (col >= 0 && row >= 0 && col < width && row < height && 
@@ -2982,6 +2985,8 @@ sample_depth_texture( GLcontext *ctx,
                                    texcoords[i][1], &j0, &j1, &b);
             slice = clamp_rect_coord_nearest(tObj->WrapR, texcoords[i][2], depth);
             break;
+         default:
+            slice = 0;
          }
 
          useBorderTexel = 0;
diff --git a/src/mesa/swrast/s_triangle.c b/src/mesa/swrast/s_triangle.c
index 2033ab5529..a2e8433e27 100644
--- a/src/mesa/swrast/s_triangle.c
+++ b/src/mesa/swrast/s_triangle.c
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.3
+ * Version:  7.3
  *
  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
  *
@@ -263,6 +263,10 @@ affine_span(GLcontext *ctx, SWspan *span,
             struct affine_info *info)
 {
    GLchan sample[4];  /* the filtered texture sample */
+   const GLuint texEnableSave = ctx->Texture._EnabledUnits;
+
+   /* Disable tex units so they're not re-applied in swrast_write_rgba_span */
+   ctx->Texture._EnabledUnits = 0x0;
 
    /* Instead of defining a function for each mode, a test is done
     * between the outer and inner loops. This is to reduce code size
@@ -493,8 +497,12 @@ affine_span(GLcontext *ctx, SWspan *span,
    }
    span->interpMask &= ~SPAN_RGBA;
    ASSERT(span->arrayMask & SPAN_RGBA);
+
    _swrast_write_rgba_span(ctx, span);
 
+   /* re-enable texture units */
+   ctx->Texture._EnabledUnits = texEnableSave;
+
 #undef SPAN_NEAREST
 #undef SPAN_LINEAR
 }
diff --git a/src/mesa/tnl/t_vertex.c b/src/mesa/tnl/t_vertex.c
index b661524c87..10b78f820e 100644
--- a/src/mesa/tnl/t_vertex.c
+++ b/src/mesa/tnl/t_vertex.c
@@ -376,6 +376,22 @@ void _tnl_notify_pipeline_output_change( GLcontext *ctx )
    invalidate_funcs(vtx);
 }
 
+
+static void adjust_input_ptrs( GLcontext *ctx, GLint diff)
+{
+   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
+   struct tnl_clipspace *vtx = GET_VERTEX_STATE(ctx);
+   struct tnl_clipspace_attr *a = vtx->attr;
+   const GLuint count = vtx->attr_count;
+   int j;
+
+   diff -= 1;
+   for (j=0; j<count; ++j) {
+           register GLvector4f *vptr = VB->AttribPtr[a->attrib];
+	   (a++)->inputptr += diff*vptr->stride;
+   }
+}
+
 static void update_input_ptrs( GLcontext *ctx, GLuint start )
 {
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
@@ -431,13 +447,40 @@ void *_tnl_emit_vertices_to_buffer( GLcontext *ctx,
    struct tnl_clipspace *vtx = GET_VERTEX_STATE(ctx);
 
    update_input_ptrs(ctx, start);
-
    /* Note: dest should not be adjusted for non-zero 'start' values:
     */
    vtx->emit( ctx, end - start, (GLubyte*) dest );	
    return (void *)((GLubyte *)dest + vtx->vertex_size * (end - start));
 }
 
+/* Emit indexed VB vertices start..end to dest.  Note that VB vertex at
+ * postion start will be emitted to dest at position zero.
+ */
+
+void *_tnl_emit_indexed_vertices_to_buffer( GLcontext *ctx,
+					    const GLuint *elts,
+					    GLuint start,
+					    GLuint end,
+					    void *dest )
+{
+   struct tnl_clipspace *vtx = GET_VERTEX_STATE(ctx);
+   GLuint oldIndex;
+   GLubyte *cdest = dest;
+
+   update_input_ptrs(ctx, oldIndex = elts[start++]);
+   vtx->emit( ctx, 1, cdest );
+   cdest += vtx->vertex_size;
+
+   for (; start < end; ++start) {
+      adjust_input_ptrs(ctx, elts[start] - oldIndex);
+      oldIndex = elts[start];
+      vtx->emit( ctx, 1, cdest);
+      cdest += vtx->vertex_size;
+   }
+
+   return (void *) cdest;
+}
+
 
 void _tnl_init_vertices( GLcontext *ctx, 
 			GLuint vb_size,
diff --git a/src/mesa/tnl/t_vertex.h b/src/mesa/tnl/t_vertex.h
index 712311a146..2dfd7b57f0 100644
--- a/src/mesa/tnl/t_vertex.h
+++ b/src/mesa/tnl/t_vertex.h
@@ -119,6 +119,18 @@ extern void *_tnl_emit_vertices_to_buffer( GLcontext *ctx,
 					   GLuint end,
 					   void *dest );
 
+/* This function isn't optimal. Check out 
+ * gallium/auxilary/translate for a more comprehensive implementation of
+ * the same functionality.
+ */
+  
+extern void *_tnl_emit_indexed_vertices_to_buffer( GLcontext *ctx,
+						   const GLuint *elts,
+						   GLuint start,
+						   GLuint end,
+						   void *dest );
+
+
 extern void _tnl_build_vertices( GLcontext *ctx,
 				 GLuint start,
 				 GLuint end,