Merge ../mesa into vulkan

120 files changed, 1150 insertions, 1428 deletions

diff --git a/src/mesa/drivers/common/driverfuncs.c b/src/mesa/drivers/common/driverfuncs.c
index 6fe42b1775c..3d1fccb3ab4 100644
--- a/src/mesa/drivers/common/driverfuncs.c
+++ b/src/mesa/drivers/common/driverfuncs.c
@@ -75,14 +75,12 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
 
    driver->GetString = NULL;  /* REQUIRED! */
    driver->UpdateState = NULL;  /* REQUIRED! */
-   driver->ResizeBuffers = _mesa_resize_framebuffer;
 
    driver->Finish = NULL;
    driver->Flush = NULL;
 
    /* framebuffer/image functions */
    driver->Clear = _swrast_Clear;
-   driver->Accum = _mesa_accum;
    driver->RasterPos = _tnl_RasterPos;
    driver->DrawPixels = _swrast_DrawPixels;
    driver->ReadPixels = _mesa_readpixels;
@@ -135,7 +133,6 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->DepthRange = NULL;
    driver->Enable = NULL;
    driver->Fogfv = NULL;
-   driver->Hint = NULL;
    driver->Lightfv = NULL;
    driver->LightModelfv = NULL;
    driver->LineStipple = NULL;
@@ -179,16 +176,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->DiscardFramebuffer = NULL;
 
    _mesa_init_texture_barrier_functions(driver);
-
-   /* APPLE_vertex_array_object */
-   driver->NewArrayObject = _mesa_new_vao;
-   driver->DeleteArrayObject = _mesa_delete_vao;
-   driver->BindArrayObject = NULL;
-
    _mesa_init_shader_object_functions(driver);
-
    _mesa_init_transform_feedback_functions(driver);
-
    _mesa_init_sampler_object_functions(driver);
 
    /* T&L stuff */
@@ -198,17 +187,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->SaveNeedFlush = 0;
 
    driver->ProgramStringNotify = _tnl_program_string;
-   driver->FlushVertices = NULL;
-   driver->SaveFlushVertices = NULL;
-   driver->NotifySaveBegin = NULL;
    driver->LightingSpaceChange = NULL;
 
-   /* display list */
-   driver->NewList = NULL;
-   driver->EndList = NULL;
-   driver->BeginCallList = NULL;
-   driver->EndCallList = NULL;
-
    /* GL_ARB_texture_storage */
    driver->AllocTextureStorage = _mesa_AllocTextureStorage_sw;
 
@@ -242,23 +222,11 @@ _mesa_init_driver_state(struct gl_context *ctx)
                                  ctx->Color.Blend[0].SrcA,
                                  ctx->Color.Blend[0].DstA);
 
-   if (ctx->Driver.ColorMaskIndexed) {
-      GLuint i;
-      for (i = 0; i < ctx->Const.MaxDrawBuffers; i++) {
-         ctx->Driver.ColorMaskIndexed(ctx, i,
-                                      ctx->Color.ColorMask[i][RCOMP],
-                                      ctx->Color.ColorMask[i][GCOMP],
-                                      ctx->Color.ColorMask[i][BCOMP],
-                                      ctx->Color.ColorMask[i][ACOMP]);
-      }
-   }
-   else {
-      ctx->Driver.ColorMask(ctx,
-                            ctx->Color.ColorMask[0][RCOMP],
-                            ctx->Color.ColorMask[0][GCOMP],
-                            ctx->Color.ColorMask[0][BCOMP],
-                            ctx->Color.ColorMask[0][ACOMP]);
-   }
+   ctx->Driver.ColorMask(ctx,
+                         ctx->Color.ColorMask[0][RCOMP],
+                         ctx->Color.ColorMask[0][GCOMP],
+                         ctx->Color.ColorMask[0][BCOMP],
+                         ctx->Color.ColorMask[0][ACOMP]);
 
    ctx->Driver.CullFace(ctx, ctx->Polygon.CullFaceMode);
    ctx->Driver.DepthFunc(ctx, ctx->Depth.Func);
diff --git a/src/mesa/drivers/common/meta_generate_mipmap.c b/src/mesa/drivers/common/meta_generate_mipmap.c
index 5dc40a2aa33..4800278a467 100644
--- a/src/mesa/drivers/common/meta_generate_mipmap.c
+++ b/src/mesa/drivers/common/meta_generate_mipmap.c
@@ -150,8 +150,7 @@ prepare_mipmap_level(struct gl_context *ctx,
 
 /**
  * Called via ctx->Driver.GenerateMipmap()
- * Note: We don't yet support 3D textures, 1D/2D array textures or texture
- * borders.
+ * Note: We don't yet support 3D textures, or texture borders.
  */
 void
 _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target,
diff --git a/src/mesa/drivers/dri/common/dri_util.c b/src/mesa/drivers/dri/common/dri_util.c
index d35ac263a45..5cfa2f8ca4f 100644
--- a/src/mesa/drivers/dri/common/dri_util.c
+++ b/src/mesa/drivers/dri/common/dri_util.c
@@ -44,6 +44,7 @@
 #include "utils.h"
 #include "xmlpool.h"
 #include "main/mtypes.h"
+#include "main/framebuffer.h"
 #include "main/version.h"
 #include "main/errors.h"
 #include "main/macros.h"
@@ -793,7 +794,7 @@ driUpdateFramebufferSize(struct gl_context *ctx, const __DRIdrawable *dPriv)
 {
    struct gl_framebuffer *fb = (struct gl_framebuffer *) dPriv->driverPrivate;
    if (fb && (dPriv->w != fb->Width || dPriv->h != fb->Height)) {
-      ctx->Driver.ResizeBuffers(ctx, fb, dPriv->w, dPriv->h);
+      _mesa_resize_framebuffer(ctx, fb, dPriv->w, dPriv->h);
       /* if the driver needs the hw lock for ResizeBuffers, the drawable
          might have changed again by now */
       assert(fb->Width == dPriv->w);
diff --git a/src/mesa/drivers/dri/common/utils.c b/src/mesa/drivers/dri/common/utils.c
index 1246bec6e02..ae8fcabc14e 100644
--- a/src/mesa/drivers/dri/common/utils.c
+++ b/src/mesa/drivers/dri/common/utils.c
@@ -41,36 +41,6 @@
 #include "utils.h"
 #include "dri_util.h"
 
-
-uint64_t
-driParseDebugString(const char *debug,
-                    const struct dri_debug_control *control)
-{
-   uint64_t flag = 0;
-
-   if (debug != NULL) {
-      for (; control->string != NULL; control++) {
-         if (!strcmp(debug, "all")) {
-            flag |= control->flag;
-
-         } else {
-            const char *s = debug;
-            unsigned n;
-
-            for (; n = strcspn(s, ", "), *s; s += MAX2(1, n)) {
-               if (strlen(control->string) == n &&
-                   !strncmp(control->string, s, n))
-                  flag |= control->flag;
-            }
-         }
-      }
-   }
-
-   return flag;
-}
-
-
-
 /**
  * Create the \c GL_RENDERER string for DRI drivers.
  * 
diff --git a/src/mesa/drivers/dri/common/utils.h b/src/mesa/drivers/dri/common/utils.h
index 3760c38fcaf..f6b8d7c3a21 100644
--- a/src/mesa/drivers/dri/common/utils.h
+++ b/src/mesa/drivers/dri/common/utils.h
@@ -32,14 +32,6 @@
 #include <GL/internal/dri_interface.h>
 #include "main/context.h"
 
-struct dri_debug_control {
-    const char * string;
-    uint64_t     flag;
-};
-
-extern uint64_t driParseDebugString( const char * debug,
-    const struct dri_debug_control * control );
-
 extern unsigned driGetRendererString( char * buffer,
     const char * hardware_name, GLuint agp_mode );
 
diff --git a/src/mesa/drivers/dri/i915/i830_state.c b/src/mesa/drivers/dri/i915/i830_state.c
index 906e942b020..7ce5ef7e41a 100644
--- a/src/mesa/drivers/dri/i915/i830_state.c
+++ b/src/mesa/drivers/dri/i915/i830_state.c
@@ -601,6 +601,8 @@ i830CullFaceFrontFace(struct gl_context * ctx, GLenum unused)
    else if (ctx->Polygon.CullFaceMode != GL_FRONT_AND_BACK) {
       mode = CULLMODE_CW;
 
+      if (ctx->DrawBuffer && _mesa_is_user_fbo(ctx->DrawBuffer))
+         mode ^= (CULLMODE_CW ^ CULLMODE_CCW);
       if (ctx->Polygon.CullFaceMode == GL_FRONT)
          mode ^= (CULLMODE_CW ^ CULLMODE_CCW);
       if (ctx->Polygon.FrontFace != GL_CCW)
diff --git a/src/mesa/drivers/dri/i915/i915_vtbl.c b/src/mesa/drivers/dri/i915/i915_vtbl.c
index 80bd249fa7b..c41cd37bcc2 100644
--- a/src/mesa/drivers/dri/i915/i915_vtbl.c
+++ b/src/mesa/drivers/dri/i915/i915_vtbl.c
@@ -31,7 +31,6 @@
 #include "main/mtypes.h"
 #include "main/imports.h"
 #include "main/macros.h"
-#include "main/colormac.h"
 #include "main/renderbuffer.h"
 #include "main/framebuffer.h"
 
diff --git a/src/mesa/drivers/dri/i915/intel_buffers.c b/src/mesa/drivers/dri/i915/intel_buffers.c
index 51eaea43a50..386e032443a 100644
--- a/src/mesa/drivers/dri/i915/intel_buffers.c
+++ b/src/mesa/drivers/dri/i915/intel_buffers.c
@@ -55,20 +55,14 @@ intel_check_front_buffer_rendering(struct intel_context *intel)
 static void
 intelDrawBuffer(struct gl_context * ctx, GLenum mode)
 {
-   if (ctx->DrawBuffer && _mesa_is_winsys_fbo(ctx->DrawBuffer)) {
+   if (_mesa_is_front_buffer_drawing(ctx->DrawBuffer)) {
       struct intel_context *const intel = intel_context(ctx);
-      const bool was_front_buffer_rendering =
-	intel->is_front_buffer_rendering;
 
-      intel->is_front_buffer_rendering = (mode == GL_FRONT_LEFT)
-	|| (mode == GL_FRONT) || (mode == GL_FRONT_AND_BACK);
-
-      /* If we weren't front-buffer rendering before but we are now,
-       * invalidate our DRI drawable so we'll ask for new buffers
+      /* If we might be front-buffer rendering on this buffer for the first
+       * time, invalidate our DRI drawable so we'll ask for new buffers
        * (including the fake front) before we start rendering again.
        */
-      if (!was_front_buffer_rendering && intel->is_front_buffer_rendering)
-	 dri2InvalidateDrawable(intel->driContext->driDrawablePriv);
+      dri2InvalidateDrawable(intel->driContext->driDrawablePriv);
    }
 
    intel_draw_buffer(ctx);
@@ -78,20 +72,14 @@ intelDrawBuffer(struct gl_context * ctx, GLenum mode)
 static void
 intelReadBuffer(struct gl_context * ctx, GLenum mode)
 {
-   if (ctx->ReadBuffer && _mesa_is_winsys_fbo(ctx->ReadBuffer)) {
+   if (_mesa_is_front_buffer_reading(ctx->ReadBuffer)) {
       struct intel_context *const intel = intel_context(ctx);
-      const bool was_front_buffer_reading =
-	intel->is_front_buffer_reading;
-
-      intel->is_front_buffer_reading = (mode == GL_FRONT_LEFT)
-	|| (mode == GL_FRONT);
 
-      /* If we weren't front-buffer reading before but we are now,
-       * invalidate our DRI drawable so we'll ask for new buffers
+      /* If we might be front-buffer reading on this buffer for the first
+       * time, invalidate our DRI drawable so we'll ask for new buffers
        * (including the fake front) before we start reading again.
        */
-      if (!was_front_buffer_reading && intel->is_front_buffer_reading)
-	 dri2InvalidateDrawable(intel->driContext->driReadablePriv);
+      dri2InvalidateDrawable(intel->driContext->driReadablePriv);
    }
 }
 
diff --git a/src/mesa/drivers/dri/i915/intel_buffers.h b/src/mesa/drivers/dri/i915/intel_buffers.h
index 42d84bcfcb7..6e22f5125b9 100644
--- a/src/mesa/drivers/dri/i915/intel_buffers.h
+++ b/src/mesa/drivers/dri/i915/intel_buffers.h
@@ -34,7 +34,6 @@
 #include "intel_context.h"
 
 struct intel_context;
-struct intel_framebuffer;
 
 extern void intel_check_front_buffer_rendering(struct intel_context *intel);
 
diff --git a/src/mesa/drivers/dri/i915/intel_context.c b/src/mesa/drivers/dri/i915/intel_context.c
index c780103228f..3f429f25d10 100644
--- a/src/mesa/drivers/dri/i915/intel_context.c
+++ b/src/mesa/drivers/dri/i915/intel_context.c
@@ -56,6 +56,7 @@
 #include "intel_mipmap_tree.h"
 
 #include "utils.h"
+#include "util/debug.h"
 #include "util/ralloc.h"
 
 int INTEL_DEBUG = (0);
@@ -243,7 +244,7 @@ intel_prepare_render(struct intel_context *intel)
     * that will happen next will probably dirty the front buffer.  So
     * mark it as dirty here.
     */
-   if (intel->is_front_buffer_rendering)
+   if (_mesa_is_front_buffer_drawing(intel->ctx.DrawBuffer))
       intel->front_buffer_dirty = true;
 
    /* Wait for the swapbuffers before the one we just emitted, so we
@@ -290,7 +291,7 @@ intel_viewport(struct gl_context *ctx)
     intelCalcViewport(ctx);
 }
 
-static const struct dri_debug_control debug_control[] = {
+static const struct debug_control debug_control[] = {
    { "tex",   DEBUG_TEXTURE},
    { "state", DEBUG_STATE},
    { "blit",  DEBUG_BLIT},
@@ -356,7 +357,7 @@ intel_glFlush(struct gl_context *ctx)
 
    intel_flush(ctx);
    intel_flush_front(ctx);
-   if (intel->is_front_buffer_rendering)
+   if (_mesa_is_front_buffer_drawing(ctx->DrawBuffer))
       intel->need_throttle = true;
 }
 
@@ -474,8 +475,8 @@ intelInitContext(struct intel_context *intel,
 
    ctx->Const.MinLineWidth = 1.0;
    ctx->Const.MinLineWidthAA = 1.0;
-   ctx->Const.MaxLineWidth = 5.0;
-   ctx->Const.MaxLineWidthAA = 5.0;
+   ctx->Const.MaxLineWidth = 7.0;
+   ctx->Const.MaxLineWidthAA = 7.0;
    ctx->Const.LineWidthGranularity = 0.5;
 
    ctx->Const.MinPointSize = 1.0;
@@ -512,7 +513,7 @@ intelInitContext(struct intel_context *intel,
 
    intelInitExtensions(ctx);
 
-   INTEL_DEBUG = driParseDebugString(getenv("INTEL_DEBUG"), debug_control);
+   INTEL_DEBUG = parse_debug_string(getenv("INTEL_DEBUG"), debug_control);
    if (INTEL_DEBUG & DEBUG_BUFMGR)
       dri_bufmgr_set_debug(intel->bufmgr, true);
    if (INTEL_DEBUG & DEBUG_PERF)
@@ -700,8 +701,8 @@ intel_query_dri2_buffers(struct intel_context *intel,
    back_rb = intel_get_renderbuffer(fb, BUFFER_BACK_LEFT);
 
    memset(attachments, 0, sizeof(attachments));
-   if ((intel->is_front_buffer_rendering ||
-	intel->is_front_buffer_reading ||
+   if ((_mesa_is_front_buffer_drawing(fb) ||
+        _mesa_is_front_buffer_reading(fb) ||
 	!back_rb) && front_rb) {
       /* If a fake front buffer is in use, then querying for
        * __DRI_BUFFER_FRONT_LEFT will cause the server to copy the image from
@@ -866,8 +867,10 @@ intel_update_image_buffers(struct intel_context *intel, __DRIdrawable *drawable)
    else
       return;
 
-   if ((intel->is_front_buffer_rendering || intel->is_front_buffer_reading || !back_rb) && front_rb)
+   if (front_rb && (_mesa_is_front_buffer_drawing(fb) ||
+                    _mesa_is_front_buffer_reading(fb) || !back_rb)) {
       buffer_mask |= __DRI_IMAGE_BUFFER_FRONT;
+   }
 
    if (back_rb)
       buffer_mask |= __DRI_IMAGE_BUFFER_BACK;
diff --git a/src/mesa/drivers/dri/i915/intel_context.h b/src/mesa/drivers/dri/i915/intel_context.h
index 4ec4015d453..aecd7c23f45 100644
--- a/src/mesa/drivers/dri/i915/intel_context.h
+++ b/src/mesa/drivers/dri/i915/intel_context.h
@@ -255,22 +255,6 @@ struct intel_context
     */
    bool front_buffer_dirty;
 
-   /**
-    * Track whether front-buffer rendering is currently enabled
-    *
-    * A separate flag is used to track this in order to support MRT more
-    * easily.
-    */
-   bool is_front_buffer_rendering;
-   /**
-    * Track whether front-buffer is the current read target.
-    *
-    * This is closely associated with is_front_buffer_rendering, but may
-    * be set separately.  The DRI2 fake front buffer must be referenced
-    * either way.
-    */
-   bool is_front_buffer_reading;
-
    bool use_early_z;
 
    __DRIcontext *driContext;
diff --git a/src/mesa/drivers/dri/i915/intel_fbo.c b/src/mesa/drivers/dri/i915/intel_fbo.c
index 12cc7e3a71b..8750c601b84 100644
--- a/src/mesa/drivers/dri/i915/intel_fbo.c
+++ b/src/mesa/drivers/dri/i915/intel_fbo.c
@@ -64,19 +64,6 @@ intel_get_rb_region(struct gl_framebuffer *fb, GLuint attIndex)
       return NULL;
 }
 
-/**
- * Create a new framebuffer object.
- */
-static struct gl_framebuffer *
-intel_new_framebuffer(struct gl_context * ctx, GLuint name)
-{
-   /* Only drawable state in intel_framebuffer at this time, just use Mesa's
-    * class
-    */
-   return _mesa_new_framebuffer(ctx, name);
-}
-
-
 /** Called by gl_renderbuffer::Delete() */
 static void
 intel_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
@@ -770,7 +757,6 @@ intel_blit_framebuffer(struct gl_context *ctx,
 void
 intel_fbo_init(struct intel_context *intel)
 {
-   intel->ctx.Driver.NewFramebuffer = intel_new_framebuffer;
    intel->ctx.Driver.NewRenderbuffer = intel_new_renderbuffer;
    intel->ctx.Driver.MapRenderbuffer = intel_map_renderbuffer;
    intel->ctx.Driver.UnmapRenderbuffer = intel_unmap_renderbuffer;
diff --git a/src/mesa/drivers/dri/i915/intel_pixel.c b/src/mesa/drivers/dri/i915/intel_pixel.c
index 6f139e19e11..feb1a3f97e8 100644
--- a/src/mesa/drivers/dri/i915/intel_pixel.c
+++ b/src/mesa/drivers/dri/i915/intel_pixel.c
@@ -126,7 +126,6 @@ intel_check_blit_fragment_ops(struct gl_context * ctx, bool src_alpha_is_one)
 void
 intelInitPixelFuncs(struct dd_function_table *functions)
 {
-   functions->Accum = _mesa_accum;
    functions->Bitmap = intelBitmap;
    functions->CopyPixels = intelCopyPixels;
    functions->DrawPixels = intelDrawPixels;
diff --git a/src/mesa/drivers/dri/i915/intel_render.c b/src/mesa/drivers/dri/i915/intel_render.c
index 6c2ad6c6c95..c1603565cc2 100644
--- a/src/mesa/drivers/dri/i915/intel_render.c
+++ b/src/mesa/drivers/dri/i915/intel_render.c
@@ -54,9 +54,7 @@
  * dma buffers.  Use strip/fan hardware primitives where possible.
  * Try to simulate missing primitives with indexed vertices.
  */
-#define HAVE_POINTS      0      /* Has it, but can't use because subpixel has to
-                                 * be adjusted for points on the INTEL/I845G
-                                 */
+#define HAVE_POINTS      1
 #define HAVE_LINES       1
 #define HAVE_LINE_STRIPS 1
 #define HAVE_TRIANGLES   1
@@ -67,42 +65,42 @@
 #define HAVE_ELTS        0
 
 static const uint32_t hw_prim[GL_POLYGON + 1] = {
-   0,
-   PRIM3D_LINELIST,
-   PRIM3D_LINESTRIP,
-   PRIM3D_LINESTRIP,
-   PRIM3D_TRILIST,
-   PRIM3D_TRISTRIP,
-   PRIM3D_TRIFAN,
-   0,
-   0,
-   PRIM3D_POLY
+   [GL_POINTS] = PRIM3D_POINTLIST,
+   [GL_LINES ] = PRIM3D_LINELIST,
+   [GL_LINE_LOOP] = PRIM3D_LINESTRIP,
+   [GL_LINE_STRIP] = PRIM3D_LINESTRIP,
+   [GL_TRIANGLES] = PRIM3D_TRILIST,
+   [GL_TRIANGLE_STRIP] = PRIM3D_TRISTRIP,
+   [GL_TRIANGLE_FAN] = PRIM3D_TRIFAN,
+   [GL_QUADS] = 0,
+   [GL_QUAD_STRIP] = 0,
+   [GL_POLYGON] = PRIM3D_POLY,
 };
 
 static const GLenum reduced_prim[GL_POLYGON + 1] = {
-   GL_POINTS,
-   GL_LINES,
-   GL_LINES,
-   GL_LINES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES
+   [GL_POINTS] = GL_POINTS,
+   [GL_LINES] = GL_LINES,
+   [GL_LINE_LOOP] = GL_LINES,
+   [GL_LINE_STRIP] = GL_LINES,
+   [GL_TRIANGLES] = GL_TRIANGLES,
+   [GL_TRIANGLE_STRIP] = GL_TRIANGLES,
+   [GL_TRIANGLE_FAN] = GL_TRIANGLES,
+   [GL_QUADS] = GL_TRIANGLES,
+   [GL_QUAD_STRIP] = GL_TRIANGLES,
+   [GL_POLYGON] = GL_TRIANGLES,
 };
 
 static const int scale_prim[GL_POLYGON + 1] = {
-   0,                           /* fallback case */
-   1,
-   2,
-   2,
-   1,
-   3,
-   3,
-   0,                           /* fallback case */
-   0,                           /* fallback case */
-   3
+   [GL_POINTS] = 1,
+   [GL_LINES] = 1,
+   [GL_LINE_LOOP] = 2,
+   [GL_LINE_STRIP] = 2,
+   [GL_TRIANGLES] = 1,
+   [GL_TRIANGLE_STRIP] = 3,
+   [GL_TRIANGLE_FAN] = 3,
+   [GL_QUADS] = 0,              /* fallback case */
+   [GL_QUAD_STRIP] = 0,         /* fallback case */
+   [GL_POLYGON] = 3,
 };
 
 
diff --git a/src/mesa/drivers/dri/i915/intel_state.c b/src/mesa/drivers/dri/i915/intel_state.c
index c951ff731b8..3de9d50a4d7 100644
--- a/src/mesa/drivers/dri/i915/intel_state.c
+++ b/src/mesa/drivers/dri/i915/intel_state.c
@@ -30,7 +30,6 @@
 #include "main/context.h"
 #include "main/macros.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/dd.h"
 
 #include "intel_screen.h"
diff --git a/src/mesa/drivers/dri/i915/intel_tris.c b/src/mesa/drivers/dri/i915/intel_tris.c
index ae62a800fb7..a093a0a7bd2 100644
--- a/src/mesa/drivers/dri/i915/intel_tris.c
+++ b/src/mesa/drivers/dri/i915/intel_tris.c
@@ -426,11 +426,7 @@ intel_draw_point(struct intel_context *intel, intelVertexPtr v0)
    GLuint *vb = intel_get_prim_space(intel, 1);
    int j;
 
-   /* Adjust for sub pixel position -- still required for conform. */
-   *(float *) &vb[0] = v0->v.x;
-   *(float *) &vb[1] = v0->v.y;
-   for (j = 2; j < vertsize; j++)
-      vb[j] = v0->ui[j];
+   COPY_DWORDS(j, vb, vertsize, v0);
 }
 
 
@@ -670,16 +666,16 @@ do {							\
  ***********************************************************************/
 
 static const GLuint hw_prim[GL_POLYGON + 1] = {
-   PRIM3D_POINTLIST,
-   PRIM3D_LINELIST,
-   PRIM3D_LINELIST,
-   PRIM3D_LINELIST,
-   PRIM3D_TRILIST,
-   PRIM3D_TRILIST,
-   PRIM3D_TRILIST,
-   PRIM3D_TRILIST,
-   PRIM3D_TRILIST,
-   PRIM3D_TRILIST
+   [GL_POINTS] = PRIM3D_POINTLIST,
+   [GL_LINES] = PRIM3D_LINELIST,
+   [GL_LINE_LOOP] = PRIM3D_LINELIST,
+   [GL_LINE_STRIP] = PRIM3D_LINELIST,
+   [GL_TRIANGLES] = PRIM3D_TRILIST,
+   [GL_TRIANGLE_STRIP] = PRIM3D_TRILIST,
+   [GL_TRIANGLE_FAN] = PRIM3D_TRILIST,
+   [GL_QUADS] = PRIM3D_TRILIST,
+   [GL_QUAD_STRIP] = PRIM3D_TRILIST,
+   [GL_POLYGON] = PRIM3D_TRILIST,
 };
 
 #define RASTERIZE(x) intelRasterPrimitive( ctx, x, hw_prim[x] )
@@ -891,18 +887,11 @@ intelRenderClippedPoly(struct gl_context * ctx, const GLuint * elts, GLuint n)
 {
    struct intel_context *intel = intel_context(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
    GLuint prim = intel->render_primitive;
 
    /* Render the new vertices as an unclipped polygon.
     */
-   {
-      GLuint *tmp = VB->Elts;
-      VB->Elts = (GLuint *) elts;
-      tnl->Driver.Render.PrimTabElts[GL_POLYGON] (ctx, 0, n,
-                                                  PRIM_BEGIN | PRIM_END);
-      VB->Elts = tmp;
-   }
+   _tnl_RenderClippedPolygon(ctx, elts, n);
 
    /* Restore the render primitive
     */
@@ -911,14 +900,6 @@ intelRenderClippedPoly(struct gl_context * ctx, const GLuint * elts, GLuint n)
 }
 
 static void
-intelRenderClippedLine(struct gl_context * ctx, GLuint ii, GLuint jj)
-{
-   TNLcontext *tnl = TNL_CONTEXT(ctx);
-
-   tnl->Driver.Render.Line(ctx, ii, jj);
-}
-
-static void
 intelFastRenderClippedPoly(struct gl_context * ctx, const GLuint * elts, GLuint n)
 {
    struct intel_context *intel = intel_context(ctx);
@@ -928,10 +909,18 @@ intelFastRenderClippedPoly(struct gl_context * ctx, const GLuint * elts, GLuint
    const GLuint *start = (const GLuint *) V(elts[0]);
    int i, j;
 
-   for (i = 2; i < n; i++) {
-      COPY_DWORDS(j, vb, vertsize, V(elts[i - 1]));
-      COPY_DWORDS(j, vb, vertsize, V(elts[i]));
-      COPY_DWORDS(j, vb, vertsize, start);
+   if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION) {
+      for (i = 2; i < n; i++) {
+         COPY_DWORDS(j, vb, vertsize, V(elts[i - 1]));
+         COPY_DWORDS(j, vb, vertsize, V(elts[i]));
+         COPY_DWORDS(j, vb, vertsize, start);
+      }
+   } else {
+      for (i = 2; i < n; i++) {
+         COPY_DWORDS(j, vb, vertsize, start);
+         COPY_DWORDS(j, vb, vertsize, V(elts[i - 1]));
+         COPY_DWORDS(j, vb, vertsize, V(elts[i]));
+      }
    }
 }
 
@@ -1036,23 +1025,23 @@ intelChooseRenderState(struct gl_context * ctx)
       else {
          tnl->Driver.Render.PrimTabVerts = _tnl_render_tab_verts;
          tnl->Driver.Render.PrimTabElts = _tnl_render_tab_elts;
-         tnl->Driver.Render.ClippedLine = intelRenderClippedLine;
+         tnl->Driver.Render.ClippedLine = _tnl_RenderClippedLine;
          tnl->Driver.Render.ClippedPolygon = intelRenderClippedPoly;
       }
    }
 }
 
 static const GLenum reduced_prim[GL_POLYGON + 1] = {
-   GL_POINTS,
-   GL_LINES,
-   GL_LINES,
-   GL_LINES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES
+   [GL_POINTS] = GL_POINTS,
+   [GL_LINES] = GL_LINES,
+   [GL_LINE_LOOP] = GL_LINES,
+   [GL_LINE_STRIP] = GL_LINES,
+   [GL_TRIANGLES] = GL_TRIANGLES,
+   [GL_TRIANGLE_STRIP] = GL_TRIANGLES,
+   [GL_TRIANGLE_FAN] = GL_TRIANGLES,
+   [GL_QUADS] = GL_TRIANGLES,
+   [GL_QUAD_STRIP] = GL_TRIANGLES,
+   [GL_POLYGON] = GL_TRIANGLES
 };
 
 
diff --git a/src/mesa/drivers/dri/i965/Makefile.am b/src/mesa/drivers/dri/i965/Makefile.am
index 3c8197226fe..2e241511049 100644
--- a/src/mesa/drivers/dri/i965/Makefile.am
+++ b/src/mesa/drivers/dri/i965/Makefile.am
@@ -43,28 +43,16 @@ AM_CXXFLAGS = $(AM_CFLAGS)
 
 noinst_LTLIBRARIES = libi965_dri.la libi965_compiler.la
 libi965_dri_la_SOURCES = $(i965_FILES)
-libi965_dri_la_LIBADD = $(INTEL_LIBS)
+libi965_dri_la_LIBADD = libi965_compiler.la $(INTEL_LIBS)
+
+libi965_compiler_la_SOURCES = $(i965_compiler_FILES)
 
 TEST_LIBS = \
-	libi965_dri.la \
-	../common/libdricommon.la \
-	../common/libxmlconfig.la \
-	../common/libmegadriver_stub.la \
+	libi965_compiler.la \
         ../../../libmesa.la \
-	$(DRI_LIB_DEPS) \
-        $(CLOCK_LIB) \
+	-lpthread -ldl \
 	../common/libdri_test_stubs.la
 
-libi965_compiler_la_SOURCES = $(i965_FILES)
-libi965_compiler_la_LIBADD = $(INTEL_LIBS) \
-	../common/libdricommon.la \
-	../common/libxmlconfig.la \
-	../common/libmegadriver_stub.la \
-        ../../../libmesa.la \
-	$(DRI_LIB_DEPS) \
-        $(CLOCK_LIB) \
-	../common/libdri_test_stubs.la -lm
-
 TESTS = \
 	test_fs_cmod_propagation \
 	test_fs_saturate_propagation \
diff --git a/src/mesa/drivers/dri/i965/Makefile.sources b/src/mesa/drivers/dri/i965/Makefile.sources
index eb8196d4845..81ef6283fa1 100644
--- a/src/mesa/drivers/dri/i965/Makefile.sources
+++ b/src/mesa/drivers/dri/i965/Makefile.sources
@@ -1,47 +1,18 @@
-i965_FILES = \
-	brw_binding_tables.c \
-	brw_blorp_blit.cpp \
-	brw_blorp_blit_eu.cpp \
-	brw_blorp_blit_eu.h \
-	brw_blorp.cpp \
-	brw_blorp.h \
-	brw_cc.c \
+i965_compiler_FILES = \
 	brw_cfg.cpp \
 	brw_cfg.h \
-	brw_clear.c \
-	brw_clip.c \
-	brw_clip.h \
-	brw_clip_line.c \
-	brw_clip_point.c \
-	brw_clip_state.c \
-	brw_clip_tri.c \
-	brw_clip_unfilled.c \
-	brw_clip_util.c \
-	brw_compute.c \
-	brw_conditional_render.c \
-	brw_context.c \
-	brw_context.h \
-	brw_cs.c \
-	brw_cs.h \
 	brw_cubemap_normalize.cpp \
-	brw_curbe.c \
 	brw_dead_control_flow.cpp \
 	brw_dead_control_flow.h \
 	brw_defines.h \
 	brw_device_info.c \
 	brw_device_info.h \
 	brw_disasm.c \
-	brw_draw.c \
-	brw_draw.h \
-	brw_draw_upload.c \
 	brw_eu.c \
 	brw_eu_compact.c \
 	brw_eu_emit.c \
 	brw_eu.h \
 	brw_eu_util.c \
-	brw_ff_gs.c \
-	brw_ff_gs_emit.c \
-	brw_ff_gs.h \
 	brw_fs_builder.h \
 	brw_fs_channel_expressions.cpp \
 	brw_fs_cmod_propagation.cpp \
@@ -55,7 +26,6 @@ i965_FILES = \
 	brw_fs_live_variables.cpp \
 	brw_fs_live_variables.h \
 	brw_fs_nir.cpp \
-	brw_fs_peephole_predicated_break.cpp \
 	brw_fs_reg_allocate.cpp \
 	brw_fs_register_coalesce.cpp \
 	brw_fs_saturate_propagation.cpp \
@@ -65,10 +35,6 @@ i965_FILES = \
 	brw_fs_validate.cpp \
 	brw_fs_vector_splitting.cpp \
 	brw_fs_visitor.cpp \
-	brw_gs.c \
-	brw_gs.h \
-	brw_gs_state.c \
-	brw_gs_surface_state.c \
 	brw_inst.h \
 	brw_interpolation_map.c \
 	brw_ir_allocator.h \
@@ -76,45 +42,17 @@ i965_FILES = \
 	brw_ir_vec4.h \
 	brw_lower_texture_gradients.cpp \
 	brw_lower_unnormalized_offset.cpp \
-	brw_meta_fast_clear.c \
-	brw_meta_stencil_blit.c \
-	brw_meta_updownsample.c \
-	brw_meta_util.c \
-	brw_meta_util.h \
-	brw_misc_state.c \
-	brw_multisample_state.h \
 	brw_nir.h \
 	brw_nir.c \
 	brw_nir_analyze_boolean_resolves.c \
 	brw_nir_uniforms.cpp \
-	brw_object_purgeable.c \
 	brw_packed_float.c \
-	brw_performance_monitor.c \
-	brw_pipe_control.c \
-	brw_primitive_restart.c \
-	brw_program.c \
-	brw_program.h \
-	brw_queryobj.c \
+	brw_predicated_break.cpp \
 	brw_reg.h \
-	brw_reset.c \
-	brw_sampler_state.c \
 	brw_schedule_instructions.cpp \
-	brw_sf.c \
-	brw_sf_emit.c \
-	brw_sf.h \
-	brw_sf_state.c \
 	brw_shader.cpp \
 	brw_shader.h \
-	brw_state_batch.c \
-	brw_state_cache.c \
-	brw_state_dump.c \
-	brw_state.h \
-	brw_state_upload.c \
-	brw_structs.h \
 	brw_surface_formats.c \
-	brw_tex.c \
-	brw_tex_layout.c \
-	brw_urb.c \
 	brw_util.c \
 	brw_util.h \
 	brw_vec4_builder.h \
@@ -135,14 +73,88 @@ i965_FILES = \
 	brw_vec4_surface_builder.h \
 	brw_vec4_visitor.cpp \
 	brw_vec4_vs_visitor.cpp \
+	brw_vue_map.c \
+	brw_wm_iz.cpp \
+	gen6_gs_visitor.cpp \
+	gen6_gs_visitor.h \
+	intel_asm_annotation.c \
+	intel_asm_annotation.h \
+	intel_debug.c \
+	intel_debug.h \
+	intel_reg.h \
+	intel_resolve_map.c \
+	intel_resolve_map.h
+
+i965_FILES = \
+	brw_binding_tables.c \
+	brw_blorp_blit.cpp \
+	brw_blorp_blit_eu.cpp \
+	brw_blorp_blit_eu.h \
+	brw_blorp.cpp \
+	brw_blorp.h \
+	brw_cc.c \
+	brw_clear.c \
+	brw_clip.c \
+	brw_clip.h \
+	brw_clip_line.c \
+	brw_clip_point.c \
+	brw_clip_state.c \
+	brw_clip_tri.c \
+	brw_clip_unfilled.c \
+	brw_clip_util.c \
+	brw_compute.c \
+	brw_conditional_render.c \
+	brw_context.c \
+	brw_context.h \
+	brw_cs.c \
+	brw_cs.h \
+	brw_curbe.c \
+	brw_draw.c \
+	brw_draw.h \
+	brw_draw_upload.c \
+	brw_ff_gs.c \
+	brw_ff_gs_emit.c \
+	brw_ff_gs.h \
+	brw_gs.c \
+	brw_gs.h \
+	brw_gs_state.c \
+	brw_gs_surface_state.c \
+	brw_link.cpp \
+	brw_meta_fast_clear.c \
+	brw_meta_stencil_blit.c \
+	brw_meta_updownsample.c \
+	brw_meta_util.c \
+	brw_meta_util.h \
+	brw_misc_state.c \
+	brw_multisample_state.h \
+	brw_object_purgeable.c \
+	brw_performance_monitor.c \
+	brw_pipe_control.c \
+	brw_program.c \
+	brw_program.h \
+	brw_primitive_restart.c \
+	brw_queryobj.c \
+	brw_reset.c \
+	brw_sampler_state.c \
+	brw_sf.c \
+	brw_sf_emit.c \
+	brw_sf.h \
+	brw_sf_state.c \
+	brw_state_batch.c \
+	brw_state_cache.c \
+	brw_state_dump.c \
+	brw_state.h \
+	brw_state_upload.c \
+	brw_structs.h \
+	brw_tex.c \
+	brw_tex_layout.c \
+	brw_urb.c \
 	brw_vs.c \
 	brw_vs.h \
 	brw_vs_state.c \
 	brw_vs_surface_state.c \
-	brw_vue_map.c \
 	brw_wm.c \
 	brw_wm.h \
-	brw_wm_iz.cpp \
 	brw_wm_state.c \
 	brw_wm_surface_state.c \
 	gen6_blorp.cpp \
@@ -152,8 +164,6 @@ i965_FILES = \
 	gen6_depth_state.c \
 	gen6_depthstencil.c \
 	gen6_gs_state.c \
-	gen6_gs_visitor.cpp \
-	gen6_gs_visitor.h \
 	gen6_multisample_state.c \
 	gen6_queryobj.c \
 	gen6_sampler_state.c \
@@ -192,8 +202,6 @@ i965_FILES = \
 	gen8_viewport_state.c \
 	gen8_vs_state.c \
 	gen8_wm_depth_stencil.c \
-	intel_asm_annotation.c \
-	intel_asm_annotation.h \
 	intel_batchbuffer.c \
 	intel_batchbuffer.h \
 	intel_blit.c \
@@ -203,8 +211,6 @@ i965_FILES = \
 	intel_buffers.c \
 	intel_buffers.h \
 	intel_copy_image.c \
-	intel_debug.c \
-	intel_debug.h \
 	intel_extensions.c \
 	intel_fbo.c \
 	intel_fbo.h \
@@ -217,9 +223,6 @@ i965_FILES = \
 	intel_pixel_draw.c \
 	intel_pixel.h \
 	intel_pixel_read.c \
-	intel_reg.h \
-	intel_resolve_map.c \
-	intel_resolve_map.h \
 	intel_screen.c \
 	intel_screen.h \
 	intel_state.c \
diff --git a/src/mesa/drivers/dri/i965/brw_clip.c b/src/mesa/drivers/dri/i965/brw_clip.c
index 3a73c64a88b..2d5abc70cb9 100644
--- a/src/mesa/drivers/dri/i965/brw_clip.c
+++ b/src/mesa/drivers/dri/i965/brw_clip.c
@@ -223,6 +223,7 @@ brw_upload_clip_prog(struct brw_context *brw)
 	       /* _NEW_POLYGON, _NEW_BUFFERS */
 	       key.offset_units = ctx->Polygon.OffsetUnits * ctx->DrawBuffer->_MRD * 2;
 	       key.offset_factor = ctx->Polygon.OffsetFactor * ctx->DrawBuffer->_MRD;
+	       key.offset_clamp = ctx->Polygon.OffsetClamp * ctx->DrawBuffer->_MRD;
 	    }
 
 	    if (!ctx->Polygon._FrontBit) {
diff --git a/src/mesa/drivers/dri/i965/brw_clip.h b/src/mesa/drivers/dri/i965/brw_clip.h
index 4e38f2f2ed6..54c76822e22 100644
--- a/src/mesa/drivers/dri/i965/brw_clip.h
+++ b/src/mesa/drivers/dri/i965/brw_clip.h
@@ -62,6 +62,7 @@ struct brw_clip_prog_key {
 
    GLfloat offset_factor;
    GLfloat offset_units;
+   GLfloat offset_clamp;
 };
 
 
diff --git a/src/mesa/drivers/dri/i965/brw_clip_unfilled.c b/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
index 6baf620a1a7..9a4d2a9d6f9 100644
--- a/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
+++ b/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
@@ -188,6 +188,12 @@ static void copy_bfc( struct brw_clip_compile *c )
   GLfloat bc	= dir.y * iz;
   offset = ctx->Polygon.OffsetUnits * DEPTH_SCALE;
   offset += MAX2( abs(ac), abs(bc) ) * ctx->Polygon.OffsetFactor;
+  if (ctx->Polygon.OffsetClamp && isfinite(ctx->Polygon.OffsetClamp)) {
+    if (ctx->Polygon.OffsetClamp < 0)
+      offset = MAX2( offset, ctx->Polygon.OffsetClamp );
+    else
+      offset = MIN2( offset, ctx->Polygon.OffsetClamp );
+  }
   offset *= MRD;
 */
 static void compute_offset( struct brw_clip_compile *c )
@@ -211,6 +217,14 @@ static void compute_offset( struct brw_clip_compile *c )
 
    brw_MUL(p, vec1(off), vec1(off), brw_imm_f(c->key.offset_factor));
    brw_ADD(p, vec1(off), vec1(off), brw_imm_f(c->key.offset_units));
+   if (c->key.offset_clamp && isfinite(c->key.offset_clamp)) {
+      brw_CMP(p,
+              vec1(brw_null_reg()),
+              c->key.offset_clamp < 0 ? BRW_CONDITIONAL_GE : BRW_CONDITIONAL_L,
+              vec1(off),
+              brw_imm_f(c->key.offset_clamp));
+      brw_SEL(p, vec1(off), vec1(off), brw_imm_f(c->key.offset_clamp));
+   }
 }
 
 
diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index 4d499295730..6b2bbd21703 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -41,6 +41,7 @@
 #include "main/version.h"
 #include "main/vtxfmt.h"
 #include "main/texobj.h"
+#include "main/framebuffer.h"
 
 #include "vbo/vbo_context.h"
 
@@ -326,7 +327,7 @@ brw_initialize_context_constants(struct brw_context *brw)
    ctx->Const.MaxUniformBlockSize = 65536;
    for (int i = 0; i < MESA_SHADER_STAGES; i++) {
       struct gl_program_constants *prog = &ctx->Const.Program[i];
-      prog->MaxUniformBlocks = 12;
+      prog->MaxUniformBlocks = BRW_MAX_UBO;
       prog->MaxCombinedUniformComponents =
          prog->MaxUniformComponents +
          ctx->Const.MaxUniformBlockSize / 4 * prog->MaxUniformBlocks;
@@ -346,7 +347,7 @@ brw_initialize_context_constants(struct brw_context *brw)
       ctx->Const.Program[MESA_SHADER_GEOMETRY].MaxTextureImageUnits = 0;
    if (_mesa_extension_override_enables.ARB_compute_shader) {
       ctx->Const.Program[MESA_SHADER_COMPUTE].MaxTextureImageUnits = BRW_MAX_TEX_UNIT;
-      ctx->Const.MaxUniformBufferBindings += 12;
+      ctx->Const.MaxUniformBufferBindings += BRW_MAX_UBO;
    } else {
       ctx->Const.Program[MESA_SHADER_COMPUTE].MaxTextureImageUnits = 0;
    }
@@ -582,17 +583,17 @@ brw_initialize_context_constants(struct brw_context *brw)
    /* FIXME: Tessellation stages are not yet supported in i965, so
     * MaxCombinedShaderStorageBlocks doesn't take them into account.
     */
-   ctx->Const.Program[MESA_SHADER_VERTEX].MaxShaderStorageBlocks = 12;
-   ctx->Const.Program[MESA_SHADER_GEOMETRY].MaxShaderStorageBlocks = 12;
+   ctx->Const.Program[MESA_SHADER_VERTEX].MaxShaderStorageBlocks = BRW_MAX_SSBO;
+   ctx->Const.Program[MESA_SHADER_GEOMETRY].MaxShaderStorageBlocks = BRW_MAX_SSBO;
    ctx->Const.Program[MESA_SHADER_TESS_EVAL].MaxShaderStorageBlocks = 0;
    ctx->Const.Program[MESA_SHADER_TESS_CTRL].MaxShaderStorageBlocks = 0;
-   ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxShaderStorageBlocks = 12;
-   ctx->Const.Program[MESA_SHADER_COMPUTE].MaxShaderStorageBlocks = 12;
-   ctx->Const.MaxCombinedShaderStorageBlocks = 12 * 3;
-   ctx->Const.MaxShaderStorageBufferBindings = 36;
+   ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxShaderStorageBlocks = BRW_MAX_SSBO;
+   ctx->Const.Program[MESA_SHADER_COMPUTE].MaxShaderStorageBlocks = BRW_MAX_SSBO;
+   ctx->Const.MaxCombinedShaderStorageBlocks = BRW_MAX_SSBO * 3;
+   ctx->Const.MaxShaderStorageBufferBindings = BRW_MAX_SSBO * 3;
 
    if (_mesa_extension_override_enables.ARB_compute_shader)
-      ctx->Const.MaxShaderStorageBufferBindings += 12;
+      ctx->Const.MaxShaderStorageBufferBindings += BRW_MAX_SSBO;
 
    if (brw->gen >= 6) {
       ctx->Const.MaxVarying = 32;
@@ -1298,7 +1299,7 @@ intel_prepare_render(struct brw_context *brw)
     * that will happen next will probably dirty the front buffer.  So
     * mark it as dirty here.
     */
-   if (brw_is_front_buffer_drawing(ctx->DrawBuffer))
+   if (_mesa_is_front_buffer_drawing(ctx->DrawBuffer))
       brw->front_buffer_dirty = true;
 }
 
@@ -1337,8 +1338,8 @@ intel_query_dri2_buffers(struct brw_context *brw,
    back_rb = intel_get_renderbuffer(fb, BUFFER_BACK_LEFT);
 
    memset(attachments, 0, sizeof(attachments));
-   if ((brw_is_front_buffer_drawing(fb) ||
-        brw_is_front_buffer_reading(fb) ||
+   if ((_mesa_is_front_buffer_drawing(fb) ||
+        _mesa_is_front_buffer_reading(fb) ||
         !back_rb) && front_rb) {
       /* If a fake front buffer is in use, then querying for
        * __DRI_BUFFER_FRONT_LEFT will cause the server to copy the image from
@@ -1452,7 +1453,7 @@ intel_process_dri2_buffer(struct brw_context *brw,
                                             drawable->w, drawable->h,
                                             buffer->pitch);
 
-   if (brw_is_front_buffer_drawing(fb) &&
+   if (_mesa_is_front_buffer_drawing(fb) &&
        (buffer->attachment == __DRI_BUFFER_FRONT_LEFT ||
         buffer->attachment == __DRI_BUFFER_FAKE_FRONT_LEFT) &&
        rb->Base.Base.NumSamples > 1) {
@@ -1510,7 +1511,7 @@ intel_update_image_buffer(struct brw_context *intel,
                                             buffer->width, buffer->height,
                                             buffer->pitch);
 
-   if (brw_is_front_buffer_drawing(fb) &&
+   if (_mesa_is_front_buffer_drawing(fb) &&
        buffer_type == __DRI_IMAGE_BUFFER_FRONT &&
        rb->Base.Base.NumSamples > 1) {
       intel_renderbuffer_upsample(intel, rb);
@@ -1538,8 +1539,8 @@ intel_update_image_buffers(struct brw_context *brw, __DRIdrawable *drawable)
    else
       return;
 
-   if (front_rb && (brw_is_front_buffer_drawing(fb) ||
-                    brw_is_front_buffer_reading(fb) || !back_rb)) {
+   if (front_rb && (_mesa_is_front_buffer_drawing(fb) ||
+                    _mesa_is_front_buffer_reading(fb) || !back_rb)) {
       buffer_mask |= __DRI_IMAGE_BUFFER_FRONT;
    }
 
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index d921a9bb7dd..a66c61e58f3 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -490,6 +490,7 @@ struct brw_cs_prog_data {
    unsigned simd_size;
    bool uses_barrier;
    bool uses_num_work_groups;
+   unsigned local_invocation_id_regs;
 
    struct {
       /** @{
@@ -714,6 +715,15 @@ struct brw_vs_prog_data {
 /** Max number of render targets in a shader */
 #define BRW_MAX_DRAW_BUFFERS 8
 
+/** Max number of UBOs in a shader */
+#define BRW_MAX_UBO 12
+
+/** Max number of SSBOs in a shader */
+#define BRW_MAX_SSBO 12
+
+/** Max number of combined UBOs and SSBOs in a shader */
+#define BRW_MAX_COMBINED_UBO_SSBO (BRW_MAX_UBO + BRW_MAX_SSBO)
+
 /** Max number of atomic counter buffer objects in a shader */
 #define BRW_MAX_ABO 16
 
@@ -750,7 +760,8 @@ struct brw_vs_prog_data {
 
 #define BRW_MAX_SURFACES   (BRW_MAX_DRAW_BUFFERS +                      \
                             BRW_MAX_TEX_UNIT * 2 + /* normal, gather */ \
-                            12 + /* ubo */                              \
+                            BRW_MAX_UBO +                               \
+                            BRW_MAX_SSBO +                              \
                             BRW_MAX_ABO +                               \
                             BRW_MAX_IMAGES +                            \
                             2 + /* shader time, pull constants */       \
@@ -1453,6 +1464,8 @@ struct brw_context
        */
       drm_intel_bo *multisampled_null_render_target_bo;
       uint32_t fast_clear_op;
+
+      float offset_clamp;
    } wm;
 
    struct {
@@ -1716,7 +1729,12 @@ void brw_validate_textures( struct brw_context *brw );
  */
 void brwInitFragProgFuncs( struct dd_function_table *functions );
 
-int brw_get_scratch_size(int size);
+/* Per-thread scratch space is a power-of-two multiple of 1KB. */
+static inline int
+brw_get_scratch_size(int size)
+{
+   return util_next_power_of_two(size | 1023);
+}
 void brw_get_scratch_bo(struct brw_context *brw,
 			drm_intel_bo **scratch_bo, int size);
 void brw_init_shader_time(struct brw_context *brw);
diff --git a/src/mesa/drivers/dri/i965/brw_cs.c b/src/mesa/drivers/dri/i965/brw_cs.c
index 6b64030a868..45fb816c160 100644
--- a/src/mesa/drivers/dri/i965/brw_cs.c
+++ b/src/mesa/drivers/dri/i965/brw_cs.c
@@ -98,8 +98,15 @@ brw_codegen_cs_prog(struct brw_context *brw,
       start_time = get_time();
    }
 
+   if (unlikely(INTEL_DEBUG & DEBUG_CS))
+      brw_dump_ir("compute", prog, &cs->base, &cp->program.Base);
+
+   int st_index = -1;
+   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
+      st_index = brw_get_shader_time_index(brw, prog, &cp->program.Base, ST_CS);
+
    program = brw_cs_emit(brw, mem_ctx, key, &prog_data,
-                         &cp->program, prog, &program_size);
+                         &cp->program, prog, st_index, &program_size);
    if (program == NULL) {
       ralloc_free(mem_ctx);
       return false;
diff --git a/src/mesa/drivers/dri/i965/brw_cs.h b/src/mesa/drivers/dri/i965/brw_cs.h
index 0c0ed2bc909..17c2ff9871a 100644
--- a/src/mesa/drivers/dri/i965/brw_cs.h
+++ b/src/mesa/drivers/dri/i965/brw_cs.h
@@ -46,10 +46,12 @@ brw_cs_emit(struct brw_context *brw,
             struct brw_cs_prog_data *prog_data,
             struct gl_compute_program *cp,
             struct gl_shader_program *prog,
+            int shader_time_index,
             unsigned *final_assembly_size);
 
-unsigned
-brw_cs_prog_local_id_payload_dwords(unsigned dispatch_width);
+void
+brw_cs_fill_local_id_payload(const struct brw_cs_prog_data *cs_prog_data,
+                             void *buffer, uint32_t threads, uint32_t stride);
 
 #ifdef __cplusplus
 }
diff --git a/src/mesa/drivers/dri/i965/brw_draw.c b/src/mesa/drivers/dri/i965/brw_draw.c
index 6a75e067915..39a26b05201 100644
--- a/src/mesa/drivers/dri/i965/brw_draw.c
+++ b/src/mesa/drivers/dri/i965/brw_draw.c
@@ -33,6 +33,7 @@
 #include "main/enums.h"
 #include "main/macros.h"
 #include "main/transformfeedback.h"
+#include "main/framebuffer.h"
 #include "tnl/tnl.h"
 #include "vbo/vbo_context.h"
 #include "swrast/swrast.h"
@@ -55,34 +56,34 @@
 #define FILE_DEBUG_FLAG DEBUG_PRIMS
 
 static const GLuint prim_to_hw_prim[GL_TRIANGLE_STRIP_ADJACENCY+1] = {
-   _3DPRIM_POINTLIST,
-   _3DPRIM_LINELIST,
-   _3DPRIM_LINELOOP,
-   _3DPRIM_LINESTRIP,
-   _3DPRIM_TRILIST,
-   _3DPRIM_TRISTRIP,
-   _3DPRIM_TRIFAN,
-   _3DPRIM_QUADLIST,
-   _3DPRIM_QUADSTRIP,
-   _3DPRIM_POLYGON,
-   _3DPRIM_LINELIST_ADJ,
-   _3DPRIM_LINESTRIP_ADJ,
-   _3DPRIM_TRILIST_ADJ,
-   _3DPRIM_TRISTRIP_ADJ,
+   [GL_POINTS] =_3DPRIM_POINTLIST,
+   [GL_LINES] = _3DPRIM_LINELIST,
+   [GL_LINE_LOOP] = _3DPRIM_LINELOOP,
+   [GL_LINE_STRIP] = _3DPRIM_LINESTRIP,
+   [GL_TRIANGLES] = _3DPRIM_TRILIST,
+   [GL_TRIANGLE_STRIP] = _3DPRIM_TRISTRIP,
+   [GL_TRIANGLE_FAN] = _3DPRIM_TRIFAN,
+   [GL_QUADS] = _3DPRIM_QUADLIST,
+   [GL_QUAD_STRIP] = _3DPRIM_QUADSTRIP,
+   [GL_POLYGON] = _3DPRIM_POLYGON,
+   [GL_LINES_ADJACENCY] = _3DPRIM_LINELIST_ADJ,
+   [GL_LINE_STRIP_ADJACENCY] = _3DPRIM_LINESTRIP_ADJ,
+   [GL_TRIANGLES_ADJACENCY] = _3DPRIM_TRILIST_ADJ,
+   [GL_TRIANGLE_STRIP_ADJACENCY] = _3DPRIM_TRISTRIP_ADJ,
 };
 
 
 static const GLenum reduced_prim[GL_POLYGON+1] = {
-   GL_POINTS,
-   GL_LINES,
-   GL_LINES,
-   GL_LINES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES,
-   GL_TRIANGLES
+   [GL_POINTS] = GL_POINTS,
+   [GL_LINES] = GL_LINES,
+   [GL_LINE_LOOP] = GL_LINES,
+   [GL_LINE_STRIP] = GL_LINES,
+   [GL_TRIANGLES] = GL_TRIANGLES,
+   [GL_TRIANGLE_STRIP] = GL_TRIANGLES,
+   [GL_TRIANGLE_FAN] = GL_TRIANGLES,
+   [GL_QUADS] = GL_TRIANGLES,
+   [GL_QUAD_STRIP] = GL_TRIANGLES,
+   [GL_POLYGON] = GL_TRIANGLES
 };
 
 uint32_t
@@ -364,7 +365,7 @@ brw_postdraw_set_buffers_need_resolve(struct brw_context *brw)
    struct intel_renderbuffer *stencil_irb = intel_get_renderbuffer(fb, BUFFER_STENCIL);
    struct gl_renderbuffer_attachment *depth_att = &fb->Attachment[BUFFER_DEPTH];
 
-   if (brw_is_front_buffer_drawing(fb))
+   if (_mesa_is_front_buffer_drawing(fb))
       front_irb = intel_get_renderbuffer(fb, BUFFER_FRONT_LEFT);
 
    if (front_irb)
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
index 0a6b23284d9..5049851c617 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -210,7 +210,7 @@ fs_visitor::VARYING_PULL_CONSTANT_LOAD(const fs_builder &bld,
    inst->regs_written = regs_written;
 
    if (devinfo->gen < 7) {
-      inst->base_mrf = 13;
+      inst->base_mrf = FIRST_PULL_LOAD_MRF(devinfo->gen);
       inst->header_size = 1;
       if (devinfo->gen == 4)
          inst->mlen = 3;
@@ -2698,7 +2698,7 @@ fs_visitor::emit_repclear_shader()
 bool
 fs_visitor::remove_duplicate_mrf_writes()
 {
-   fs_inst *last_mrf_move[16];
+   fs_inst *last_mrf_move[BRW_MAX_MRF(devinfo->gen)];
    bool progress = false;
 
    /* Need to update the MRF tracking for compressed instructions. */
@@ -3019,7 +3019,7 @@ fs_visitor::lower_uniform_pull_constant_loads()
           * else does except for register spill/unspill, which generates and
           * uses its MRF within a single IR instruction.
           */
-         inst->base_mrf = 14;
+         inst->base_mrf = FIRST_PULL_LOAD_MRF(devinfo->gen) + 1;
          inst->mlen = 1;
       }
    }
@@ -4738,20 +4738,43 @@ fs_visitor::setup_vs_payload()
    payload.num_regs = 2;
 }
 
+/**
+ * We are building the local ID push constant data using the simplest possible
+ * method. We simply push the local IDs directly as they should appear in the
+ * registers for the uvec3 gl_LocalInvocationID variable.
+ *
+ * Therefore, for SIMD8, we use 3 full registers, and for SIMD16 we use 6
+ * registers worth of push constant space.
+ *
+ * Note: Any updates to brw_cs_prog_local_id_payload_dwords,
+ * fill_local_id_payload or fs_visitor::emit_cs_local_invocation_id_setup need
+ * to coordinated.
+ *
+ * FINISHME: There are a few easy optimizations to consider.
+ *
+ * 1. If gl_WorkGroupSize x, y or z is 1, we can just use zero, and there is
+ *    no need for using push constant space for that dimension.
+ *
+ * 2. Since GL_MAX_COMPUTE_WORK_GROUP_SIZE is currently 1024 or less, we can
+ *    easily use 16-bit words rather than 32-bit dwords in the push constant
+ *    data.
+ *
+ * 3. If gl_WorkGroupSize x, y or z is small, then we can use bytes for
+ *    conveying the data, and thereby reduce push constant usage.
+ *
+ */
 void
 fs_visitor::setup_cs_payload()
 {
    assert(devinfo->gen >= 7);
+   brw_cs_prog_data *prog_data = (brw_cs_prog_data*) this->prog_data;
 
    payload.num_regs = 1;
 
    if (nir->info.system_values_read & SYSTEM_BIT_LOCAL_INVOCATION_ID) {
-      const unsigned local_id_dwords =
-         brw_cs_prog_local_id_payload_dwords(dispatch_width);
-      assert((local_id_dwords & 0x7) == 0);
-      const unsigned local_id_regs = local_id_dwords / 8;
+      prog_data->local_invocation_id_regs = dispatch_width * 3 / 8;
       payload.local_invocation_id_reg = payload.num_regs;
-      payload.num_regs += local_id_regs;
+      payload.num_regs += prog_data->local_invocation_id_regs;
    }
 }
 
@@ -4843,7 +4866,7 @@ fs_visitor::optimize()
       OPT(opt_algebraic);
       OPT(opt_cse);
       OPT(opt_copy_propagate);
-      OPT(opt_peephole_predicated_break);
+      OPT(opt_predicated_break, this);
       OPT(opt_cmod_propagation);
       OPT(dead_code_eliminate);
       OPT(opt_peephole_sel);
@@ -5118,25 +5141,13 @@ brw_wm_fs_emit(struct brw_context *brw,
                struct brw_wm_prog_data *prog_data,
                struct gl_fragment_program *fp,
                struct gl_shader_program *prog,
+               int shader_time_index8, int shader_time_index16,
                unsigned *final_assembly_size)
 {
-   struct brw_shader *shader = NULL;
-   if (prog)
-      shader = (brw_shader *) prog->_LinkedShaders[MESA_SHADER_FRAGMENT];
-
-   if (unlikely(INTEL_DEBUG & DEBUG_WM) && shader->base.ir)
-      brw_dump_ir("fragment", prog, &shader->base, &fp->Base);
-
-   int st_index8 = -1, st_index16 = -1;
-   if (INTEL_DEBUG & DEBUG_SHADER_TIME) {
-      st_index8 = brw_get_shader_time_index(brw, prog, &fp->Base, ST_FS8);
-      st_index16 = brw_get_shader_time_index(brw, prog, &fp->Base, ST_FS16);
-   }
-
    /* Now the main event: Visit the shader IR and generate our FS IR for it.
     */
    fs_visitor v(brw->intelScreen->compiler, brw, mem_ctx, key,
-                &prog_data->base, &fp->Base, fp->Base.nir, 8, st_index8);
+                &prog_data->base, &fp->Base, fp->Base.nir, 8, shader_time_index8);
    if (!v.run_fs(false /* do_rep_send */)) {
       if (prog) {
          prog->LinkStatus = false;
@@ -5151,7 +5162,7 @@ brw_wm_fs_emit(struct brw_context *brw,
 
    cfg_t *simd16_cfg = NULL;
    fs_visitor v2(brw->intelScreen->compiler, brw, mem_ctx, key,
-                 &prog_data->base, &fp->Base, fp->Base.nir, 16, st_index16);
+                 &prog_data->base, &fp->Base, fp->Base.nir, 16, shader_time_index16);
    if (likely(!(INTEL_DEBUG & DEBUG_NO16) || brw->use_rep_send)) {
       if (!v.simd16_unsupported) {
          /* Try a SIMD16 compile */
@@ -5198,6 +5209,42 @@ brw_wm_fs_emit(struct brw_context *brw,
    return g.get_assembly(final_assembly_size);
 }
 
+void
+brw_cs_fill_local_id_payload(const struct brw_cs_prog_data *prog_data,
+                             void *buffer, uint32_t threads, uint32_t stride)
+{
+   if (prog_data->local_invocation_id_regs == 0)
+      return;
+
+   /* 'stride' should be an integer number of registers, that is, a multiple
+    * of 32 bytes.
+    */
+   assert(stride % 32 == 0);
+
+   unsigned x = 0, y = 0, z = 0;
+   for (unsigned t = 0; t < threads; t++) {
+      uint32_t *param = (uint32_t *) buffer + stride * t / 4;
+
+      for (unsigned i = 0; i < prog_data->simd_size; i++) {
+         param[0 * prog_data->simd_size + i] = x;
+         param[1 * prog_data->simd_size + i] = y;
+         param[2 * prog_data->simd_size + i] = z;
+
+         x++;
+         if (x == prog_data->local_size[0]) {
+            x = 0;
+            y++;
+            if (y == prog_data->local_size[1]) {
+               y = 0;
+               z++;
+               if (z == prog_data->local_size[2])
+                  z = 0;
+            }
+         }
+      }
+   }
+}
+
 fs_reg *
 fs_visitor::emit_cs_local_invocation_id_setup()
 {
@@ -5242,43 +5289,35 @@ brw_cs_emit(struct brw_context *brw,
             struct brw_cs_prog_data *prog_data,
             struct gl_compute_program *cp,
             struct gl_shader_program *prog,
+            int shader_time_index,
             unsigned *final_assembly_size)
 {
-   struct brw_shader *shader =
-      (struct brw_shader *) prog->_LinkedShaders[MESA_SHADER_COMPUTE];
-
-   if (unlikely(INTEL_DEBUG & DEBUG_CS))
-      brw_dump_ir("compute", prog, &shader->base, &cp->Base);
-
    prog_data->local_size[0] = cp->LocalSize[0];
    prog_data->local_size[1] = cp->LocalSize[1];
    prog_data->local_size[2] = cp->LocalSize[2];
    unsigned local_workgroup_size =
       cp->LocalSize[0] * cp->LocalSize[1] * cp->LocalSize[2];
+   unsigned max_cs_threads = brw->intelScreen->compiler->devinfo->max_cs_threads;
 
    cfg_t *cfg = NULL;
    const char *fail_msg = NULL;
 
-   int st_index = -1;
-   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
-      st_index = brw_get_shader_time_index(brw, prog, &cp->Base, ST_CS);
-
    /* Now the main event: Visit the shader IR and generate our CS IR for it.
     */
    fs_visitor v8(brw->intelScreen->compiler, brw, mem_ctx, key,
-                 &prog_data->base, &cp->Base, cp->Base.nir, 8, st_index);
+                 &prog_data->base, &cp->Base, cp->Base.nir, 8, shader_time_index);
    if (!v8.run_cs()) {
       fail_msg = v8.fail_msg;
-   } else if (local_workgroup_size <= 8 * brw->max_cs_threads) {
+   } else if (local_workgroup_size <= 8 * max_cs_threads) {
       cfg = v8.cfg;
       prog_data->simd_size = 8;
    }
 
    fs_visitor v16(brw->intelScreen->compiler, brw, mem_ctx, key,
-                  &prog_data->base, &cp->Base, cp->Base.nir, 16, st_index);
+                  &prog_data->base, &cp->Base, cp->Base.nir, 16, shader_time_index);
    if (likely(!(INTEL_DEBUG & DEBUG_NO16)) &&
        !fail_msg && !v8.simd16_unsupported &&
-       local_workgroup_size <= 16 * brw->max_cs_threads) {
+       local_workgroup_size <= 16 * max_cs_threads) {
       /* Try a SIMD16 compile */
       v16.import_uniforms(&v8);
       if (!v16.run_cs()) {
diff --git a/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp b/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
index 6900cee86f4..c3a037be4b1 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
@@ -30,8 +30,6 @@
 #include "glsl/glsl_types.h"
 #include "glsl/ir_optimization.h"
 
-#define FIRST_SPILL_MRF(gen) (gen == 6 ? 21 : 13)
-
 using namespace brw;
 
 static void
diff --git a/src/mesa/drivers/dri/i965/brw_fs_sel_peephole.cpp b/src/mesa/drivers/dri/i965/brw_fs_sel_peephole.cpp
index d190d8eb6b4..8613725f6b9 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_sel_peephole.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_sel_peephole.cpp
@@ -155,18 +155,6 @@ fs_visitor::opt_peephole_sel()
       if (movs == 0)
          continue;
 
-      enum brw_predicate predicate;
-      bool predicate_inverse;
-      if (devinfo->gen == 6 && if_inst->conditional_mod) {
-         /* For Sandybridge with IF with embedded comparison */
-         predicate = BRW_PREDICATE_NORMAL;
-         predicate_inverse = false;
-      } else {
-         /* Separate CMP and IF instructions */
-         predicate = if_inst->predicate;
-         predicate_inverse = if_inst->predicate_inverse;
-      }
-
       /* Generate SEL instructions for pairs of MOVs to a common destination. */
       for (int i = 0; i < movs; i++) {
          if (!then_mov[i] || !else_mov[i])
@@ -195,13 +183,6 @@ fs_visitor::opt_peephole_sel()
       if (movs == 0)
          continue;
 
-      /* Emit a CMP if our IF used the embedded comparison */
-      if (devinfo->gen == 6 && if_inst->conditional_mod) {
-         const fs_builder ibld(this, block, if_inst);
-         ibld.CMP(ibld.null_reg_d(), if_inst->src[0], if_inst->src[1],
-                  if_inst->conditional_mod);
-      }
-
       for (int i = 0; i < movs; i++) {
          const fs_builder ibld = fs_builder(this, then_block, then_mov[i])
                                  .at(block, if_inst);
@@ -220,7 +201,7 @@ fs_visitor::opt_peephole_sel()
                ibld.MOV(src0, then_mov[i]->src[0]);
             }
 
-            set_predicate_inv(predicate, predicate_inverse,
+            set_predicate_inv(if_inst->predicate, if_inst->predicate_inverse,
                               ibld.SEL(then_mov[i]->dst, src0,
                                        else_mov[i]->src[0]));
          }
diff --git a/src/mesa/drivers/dri/i965/brw_gs.c b/src/mesa/drivers/dri/i965/brw_gs.c
index 4d0b125ffe4..e0165fb4a23 100644
--- a/src/mesa/drivers/dri/i965/brw_gs.c
+++ b/src/mesa/drivers/dri/i965/brw_gs.c
@@ -52,22 +52,17 @@ assign_gs_binding_table_offsets(const struct brw_device_info *devinfo,
 }
 
 bool
-brw_compile_gs_prog(struct brw_context *brw,
+brw_codegen_gs_prog(struct brw_context *brw,
                     struct gl_shader_program *prog,
                     struct brw_geometry_program *gp,
-                    struct brw_gs_prog_key *key,
-                    struct brw_gs_compile_output *output)
+                    struct brw_gs_prog_key *key)
 {
+   struct brw_stage_state *stage_state = &brw->gs.base;
    struct brw_gs_compile c;
    memset(&c, 0, sizeof(c));
    c.key = *key;
    c.gp = gp;
 
-   /* We get the bind map as input in the output struct...*/
-   c.prog_data.base.base.map_entries = output->prog_data.base.base.map_entries;
-   memcpy(c.prog_data.base.base.bind_map, output->prog_data.base.base.bind_map,
-          sizeof(c.prog_data.base.base.bind_map));
-
    c.prog_data.include_primitive_id =
       (gp->program.Base.InputsRead & VARYING_BIT_PRIMITIVE_ID) != 0;
 
@@ -296,48 +291,35 @@ brw_compile_gs_prog(struct brw_context *brw,
     */
    c.prog_data.base.urb_read_length = (c.input_vue_map.num_slots + 1) / 2;
 
+   if (unlikely(INTEL_DEBUG & DEBUG_GS))
+      brw_dump_ir("geometry", prog, gs, NULL);
+
+   int st_index = -1;
+   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
+      st_index = brw_get_shader_time_index(brw, prog, NULL, ST_GS);
+
    void *mem_ctx = ralloc_context(NULL);
    unsigned program_size;
    const unsigned *program =
-      brw_gs_emit(brw, prog, &c, mem_ctx, &program_size);
+      brw_gs_emit(brw, prog, &c, mem_ctx, st_index, &program_size);
    if (program == NULL) {
       ralloc_free(mem_ctx);
       return false;
    }
 
-   output->mem_ctx = mem_ctx;
-   output->program = program;
-   output->program_size = program_size;
-   memcpy(&output->prog_data, &c.prog_data,
-          sizeof(output->prog_data));
-
-   return true;
-}
-
-bool
-brw_codegen_gs_prog(struct brw_context *brw,
-                    struct gl_shader_program *prog,
-                    struct brw_geometry_program *gp,
-                    struct brw_gs_prog_key *key)
-{
-   struct brw_gs_compile_output output;
-   struct brw_stage_state *stage_state = &brw->gs.base;
-
-   if (brw_compile_gs_prog(brw, prog, gp, key, &output))
-      return false;
-
-   if (output.prog_data.base.base.total_scratch) {
+   /* Scratch space is used for register spilling */
+   if (c.prog_data.base.base.total_scratch) {
       brw_get_scratch_bo(brw, &stage_state->scratch_bo,
-			 output.prog_data.base.base.total_scratch *
+			 c.prog_data.base.base.total_scratch *
                          brw->max_gs_threads);
    }
 
    brw_upload_cache(&brw->cache, BRW_CACHE_GS_PROG,
-                    key, sizeof(*key),
-                    output.program, output.program_size,
-                    &output.prog_data, sizeof(output.prog_data),
+                    &c.key, sizeof(c.key),
+                    program, program_size,
+                    &c.prog_data, sizeof(c.prog_data),
                     &stage_state->prog_offset, &brw->gs.prog_data);
-   ralloc_free(output.mem_ctx);
+   ralloc_free(mem_ctx);
 
    return true;
 }
diff --git a/src/mesa/drivers/dri/i965/brw_gs.h b/src/mesa/drivers/dri/i965/brw_gs.h
index 573bbdb16f8..9879f3d6464 100644
--- a/src/mesa/drivers/dri/i965/brw_gs.h
+++ b/src/mesa/drivers/dri/i965/brw_gs.h
@@ -37,24 +37,6 @@ struct gl_context;
 struct gl_shader_program;
 struct gl_program;
 
-struct brw_gs_compile_output {
-   void *mem_ctx;
-   const void *program;
-   uint32_t program_size;
-   struct brw_gs_prog_data prog_data;
-};
-
-struct brw_gs_prog_key;
-
-bool
-brw_compile_gs_prog(struct brw_context *brw,
-                    struct gl_shader_program *prog,
-                    struct brw_geometry_program *gp,
-                    struct brw_gs_prog_key *key,
-                    struct brw_gs_compile_output *output);
-
-bool brw_gs_prog_data_compare(const void *a, const void *b);
-
 void
 brw_upload_gs_prog(struct brw_context *brw);
 
diff --git a/src/mesa/drivers/dri/i965/brw_inst.h b/src/mesa/drivers/dri/i965/brw_inst.h
index c5132ba15ed..ab37b709d65 100644
--- a/src/mesa/drivers/dri/i965/brw_inst.h
+++ b/src/mesa/drivers/dri/i965/brw_inst.h
@@ -42,6 +42,12 @@ extern "C" {
 /** Maximum SEND message length */
 #define BRW_MAX_MSG_LENGTH 15
 
+/** First MRF register used by pull loads */
+#define FIRST_SPILL_MRF(gen) ((gen) == 6 ? 21 : 13)
+
+/** First MRF register used by spills */
+#define FIRST_PULL_LOAD_MRF(gen) ((gen) == 6 ? 16 : 13)
+
 /* brw_context.h has a forward declaration of brw_inst, so name the struct. */
 typedef struct brw_inst {
    uint64_t data[2];
diff --git a/src/mesa/drivers/dri/i965/brw_link.cpp b/src/mesa/drivers/dri/i965/brw_link.cpp
new file mode 100644
index 00000000000..fc9bee43d80
--- /dev/null
+++ b/src/mesa/drivers/dri/i965/brw_link.cpp
@@ -0,0 +1,280 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "main/macros.h"
+#include "brw_context.h"
+#include "brw_vs.h"
+#include "brw_gs.h"
+#include "brw_fs.h"
+#include "brw_cfg.h"
+#include "brw_nir.h"
+#include "glsl/ir_optimization.h"
+#include "glsl/glsl_parser_extras.h"
+#include "main/shaderapi.h"
+
+/**
+ * Performs a compile of the shader stages even when we don't know
+ * what non-orthogonal state will be set, in the hope that it reflects
+ * the eventual NOS used, and thus allows us to produce link failures.
+ */
+static bool
+brw_shader_precompile(struct gl_context *ctx,
+                      struct gl_shader_program *sh_prog)
+{
+   struct gl_shader *vs = sh_prog->_LinkedShaders[MESA_SHADER_VERTEX];
+   struct gl_shader *gs = sh_prog->_LinkedShaders[MESA_SHADER_GEOMETRY];
+   struct gl_shader *fs = sh_prog->_LinkedShaders[MESA_SHADER_FRAGMENT];
+   struct gl_shader *cs = sh_prog->_LinkedShaders[MESA_SHADER_COMPUTE];
+
+   if (fs && !brw_fs_precompile(ctx, sh_prog, fs->Program))
+      return false;
+
+   if (gs && !brw_gs_precompile(ctx, sh_prog, gs->Program))
+      return false;
+
+   if (vs && !brw_vs_precompile(ctx, sh_prog, vs->Program))
+      return false;
+
+   if (cs && !brw_cs_precompile(ctx, sh_prog, cs->Program))
+      return false;
+
+   return true;
+}
+
+static void
+brw_lower_packing_builtins(struct brw_context *brw,
+                           gl_shader_stage shader_type,
+                           exec_list *ir)
+{
+   int ops = LOWER_PACK_SNORM_2x16
+           | LOWER_UNPACK_SNORM_2x16
+           | LOWER_PACK_UNORM_2x16
+           | LOWER_UNPACK_UNORM_2x16;
+
+   if (is_scalar_shader_stage(brw->intelScreen->compiler, shader_type)) {
+      ops |= LOWER_UNPACK_UNORM_4x8
+           | LOWER_UNPACK_SNORM_4x8
+           | LOWER_PACK_UNORM_4x8
+           | LOWER_PACK_SNORM_4x8;
+   }
+
+   if (brw->gen >= 7) {
+      /* Gen7 introduced the f32to16 and f16to32 instructions, which can be
+       * used to execute packHalf2x16 and unpackHalf2x16. For AOS code, no
+       * lowering is needed. For SOA code, the Half2x16 ops must be
+       * scalarized.
+       */
+      if (is_scalar_shader_stage(brw->intelScreen->compiler, shader_type)) {
+         ops |= LOWER_PACK_HALF_2x16_TO_SPLIT
+             |  LOWER_UNPACK_HALF_2x16_TO_SPLIT;
+      }
+   } else {
+      ops |= LOWER_PACK_HALF_2x16
+          |  LOWER_UNPACK_HALF_2x16;
+   }
+
+   lower_packing_builtins(ir, ops);
+}
+
+static void
+process_glsl_ir(gl_shader_stage stage,
+                struct brw_context *brw,
+                struct gl_shader_program *shader_prog,
+                struct gl_shader *shader)
+{
+   struct gl_context *ctx = &brw->ctx;
+   const struct gl_shader_compiler_options *options =
+      &ctx->Const.ShaderCompilerOptions[shader->Stage];
+
+   /* Temporary memory context for any new IR. */
+   void *mem_ctx = ralloc_context(NULL);
+
+   ralloc_adopt(mem_ctx, shader->ir);
+
+   /* lower_packing_builtins() inserts arithmetic instructions, so it
+    * must precede lower_instructions().
+    */
+   brw_lower_packing_builtins(brw, shader->Stage, shader->ir);
+   do_mat_op_to_vec(shader->ir);
+   const int bitfield_insert = brw->gen >= 7 ? BITFIELD_INSERT_TO_BFM_BFI : 0;
+   lower_instructions(shader->ir,
+                      MOD_TO_FLOOR |
+                      DIV_TO_MUL_RCP |
+                      SUB_TO_ADD_NEG |
+                      EXP_TO_EXP2 |
+                      LOG_TO_LOG2 |
+                      bitfield_insert |
+                      LDEXP_TO_ARITH |
+                      CARRY_TO_ARITH |
+                      BORROW_TO_ARITH);
+
+   /* Pre-gen6 HW can only nest if-statements 16 deep.  Beyond this,
+    * if-statements need to be flattened.
+    */
+   if (brw->gen < 6)
+      lower_if_to_cond_assign(shader->ir, 16);
+
+   do_lower_texture_projection(shader->ir);
+   brw_lower_texture_gradients(brw, shader->ir);
+   do_vec_index_to_cond_assign(shader->ir);
+   lower_vector_insert(shader->ir, true);
+   lower_offset_arrays(shader->ir);
+   brw_do_lower_unnormalized_offset(shader->ir);
+   lower_noise(shader->ir);
+   lower_quadop_vector(shader->ir, false);
+
+   bool lowered_variable_indexing =
+      lower_variable_index_to_cond_assign((gl_shader_stage)stage,
+                                          shader->ir,
+                                          options->EmitNoIndirectInput,
+                                          options->EmitNoIndirectOutput,
+                                          options->EmitNoIndirectTemp,
+                                          options->EmitNoIndirectUniform);
+
+   if (unlikely(brw->perf_debug && lowered_variable_indexing)) {
+      perf_debug("Unsupported form of variable indexing in %s; falling "
+                 "back to very inefficient code generation\n",
+                 _mesa_shader_stage_to_abbrev(shader->Stage));
+   }
+
+   lower_ubo_reference(shader, shader->ir);
+
+   bool progress;
+   do {
+      progress = false;
+
+      if (is_scalar_shader_stage(brw->intelScreen->compiler, shader->Stage)) {
+         brw_do_channel_expressions(shader->ir);
+         brw_do_vector_splitting(shader->ir);
+      }
+
+      progress = do_lower_jumps(shader->ir, true, true,
+                                true, /* main return */
+                                false, /* continue */
+                                false /* loops */
+                                ) || progress;
+
+      progress = do_common_optimization(shader->ir, true, true,
+                                        options, ctx->Const.NativeIntegers) || progress;
+   } while (progress);
+
+   validate_ir_tree(shader->ir);
+
+   /* Now that we've finished altering the linked IR, reparent any live IR back
+    * to the permanent memory context, and free the temporary one (discarding any
+    * junk we optimized away).
+    */
+   reparent_ir(shader->ir, shader->ir);
+   ralloc_free(mem_ctx);
+
+   if (ctx->_Shader->Flags & GLSL_DUMP) {
+      fprintf(stderr, "\n");
+      fprintf(stderr, "GLSL IR for linked %s program %d:\n",
+              _mesa_shader_stage_to_string(shader->Stage),
+              shader_prog->Name);
+      _mesa_print_ir(stderr, shader->ir, NULL);
+      fprintf(stderr, "\n");
+   }
+}
+
+GLboolean
+brw_link_shader(struct gl_context *ctx, struct gl_shader_program *shProg)
+{
+   struct brw_context *brw = brw_context(ctx);
+   const struct brw_compiler *compiler = brw->intelScreen->compiler;
+   unsigned int stage;
+
+   for (stage = 0; stage < ARRAY_SIZE(shProg->_LinkedShaders); stage++) {
+      struct gl_shader *shader = shProg->_LinkedShaders[stage];
+      if (!shader)
+	 continue;
+
+      struct gl_program *prog =
+	 ctx->Driver.NewProgram(ctx, _mesa_shader_stage_to_program(stage),
+                                shader->Name);
+      if (!prog)
+	return false;
+      prog->Parameters = _mesa_new_parameter_list();
+
+      _mesa_copy_linked_program_data((gl_shader_stage) stage, shProg, prog);
+
+      process_glsl_ir((gl_shader_stage) stage, brw, shProg, shader);
+
+      /* Make a pass over the IR to add state references for any built-in
+       * uniforms that are used.  This has to be done now (during linking).
+       * Code generation doesn't happen until the first time this shader is
+       * used for rendering.  Waiting until then to generate the parameters is
+       * too late.  At that point, the values for the built-in uniforms won't
+       * get sent to the shader.
+       */
+      foreach_in_list(ir_instruction, node, shader->ir) {
+	 ir_variable *var = node->as_variable();
+
+	 if ((var == NULL) || (var->data.mode != ir_var_uniform)
+	     || (strncmp(var->name, "gl_", 3) != 0))
+	    continue;
+
+	 const ir_state_slot *const slots = var->get_state_slots();
+	 assert(slots != NULL);
+
+	 for (unsigned int i = 0; i < var->get_num_state_slots(); i++) {
+	    _mesa_add_state_reference(prog->Parameters,
+				      (gl_state_index *) slots[i].tokens);
+	 }
+      }
+
+      do_set_program_inouts(shader->ir, prog, shader->Stage);
+
+      prog->SamplersUsed = shader->active_samplers;
+      prog->ShadowSamplers = shader->shadow_samplers;
+      _mesa_update_shader_textures_used(shProg, prog);
+
+      _mesa_reference_program(ctx, &shader->Program, prog);
+
+      brw_add_texrect_params(prog);
+
+      prog->nir = brw_create_nir(brw, shProg, prog, (gl_shader_stage) stage,
+                                 is_scalar_shader_stage(compiler, stage));
+
+      _mesa_reference_program(ctx, &prog, NULL);
+   }
+
+   if ((ctx->_Shader->Flags & GLSL_DUMP) && shProg->Name != 0) {
+      for (unsigned i = 0; i < shProg->NumShaders; i++) {
+         const struct gl_shader *sh = shProg->Shaders[i];
+         if (!sh)
+            continue;
+
+         fprintf(stderr, "GLSL %s shader %d source for linked program %d:\n",
+                 _mesa_shader_stage_to_string(sh->Stage),
+                 i, shProg->Name);
+         fprintf(stderr, "%s", sh->Source);
+         fprintf(stderr, "\n");
+      }
+   }
+
+   if (brw->precompile && !brw_shader_precompile(ctx, shProg))
+      return false;
+
+   return true;
+}
diff --git a/src/mesa/drivers/dri/i965/brw_misc_state.c b/src/mesa/drivers/dri/i965/brw_misc_state.c
index 7d17edb9023..cf6ba5b4aeb 100644
--- a/src/mesa/drivers/dri/i965/brw_misc_state.c
+++ b/src/mesa/drivers/dri/i965/brw_misc_state.c
@@ -887,14 +887,6 @@ brw_upload_invariant_state(struct brw_context *brw)
    brw_emit_select_pipeline(brw, BRW_RENDER_PIPELINE);
    brw->last_pipeline = BRW_RENDER_PIPELINE;
 
-   if (brw->gen < 6) {
-      /* Disable depth offset clamping. */
-      BEGIN_BATCH(2);
-      OUT_BATCH(_3DSTATE_GLOBAL_DEPTH_OFFSET_CLAMP << 16 | (2 - 2));
-      OUT_BATCH_F(0.0);
-      ADVANCE_BATCH();
-   }
-
    if (brw->gen >= 8) {
       BEGIN_BATCH(3);
       OUT_BATCH(CMD_STATE_SIP << 16 | (3 - 2));
diff --git a/src/mesa/drivers/dri/i965/brw_nir.c b/src/mesa/drivers/dri/i965/brw_nir.c
index 7ab6afa8458..19206600e64 100644
--- a/src/mesa/drivers/dri/i965/brw_nir.c
+++ b/src/mesa/drivers/dri/i965/brw_nir.c
@@ -30,8 +30,17 @@
 static void
 brw_nir_lower_inputs(nir_shader *nir, bool is_scalar)
 {
-   nir_assign_var_locations(&nir->inputs, &nir->num_inputs,
-                            is_scalar ? type_size_scalar : type_size_vec4);
+   switch (nir->stage) {
+   case MESA_SHADER_GEOMETRY:
+      foreach_list_typed(nir_variable, var, node, &nir->inputs) {
+         var->data.driver_location = var->data.location;
+      }
+      break;
+   default:
+      nir_assign_var_locations(&nir->inputs, &nir->num_inputs,
+                               is_scalar ? type_size_scalar : type_size_vec4);
+      break;
+   }
 }
 
 static void
diff --git a/src/mesa/drivers/dri/i965/brw_fs_peephole_predicated_break.cpp b/src/mesa/drivers/dri/i965/brw_predicated_break.cpp
index b75f40ba5a1..607715dace4 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_peephole_predicated_break.cpp
+++ b/src/mesa/drivers/dri/i965/brw_predicated_break.cpp
@@ -21,12 +21,11 @@
  * IN THE SOFTWARE.
  */
 
-#include "brw_fs.h"
 #include "brw_cfg.h"
 
 using namespace brw;
 
-/** @file brw_fs_peephole_predicated_break.cpp
+/** @file brw_predicated_break.cpp
  *
  * Loops are often structured as
  *
@@ -55,27 +54,27 @@ using namespace brw;
  */
 
 bool
-fs_visitor::opt_peephole_predicated_break()
+opt_predicated_break(backend_shader *s)
 {
    bool progress = false;
 
-   foreach_block (block, cfg) {
+   foreach_block (block, s->cfg) {
       if (block->start_ip != block->end_ip)
          continue;
 
       /* BREAK and CONTINUE instructions, by definition, can only be found at
        * the ends of basic blocks.
        */
-      fs_inst *jump_inst = (fs_inst *)block->end();
+      backend_instruction *jump_inst = block->end();
       if (jump_inst->opcode != BRW_OPCODE_BREAK &&
           jump_inst->opcode != BRW_OPCODE_CONTINUE)
          continue;
 
-      fs_inst *if_inst = (fs_inst *)block->prev()->end();
+      backend_instruction *if_inst = block->prev()->end();
       if (if_inst->opcode != BRW_OPCODE_IF)
          continue;
 
-      fs_inst *endif_inst = (fs_inst *)block->next()->start();
+      backend_instruction *endif_inst = block->next()->start();
       if (endif_inst->opcode != BRW_OPCODE_ENDIF)
          continue;
 
@@ -83,18 +82,8 @@ fs_visitor::opt_peephole_predicated_break()
       bblock_t *if_block = jump_block->prev();
       bblock_t *endif_block = jump_block->next();
 
-      /* For Sandybridge with IF with embedded comparison we need to emit an
-       * instruction to set the flag register.
-       */
-      if (devinfo->gen == 6 && if_inst->conditional_mod) {
-         const fs_builder ibld(this, if_block, if_inst);
-         ibld.CMP(ibld.null_reg_d(), if_inst->src[0], if_inst->src[1],
-                  if_inst->conditional_mod);
-         jump_inst->predicate = BRW_PREDICATE_NORMAL;
-      } else {
-         jump_inst->predicate = if_inst->predicate;
-         jump_inst->predicate_inverse = if_inst->predicate_inverse;
-      }
+      jump_inst->predicate = if_inst->predicate;
+      jump_inst->predicate_inverse = if_inst->predicate_inverse;
 
       bblock_t *earlier_block = if_block;
       if (if_block->start_ip == if_block->end_ip) {
@@ -111,13 +100,13 @@ fs_visitor::opt_peephole_predicated_break()
 
       if (!earlier_block->ends_with_control_flow()) {
          earlier_block->children.make_empty();
-         earlier_block->add_successor(cfg->mem_ctx, jump_block);
+         earlier_block->add_successor(s->cfg->mem_ctx, jump_block);
       }
 
       if (!later_block->starts_with_control_flow()) {
          later_block->parents.make_empty();
       }
-      jump_block->add_successor(cfg->mem_ctx, later_block);
+      jump_block->add_successor(s->cfg->mem_ctx, later_block);
 
       if (earlier_block->can_combine_with(jump_block)) {
          earlier_block->combine_with(jump_block);
@@ -130,7 +119,7 @@ fs_visitor::opt_peephole_predicated_break()
        * the two basic blocks.
        */
       bblock_t *while_block = earlier_block->next();
-      fs_inst *while_inst = (fs_inst *)while_block->start();
+      backend_instruction *while_inst = while_block->start();
 
       if (jump_inst->opcode == BRW_OPCODE_BREAK &&
           while_inst->opcode == BRW_OPCODE_WHILE &&
@@ -140,20 +129,20 @@ fs_visitor::opt_peephole_predicated_break()
          while_inst->predicate_inverse = !jump_inst->predicate_inverse;
 
          earlier_block->children.make_empty();
-         earlier_block->add_successor(cfg->mem_ctx, while_block);
+         earlier_block->add_successor(s->cfg->mem_ctx, while_block);
 
          assert(earlier_block->can_combine_with(while_block));
          earlier_block->combine_with(while_block);
 
          earlier_block->next()->parents.make_empty();
-         earlier_block->add_successor(cfg->mem_ctx, earlier_block->next());
+         earlier_block->add_successor(s->cfg->mem_ctx, earlier_block->next());
       }
 
       progress = true;
    }
 
    if (progress)
-      invalidate_live_intervals();
+      s->invalidate_live_intervals();
 
    return progress;
 }
diff --git a/src/mesa/drivers/dri/i965/brw_program.c b/src/mesa/drivers/dri/i965/brw_program.c
index 0a9a99edf2d..dbd0e50228b 100644
--- a/src/mesa/drivers/dri/i965/brw_program.c
+++ b/src/mesa/drivers/dri/i965/brw_program.c
@@ -242,18 +242,6 @@ brw_add_texrect_params(struct gl_program *prog)
    }
 }
 
-/* Per-thread scratch space is a power-of-two multiple of 1KB. */
-int
-brw_get_scratch_size(int size)
-{
-   int i;
-
-   for (i = 1024; i < size; i *= 2)
-      ;
-
-   return i;
-}
-
 void
 brw_get_scratch_bo(struct brw_context *brw,
 		   drm_intel_bo **scratch_bo, int size)
@@ -531,16 +519,6 @@ brw_destroy_shader_time(struct brw_context *brw)
 }
 
 void
-brw_mark_surface_used(struct brw_stage_prog_data *prog_data,
-                      unsigned surf_index)
-{
-   assert(surf_index < BRW_MAX_SURFACES);
-
-   prog_data->binding_table.size_bytes =
-      MAX2(prog_data->binding_table.size_bytes, (surf_index + 1) * 4);
-}
-
-void
 brw_stage_prog_data_free(const void *p)
 {
    struct brw_stage_prog_data *prog_data = (struct brw_stage_prog_data *)p;
diff --git a/src/mesa/drivers/dri/i965/brw_shader.cpp b/src/mesa/drivers/dri/i965/brw_shader.cpp
index 8bc7d9d6aac..3a58a58a00b 100644
--- a/src/mesa/drivers/dri/i965/brw_shader.cpp
+++ b/src/mesa/drivers/dri/i965/brw_shader.cpp
@@ -72,7 +72,7 @@ shader_perf_log_mesa(void *data, const char *fmt, ...)
    va_end(args);
 }
 
-static bool
+bool
 is_scalar_shader_stage(const struct brw_compiler *compiler, int stage)
 {
    switch (stage) {
@@ -166,254 +166,16 @@ brw_new_shader(struct gl_context *ctx, GLuint name, GLuint type)
    return &shader->base;
 }
 
-/**
- * Performs a compile of the shader stages even when we don't know
- * what non-orthogonal state will be set, in the hope that it reflects
- * the eventual NOS used, and thus allows us to produce link failures.
- */
-static bool
-brw_shader_precompile(struct gl_context *ctx,
-                      struct gl_shader_program *sh_prog)
-{
-   struct gl_shader *vs = sh_prog->_LinkedShaders[MESA_SHADER_VERTEX];
-   struct gl_shader *gs = sh_prog->_LinkedShaders[MESA_SHADER_GEOMETRY];
-   struct gl_shader *fs = sh_prog->_LinkedShaders[MESA_SHADER_FRAGMENT];
-   struct gl_shader *cs = sh_prog->_LinkedShaders[MESA_SHADER_COMPUTE];
-
-   if (fs && !brw_fs_precompile(ctx, sh_prog, fs->Program))
-      return false;
-
-   if (gs && !brw_gs_precompile(ctx, sh_prog, gs->Program))
-      return false;
-
-   if (vs && !brw_vs_precompile(ctx, sh_prog, vs->Program))
-      return false;
-
-   if (cs && !brw_cs_precompile(ctx, sh_prog, cs->Program))
-      return false;
-
-   return true;
-}
-
-static void
-brw_lower_packing_builtins(struct brw_context *brw,
-                           gl_shader_stage shader_type,
-                           exec_list *ir)
-{
-   int ops = LOWER_PACK_SNORM_2x16
-           | LOWER_UNPACK_SNORM_2x16
-           | LOWER_PACK_UNORM_2x16
-           | LOWER_UNPACK_UNORM_2x16;
-
-   if (is_scalar_shader_stage(brw->intelScreen->compiler, shader_type)) {
-      ops |= LOWER_UNPACK_UNORM_4x8
-           | LOWER_UNPACK_SNORM_4x8
-           | LOWER_PACK_UNORM_4x8
-           | LOWER_PACK_SNORM_4x8;
-   }
-
-   if (brw->gen >= 7) {
-      /* Gen7 introduced the f32to16 and f16to32 instructions, which can be
-       * used to execute packHalf2x16 and unpackHalf2x16. For AOS code, no
-       * lowering is needed. For SOA code, the Half2x16 ops must be
-       * scalarized.
-       */
-      if (is_scalar_shader_stage(brw->intelScreen->compiler, shader_type)) {
-         ops |= LOWER_PACK_HALF_2x16_TO_SPLIT
-             |  LOWER_UNPACK_HALF_2x16_TO_SPLIT;
-      }
-   } else {
-      ops |= LOWER_PACK_HALF_2x16
-          |  LOWER_UNPACK_HALF_2x16;
-   }
-
-   lower_packing_builtins(ir, ops);
-}
-
-static void
-process_glsl_ir(gl_shader_stage stage,
-                struct brw_context *brw,
-                struct gl_shader_program *shader_prog,
-                struct gl_shader *shader)
-{
-   struct gl_context *ctx = &brw->ctx;
-   const struct gl_shader_compiler_options *options =
-      &ctx->Const.ShaderCompilerOptions[shader->Stage];
-
-   /* Temporary memory context for any new IR. */
-   void *mem_ctx = ralloc_context(NULL);
-
-   ralloc_adopt(mem_ctx, shader->ir);
-
-   /* lower_packing_builtins() inserts arithmetic instructions, so it
-    * must precede lower_instructions().
-    */
-   brw_lower_packing_builtins(brw, shader->Stage, shader->ir);
-   do_mat_op_to_vec(shader->ir);
-   const int bitfield_insert = brw->gen >= 7 ? BITFIELD_INSERT_TO_BFM_BFI : 0;
-   lower_instructions(shader->ir,
-                      MOD_TO_FLOOR |
-                      DIV_TO_MUL_RCP |
-                      SUB_TO_ADD_NEG |
-                      EXP_TO_EXP2 |
-                      LOG_TO_LOG2 |
-                      bitfield_insert |
-                      LDEXP_TO_ARITH |
-                      CARRY_TO_ARITH |
-                      BORROW_TO_ARITH);
-
-   /* Pre-gen6 HW can only nest if-statements 16 deep.  Beyond this,
-    * if-statements need to be flattened.
-    */
-   if (brw->gen < 6)
-      lower_if_to_cond_assign(shader->ir, 16);
-
-   do_lower_texture_projection(shader->ir);
-   brw_lower_texture_gradients(brw, shader->ir);
-   do_vec_index_to_cond_assign(shader->ir);
-   lower_vector_insert(shader->ir, true);
-   lower_offset_arrays(shader->ir);
-   brw_do_lower_unnormalized_offset(shader->ir);
-   lower_noise(shader->ir);
-   lower_quadop_vector(shader->ir, false);
-
-   bool lowered_variable_indexing =
-      lower_variable_index_to_cond_assign((gl_shader_stage)stage,
-                                          shader->ir,
-                                          options->EmitNoIndirectInput,
-                                          options->EmitNoIndirectOutput,
-                                          options->EmitNoIndirectTemp,
-                                          options->EmitNoIndirectUniform);
-
-   if (unlikely(brw->perf_debug && lowered_variable_indexing)) {
-      perf_debug("Unsupported form of variable indexing in %s; falling "
-                 "back to very inefficient code generation\n",
-                 _mesa_shader_stage_to_abbrev(shader->Stage));
-   }
-
-   lower_ubo_reference(shader, shader->ir);
-
-   bool progress;
-   do {
-      progress = false;
-
-      if (is_scalar_shader_stage(brw->intelScreen->compiler, shader->Stage)) {
-         brw_do_channel_expressions(shader->ir);
-         brw_do_vector_splitting(shader->ir);
-      }
-
-      progress = do_lower_jumps(shader->ir, true, true,
-                                true, /* main return */
-                                false, /* continue */
-                                false /* loops */
-                                ) || progress;
-
-      progress = do_common_optimization(shader->ir, true, true,
-                                        options, ctx->Const.NativeIntegers) || progress;
-   } while (progress);
-
-   validate_ir_tree(shader->ir);
-
-   /* Now that we've finished altering the linked IR, reparent any live IR back
-    * to the permanent memory context, and free the temporary one (discarding any
-    * junk we optimized away).
-    */
-   reparent_ir(shader->ir, shader->ir);
-   ralloc_free(mem_ctx);
-
-   if (ctx->_Shader->Flags & GLSL_DUMP) {
-      fprintf(stderr, "\n");
-      fprintf(stderr, "GLSL IR for linked %s program %d:\n",
-              _mesa_shader_stage_to_string(shader->Stage),
-              shader_prog->Name);
-      _mesa_print_ir(stderr, shader->ir, NULL);
-      fprintf(stderr, "\n");
-   }
-}
-
-GLboolean
-brw_link_shader(struct gl_context *ctx, struct gl_shader_program *shProg)
+void
+brw_mark_surface_used(struct brw_stage_prog_data *prog_data,
+                      unsigned surf_index)
 {
-   struct brw_context *brw = brw_context(ctx);
-   const struct brw_compiler *compiler = brw->intelScreen->compiler;
-   unsigned int stage;
-
-   for (stage = 0; stage < ARRAY_SIZE(shProg->_LinkedShaders); stage++) {
-      struct gl_shader *shader = shProg->_LinkedShaders[stage];
-      if (!shader)
-	 continue;
-
-      struct gl_program *prog =
-	 ctx->Driver.NewProgram(ctx, _mesa_shader_stage_to_program(stage),
-                                shader->Name);
-      if (!prog)
-	return false;
-      prog->Parameters = _mesa_new_parameter_list();
-
-      _mesa_copy_linked_program_data((gl_shader_stage) stage, shProg, prog);
-
-      process_glsl_ir((gl_shader_stage) stage, brw, shProg, shader);
-
-      /* Make a pass over the IR to add state references for any built-in
-       * uniforms that are used.  This has to be done now (during linking).
-       * Code generation doesn't happen until the first time this shader is
-       * used for rendering.  Waiting until then to generate the parameters is
-       * too late.  At that point, the values for the built-in uniforms won't
-       * get sent to the shader.
-       */
-      foreach_in_list(ir_instruction, node, shader->ir) {
-	 ir_variable *var = node->as_variable();
-
-	 if ((var == NULL) || (var->data.mode != ir_var_uniform)
-	     || (strncmp(var->name, "gl_", 3) != 0))
-	    continue;
-
-	 const ir_state_slot *const slots = var->get_state_slots();
-	 assert(slots != NULL);
-
-	 for (unsigned int i = 0; i < var->get_num_state_slots(); i++) {
-	    _mesa_add_state_reference(prog->Parameters,
-				      (gl_state_index *) slots[i].tokens);
-	 }
-      }
-
-      do_set_program_inouts(shader->ir, prog, shader->Stage);
-
-      prog->SamplersUsed = shader->active_samplers;
-      prog->ShadowSamplers = shader->shadow_samplers;
-      _mesa_update_shader_textures_used(shProg, prog);
-
-      _mesa_reference_program(ctx, &shader->Program, prog);
-
-      brw_add_texrect_params(prog);
+   assert(surf_index < BRW_MAX_SURFACES);
 
-      prog->nir = brw_create_nir(brw, shProg, prog, (gl_shader_stage) stage,
-                                 is_scalar_shader_stage(compiler, stage));
-
-      _mesa_reference_program(ctx, &prog, NULL);
-   }
-
-   if ((ctx->_Shader->Flags & GLSL_DUMP) && shProg->Name != 0) {
-      for (unsigned i = 0; i < shProg->NumShaders; i++) {
-         const struct gl_shader *sh = shProg->Shaders[i];
-         if (!sh)
-            continue;
-
-         fprintf(stderr, "GLSL %s shader %d source for linked program %d:\n",
-                 _mesa_shader_stage_to_string(sh->Stage),
-                 i, shProg->Name);
-         fprintf(stderr, "%s", sh->Source);
-         fprintf(stderr, "\n");
-      }
-   }
-
-   if (brw->precompile && !brw_shader_precompile(ctx, shProg))
-      return false;
-
-   return true;
+   prog_data->binding_table.size_bytes =
+      MAX2(prog_data->binding_table.size_bytes, (surf_index + 1) * 4);
 }
 
-
 enum brw_reg_type
 brw_type_for_base_type(const struct glsl_type *type)
 {
@@ -1310,13 +1072,15 @@ backend_shader::dump_instructions(const char *name)
    if (cfg) {
       int ip = 0;
       foreach_block_and_inst(block, backend_instruction, inst, cfg) {
-         fprintf(file, "%4d: ", ip++);
+         if (!unlikely(INTEL_DEBUG & DEBUG_OPTIMIZER))
+            fprintf(file, "%4d: ", ip++);
          dump_instruction(inst, file);
       }
    } else {
       int ip = 0;
       foreach_in_list(backend_instruction, inst, &instructions) {
-         fprintf(file, "%4d: ", ip++);
+         if (!unlikely(INTEL_DEBUG & DEBUG_OPTIMIZER))
+            fprintf(file, "%4d: ", ip++);
          dump_instruction(inst, file);
       }
    }
@@ -1367,6 +1131,7 @@ brw_assign_common_binding_table_offsets(gl_shader_stage stage,
    next_binding_table_offset += num_textures;
 
    if (shader) {
+      assert(shader->NumUniformBlocks <= BRW_MAX_COMBINED_UBO_SSBO);
       stage_prog_data->binding_table.ubo_start = next_binding_table_offset;
       next_binding_table_offset += shader->NumUniformBlocks;
    } else {
@@ -1439,7 +1204,7 @@ brw_setup_image_uniform_values(gl_shader_stage stage,
       &stage_prog_data->param[param_start_index];
 
    for (unsigned i = 0; i < MAX2(storage->array_elements, 1); i++) {
-      const unsigned image_idx = storage->image[stage].index + i;
+      const unsigned image_idx = storage->opaque[stage].index + i;
       const brw_image_param *image_param =
          &stage_prog_data->image_param[image_idx];
 
@@ -1465,3 +1230,29 @@ brw_setup_image_uniform_values(gl_shader_stage stage,
          stage_prog_data->binding_table.image_start + image_idx);
    }
 }
+
+/**
+ * Decide which set of clip planes should be used when clipping via
+ * gl_Position or gl_ClipVertex.
+ */
+gl_clip_plane *brw_select_clip_planes(struct gl_context *ctx)
+{
+   if (ctx->_Shader->CurrentProgram[MESA_SHADER_VERTEX]) {
+      /* There is currently a GLSL vertex shader, so clip according to GLSL
+       * rules, which means compare gl_ClipVertex (or gl_Position, if
+       * gl_ClipVertex wasn't assigned) against the eye-coordinate clip planes
+       * that were stored in EyeUserPlane at the time the clip planes were
+       * specified.
+       */
+      return ctx->Transform.EyeUserPlane;
+   } else {
+      /* Either we are using fixed function or an ARB vertex program.  In
+       * either case the clip planes are going to be compared against
+       * gl_Position (which is in clip coordinates) so we have to clip using
+       * _ClipUserPlane, which was transformed into clip coordinates by Mesa
+       * core.
+       */
+      return ctx->Transform._ClipUserPlane;
+   }
+}
+
diff --git a/src/mesa/drivers/dri/i965/brw_shader.h b/src/mesa/drivers/dri/i965/brw_shader.h
index fd96740526b..ad2de5eae2d 100644
--- a/src/mesa/drivers/dri/i965/brw_shader.h
+++ b/src/mesa/drivers/dri/i965/brw_shader.h
@@ -219,7 +219,7 @@ enum instruction_scheduler_mode {
    SCHEDULE_POST,
 };
 
-class backend_shader {
+struct backend_shader {
 protected:
 
    backend_shader(const struct brw_compiler *compiler,
@@ -273,6 +273,8 @@ void brw_setup_image_uniform_values(gl_shader_stage stage,
                                     unsigned param_start_index,
                                     const gl_uniform_storage *storage);
 
+#else
+struct backend_shader;
 #endif /* __cplusplus */
 
 enum brw_reg_type brw_type_for_base_type(const struct glsl_type *type);
@@ -283,6 +285,8 @@ bool brw_saturate_immediate(enum brw_reg_type type, struct brw_reg *reg);
 bool brw_negate_immediate(enum brw_reg_type type, struct brw_reg *reg);
 bool brw_abs_immediate(enum brw_reg_type type, struct brw_reg *reg);
 
+bool opt_predicated_break(struct backend_shader *s);
+
 #ifdef __cplusplus
 extern "C" {
 #endif
@@ -314,6 +318,8 @@ bool brw_cs_precompile(struct gl_context *ctx,
 int type_size_scalar(const struct glsl_type *type);
 int type_size_vec4(const struct glsl_type *type);
 
+bool is_scalar_shader_stage(const struct brw_compiler *compiler, int stage);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/src/mesa/drivers/dri/i965/brw_vec4.cpp b/src/mesa/drivers/dri/i965/brw_vec4.cpp
index 689c767d2d7..e966b96a5ca 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4.cpp
@@ -1862,6 +1862,7 @@ vec4_visitor::run()
       pass_num = 0;
       iteration++;
 
+      OPT(opt_predicated_break, this);
       OPT(opt_reduce_swizzle);
       OPT(dead_code_eliminate);
       OPT(dead_control_flow_eliminate, this);
@@ -1942,28 +1943,18 @@ brw_vs_emit(struct brw_context *brw,
             struct brw_vs_prog_data *prog_data,
             struct gl_vertex_program *vp,
             struct gl_shader_program *prog,
+            int shader_time_index,
             unsigned *final_assembly_size)
 {
    const unsigned *assembly = NULL;
 
-   struct brw_shader *shader = NULL;
-   if (prog)
-      shader = (brw_shader *) prog->_LinkedShaders[MESA_SHADER_VERTEX];
-
-   int st_index = -1;
-   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
-      st_index = brw_get_shader_time_index(brw, prog, &vp->Base, ST_VS);
-
-   if (unlikely(INTEL_DEBUG & DEBUG_VS) && shader->base.ir)
-      brw_dump_ir("vertex", prog, &shader->base, &vp->Base);
-
    if (brw->intelScreen->compiler->scalar_vs) {
       prog_data->base.dispatch_mode = DISPATCH_MODE_SIMD8;
 
       fs_visitor v(brw->intelScreen->compiler, brw,
                    mem_ctx, key, &prog_data->base.base,
                    NULL, /* prog; Only used for TEXTURE_RECTANGLE on gen < 8 */
-                   vp->Base.nir, 8, st_index);
+                   vp->Base.nir, 8, shader_time_index);
       if (!v.run_vs(brw_select_clip_planes(&brw->ctx))) {
          if (prog) {
             prog->LinkStatus = false;
@@ -2001,7 +1992,7 @@ brw_vs_emit(struct brw_context *brw,
 
       vec4_vs_visitor v(brw->intelScreen->compiler, brw, key, prog_data,
                         vp->Base.nir, brw_select_clip_planes(&brw->ctx),
-                        mem_ctx, st_index,
+                        mem_ctx, shader_time_index,
                         !_mesa_is_gles3(&brw->ctx));
       if (!v.run()) {
          if (prog) {
diff --git a/src/mesa/drivers/dri/i965/brw_vec4.h b/src/mesa/drivers/dri/i965/brw_vec4.h
index 51b3161f659..5e3500c0c9a 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4.h
+++ b/src/mesa/drivers/dri/i965/brw_vec4.h
@@ -76,7 +76,7 @@ public:
 		void *mem_ctx,
                 bool no_spills,
                 int shader_time_index);
-   ~vec4_visitor();
+   virtual ~vec4_visitor();
 
    dst_reg dst_null_f()
    {
@@ -223,9 +223,6 @@ public:
 
    int implied_mrf_writes(vec4_instruction *inst);
 
-   void emit_vp_sop(enum brw_conditional_mod condmod, dst_reg dst,
-                    src_reg src0, src_reg src1, src_reg one);
-
    vec4_instruction *emit_minmax(enum brw_conditional_mod conditionalmod, dst_reg dst,
                                  src_reg src0, src_reg src1);
 
@@ -238,11 +235,6 @@ public:
     */
    src_reg emit_uniformize(const src_reg &src);
 
-   /**
-    * Emit the correct dot-product instruction for the type of arguments
-    */
-   void emit_dp(dst_reg dst, src_reg src0, src_reg src1, unsigned elements);
-
    src_reg fix_3src_operand(const src_reg &src);
    src_reg resolve_source_modifiers(const src_reg &src);
 
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_gs_nir.cpp b/src/mesa/drivers/dri/i965/brw_vec4_gs_nir.cpp
index af4c102c026..1b929b3df2c 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_gs_nir.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_gs_nir.cpp
@@ -29,41 +29,6 @@ namespace brw {
 void
 vec4_gs_visitor::nir_setup_inputs()
 {
-   nir_inputs = ralloc_array(mem_ctx, src_reg, nir->num_inputs);
-
-   foreach_list_typed(nir_variable, var, node, &nir->inputs) {
-      int offset = var->data.driver_location;
-      if (var->type->base_type == GLSL_TYPE_ARRAY) {
-         /* Geometry shader inputs are arrays, but they use an unusual array
-          * layout: instead of all array elements for a given geometry shader
-          * input being stored consecutively, all geometry shader inputs are
-          * interleaved into one giant array. At this stage of compilation, we
-          * assume that the stride of the array is BRW_VARYING_SLOT_COUNT.
-          * Later, setup_attributes() will remap our accesses to the actual
-          * input array.
-          */
-         assert(var->type->length > 0);
-         int length = var->type->length;
-         int size = type_size_vec4(var->type) / length;
-         for (int i = 0; i < length; i++) {
-            int location = var->data.location + i * BRW_VARYING_SLOT_COUNT;
-            for (int j = 0; j < size; j++) {
-               src_reg src = src_reg(ATTR, location + j, var->type);
-               src = retype(src, brw_type_for_base_type(var->type));
-               nir_inputs[offset] = src;
-               offset++;
-            }
-         }
-      } else {
-         int size = type_size_vec4(var->type);
-         for (int i = 0; i < size; i++) {
-            src_reg src = src_reg(ATTR, var->data.location + i, var->type);
-            src = retype(src, brw_type_for_base_type(var->type));
-            nir_inputs[offset] = src;
-            offset++;
-         }
-      }
-   }
 }
 
 void
@@ -96,6 +61,29 @@ vec4_gs_visitor::nir_emit_intrinsic(nir_intrinsic_instr *instr)
    src_reg src;
 
    switch (instr->intrinsic) {
+   case nir_intrinsic_load_per_vertex_input_indirect:
+      assert(!"EmitNoIndirectInput should prevent this.");
+   case nir_intrinsic_load_per_vertex_input: {
+      /* The EmitNoIndirectInput flag guarantees our vertex index will
+       * be constant.  We should handle indirects someday.
+       */
+      nir_const_value *vertex = nir_src_as_const_value(instr->src[0]);
+
+      /* Make up a type...we have no way of knowing... */
+      const glsl_type *const type = glsl_type::ivec(instr->num_components);
+
+      src = src_reg(ATTR, BRW_VARYING_SLOT_COUNT * vertex->u[0] +
+                          instr->const_index[0], type);
+      dest = get_nir_dest(instr->dest, src.type);
+      dest.writemask = brw_writemask_for_size(instr->num_components);
+      emit(MOV(dest, src));
+      break;
+   }
+
+   case nir_intrinsic_load_input:
+   case nir_intrinsic_load_input_indirect:
+      unreachable("nir_lower_io should have produced per_vertex intrinsics");
+
    case nir_intrinsic_emit_vertex_with_counter: {
       this->vertex_count =
          retype(get_nir_src(instr->src[0], 1), BRW_REGISTER_TYPE_UD);
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
index c673ccd137c..4ce471e0669 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
@@ -37,7 +37,6 @@ namespace brw {
 vec4_gs_visitor::vec4_gs_visitor(const struct brw_compiler *compiler,
                                  void *log_data,
                                  struct brw_gs_compile *c,
-                                 struct gl_shader_program *prog,
                                  nir_shader *shader,
                                  void *mem_ctx,
                                  bool no_spills,
@@ -45,7 +44,6 @@ vec4_gs_visitor::vec4_gs_visitor(const struct brw_compiler *compiler,
    : vec4_visitor(compiler, log_data, &c->key.tex,
                   &c->prog_data.base, shader,  mem_ctx,
                   no_spills, shader_time_index),
-     shader_prog(prog),
      c(c)
 {
 }
@@ -473,7 +471,7 @@ vec4_gs_visitor::gs_emit_vertex(int stream_id)
     * be recorded by transform feedback, we can simply discard all geometry
     * bound to these streams when transform feedback is disabled.
     */
-   if (stream_id > 0 && shader_prog->TransformFeedback.NumVarying == 0)
+   if (stream_id > 0 && !nir->info.has_transform_feedback_varyings)
       return;
 
    /* If we're outputting 32 control data bits or less, then we can wait
@@ -620,17 +618,11 @@ brw_gs_emit(struct brw_context *brw,
             struct gl_shader_program *prog,
             struct brw_gs_compile *c,
             void *mem_ctx,
+            int shader_time_index,
             unsigned *final_assembly_size)
 {
    struct gl_shader *shader = prog->_LinkedShaders[MESA_SHADER_GEOMETRY];
 
-   if (unlikely(INTEL_DEBUG & DEBUG_GS))
-      brw_dump_ir("geometry", prog, shader, NULL);
-
-   int st_index = -1;
-   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
-      st_index = brw_get_shader_time_index(brw, prog, NULL, ST_GS);
-
    if (brw->gen >= 7) {
       /* Compile the geometry shader in DUAL_OBJECT dispatch mode, if we can do
        * so without spilling. If the GS invocations count > 1, then we can't use
@@ -641,8 +633,8 @@ brw_gs_emit(struct brw_context *brw,
          c->prog_data.base.dispatch_mode = DISPATCH_MODE_4X2_DUAL_OBJECT;
 
          vec4_gs_visitor v(brw->intelScreen->compiler, brw,
-                           c, prog, shader->Program->nir,
-                           mem_ctx, true /* no_spills */, st_index);
+                           c, shader->Program->nir,
+                           mem_ctx, true /* no_spills */, shader_time_index);
          if (v.run()) {
             return generate_assembly(brw, prog, &c->gp->program.Base,
                                      &c->prog_data.base, mem_ctx, v.cfg,
@@ -684,14 +676,14 @@ brw_gs_emit(struct brw_context *brw,
 
    if (brw->gen >= 7)
       gs = new vec4_gs_visitor(brw->intelScreen->compiler, brw,
-                               c, prog, shader->Program->nir,
+                               c, shader->Program->nir,
                                mem_ctx, false /* no_spills */,
-                               st_index);
+                               shader_time_index);
    else
       gs = new gen6_gs_visitor(brw->intelScreen->compiler, brw,
                                c, prog, shader->Program->nir,
                                mem_ctx, false /* no_spills */,
-                               st_index);
+                               shader_time_index);
 
    if (!gs->run()) {
       prog->LinkStatus = false;
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.h b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.h
index 85d80b8fc63..3ff195c3e68 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.h
+++ b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.h
@@ -55,6 +55,7 @@ const unsigned *brw_gs_emit(struct brw_context *brw,
                             struct gl_shader_program *prog,
                             struct brw_gs_compile *c,
                             void *mem_ctx,
+                            int shader_time_index,
                             unsigned *final_assembly_size);
 
 #ifdef __cplusplus
@@ -70,7 +71,6 @@ public:
    vec4_gs_visitor(const struct brw_compiler *compiler,
                    void *log_data,
                    struct brw_gs_compile *c,
-                   struct gl_shader_program *prog,
                    nir_shader *shader,
                    void *mem_ctx,
                    bool no_spills,
@@ -97,8 +97,6 @@ protected:
    void emit_control_data_bits();
    void set_stream_control_data_bits(unsigned stream_id);
 
-   struct gl_shader_program *shader_prog;
-
    src_reg vertex_count;
    src_reg control_data_bits;
    const struct brw_gs_compile * const c;
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
index f80425a5b00..98ea9be6ee4 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
@@ -26,8 +26,6 @@
 #include "glsl/ir_uniform.h"
 #include "program/sampler.h"
 
-#define FIRST_SPILL_MRF(gen) (gen == 6 ? 21 : 13)
-
 namespace brw {
 
 vec4_instruction::vec4_instruction(enum opcode opcode, const dst_reg &dst,
@@ -278,16 +276,6 @@ vec4_visitor::SCRATCH_WRITE(const dst_reg &dst, const src_reg &src,
    return inst;
 }
 
-void
-vec4_visitor::emit_dp(dst_reg dst, src_reg src0, src_reg src1, unsigned elements)
-{
-   static enum opcode dot_opcodes[] = {
-      BRW_OPCODE_DP2, BRW_OPCODE_DP3, BRW_OPCODE_DP4
-   };
-
-   emit(dot_opcodes[elements - 2], dst, src0, src1);
-}
-
 src_reg
 vec4_visitor::fix_3src_operand(const src_reg &src)
 {
@@ -803,7 +791,7 @@ vec4_visitor::emit_pull_constant_load_reg(dst_reg dst,
                                            dst,
                                            surf_index,
                                            offset_reg);
-      pull->base_mrf = FIRST_SPILL_MRF(devinfo->gen) + 1;
+      pull->base_mrf = FIRST_PULL_LOAD_MRF(devinfo->gen) + 1;
       pull->mlen = 1;
    }
 
diff --git a/src/mesa/drivers/dri/i965/brw_vs.c b/src/mesa/drivers/dri/i965/brw_vs.c
index 3c6ee0a7a03..38de98fab86 100644
--- a/src/mesa/drivers/dri/i965/brw_vs.c
+++ b/src/mesa/drivers/dri/i965/brw_vs.c
@@ -41,31 +41,6 @@
 
 #include "util/ralloc.h"
 
-/**
- * Decide which set of clip planes should be used when clipping via
- * gl_Position or gl_ClipVertex.
- */
-gl_clip_plane *brw_select_clip_planes(struct gl_context *ctx)
-{
-   if (ctx->_Shader->CurrentProgram[MESA_SHADER_VERTEX]) {
-      /* There is currently a GLSL vertex shader, so clip according to GLSL
-       * rules, which means compare gl_ClipVertex (or gl_Position, if
-       * gl_ClipVertex wasn't assigned) against the eye-coordinate clip planes
-       * that were stored in EyeUserPlane at the time the clip planes were
-       * specified.
-       */
-      return ctx->Transform.EyeUserPlane;
-   } else {
-      /* Either we are using fixed function or an ARB vertex program.  In
-       * either case the clip planes are going to be compared against
-       * gl_Position (which is in clip coordinates) so we have to clip using
-       * _ClipUserPlane, which was transformed into clip coordinates by Mesa
-       * core.
-       */
-      return ctx->Transform._ClipUserPlane;
-   }
-}
-
 bool
 brw_codegen_vs_prog(struct brw_context *brw,
                     struct gl_shader_program *prog,
@@ -195,10 +170,17 @@ brw_codegen_vs_prog(struct brw_context *brw,
       start_time = get_time();
    }
 
+   if (unlikely(INTEL_DEBUG & DEBUG_VS))
+      brw_dump_ir("vertex", prog, &vs->base, &vp->program.Base);
+
+   int st_index = -1;
+   if (INTEL_DEBUG & DEBUG_SHADER_TIME)
+      st_index = brw_get_shader_time_index(brw, prog, &vp->program.Base, ST_VS);
+
    /* Emit GEN4 code.
     */
    program = brw_vs_emit(brw, mem_ctx, key, &prog_data,
-                         &vp->program, prog, &program_size);
+                         &vp->program, prog, st_index, &program_size);
    if (program == NULL) {
       ralloc_free(mem_ctx);
       return false;
diff --git a/src/mesa/drivers/dri/i965/brw_vs.h b/src/mesa/drivers/dri/i965/brw_vs.h
index 96d2435a515..f1242f61b33 100644
--- a/src/mesa/drivers/dri/i965/brw_vs.h
+++ b/src/mesa/drivers/dri/i965/brw_vs.h
@@ -60,6 +60,7 @@ const unsigned *brw_vs_emit(struct brw_context *brw,
                             struct brw_vs_prog_data *prog_data,
                             struct gl_vertex_program *vp,
                             struct gl_shader_program *shader_prog,
+                            int shader_time_index,
                             unsigned *program_size);
 void brw_vs_debug_recompile(struct brw_context *brw,
                             struct gl_shader_program *prog,
@@ -105,16 +106,11 @@ protected:
 
 private:
    int setup_attributes(int payload_reg);
-   void setup_vp_regs();
    void setup_uniform_clipplane_values();
    void emit_clip_distances(dst_reg reg, int offset);
-   dst_reg get_vp_dst_reg(const prog_dst_register &dst);
-   src_reg get_vp_src_reg(const prog_src_register &src);
 
    const struct brw_vs_prog_key *const key;
    struct brw_vs_prog_data * const vs_prog_data;
-   src_reg *vp_temp_regs;
-   src_reg vp_addr_reg;
 
    gl_clip_plane *clip_planes;
 
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 21048885755..4d5e7f67bd6 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -43,7 +43,7 @@
  * Return a bitfield where bit n is set if barycentric interpolation mode n
  * (see enum brw_wm_barycentric_interp_mode) is needed by the fragment shader.
  */
-unsigned
+static unsigned
 brw_compute_barycentric_interp_modes(const struct brw_device_info *devinfo,
                                      bool shade_model_flat,
                                      bool persample_shading,
@@ -221,8 +221,17 @@ brw_codegen_wm_prog(struct brw_context *brw,
       start_time = get_time();
    }
 
+   if (unlikely(INTEL_DEBUG & DEBUG_WM))
+      brw_dump_ir("fragment", prog, &fs->base, &fp->program.Base);
+
+   int st_index8 = -1, st_index16 = -1;
+   if (INTEL_DEBUG & DEBUG_SHADER_TIME) {
+      st_index8 = brw_get_shader_time_index(brw, prog, &fp->program.Base, ST_FS8);
+      st_index16 = brw_get_shader_time_index(brw, prog, &fp->program.Base, ST_FS16);
+   }
+
    program = brw_wm_fs_emit(brw, mem_ctx, key, &prog_data,
-                            &fp->program, prog, &program_size);
+                            &fp->program, prog, st_index8, st_index16, &program_size);
    if (program == NULL) {
       ralloc_free(mem_ctx);
       return false;
diff --git a/src/mesa/drivers/dri/i965/brw_wm.h b/src/mesa/drivers/dri/i965/brw_wm.h
index 053f2ee62dd..6ee22b2f907 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.h
+++ b/src/mesa/drivers/dri/i965/brw_wm.h
@@ -72,6 +72,8 @@ const unsigned *brw_wm_fs_emit(struct brw_context *brw,
                                struct brw_wm_prog_data *prog_data,
                                struct gl_fragment_program *fp,
                                struct gl_shader_program *prog,
+                               int shader_time_index8,
+                               int shader_time_index16,
                                unsigned *final_assembly_size);
 
 GLboolean brw_link_shader(struct gl_context *ctx, struct gl_shader_program *prog);
@@ -89,14 +91,6 @@ void brw_wm_debug_recompile(struct brw_context *brw,
 void
 brw_upload_wm_prog(struct brw_context *brw);
 
-struct nir_shader;
-
-unsigned
-brw_compute_barycentric_interp_modes(const struct brw_device_info *devinfo,
-                                     bool shade_model_flat,
-                                     bool persample_shading,
-                                     struct nir_shader *shader);
-
 #ifdef __cplusplus
 } // extern "C"
 #endif
diff --git a/src/mesa/drivers/dri/i965/brw_wm_state.c b/src/mesa/drivers/dri/i965/brw_wm_state.c
index cd0b56ba60c..ec54ef2acd9 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_state.c
@@ -31,6 +31,7 @@
 
 
 
+#include "intel_batchbuffer.h"
 #include "intel_fbo.h"
 #include "brw_context.h"
 #include "brw_state.h"
@@ -251,6 +252,16 @@ brw_upload_wm_unit(struct brw_context *brw)
    }
 
    brw->ctx.NewDriverState |= BRW_NEW_GEN4_UNIT_STATE;
+
+   /* _NEW_POLGYON */
+   if (brw->wm.offset_clamp != ctx->Polygon.OffsetClamp) {
+      BEGIN_BATCH(2);
+      OUT_BATCH(_3DSTATE_GLOBAL_DEPTH_OFFSET_CLAMP << 16 | (2 - 2));
+      OUT_BATCH_F(ctx->Polygon.OffsetClamp);
+      ADVANCE_BATCH();
+
+      brw->wm.offset_clamp = ctx->Polygon.OffsetClamp;
+   }
 }
 
 const struct brw_tracked_state brw_wm_unit = {
diff --git a/src/mesa/drivers/dri/i965/gen6_gs_visitor.cpp b/src/mesa/drivers/dri/i965/gen6_gs_visitor.cpp
index def21d80b24..59a76559103 100644
--- a/src/mesa/drivers/dri/i965/gen6_gs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/gen6_gs_visitor.cpp
@@ -273,6 +273,18 @@ gen6_gs_visitor::emit_urb_write_header(int mrf)
    emit(GS_OPCODE_SET_DWORD_2, dst_reg(MRF, mrf), flags_data);
 }
 
+static int
+align_interleaved_urb_mlen(int mlen)
+{
+   /* URB data written (does not include the message header reg) must
+    * be a multiple of 256 bits, or 2 VS registers.  See vol5c.5,
+    * section 5.4.3.2.2: URB_INTERLEAVED.
+    */
+   if ((mlen % 2) != 1)
+      mlen++;
+   return mlen;
+}
+
 void
 gen6_gs_visitor::emit_urb_write_opcode(bool complete, int base_mrf,
                                        int last_mrf, int urb_offset)
@@ -299,14 +311,7 @@ gen6_gs_visitor::emit_urb_write_opcode(bool complete, int base_mrf,
    }
 
    inst->base_mrf = base_mrf;
-   /* URB data written (does not include the message header reg) must
-    * be a multiple of 256 bits, or 2 VS registers.  See vol5c.5,
-    * section 5.4.3.2.2: URB_INTERLEAVED.
-    */
-   int mlen = last_mrf - base_mrf;
-   if ((mlen % 2) != 1)
-      mlen++;
-   inst->mlen = mlen;
+   inst->mlen = align_interleaved_urb_mlen(last_mrf - base_mrf);
    inst->offset = urb_offset;
 }
 
@@ -339,9 +344,9 @@ gen6_gs_visitor::emit_thread_end()
 
    /* In the process of generating our URB write message contents, we
     * may need to unspill a register or load from an array.  Those
-    * reads would use MRFs 14-15.
+    * reads would use MRFs 21..23
     */
-   int max_usable_mrf = 13;
+   int max_usable_mrf = FIRST_SPILL_MRF(devinfo->gen);
 
    /* Issue the FF_SYNC message and obtain the initial VUE handle. */
    emit(CMP(dst_null_d(), this->vertex_count, 0u, BRW_CONDITIONAL_G));
@@ -416,9 +421,10 @@ gen6_gs_visitor::emit_thread_end()
                         this->vertex_output_offset, 1u));
 
                /* If this was max_usable_mrf, we can't fit anything more into
-                * this URB WRITE.
+                * this URB WRITE. Same if we reached the max. message length.
                 */
-               if (mrf > max_usable_mrf) {
+               if (mrf > max_usable_mrf ||
+                   align_interleaved_urb_mlen(mrf - base_mrf + 1) > BRW_MAX_MSG_LENGTH) {
                   slot++;
                   break;
                }
diff --git a/src/mesa/drivers/dri/i965/gen6_gs_visitor.h b/src/mesa/drivers/dri/i965/gen6_gs_visitor.h
index 41c6d183acd..e75d6aa10b8 100644
--- a/src/mesa/drivers/dri/i965/gen6_gs_visitor.h
+++ b/src/mesa/drivers/dri/i965/gen6_gs_visitor.h
@@ -43,8 +43,11 @@ public:
                    void *mem_ctx,
                    bool no_spills,
                    int shader_time_index) :
-      vec4_gs_visitor(comp, log_data, c, prog, shader, mem_ctx, no_spills,
-                      shader_time_index) {}
+      vec4_gs_visitor(comp, log_data, c, shader, mem_ctx, no_spills,
+                      shader_time_index),
+      shader_prog(prog)
+      {
+      }
 
 protected:
    virtual void emit_prolog();
@@ -64,6 +67,8 @@ private:
    void xfb_setup();
    int get_vertex_output_offset_for_varying(int vertex, int varying);
 
+   const struct gl_shader_program *shader_prog;
+
    src_reg vertex_output;
    src_reg vertex_output_offset;
    src_reg temp;
diff --git a/src/mesa/drivers/dri/i965/gen7_cs_state.c b/src/mesa/drivers/dri/i965/gen7_cs_state.c
index 5edc4fc9842..6aeb0cb243f 100644
--- a/src/mesa/drivers/dri/i965/gen7_cs_state.c
+++ b/src/mesa/drivers/dri/i965/gen7_cs_state.c
@@ -70,10 +70,8 @@ brw_upload_cs_state(struct brw_context *brw)
 
    unsigned local_id_dwords = 0;
 
-   if (prog->SystemValuesRead & SYSTEM_BIT_LOCAL_INVOCATION_ID) {
-      local_id_dwords =
-         brw_cs_prog_local_id_payload_dwords(cs_prog_data->simd_size);
-   }
+   if (prog->SystemValuesRead & SYSTEM_BIT_LOCAL_INVOCATION_ID)
+      local_id_dwords = cs_prog_data->local_invocation_id_regs * 8;
 
    unsigned push_constant_data_size =
       (prog_data->nr_params + local_id_dwords) * sizeof(gl_constant_value);
@@ -191,63 +189,6 @@ const struct brw_tracked_state brw_cs_state = {
 
 
 /**
- * We are building the local ID push constant data using the simplest possible
- * method. We simply push the local IDs directly as they should appear in the
- * registers for the uvec3 gl_LocalInvocationID variable.
- *
- * Therefore, for SIMD8, we use 3 full registers, and for SIMD16 we use 6
- * registers worth of push constant space.
- *
- * Note: Any updates to brw_cs_prog_local_id_payload_dwords,
- * fill_local_id_payload or fs_visitor::emit_cs_local_invocation_id_setup need
- * to coordinated.
- *
- * FINISHME: There are a few easy optimizations to consider.
- *
- * 1. If gl_WorkGroupSize x, y or z is 1, we can just use zero, and there is
- *    no need for using push constant space for that dimension.
- *
- * 2. Since GL_MAX_COMPUTE_WORK_GROUP_SIZE is currently 1024 or less, we can
- *    easily use 16-bit words rather than 32-bit dwords in the push constant
- *    data.
- *
- * 3. If gl_WorkGroupSize x, y or z is small, then we can use bytes for
- *    conveying the data, and thereby reduce push constant usage.
- *
- */
-unsigned
-brw_cs_prog_local_id_payload_dwords(unsigned dispatch_width)
-{
-   return 3 * dispatch_width;
-}
-
-
-static void
-fill_local_id_payload(const struct brw_cs_prog_data *cs_prog_data,
-                      void *buffer, unsigned *x, unsigned *y, unsigned *z)
-{
-   uint32_t *param = (uint32_t *)buffer;
-   for (unsigned i = 0; i < cs_prog_data->simd_size; i++) {
-      param[0 * cs_prog_data->simd_size + i] = *x;
-      param[1 * cs_prog_data->simd_size + i] = *y;
-      param[2 * cs_prog_data->simd_size + i] = *z;
-
-      (*x)++;
-      if (*x == cs_prog_data->local_size[0]) {
-         *x = 0;
-         (*y)++;
-         if (*y == cs_prog_data->local_size[1]) {
-            *y = 0;
-            (*z)++;
-            if (*z == cs_prog_data->local_size[2])
-               *z = 0;
-         }
-      }
-   }
-}
-
-
-/**
  * Creates a region containing the push constants for the CS on gen7+.
  *
  * Push constants are constant values (such as GLSL uniforms) that are
@@ -269,10 +210,8 @@ brw_upload_cs_push_constants(struct brw_context *brw,
       (struct brw_stage_prog_data*) cs_prog_data;
    unsigned local_id_dwords = 0;
 
-   if (prog->SystemValuesRead & SYSTEM_BIT_LOCAL_INVOCATION_ID) {
-      local_id_dwords =
-         brw_cs_prog_local_id_payload_dwords(cs_prog_data->simd_size);
-   }
+   if (prog->SystemValuesRead & SYSTEM_BIT_LOCAL_INVOCATION_ID)
+      local_id_dwords = cs_prog_data->local_invocation_id_regs * 8;
 
    /* Updates the ParamaterValues[i] pointers for all parameters of the
     * basic type of PROGRAM_STATE_VAR.
@@ -302,14 +241,13 @@ brw_upload_cs_push_constants(struct brw_context *brw,
 
       STATIC_ASSERT(sizeof(gl_constant_value) == sizeof(float));
 
+      brw_cs_fill_local_id_payload(cs_prog_data, param, threads,
+                                   reg_aligned_constant_size);
+
       /* _NEW_PROGRAM_CONSTANTS */
-      unsigned x = 0, y = 0, z = 0;
       for (t = 0; t < threads; t++) {
-         gl_constant_value *next_param = &param[t * param_aligned_count];
-         if (local_id_dwords > 0) {
-            fill_local_id_payload(cs_prog_data, (void*)next_param, &x, &y, &z);
-            next_param += local_id_dwords;
-         }
+         gl_constant_value *next_param =
+            &param[t * param_aligned_count + local_id_dwords];
          for (i = 0; i < prog_data->nr_params; i++) {
             next_param[i] = *prog_data->param[i];
          }
diff --git a/src/mesa/drivers/dri/i965/intel_blit.c b/src/mesa/drivers/dri/i965/intel_blit.c
index 46fccc8d6ce..bd204aa3ce8 100644
--- a/src/mesa/drivers/dri/i965/intel_blit.c
+++ b/src/mesa/drivers/dri/i965/intel_blit.c
@@ -27,7 +27,6 @@
 #include "main/blit.h"
 #include "main/context.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/fbobject.h"
 
 #include "brw_context.h"
diff --git a/src/mesa/drivers/dri/i965/intel_buffers.c b/src/mesa/drivers/dri/i965/intel_buffers.c
index c98e19382c3..fd522cc4f4d 100644
--- a/src/mesa/drivers/dri/i965/intel_buffers.c
+++ b/src/mesa/drivers/dri/i965/intel_buffers.c
@@ -32,30 +32,10 @@
 #include "main/framebuffer.h"
 #include "main/renderbuffer.h"
 
-
-bool
-brw_is_front_buffer_reading(struct gl_framebuffer *fb)
-{
-   if (!fb || _mesa_is_user_fbo(fb))
-      return false;
-
-   return fb->_ColorReadBufferIndex == BUFFER_FRONT_LEFT;
-}
-
-bool
-brw_is_front_buffer_drawing(struct gl_framebuffer *fb)
-{
-   if (!fb || _mesa_is_user_fbo(fb))
-      return false;
-
-   return (fb->_NumColorDrawBuffers >= 1 &&
-           fb->_ColorDrawBufferIndexes[0] == BUFFER_FRONT_LEFT);
-}
-
 static void
 intelDrawBuffer(struct gl_context * ctx, GLenum mode)
 {
-   if (brw_is_front_buffer_drawing(ctx->DrawBuffer)) {
+   if (_mesa_is_front_buffer_drawing(ctx->DrawBuffer)) {
       struct brw_context *const brw = brw_context(ctx);
 
       /* If we might be front-buffer rendering on this buffer for the first
@@ -71,7 +51,7 @@ intelDrawBuffer(struct gl_context * ctx, GLenum mode)
 static void
 intelReadBuffer(struct gl_context * ctx, GLenum mode)
 {
-   if (brw_is_front_buffer_reading(ctx->ReadBuffer)) {
+   if (_mesa_is_front_buffer_reading(ctx->ReadBuffer)) {
       struct brw_context *const brw = brw_context(ctx);
 
       /* If we might be front-buffer reading on this buffer for the first
diff --git a/src/mesa/drivers/dri/i965/intel_buffers.h b/src/mesa/drivers/dri/i965/intel_buffers.h
index 85f54b2c653..0e0d9c31f2b 100644
--- a/src/mesa/drivers/dri/i965/intel_buffers.h
+++ b/src/mesa/drivers/dri/i965/intel_buffers.h
@@ -30,11 +30,6 @@
 #include "drm.h"
 #include "brw_context.h"
 
-struct intel_framebuffer;
-
 extern void intelInitBufferFuncs(struct dd_function_table *functions);
 
-bool brw_is_front_buffer_reading(struct gl_framebuffer *fb);
-bool brw_is_front_buffer_drawing(struct gl_framebuffer *fb);
-
 #endif /* INTEL_BUFFERS_H */
diff --git a/src/mesa/drivers/dri/i965/intel_debug.c b/src/mesa/drivers/dri/i965/intel_debug.c
index 5a9c9533fde..f7c02c8a38d 100644
--- a/src/mesa/drivers/dri/i965/intel_debug.c
+++ b/src/mesa/drivers/dri/i965/intel_debug.c
@@ -33,10 +33,11 @@
 #include "intel_debug.h"
 #include "utils.h"
 #include "util/u_atomic.h" /* for p_atomic_cmpxchg */
+#include "util/debug.h"
 
 uint64_t INTEL_DEBUG = 0;
 
-static const struct dri_debug_control debug_control[] = {
+static const struct debug_control debug_control[] = {
    { "tex",         DEBUG_TEXTURE},
    { "state",       DEBUG_STATE},
    { "blit",        DEBUG_BLIT},
@@ -91,22 +92,10 @@ intel_debug_flag_for_shader_stage(gl_shader_stage stage)
 }
 
 void
-brw_process_intel_debug_variable(struct intel_screen *screen)
+brw_process_intel_debug_variable(void)
 {
-   uint64_t intel_debug = driParseDebugString(getenv("INTEL_DEBUG"), debug_control);
+   uint64_t intel_debug = parse_debug_string(getenv("INTEL_DEBUG"), debug_control);
    (void) p_atomic_cmpxchg(&INTEL_DEBUG, 0, intel_debug);
-
-   if (INTEL_DEBUG & DEBUG_BUFMGR)
-      dri_bufmgr_set_debug(screen->bufmgr, true);
-
-   if ((INTEL_DEBUG & DEBUG_SHADER_TIME) && screen->devinfo->gen < 7) {
-      fprintf(stderr,
-              "shader_time debugging requires gen7 (Ivybridge) or better.\n");
-      INTEL_DEBUG &= ~DEBUG_SHADER_TIME;
-   }
-
-   if (INTEL_DEBUG & DEBUG_AUB)
-      drm_intel_bufmgr_gem_set_aub_dump(screen->bufmgr, true);
 }
 
 /**
diff --git a/src/mesa/drivers/dri/i965/intel_debug.h b/src/mesa/drivers/dri/i965/intel_debug.h
index b7d0c823fa8..0a6e1b90b98 100644
--- a/src/mesa/drivers/dri/i965/intel_debug.h
+++ b/src/mesa/drivers/dri/i965/intel_debug.h
@@ -115,8 +115,6 @@ extern uint64_t INTEL_DEBUG;
 
 extern uint64_t intel_debug_flag_for_shader_stage(gl_shader_stage stage);
 
-struct intel_screen;
-
-extern void brw_process_intel_debug_variable(struct intel_screen *);
+extern void brw_process_intel_debug_variable(void);
 
 extern bool brw_env_var_as_boolean(const char *var_name, bool default_value);
diff --git a/src/mesa/drivers/dri/i965/intel_extensions.c b/src/mesa/drivers/dri/i965/intel_extensions.c
index 3c77f4773c6..3f9afd16c71 100644
--- a/src/mesa/drivers/dri/i965/intel_extensions.c
+++ b/src/mesa/drivers/dri/i965/intel_extensions.c
@@ -229,6 +229,7 @@ intelInitExtensions(struct gl_context *ctx)
    ctx->Extensions.EXT_packed_float = true;
    ctx->Extensions.EXT_pixel_buffer_object = true;
    ctx->Extensions.EXT_point_parameters = true;
+   ctx->Extensions.EXT_polygon_offset_clamp = true;
    ctx->Extensions.EXT_provoking_vertex = true;
    ctx->Extensions.EXT_stencil_two_side = true;
    ctx->Extensions.EXT_texture_array = true;
@@ -300,7 +301,6 @@ intelInitExtensions(struct gl_context *ctx)
       ctx->Extensions.AMD_vertex_shader_layer = true;
       ctx->Extensions.EXT_framebuffer_multisample = true;
       ctx->Extensions.EXT_framebuffer_multisample_blit_scaled = true;
-      ctx->Extensions.EXT_polygon_offset_clamp = true;
       ctx->Extensions.EXT_transform_feedback = true;
       ctx->Extensions.OES_depth_texture_cube_map = true;
 
diff --git a/src/mesa/drivers/dri/i965/intel_fbo.c b/src/mesa/drivers/dri/i965/intel_fbo.c
index 6b2349e8b69..5a6b0dd1ec5 100644
--- a/src/mesa/drivers/dri/i965/intel_fbo.c
+++ b/src/mesa/drivers/dri/i965/intel_fbo.c
@@ -52,19 +52,6 @@
 
 #define FILE_DEBUG_FLAG DEBUG_FBO
 
-/**
- * Create a new framebuffer object.
- */
-static struct gl_framebuffer *
-intel_new_framebuffer(struct gl_context * ctx, GLuint name)
-{
-   /* Only drawable state in intel_framebuffer at this time, just use Mesa's
-    * class
-    */
-   return _mesa_new_framebuffer(ctx, name);
-}
-
-
 /** Called by gl_renderbuffer::Delete() */
 static void
 intel_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
@@ -1093,7 +1080,6 @@ void
 intel_fbo_init(struct brw_context *brw)
 {
    struct dd_function_table *dd = &brw->ctx.Driver;
-   dd->NewFramebuffer = intel_new_framebuffer;
    dd->NewRenderbuffer = intel_new_renderbuffer;
    dd->MapRenderbuffer = intel_map_renderbuffer;
    dd->UnmapRenderbuffer = intel_unmap_renderbuffer;
diff --git a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
index ffc356c9240..a169c41790e 100644
--- a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
+++ b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
@@ -160,7 +160,7 @@ intel_get_non_msrt_mcs_alignment(struct intel_mipmap_tree *mt,
    }
 }
 
-bool
+static bool
 intel_tiling_supports_non_msrt_mcs(struct brw_context *brw, unsigned tiling)
 {
    /* From the Ivy Bridge PRM, Vol2 Part1 11.7 "MCS Buffer for Render
@@ -193,9 +193,9 @@ intel_tiling_supports_non_msrt_mcs(struct brw_context *brw, unsigned tiling)
  *     - MCS buffer for non-MSRT is supported only for RT formats 32bpp,
  *       64bpp, and 128bpp.
  */
-bool
-intel_miptree_is_fast_clear_capable(struct brw_context *brw,
-                                    struct intel_mipmap_tree *mt)
+static bool
+intel_miptree_supports_non_msrt_fast_clear(struct brw_context *brw,
+                                           struct intel_mipmap_tree *mt)
 {
    /* MCS support does not exist prior to Gen7 */
    if (brw->gen < 7)
@@ -204,6 +204,10 @@ intel_miptree_is_fast_clear_capable(struct brw_context *brw,
    if (mt->disable_aux_buffers)
       return false;
 
+   /* This function applies only to non-multisampled render targets. */
+   if (mt->num_samples > 1)
+      return false;
+
    /* MCS is only supported for color buffers */
    switch (_mesa_get_format_base_format(mt->format)) {
    case GL_DEPTH_COMPONENT:
@@ -222,7 +226,16 @@ intel_miptree_is_fast_clear_capable(struct brw_context *brw,
 
       return false;
    }
+
+   /* Check for layered surfaces. */
    if (mt->physical_depth0 != 1) {
+       /* Multisample surfaces with the CMS layout are not layered surfaces,
+        * yet still have physical_depth0 > 1. Assert that we don't
+        * accidentally reject a multisampled surface here. We should have
+        * rejected it earlier by explicitly checking the sample count.
+        */
+      assert(mt->num_samples <= 1);
+
       if (brw->gen >= 8) {
          perf_debug("Layered fast clear - giving up. (%dx%d%d)\n",
                     mt->logical_width0, mt->logical_height0,
@@ -494,7 +507,7 @@ intel_miptree_create_layout(struct brw_context *brw,
     *  7   |      ?         |        ?
     *  6   |      ?         |        ?
     */
-   if (intel_miptree_is_fast_clear_capable(brw, mt)) {
+   if (intel_miptree_supports_non_msrt_fast_clear(brw, mt)) {
       if (brw->gen >= 9 || (brw->gen == 8 && num_samples <= 1))
          layout_flags |= MIPTREE_LAYOUT_FORCE_HALIGN16;
    } else if (brw->gen >= 9 && num_samples > 1) {
@@ -692,7 +705,7 @@ intel_miptree_create(struct brw_context *brw,
     * clear actually occurs.
     */
    if (intel_tiling_supports_non_msrt_mcs(brw, mt->tiling) &&
-       intel_miptree_is_fast_clear_capable(brw, mt)) {
+       intel_miptree_supports_non_msrt_fast_clear(brw, mt)) {
       mt->fast_clear_state = INTEL_FAST_CLEAR_STATE_RESOLVED;
       assert(brw->gen < 8 || mt->halign == 16 || num_samples <= 1);
    }
@@ -800,8 +813,9 @@ intel_update_winsys_renderbuffer_miptree(struct brw_context *intel,
     * clear actually occurs.
     */
    if (intel_tiling_supports_non_msrt_mcs(intel, singlesample_mt->tiling) &&
-       intel_miptree_is_fast_clear_capable(intel, singlesample_mt))
+       intel_miptree_supports_non_msrt_fast_clear(intel, singlesample_mt)) {
       singlesample_mt->fast_clear_state = INTEL_FAST_CLEAR_STATE_RESOLVED;
+   }
 
    if (num_samples == 0) {
       intel_miptree_release(&irb->mt);
diff --git a/src/mesa/drivers/dri/i965/intel_mipmap_tree.h b/src/mesa/drivers/dri/i965/intel_mipmap_tree.h
index 486e5c6f43b..805cd714d88 100644
--- a/src/mesa/drivers/dri/i965/intel_mipmap_tree.h
+++ b/src/mesa/drivers/dri/i965/intel_mipmap_tree.h
@@ -658,11 +658,7 @@ struct intel_mipmap_tree
 void
 intel_get_non_msrt_mcs_alignment(struct intel_mipmap_tree *mt,
                                  unsigned *width_px, unsigned *height);
-bool
-intel_tiling_supports_non_msrt_mcs(struct brw_context *brw, unsigned tiling);
-bool
-intel_miptree_is_fast_clear_capable(struct brw_context *brw,
-                                    struct intel_mipmap_tree *mt);
+
 bool
 intel_miptree_alloc_non_msrt_mcs(struct brw_context *brw,
                                  struct intel_mipmap_tree *mt);
diff --git a/src/mesa/drivers/dri/i965/intel_pixel.c b/src/mesa/drivers/dri/i965/intel_pixel.c
index 30d3a521ec8..d4f86fdffe0 100644
--- a/src/mesa/drivers/dri/i965/intel_pixel.c
+++ b/src/mesa/drivers/dri/i965/intel_pixel.c
@@ -128,7 +128,6 @@ intel_check_blit_fragment_ops(struct gl_context * ctx, bool src_alpha_is_one)
 void
 intelInitPixelFuncs(struct dd_function_table *functions)
 {
-   functions->Accum = _mesa_accum;
    functions->Bitmap = intelBitmap;
    functions->CopyPixels = intelCopyPixels;
    functions->DrawPixels = intelDrawPixels;
diff --git a/src/mesa/drivers/dri/i965/intel_screen.c b/src/mesa/drivers/dri/i965/intel_screen.c
index 17838350cda..590c45d93ea 100644
--- a/src/mesa/drivers/dri/i965/intel_screen.c
+++ b/src/mesa/drivers/dri/i965/intel_screen.c
@@ -1421,7 +1421,19 @@ __DRIconfig **intelInitScreen2(__DRIscreen *psp)
    if (!intelScreen->devinfo)
       return false;
 
-   brw_process_intel_debug_variable(intelScreen);
+   brw_process_intel_debug_variable();
+
+   if (INTEL_DEBUG & DEBUG_BUFMGR)
+      dri_bufmgr_set_debug(intelScreen->bufmgr, true);
+
+   if ((INTEL_DEBUG & DEBUG_SHADER_TIME) && intelScreen->devinfo->gen < 7) {
+      fprintf(stderr,
+              "shader_time debugging requires gen7 (Ivybridge) or better.\n");
+      INTEL_DEBUG &= ~DEBUG_SHADER_TIME;
+   }
+
+   if (INTEL_DEBUG & DEBUG_AUB)
+      drm_intel_bufmgr_gem_set_aub_dump(intelScreen->bufmgr, true);
 
    intelScreen->hw_must_use_separate_stencil = intelScreen->devinfo->gen >= 7;
 
diff --git a/src/mesa/drivers/dri/i965/intel_state.c b/src/mesa/drivers/dri/i965/intel_state.c
index 498cab49ec4..2f5c901fdf8 100644
--- a/src/mesa/drivers/dri/i965/intel_state.c
+++ b/src/mesa/drivers/dri/i965/intel_state.c
@@ -27,7 +27,6 @@
 #include "main/context.h"
 #include "main/macros.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/dd.h"
 
 #include "intel_screen.h"
diff --git a/src/mesa/drivers/dri/r200/r200_context.h b/src/mesa/drivers/dri/r200/r200_context.h
index c02a4f399ee..7c6f48008a1 100644
--- a/src/mesa/drivers/dri/r200/r200_context.h
+++ b/src/mesa/drivers/dri/r200/r200_context.h
@@ -42,7 +42,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/macros.h"
 #include "main/mtypes.h"
-#include "main/colormac.h"
 #include "r200_reg.h"
 #include "r200_vertprog.h"
 
diff --git a/src/mesa/drivers/dri/r200/r200_maos_arrays.c b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
index 3cfc03d10cc..9b16cf84cf5 100644
--- a/src/mesa/drivers/dri/r200/r200_maos_arrays.c
+++ b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
@@ -34,7 +34,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/mtypes.h"
-#include "main/colormac.h"
 #include "main/imports.h"
 #include "main/macros.h"
 
diff --git a/src/mesa/drivers/dri/r200/r200_state.c b/src/mesa/drivers/dri/r200/r200_state.c
index cca176d7f9b..b4acf985ee8 100644
--- a/src/mesa/drivers/dri/r200/r200_state.c
+++ b/src/mesa/drivers/dri/r200/r200_state.c
@@ -37,7 +37,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "main/imports.h"
 #include "main/api_arrayelt.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/light.h"
 #include "main/framebuffer.h"
 #include "main/fbobject.h"
@@ -2389,7 +2388,6 @@ void r200InitStateFuncs( radeonContextPtr radeon, struct dd_function_table *func
    functions->Enable			= r200Enable;
    functions->Fogfv			= r200Fogfv;
    functions->FrontFace			= r200FrontFace;
-   functions->Hint			= NULL;
    functions->LightModelfv		= r200LightModelfv;
    functions->Lightfv			= r200Lightfv;
    functions->LineStipple		= r200LineStipple;
diff --git a/src/mesa/drivers/dri/r200/r200_state_init.c b/src/mesa/drivers/dri/r200/r200_state_init.c
index ad64f788b9f..8cffa92c10b 100644
--- a/src/mesa/drivers/dri/r200/r200_state_init.c
+++ b/src/mesa/drivers/dri/r200/r200_state_init.c
@@ -34,7 +34,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "main/glheader.h"
 #include "main/imports.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/api_arrayelt.h"
 
 #include "swrast/swrast.h"
diff --git a/src/mesa/drivers/dri/r200/r200_swtcl.c b/src/mesa/drivers/dri/r200/r200_swtcl.c
index bb9be210567..72f09ae4056 100644
--- a/src/mesa/drivers/dri/r200/r200_swtcl.c
+++ b/src/mesa/drivers/dri/r200/r200_swtcl.c
@@ -34,7 +34,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/mtypes.h"
-#include "main/colormac.h"
 #include "main/enums.h"
 #include "main/image.h"
 #include "main/imports.h"
diff --git a/src/mesa/drivers/dri/r200/r200_tcl.c b/src/mesa/drivers/dri/r200/r200_tcl.c
index 747275334b6..c042aae0ef3 100644
--- a/src/mesa/drivers/dri/r200/r200_tcl.c
+++ b/src/mesa/drivers/dri/r200/r200_tcl.c
@@ -36,7 +36,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "main/imports.h"
 #include "main/mtypes.h"
 #include "main/enums.h"
-#include "main/colormac.h"
 #include "main/light.h"
 #include "main/state.h"
 
diff --git a/src/mesa/drivers/dri/r200/r200_tex.c b/src/mesa/drivers/dri/r200/r200_tex.c
index feee0b2ba3f..ca921100c12 100644
--- a/src/mesa/drivers/dri/r200/r200_tex.c
+++ b/src/mesa/drivers/dri/r200/r200_tex.c
@@ -33,7 +33,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/imports.h"
-#include "main/colormac.h"
 #include "main/context.h"
 #include "main/enums.h"
 #include "main/image.h"
diff --git a/src/mesa/drivers/dri/radeon/radeon_context.c b/src/mesa/drivers/dri/radeon/radeon_context.c
index a9e2ab563d3..5e15b46fb32 100644
--- a/src/mesa/drivers/dri/radeon/radeon_context.c
+++ b/src/mesa/drivers/dri/radeon/radeon_context.c
@@ -341,8 +341,8 @@ r100CreateContext( gl_api api,
 
 
 #if DO_DEBUG
-   RADEON_DEBUG = driParseDebugString( getenv( "RADEON_DEBUG" ),
-				       debug_control );
+   RADEON_DEBUG = parse_debug_string( getenv( "RADEON_DEBUG" ),
+                                      debug_control );
 #endif
 
    tcl_mode = driQueryOptioni(&rmesa->radeon.optionCache, "tcl_mode");
diff --git a/src/mesa/drivers/dri/radeon/radeon_context.h b/src/mesa/drivers/dri/radeon/radeon_context.h
index badabd9508c..88a295386ca 100644
--- a/src/mesa/drivers/dri/radeon/radeon_context.h
+++ b/src/mesa/drivers/dri/radeon/radeon_context.h
@@ -49,7 +49,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "radeon_drm.h"
 #include "main/macros.h"
 #include "main/mtypes.h"
-#include "main/colormac.h"
 #include "radeon_screen.h"
 
 #include "radeon_common.h"
diff --git a/src/mesa/drivers/dri/radeon/radeon_debug.c b/src/mesa/drivers/dri/radeon/radeon_debug.c
index 7ddba1ae85f..383a5df6749 100644
--- a/src/mesa/drivers/dri/radeon/radeon_debug.c
+++ b/src/mesa/drivers/dri/radeon/radeon_debug.c
@@ -27,7 +27,7 @@
  *      Pauli Nieminen <[email protected]>
  */
 
-#include "utils.h"
+#include "util/debug.h"
 
 #include "radeon_common_context.h"
 #include "radeon_debug.h"
@@ -35,7 +35,7 @@
 #include <stdarg.h>
 #include <stdio.h>
 
-static const struct dri_debug_control debug_control[] = {
+static const struct debug_control debug_control[] = {
 	{"fall", RADEON_FALLBACKS},
 	{"tex", RADEON_TEXTURE},
 	{"ioctl", RADEON_IOCTL},
@@ -61,7 +61,7 @@ radeon_debug_type_t radeon_enabled_debug_types;
 
 void radeon_init_debug(void)
 {
-	radeon_enabled_debug_types = driParseDebugString(getenv("RADEON_DEBUG"), debug_control);
+	radeon_enabled_debug_types = parse_debug_string(getenv("RADEON_DEBUG"), debug_control);
 
 	radeon_enabled_debug_types |= RADEON_GENERAL;
 }
diff --git a/src/mesa/drivers/dri/radeon/radeon_fbo.c b/src/mesa/drivers/dri/radeon/radeon_fbo.c
index 5eece518c95..4d75d149b27 100644
--- a/src/mesa/drivers/dri/radeon/radeon_fbo.c
+++ b/src/mesa/drivers/dri/radeon/radeon_fbo.c
@@ -46,12 +46,6 @@
                 printf(__VA_ARGS__);                      \
 } while(0)
 
-static struct gl_framebuffer *
-radeon_new_framebuffer(struct gl_context *ctx, GLuint name)
-{
-  return _mesa_new_framebuffer(ctx, name);
-}
-
 static void
 radeon_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
@@ -868,7 +862,6 @@ radeon_validate_framebuffer(struct gl_context *ctx, struct gl_framebuffer *fb)
 
 void radeon_fbo_init(struct radeon_context *radeon)
 {
-  radeon->glCtx.Driver.NewFramebuffer = radeon_new_framebuffer;
   radeon->glCtx.Driver.NewRenderbuffer = radeon_new_renderbuffer;
   radeon->glCtx.Driver.MapRenderbuffer = radeon_map_renderbuffer;
   radeon->glCtx.Driver.UnmapRenderbuffer = radeon_unmap_renderbuffer;
diff --git a/src/mesa/drivers/dri/radeon/radeon_state.c b/src/mesa/drivers/dri/radeon/radeon_state.c
index 74c1fc6c902..8a1b81d8f32 100644
--- a/src/mesa/drivers/dri/radeon/radeon_state.c
+++ b/src/mesa/drivers/dri/radeon/radeon_state.c
@@ -2148,7 +2148,6 @@ void radeonInitStateFuncs( struct gl_context *ctx )
    ctx->Driver.Enable			= radeonEnable;
    ctx->Driver.Fogfv			= radeonFogfv;
    ctx->Driver.FrontFace		= radeonFrontFace;
-   ctx->Driver.Hint			= NULL;
    ctx->Driver.LightModelfv		= radeonLightModelfv;
    ctx->Driver.Lightfv			= radeonLightfv;
    ctx->Driver.LineStipple              = radeonLineStipple;
diff --git a/src/mesa/drivers/dri/radeon/radeon_swtcl.c b/src/mesa/drivers/dri/radeon/radeon_swtcl.c
index b671a3be143..1e19cf7c7c0 100644
--- a/src/mesa/drivers/dri/radeon/radeon_swtcl.c
+++ b/src/mesa/drivers/dri/radeon/radeon_swtcl.c
@@ -34,7 +34,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/mtypes.h"
-#include "main/colormac.h"
 #include "main/enums.h"
 #include "main/imports.h"
 #include "main/macros.h"
@@ -359,16 +358,16 @@ void r100_swtcl_flush(struct gl_context *ctx, uint32_t current_offset)
 #define HAVE_ELTS        0
 
 static const GLuint hw_prim[GL_POLYGON+1] = {
-   RADEON_CP_VC_CNTL_PRIM_TYPE_POINT,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
-   0,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_LINE_STRIP,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_STRIP,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_FAN,
-   0,
-   0,
-   0
+   [GL_POINTS] = RADEON_CP_VC_CNTL_PRIM_TYPE_POINT,
+   [GL_LINES] = RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
+   [GL_LINE_LOOP] = 0,
+   [GL_LINE_STRIP] = RADEON_CP_VC_CNTL_PRIM_TYPE_LINE_STRIP,
+   [GL_TRIANGLES] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_TRIANGLE_STRIP] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_STRIP,
+   [GL_TRIANGLE_FAN] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_FAN,
+   [GL_QUADS] = 0,
+   [GL_QUAD_STRIP] = 0,
+   [GL_POLYGON] = 0
 };
 
 static inline void
@@ -468,16 +467,16 @@ const struct tnl_pipeline_stage _radeon_render_stage =
 
 
 static const GLuint reduced_hw_prim[GL_POLYGON+1] = {
-   RADEON_CP_VC_CNTL_PRIM_TYPE_POINT,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
-   RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST
+   [GL_POINTS] = RADEON_CP_VC_CNTL_PRIM_TYPE_POINT,
+   [GL_LINES] = RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
+   [GL_LINE_LOOP] = RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
+   [GL_LINE_STRIP] = RADEON_CP_VC_CNTL_PRIM_TYPE_LINE,
+   [GL_TRIANGLES] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_TRIANGLE_STRIP] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_TRIANGLE_FAN] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_QUADS] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_QUAD_STRIP] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST,
+   [GL_POLYGON] = RADEON_CP_VC_CNTL_PRIM_TYPE_TRI_LIST
 };
 
 static void radeonRasterPrimitive( struct gl_context *ctx, GLuint hwprim );
diff --git a/src/mesa/drivers/dri/radeon/radeon_tex.c b/src/mesa/drivers/dri/radeon/radeon_tex.c
index 0955a135de8..d1aa1a18737 100644
--- a/src/mesa/drivers/dri/radeon/radeon_tex.c
+++ b/src/mesa/drivers/dri/radeon/radeon_tex.c
@@ -33,7 +33,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/imports.h"
-#include "main/colormac.h"
 #include "main/context.h"
 #include "main/enums.h"
 #include "main/image.h"
diff --git a/src/mesa/drivers/dri/radeon/radeon_texstate.c b/src/mesa/drivers/dri/radeon/radeon_texstate.c
index ec835f248eb..35b1538d9e1 100644
--- a/src/mesa/drivers/dri/radeon/radeon_texstate.c
+++ b/src/mesa/drivers/dri/radeon/radeon_texstate.c
@@ -35,7 +35,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/glheader.h"
 #include "main/imports.h"
-#include "main/colormac.h"
 #include "main/context.h"
 #include "main/macros.h"
 #include "main/teximage.h"
diff --git a/src/mesa/main/arrayobj.c b/src/mesa/main/arrayobj.c
index 28851434133..061e557a397 100644
--- a/src/mesa/main/arrayobj.c
+++ b/src/mesa/main/arrayobj.c
@@ -151,7 +151,6 @@ unbind_array_object_vbos(struct gl_context *ctx, struct gl_vertex_array_object *
  * Allocate and initialize a new vertex array object.
  *
  * This function is intended to be called via
- * \c dd_function_table::NewArrayObject.
  */
 struct gl_vertex_array_object *
 _mesa_new_vao(struct gl_context *ctx, GLuint name)
@@ -203,10 +202,8 @@ _mesa_reference_vao_(struct gl_context *ctx,
       deleteFlag = (oldObj->RefCount == 0);
       mtx_unlock(&oldObj->Mutex);
 
-      if (deleteFlag) {
-	 assert(ctx->Driver.DeleteArrayObject);
-         ctx->Driver.DeleteArrayObject(ctx, oldObj);
-      }
+      if (deleteFlag)
+         _mesa_delete_vao(ctx, oldObj);
 
       *ptr = NULL;
    }
@@ -408,7 +405,7 @@ bind_vertex_array(struct gl_context *ctx, GLuint id, GLboolean genRequired)
          }
 
          /* For APPLE version, generate a new array object now */
-	 newObj = (*ctx->Driver.NewArrayObject)(ctx, id);
+	 newObj = _mesa_new_vao(ctx, id);
          if (!newObj) {
             _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindVertexArrayAPPLE");
             return;
@@ -446,10 +443,6 @@ bind_vertex_array(struct gl_context *ctx, GLuint id, GLboolean genRequired)
 
    ctx->NewState |= _NEW_ARRAY;
    _mesa_reference_vao(ctx, &ctx->Array.VAO, newObj);
-
-   /* Pass BindVertexArray call to device driver */
-   if (ctx->Driver.BindArrayObject && newObj)
-      ctx->Driver.BindArrayObject(ctx, newObj);
 }
 
 
@@ -565,7 +558,7 @@ gen_vertex_arrays(struct gl_context *ctx, GLsizei n, GLuint *arrays,
       struct gl_vertex_array_object *obj;
       GLuint name = first + i;
 
-      obj = (*ctx->Driver.NewArrayObject)( ctx, name );
+      obj = _mesa_new_vao(ctx, name);
       if (!obj) {
          _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
          return;
diff --git a/src/mesa/main/blend.c b/src/mesa/main/blend.c
index 4fc32962425..dee5e29d5b8 100644
--- a/src/mesa/main/blend.c
+++ b/src/mesa/main/blend.c
@@ -303,11 +303,6 @@ _mesa_BlendFuncSeparateiARB(GLuint buf, GLenum sfactorRGB, GLenum dfactorRGB,
    ctx->Color.Blend[buf].DstA = dfactorA;
    update_uses_dual_src(ctx, buf);
    ctx->Color._BlendFuncPerBuffer = GL_TRUE;
-
-   if (ctx->Driver.BlendFuncSeparatei) {
-      ctx->Driver.BlendFuncSeparatei(ctx, buf, sfactorRGB, dfactorRGB,
-                                     sfactorA, dfactorA);
-   }
 }
 
 
@@ -406,9 +401,6 @@ _mesa_BlendEquationiARB(GLuint buf, GLenum mode)
    ctx->Color.Blend[buf].EquationRGB = mode;
    ctx->Color.Blend[buf].EquationA = mode;
    ctx->Color._BlendEquationPerBuffer = GL_TRUE;
-
-   if (ctx->Driver.BlendEquationSeparatei)
-      ctx->Driver.BlendEquationSeparatei(ctx, buf, mode, mode);
 }
 
 
@@ -503,9 +495,6 @@ _mesa_BlendEquationSeparateiARB(GLuint buf, GLenum modeRGB, GLenum modeA)
    ctx->Color.Blend[buf].EquationRGB = modeRGB;
    ctx->Color.Blend[buf].EquationA = modeA;
    ctx->Color._BlendEquationPerBuffer = GL_TRUE;
-
-   if (ctx->Driver.BlendEquationSeparatei)
-      ctx->Driver.BlendEquationSeparatei(ctx, buf, modeRGB, modeA);
 }
 
 
@@ -745,9 +734,6 @@ _mesa_ColorMaski( GLuint buf, GLboolean red, GLboolean green,
 
    FLUSH_VERTICES(ctx, _NEW_COLOR);
    COPY_4UBV(ctx->Color.ColorMask[buf], tmp);
-
-   if (ctx->Driver.ColorMaskIndexed)
-      ctx->Driver.ColorMaskIndexed(ctx, buf, red, green, blue, alpha);
 }
 
 
diff --git a/src/mesa/main/context.h b/src/mesa/main/context.h
index 0f7529ad975..1e7a12c8a84 100644
--- a/src/mesa/main/context.h
+++ b/src/mesa/main/context.h
@@ -51,6 +51,7 @@
 
 #include "imports.h"
 #include "mtypes.h"
+#include "vbo/vbo.h"
 
 
 #ifdef __cplusplus
@@ -227,7 +228,7 @@ do {								\
    if (MESA_VERBOSE & VERBOSE_STATE)				\
       _mesa_debug(ctx, "FLUSH_VERTICES in %s\n", MESA_FUNCTION);\
    if (ctx->Driver.NeedFlush & FLUSH_STORED_VERTICES)		\
-      ctx->Driver.FlushVertices(ctx, FLUSH_STORED_VERTICES);	\
+      vbo_exec_FlushVertices(ctx, FLUSH_STORED_VERTICES);	\
    ctx->NewState |= newstate;					\
 } while (0)
 
@@ -246,7 +247,7 @@ do {								\
    if (MESA_VERBOSE & VERBOSE_STATE)				\
       _mesa_debug(ctx, "FLUSH_CURRENT in %s\n", MESA_FUNCTION);	\
    if (ctx->Driver.NeedFlush & FLUSH_UPDATE_CURRENT)		\
-      ctx->Driver.FlushVertices(ctx, FLUSH_UPDATE_CURRENT);	\
+      vbo_exec_FlushVertices(ctx, FLUSH_UPDATE_CURRENT);	\
    ctx->NewState |= newstate;					\
 } while (0)
 
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index 88f37273e1e..496a14f8dad 100644
--- a/src/mesa/main/dd.h
+++ b/src/mesa/main/dd.h
@@ -94,13 +94,6 @@ struct dd_function_table {
    void (*UpdateState)( struct gl_context *ctx, GLbitfield new_state );
 
    /**
-    * Resize the given framebuffer to the given size.
-    * XXX OBSOLETE: this function will be removed in the future.
-    */
-   void (*ResizeBuffers)( struct gl_context *ctx, struct gl_framebuffer *fb,
-                          GLuint width, GLuint height);
-
-   /**
     * This is called whenever glFinish() is called.
     */
    void (*Finish)( struct gl_context *ctx );
@@ -118,12 +111,6 @@ struct dd_function_table {
    void (*Clear)( struct gl_context *ctx, GLbitfield buffers );
 
    /**
-    * Execute glAccum command.
-    */
-   void (*Accum)( struct gl_context *ctx, GLenum op, GLfloat value );
-
-
-   /**
     * Execute glRasterPos, updating the ctx->Current.Raster fields
     */
    void (*RasterPos)( struct gl_context *ctx, const GLfloat v[4] );
@@ -529,22 +516,15 @@ struct dd_function_table {
    /** Set the blend equation */
    void (*BlendEquationSeparate)(struct gl_context *ctx,
                                  GLenum modeRGB, GLenum modeA);
-   void (*BlendEquationSeparatei)(struct gl_context *ctx, GLuint buffer,
-                                  GLenum modeRGB, GLenum modeA);
    /** Specify pixel arithmetic */
    void (*BlendFuncSeparate)(struct gl_context *ctx,
                              GLenum sfactorRGB, GLenum dfactorRGB,
                              GLenum sfactorA, GLenum dfactorA);
-   void (*BlendFuncSeparatei)(struct gl_context *ctx, GLuint buffer,
-                              GLenum sfactorRGB, GLenum dfactorRGB,
-                              GLenum sfactorA, GLenum dfactorA);
    /** Specify a plane against which all geometry is clipped */
    void (*ClipPlane)(struct gl_context *ctx, GLenum plane, const GLfloat *eq);
    /** Enable and disable writing of frame buffer color components */
    void (*ColorMask)(struct gl_context *ctx, GLboolean rmask, GLboolean gmask,
                      GLboolean bmask, GLboolean amask );
-   void (*ColorMaskIndexed)(struct gl_context *ctx, GLuint buf, GLboolean rmask,
-                            GLboolean gmask, GLboolean bmask, GLboolean amask);
    /** Cause a material color to track the current color */
    void (*ColorMaterial)(struct gl_context *ctx, GLenum face, GLenum mode);
    /** Specify whether front- or back-facing facets can be culled */
@@ -565,8 +545,6 @@ struct dd_function_table {
    void (*Enable)(struct gl_context *ctx, GLenum cap, GLboolean state);
    /** Specify fog parameters */
    void (*Fogfv)(struct gl_context *ctx, GLenum pname, const GLfloat *params);
-   /** Specify implementation-specific hints */
-   void (*Hint)(struct gl_context *ctx, GLenum target, GLenum mode);
    /** Set light source parameters.
     * Note: for GL_POSITION and GL_SPOT_DIRECTION, params will have already
     * been transformed to eye-space.
@@ -769,26 +747,12 @@ struct dd_function_table {
                                 GLint *bytesWritten);
    /*@}*/
 
-
-   /**
-    * \name Vertex Array objects
-    */
-   /*@{*/
-   struct gl_vertex_array_object * (*NewArrayObject)(struct gl_context *ctx, GLuint id);
-   void (*DeleteArrayObject)(struct gl_context *ctx, struct gl_vertex_array_object *);
-   void (*BindArrayObject)(struct gl_context *ctx, struct gl_vertex_array_object *);
-   /*@}*/
-
    /**
     * \name GLSL-related functions (ARB extensions and OpenGL 2.x)
     */
    /*@{*/
    struct gl_shader *(*NewShader)(struct gl_context *ctx,
                                   GLuint name, GLenum type);
-   void (*DeleteShader)(struct gl_context *ctx, struct gl_shader *shader);
-   struct gl_shader_program *(*NewShaderProgram)(GLuint name);
-   void (*DeleteShaderProgram)(struct gl_context *ctx,
-                               struct gl_shader_program *shProg);
    void (*UseProgram)(struct gl_context *ctx, struct gl_shader_program *shProg);
    /*@}*/
 
@@ -825,67 +789,15 @@ struct dd_function_table {
     */
    GLbitfield NeedFlush;
 
-   /** Need to call SaveFlushVertices() upon state change? */
+   /** Need to call vbo_save_SaveFlushVertices() upon state change? */
    GLboolean SaveNeedFlush;
 
-   /* Called prior to any of the GLvertexformat functions being
-    * called.  Paired with Driver.FlushVertices().
-    */
-   void (*BeginVertices)( struct gl_context *ctx );
-
-   /**
-    * If inside glBegin()/glEnd(), it should assert(0).  Otherwise, if
-    * FLUSH_STORED_VERTICES bit in \p flags is set flushes any buffered
-    * vertices, if FLUSH_UPDATE_CURRENT bit is set updates
-    * __struct gl_contextRec::Current and gl_light_attrib::Material
-    *
-    * Note that the default T&L engine never clears the
-    * FLUSH_UPDATE_CURRENT bit, even after performing the update.
-    */
-   void (*FlushVertices)( struct gl_context *ctx, GLuint flags );
-   void (*SaveFlushVertices)( struct gl_context *ctx );
-
-   /**
-    * Give the driver the opportunity to hook in its own vtxfmt for
-    * compiling optimized display lists.  This is called on each valid
-    * glBegin() during list compilation.
-    */
-   GLboolean (*NotifySaveBegin)( struct gl_context *ctx, GLenum mode );
-
    /**
     * Notify driver that the special derived value _NeedEyeCoords has
     * changed.
     */
    void (*LightingSpaceChange)( struct gl_context *ctx );
 
-   /**
-    * Called by glNewList().
-    *
-    * Let the T&L component know what is going on with display lists
-    * in time to make changes to dispatch tables, etc.
-    */
-   void (*NewList)( struct gl_context *ctx, GLuint list, GLenum mode );
-   /**
-    * Called by glEndList().
-    *
-    * \sa dd_function_table::NewList.
-    */
-   void (*EndList)( struct gl_context *ctx );
-
-   /**
-    * Called by glCallList(s).
-    *
-    * Notify the T&L component before and after calling a display list.
-    */
-   void (*BeginCallList)( struct gl_context *ctx, 
-			  struct gl_display_list *dlist );
-   /**
-    * Called by glEndCallList().
-    *
-    * \sa dd_function_table::BeginCallList.
-    */
-   void (*EndCallList)( struct gl_context *ctx );
-
    /**@}*/
 
    /**
@@ -963,8 +875,6 @@ struct dd_function_table {
     */
    struct gl_sampler_object * (*NewSamplerObject)(struct gl_context *ctx,
                                                   GLuint name);
-   void (*DeleteSamplerObject)(struct gl_context *ctx,
-                               struct gl_sampler_object *samp);
 
    /**
     * \name Return a timestamp in nanoseconds as defined by GL_ARB_timer_query.
@@ -1007,12 +917,6 @@ struct dd_function_table {
     * \name GL_ARB_shader_image_load_store interface.
     */
    /** @{ */
-   void (*BindImageTexture)(struct gl_context *ctx,
-                            struct gl_image_unit *unit,
-                            struct gl_texture_object *texObj,
-                            GLint level, GLboolean layered, GLint layer,
-                            GLenum access, GLenum format);
-
    void (*MemoryBarrier)(struct gl_context *ctx, GLbitfield barriers);
    /** @} */
 
diff --git a/src/mesa/main/dlist.c b/src/mesa/main/dlist.c
index 5554738d1a3..e8059c7b260 100644
--- a/src/mesa/main/dlist.c
+++ b/src/mesa/main/dlist.c
@@ -105,13 +105,12 @@ struct gl_list_extensions
  * \param ctx GL context.
  *
  * Checks if dd_function_table::SaveNeedFlush is marked to flush
- * stored (save) vertices, and calls
- * dd_function_table::SaveFlushVertices if so.
+ * stored (save) vertices, and calls vbo_save_SaveFlushVertices if so.
  */
 #define SAVE_FLUSH_VERTICES(ctx)		\
 do {						\
    if (ctx->Driver.SaveNeedFlush)		\
-      ctx->Driver.SaveFlushVertices(ctx);	\
+      vbo_save_SaveFlushVertices(ctx);               \
 } while (0)
 
 
@@ -5466,7 +5465,7 @@ save_Begin(GLenum mode)
       /* Give the driver an opportunity to hook in an optimized
        * display list compiler.
        */
-      if (ctx->Driver.NotifySaveBegin(ctx, mode))
+      if (vbo_save_NotifyBegin(ctx, mode))
          return;
 
       SAVE_FLUSH_VERTICES(ctx);
@@ -7743,8 +7742,7 @@ execute_list(struct gl_context *ctx, GLuint list)
 
    ctx->ListState.CallDepth++;
 
-   if (ctx->Driver.BeginCallList)
-      ctx->Driver.BeginCallList(ctx, dlist);
+   vbo_save_BeginCallList(ctx, dlist);
 
    n = dlist->Head;
 
@@ -8900,8 +8898,7 @@ execute_list(struct gl_context *ctx, GLuint list)
       }
    }
 
-   if (ctx->Driver.EndCallList)
-      ctx->Driver.EndCallList(ctx);
+   vbo_save_EndCallList(ctx);
 
    ctx->ListState.CallDepth--;
 }
@@ -9029,7 +9026,7 @@ _mesa_NewList(GLuint name, GLenum mode)
    ctx->ListState.CurrentBlock = ctx->ListState.CurrentList->Head;
    ctx->ListState.CurrentPos = 0;
 
-   ctx->Driver.NewList(ctx, name, mode);
+   vbo_save_NewList(ctx, name, mode);
 
    ctx->CurrentDispatch = ctx->Save;
    _glapi_set_dispatch(ctx->CurrentDispatch);
@@ -9063,7 +9060,7 @@ _mesa_EndList(void)
    /* Call before emitting END_OF_LIST, in case the driver wants to
     * emit opcodes itself.
     */
-   ctx->Driver.EndList(ctx);
+   vbo_save_EndList(ctx);
 
    (void) alloc_instruction(ctx, OPCODE_END_OF_LIST, 0);
 
diff --git a/src/mesa/main/ff_fragment_shader.cpp b/src/mesa/main/ff_fragment_shader.cpp
index c0030bc5687..e4e2a18c1da 100644
--- a/src/mesa/main/ff_fragment_shader.cpp
+++ b/src/mesa/main/ff_fragment_shader.cpp
@@ -32,6 +32,7 @@
 #include "main/imports.h"
 #include "main/macros.h"
 #include "main/samplerobj.h"
+#include "main/shaderobj.h"
 #include "main/texenvprogram.h"
 #include "main/texobj.h"
 #include "main/uniforms.h"
@@ -1208,7 +1209,7 @@ create_new_program(struct gl_context *ctx, struct state_key *key)
    p.top_instructions = p.shader->ir;
    p.instructions = p.shader->ir;
    p.state = key;
-   p.shader_program = ctx->Driver.NewShaderProgram(0);
+   p.shader_program = _mesa_new_shader_program(0);
 
    /* Tell the linker to ignore the fact that we're building a
     * separate shader, in case we're in a GLES2 context that would
diff --git a/src/mesa/main/framebuffer.c b/src/mesa/main/framebuffer.c
index 37e2c29c89c..d18166d528e 100644
--- a/src/mesa/main/framebuffer.c
+++ b/src/mesa/main/framebuffer.c
@@ -271,8 +271,7 @@ _mesa_reference_framebuffer_(struct gl_framebuffer **ptr,
  * Resize the given framebuffer's renderbuffers to the new width and height.
  * This should only be used for window-system framebuffers, not
  * user-created renderbuffers (i.e. made with GL_EXT_framebuffer_object).
- * This will typically be called via ctx->Driver.ResizeBuffers() or directly
- * from a device driver.
+ * This will typically be called directly from a device driver.
  *
  * \note it's possible for ctx to be null since a window can be resized
  * without a currently bound rendering context.
@@ -965,3 +964,22 @@ _mesa_print_framebuffer(const struct gl_framebuffer *fb)
       }
    }
 }
+
+bool
+_mesa_is_front_buffer_reading(const struct gl_framebuffer *fb)
+{
+   if (!fb || _mesa_is_user_fbo(fb))
+      return false;
+
+   return fb->_ColorReadBufferIndex == BUFFER_FRONT_LEFT;
+}
+
+bool
+_mesa_is_front_buffer_drawing(const struct gl_framebuffer *fb)
+{
+   if (!fb || _mesa_is_user_fbo(fb))
+      return false;
+
+   return (fb->_NumColorDrawBuffers >= 1 &&
+           fb->_ColorDrawBufferIndexes[0] == BUFFER_FRONT_LEFT);
+}
diff --git a/src/mesa/main/framebuffer.h b/src/mesa/main/framebuffer.h
index 08e43222045..bfc8a0836e7 100644
--- a/src/mesa/main/framebuffer.h
+++ b/src/mesa/main/framebuffer.h
@@ -139,4 +139,10 @@ _mesa_get_read_renderbuffer_for_format(const struct gl_context *ctx,
 extern void
 _mesa_print_framebuffer(const struct gl_framebuffer *fb);
 
+extern bool
+_mesa_is_front_buffer_reading(const struct gl_framebuffer *fb);
+
+extern bool
+_mesa_is_front_buffer_drawing(const struct gl_framebuffer *fb);
+
 #endif /* FRAMEBUFFER_H */
diff --git a/src/mesa/main/get_hash_params.py b/src/mesa/main/get_hash_params.py
index 38b08b02a65..c295615b475 100644
--- a/src/mesa/main/get_hash_params.py
+++ b/src/mesa/main/get_hash_params.py
@@ -124,6 +124,15 @@ descriptor=[
 
 # GL_EXT_texture_filter_anisotropic
   [ "MAX_TEXTURE_MAX_ANISOTROPY_EXT", "CONTEXT_FLOAT(Const.MaxTextureMaxAnisotropy), extra_EXT_texture_filter_anisotropic" ],
+
+# GL_KHR_debug (GL 4.3)/ GL_ARB_debug_output
+  [ "DEBUG_LOGGED_MESSAGES", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
+  [ "DEBUG_NEXT_LOGGED_MESSAGE_LENGTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
+  [ "MAX_DEBUG_LOGGED_MESSAGES", "CONST(MAX_DEBUG_LOGGED_MESSAGES), NO_EXTRA" ],
+  [ "MAX_DEBUG_MESSAGE_LENGTH", "CONST(MAX_DEBUG_MESSAGE_LENGTH), NO_EXTRA" ],
+  [ "MAX_LABEL_LENGTH", "CONST(MAX_LABEL_LENGTH), NO_EXTRA" ],
+  [ "MAX_DEBUG_GROUP_STACK_DEPTH", "CONST(MAX_DEBUG_GROUP_STACK_DEPTH), NO_EXTRA" ],
+  [ "DEBUG_GROUP_STACK_DEPTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
 ]},
 
 # Enums in OpenGL and GLES1
@@ -791,15 +800,6 @@ descriptor=[
 # GL_ARB_robustness
   [ "RESET_NOTIFICATION_STRATEGY_ARB", "CONTEXT_ENUM(Const.ResetStrategy), NO_EXTRA" ],
 
-# GL_KHR_debug (GL 4.3)/ GL_ARB_debug_output
-  [ "DEBUG_LOGGED_MESSAGES", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
-  [ "DEBUG_NEXT_LOGGED_MESSAGE_LENGTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
-  [ "MAX_DEBUG_LOGGED_MESSAGES", "CONST(MAX_DEBUG_LOGGED_MESSAGES), NO_EXTRA" ],
-  [ "MAX_DEBUG_MESSAGE_LENGTH", "CONST(MAX_DEBUG_MESSAGE_LENGTH), NO_EXTRA" ],
-  [ "MAX_LABEL_LENGTH", "CONST(MAX_LABEL_LENGTH), NO_EXTRA" ],
-  [ "MAX_DEBUG_GROUP_STACK_DEPTH", "CONST(MAX_DEBUG_GROUP_STACK_DEPTH), NO_EXTRA" ],
-  [ "DEBUG_GROUP_STACK_DEPTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
-
   [ "MAX_DUAL_SOURCE_DRAW_BUFFERS", "CONTEXT_INT(Const.MaxDualSourceDrawBuffers), extra_ARB_blend_func_extended" ],
 
 # GL_ARB_uniform_buffer_object
diff --git a/src/mesa/main/glformats.c b/src/mesa/main/glformats.c
index 7dab33c0051..faa63825380 100644
--- a/src/mesa/main/glformats.c
+++ b/src/mesa/main/glformats.c
@@ -2678,6 +2678,7 @@ _mesa_es3_effective_internal_format_for_format_and_type(GLenum format,
        * internal formats, they do not correspond to GL constants, so the base
        * format is returned instead.
        */
+      case GL_BGRA_EXT:
       case GL_LUMINANCE_ALPHA:
       case GL_LUMINANCE:
       case GL_ALPHA:
@@ -2797,8 +2798,19 @@ _mesa_es3_error_check_format_and_type(const struct gl_context *ctx,
       if (effectiveInternalFormat == GL_NONE)
          return GL_INVALID_OPERATION;
 
-      GLenum baseInternalFormat =
-         _mesa_base_tex_format(ctx, effectiveInternalFormat);
+      GLenum baseInternalFormat;
+      if (internalFormat == GL_BGRA_EXT) {
+         /* Unfortunately, _mesa_base_tex_format returns a base format of
+          * GL_RGBA for GL_BGRA_EXT.  This makes perfect sense if you're
+          * asking the question, "what channels does this format have?"
+          * However, if we're trying to determine if two internal formats
+          * match in the ES3 sense, we actually want GL_BGRA.
+          */
+         baseInternalFormat = GL_BGRA_EXT;
+      } else {
+         baseInternalFormat =
+            _mesa_base_tex_format(ctx, effectiveInternalFormat);
+      }
 
       if (internalFormat != baseInternalFormat)
          return GL_INVALID_OPERATION;
@@ -2807,6 +2819,11 @@ _mesa_es3_error_check_format_and_type(const struct gl_context *ctx,
    }
 
    switch (format) {
+   case GL_BGRA_EXT:
+      if (type != GL_UNSIGNED_BYTE || internalFormat != GL_BGRA)
+         return GL_INVALID_OPERATION;
+      break;
+
    case GL_RGBA:
       switch (type) {
       case GL_UNSIGNED_BYTE:
diff --git a/src/mesa/main/hint.c b/src/mesa/main/hint.c
index 984239a7276..5d0c15d35ab 100644
--- a/src/mesa/main/hint.c
+++ b/src/mesa/main/hint.c
@@ -123,11 +123,6 @@ _mesa_Hint( GLenum target, GLenum mode )
       default:
          goto invalid_target;
    }
-
-   if (ctx->Driver.Hint) {
-      (*ctx->Driver.Hint)( ctx, target, mode );
-   }
-
    return;
 
 invalid_target:
diff --git a/src/mesa/main/pack.c b/src/mesa/main/pack.c
index 7147fd6e4fe..00e31b05c99 100644
--- a/src/mesa/main/pack.c
+++ b/src/mesa/main/pack.c
@@ -43,7 +43,6 @@
 
 
 #include "glheader.h"
-#include "colormac.h"
 #include "enums.h"
 #include "image.h"
 #include "imports.h"
@@ -1124,7 +1123,8 @@ _mesa_pack_depth_span( struct gl_context *ctx, GLuint n, GLvoid *dest,
       }
       break;
    default:
-      _mesa_problem(ctx, "bad type in _mesa_pack_depth_span");
+      _mesa_problem(ctx, "bad type in _mesa_pack_depth_span (%s)",
+                    _mesa_enum_to_string(dstType));
    }
 
    free(depthCopy);
diff --git a/src/mesa/main/samplerobj.c b/src/mesa/main/samplerobj.c
index c7b96664c21..676dd367b3f 100644
--- a/src/mesa/main/samplerobj.c
+++ b/src/mesa/main/samplerobj.c
@@ -72,6 +72,14 @@ lookup_samplerobj_locked(struct gl_context *ctx, GLuint name)
          _mesa_HashLookupLocked(ctx->Shared->SamplerObjects, name);
 }
 
+static void
+delete_sampler_object(struct gl_context *ctx,
+                      struct gl_sampler_object *sampObj)
+{
+   mtx_destroy(&sampObj->Mutex);
+   free(sampObj->Label);
+   free(sampObj);
+}
 
 /**
  * Handle reference counting.
@@ -94,10 +102,8 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
       deleteFlag = (oldSamp->RefCount == 0);
       mtx_unlock(&oldSamp->Mutex);
 
-      if (deleteFlag) {
-	 assert(ctx->Driver.DeleteSamplerObject);
-         ctx->Driver.DeleteSamplerObject(ctx, oldSamp);
-      }
+      if (deleteFlag)
+         delete_sampler_object(ctx, oldSamp);
 
       *ptr = NULL;
    }
@@ -162,19 +168,6 @@ _mesa_new_sampler_object(struct gl_context *ctx, GLuint name)
    return sampObj;
 }
 
-
-/**
- * Fallback for ctx->Driver.DeleteSamplerObject();
- */
-static void
-_mesa_delete_sampler_object(struct gl_context *ctx,
-                            struct gl_sampler_object *sampObj)
-{
-   mtx_destroy(&sampObj->Mutex);
-   free(sampObj->Label);
-   free(sampObj);
-}
-
 static void
 create_samplers(struct gl_context *ctx, GLsizei count, GLuint *samplers,
                 const char *caller)
@@ -628,8 +621,12 @@ static GLuint
 set_sampler_compare_mode(struct gl_context *ctx,
                          struct gl_sampler_object *samp, GLint param)
 {
+    /* If GL_ARB_shadow is not supported, don't report an error.  The
+     * sampler object extension spec isn't clear on this extension interaction.
+     * Silences errors with Wine on older GPUs such as R200.
+     */
    if (!ctx->Extensions.ARB_shadow)
-      return INVALID_PNAME;
+      return GL_FALSE;
 
    if (samp->CompareMode == param)
       return GL_FALSE;
@@ -649,8 +646,12 @@ static GLuint
 set_sampler_compare_func(struct gl_context *ctx,
                          struct gl_sampler_object *samp, GLint param)
 {
+    /* If GL_ARB_shadow is not supported, don't report an error.  The
+     * sampler object extension spec isn't clear on this extension interaction.
+     * Silences errors with Wine on older GPUs such as R200.
+     */
    if (!ctx->Extensions.ARB_shadow)
-      return INVALID_PNAME;
+      return GL_FALSE;
 
    if (samp->CompareFunc == param)
       return GL_FALSE;
@@ -1336,13 +1337,9 @@ _mesa_GetSamplerParameteriv(GLuint sampler, GLenum pname, GLint *params)
       *params = IROUND(sampObj->LodBias);
       break;
    case GL_TEXTURE_COMPARE_MODE:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareMode;
       break;
    case GL_TEXTURE_COMPARE_FUNC:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareFunc;
       break;
    case GL_TEXTURE_MAX_ANISOTROPY_EXT:
@@ -1425,13 +1422,9 @@ _mesa_GetSamplerParameterfv(GLuint sampler, GLenum pname, GLfloat *params)
       *params = sampObj->LodBias;
       break;
    case GL_TEXTURE_COMPARE_MODE:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = (GLfloat) sampObj->CompareMode;
       break;
    case GL_TEXTURE_COMPARE_FUNC:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = (GLfloat) sampObj->CompareFunc;
       break;
    case GL_TEXTURE_MAX_ANISOTROPY_EXT:
@@ -1504,13 +1497,9 @@ _mesa_GetSamplerParameterIiv(GLuint sampler, GLenum pname, GLint *params)
       *params = (GLint) sampObj->LodBias;
       break;
    case GL_TEXTURE_COMPARE_MODE:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareMode;
       break;
    case GL_TEXTURE_COMPARE_FUNC:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareFunc;
       break;
    case GL_TEXTURE_MAX_ANISOTROPY_EXT:
@@ -1583,13 +1572,9 @@ _mesa_GetSamplerParameterIuiv(GLuint sampler, GLenum pname, GLuint *params)
       *params = (GLuint) sampObj->LodBias;
       break;
    case GL_TEXTURE_COMPARE_MODE:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareMode;
       break;
    case GL_TEXTURE_COMPARE_FUNC:
-      if (!ctx->Extensions.ARB_shadow)
-         goto invalid_pname;
       *params = sampObj->CompareFunc;
       break;
    case GL_TEXTURE_MAX_ANISOTROPY_EXT:
@@ -1626,5 +1611,4 @@ void
 _mesa_init_sampler_object_functions(struct dd_function_table *driver)
 {
    driver->NewSamplerObject = _mesa_new_sampler_object;
-   driver->DeleteSamplerObject = _mesa_delete_sampler_object;
 }
diff --git a/src/mesa/main/shader_query.cpp b/src/mesa/main/shader_query.cpp
index 718967605b5..6d73e3bdcf2 100644
--- a/src/mesa/main/shader_query.cpp
+++ b/src/mesa/main/shader_query.cpp
@@ -996,7 +996,7 @@ program_resource_top_level_array_stride(struct gl_shader_program *shProg,
                const glsl_type *array_type = field->type->fields.array;
 
                if (interface->interface_packing != GLSL_INTERFACE_PACKING_STD430) {
-                  if (array_type->is_record()) {
+                  if (array_type->is_record() || array_type->is_array()) {
                      array_stride = array_type->std140_size(row_major);
                      array_stride = glsl_align(array_stride, 16);
                   } else {
diff --git a/src/mesa/main/shaderapi.c b/src/mesa/main/shaderapi.c
index 82a1ec37130..9dd1054c8ee 100644
--- a/src/mesa/main/shaderapi.c
+++ b/src/mesa/main/shaderapi.c
@@ -320,7 +320,7 @@ create_shader_program(struct gl_context *ctx)
 
    name = _mesa_HashFindFreeKeyBlock(ctx->Shared->ShaderObjects, 1);
 
-   shProg = ctx->Driver.NewShaderProgram(name);
+   shProg = _mesa_new_shader_program(name);
 
    _mesa_HashInsert(ctx->Shared->ShaderObjects, name, shProg);
 
@@ -2597,7 +2597,7 @@ _mesa_GetUniformSubroutineuiv(GLenum shadertype, GLint location,
 
    {
       struct gl_uniform_storage *uni = sh->SubroutineUniformRemapTable[location];
-      int offset = location - uni->subroutine[stage].index;
+      int offset = location - uni->opaque[stage].index;
       memcpy(params, &uni->storage[offset],
 	     sizeof(GLuint));
    }
diff --git a/src/mesa/main/shaderimage.c b/src/mesa/main/shaderimage.c
index c4bba842ca7..bd4b7c7be3b 100644
--- a/src/mesa/main/shaderimage.c
+++ b/src/mesa/main/shaderimage.c
@@ -577,10 +577,6 @@ _mesa_BindImageTexture(GLuint unit, GLuint texture, GLint level,
       u->Layered = GL_FALSE;
       u->Layer = 0;
    }
-
-   if (ctx->Driver.BindImageTexture)
-      ctx->Driver.BindImageTexture(ctx, u, u->TexObj, level, layered,
-                                   layer, access, format);
 }
 
 void GLAPIENTRY
@@ -719,11 +715,6 @@ _mesa_BindImageTextures(GLuint first, GLsizei count, const GLuint *textures)
          u->_ActualFormat = MESA_FORMAT_R_UNORM8;
          u->_Valid = GL_FALSE;
       }
-
-      /* Pass the BindImageTexture call down to the device driver */
-      if (ctx->Driver.BindImageTexture)
-         ctx->Driver.BindImageTexture(ctx, u, u->TexObj, u->Level, u->Layered,
-                                      u->Layer, u->Access, u->Format);
    }
 
    _mesa_end_texture_lookups(ctx);
diff --git a/src/mesa/main/shaderobj.c b/src/mesa/main/shaderobj.c
index 5cd37d7e4c4..4e85fda24b4 100644
--- a/src/mesa/main/shaderobj.c
+++ b/src/mesa/main/shaderobj.c
@@ -74,7 +74,7 @@ _mesa_reference_shader(struct gl_context *ctx, struct gl_shader **ptr,
       if (deleteFlag) {
 	 if (old->Name != 0)
 	    _mesa_HashRemove(ctx->Shared->ShaderObjects, old->Name);
-         ctx->Driver.DeleteShader(ctx, old);
+         _mesa_delete_shader(ctx, old);
       }
 
       *ptr = NULL;
@@ -116,9 +116,8 @@ _mesa_new_shader(struct gl_context *ctx, GLuint name, GLenum type)
 
 /**
  * Delete a shader object.
- * Called via ctx->Driver.DeleteShader().
  */
-static void
+void
 _mesa_delete_shader(struct gl_context *ctx, struct gl_shader *sh)
 {
    free((void *)sh->Source);
@@ -210,7 +209,7 @@ _mesa_reference_shader_program_(struct gl_context *ctx,
       if (deleteFlag) {
 	 if (old->Name != 0)
 	    _mesa_HashRemove(ctx->Shared->ShaderObjects, old->Name);
-         ctx->Driver.DeleteShaderProgram(ctx, old);
+         _mesa_delete_shader_program(ctx, old);
       }
 
       *ptr = NULL;
@@ -246,9 +245,8 @@ init_shader_program(struct gl_shader_program *prog)
 
 /**
  * Allocate a new gl_shader_program object, initialize it.
- * Called via ctx->Driver.NewShaderProgram()
  */
-static struct gl_shader_program *
+struct gl_shader_program *
 _mesa_new_shader_program(GLuint name)
 {
    struct gl_shader_program *shProg;
@@ -362,7 +360,7 @@ _mesa_free_shader_program_data(struct gl_context *ctx,
 
    for (sh = 0; sh < MESA_SHADER_STAGES; sh++) {
       if (shProg->_LinkedShaders[sh] != NULL) {
-	 ctx->Driver.DeleteShader(ctx, shProg->_LinkedShaders[sh]);
+	 _mesa_delete_shader(ctx, shProg->_LinkedShaders[sh]);
 	 shProg->_LinkedShaders[sh] = NULL;
       }
    }
@@ -374,10 +372,10 @@ _mesa_free_shader_program_data(struct gl_context *ctx,
 
 /**
  * Free/delete a shader program object.
- * Called via ctx->Driver.DeleteShaderProgram().
  */
-static void
-_mesa_delete_shader_program(struct gl_context *ctx, struct gl_shader_program *shProg)
+void
+_mesa_delete_shader_program(struct gl_context *ctx,
+                            struct gl_shader_program *shProg)
 {
    _mesa_free_shader_program_data(ctx, shProg);
 
@@ -439,8 +437,5 @@ void
 _mesa_init_shader_object_functions(struct dd_function_table *driver)
 {
    driver->NewShader = _mesa_new_shader;
-   driver->DeleteShader = _mesa_delete_shader;
-   driver->NewShaderProgram = _mesa_new_shader_program;
-   driver->DeleteShaderProgram = _mesa_delete_shader_program;
    driver->LinkShader = _mesa_ir_link_shader;
 }
diff --git a/src/mesa/main/shaderobj.h b/src/mesa/main/shaderobj.h
index 943044e37cd..796de470735 100644
--- a/src/mesa/main/shaderobj.h
+++ b/src/mesa/main/shaderobj.h
@@ -82,6 +82,9 @@ _mesa_init_shader(struct gl_context *ctx, struct gl_shader *shader);
 extern struct gl_shader *
 _mesa_new_shader(struct gl_context *ctx, GLuint name, GLenum type);
 
+extern void
+_mesa_delete_shader(struct gl_context *ctx, struct gl_shader *sh);
+
 extern struct gl_shader_program *
 _mesa_lookup_shader_program(struct gl_context *ctx, GLuint name);
 
@@ -89,6 +92,9 @@ extern struct gl_shader_program *
 _mesa_lookup_shader_program_err(struct gl_context *ctx, GLuint name,
                                 const char *caller);
 
+extern struct gl_shader_program *
+_mesa_new_shader_program(GLuint name);
+
 extern void
 _mesa_clear_shader_program_data(struct gl_shader_program *shProg);
 
@@ -96,6 +102,9 @@ extern void
 _mesa_free_shader_program_data(struct gl_context *ctx,
                                struct gl_shader_program *shProg);
 
+extern void
+_mesa_delete_shader_program(struct gl_context *ctx,
+                            struct gl_shader_program *shProg);
 
 
 extern void
diff --git a/src/mesa/main/shared.c b/src/mesa/main/shared.c
index d5ac9f1fb13..1acaf59f432 100644
--- a/src/mesa/main/shared.c
+++ b/src/mesa/main/shared.c
@@ -219,12 +219,12 @@ delete_shader_cb(GLuint id, void *data, void *userData)
    struct gl_context *ctx = (struct gl_context *) userData;
    struct gl_shader *sh = (struct gl_shader *) data;
    if (_mesa_validate_shader_target(ctx, sh->Type)) {
-      ctx->Driver.DeleteShader(ctx, sh);
+      _mesa_delete_shader(ctx, sh);
    }
    else {
       struct gl_shader_program *shProg = (struct gl_shader_program *) data;
       assert(shProg->Type == GL_SHADER_PROGRAM_MESA);
-      ctx->Driver.DeleteShaderProgram(ctx, shProg);
+      _mesa_delete_shader_program(ctx, shProg);
    }
 }
 
diff --git a/src/mesa/main/streaming-load-memcpy.c b/src/mesa/main/streaming-load-memcpy.c
index d7147afdc5c..32854b60eb2 100644
--- a/src/mesa/main/streaming-load-memcpy.c
+++ b/src/mesa/main/streaming-load-memcpy.c
@@ -59,6 +59,9 @@ _mesa_streaming_load_memcpy(void *restrict dst, void *restrict src, size_t len)
       len -= MIN2(bytes_before_alignment_boundary, len);
    }
 
+   if (len >= 64)
+      _mm_mfence();
+
    while (len >= 64) {
       __m128i *dst_cacheline = (__m128i *)d;
       __m128i *src_cacheline = (__m128i *)s;
diff --git a/src/mesa/main/tests/dispatch_sanity.cpp b/src/mesa/main/tests/dispatch_sanity.cpp
index b19c6d74bc0..ac2d2332df8 100644
--- a/src/mesa/main/tests/dispatch_sanity.cpp
+++ b/src/mesa/main/tests/dispatch_sanity.cpp
@@ -2039,6 +2039,19 @@ const struct function gles11_functions_possible[] = {
    { "glUnmapBufferOES", 11, -1 },
    { "glVertexPointer", 11, _gloffset_VertexPointer },
    { "glViewport", 11, _gloffset_Viewport },
+
+   /* GL_KHR_debug */
+   { "glPushDebugGroupKHR", 11, -1 },
+   { "glPopDebugGroupKHR", 11, -1 },
+   { "glDebugMessageCallbackKHR", 11, -1 },
+   { "glDebugMessageControlKHR", 11, -1 },
+   { "glDebugMessageInsertKHR", 11, -1 },
+   { "glGetDebugMessageLogKHR", 11, -1 },
+   { "glGetObjectLabelKHR", 11, -1 },
+   { "glGetObjectPtrLabelKHR", 11, -1 },
+   { "glObjectLabelKHR", 11, -1 },
+   { "glObjectPtrLabelKHR", 11, -1 },
+
    { NULL, 0, -1 }
 };
 
@@ -2262,6 +2275,18 @@ const struct function gles2_functions_possible[] = {
    { "glEndPerfQueryINTEL", 20, -1 },
    { "glGetPerfQueryDataINTEL", 20, -1 },
 
+   /* GL_KHR_debug */
+   { "glPushDebugGroupKHR", 20, -1 },
+   { "glPopDebugGroupKHR", 20, -1 },
+   { "glDebugMessageCallbackKHR", 20, -1 },
+   { "glDebugMessageControlKHR", 20, -1 },
+   { "glDebugMessageInsertKHR", 20, -1 },
+   { "glGetDebugMessageLogKHR", 20, -1 },
+   { "glGetObjectLabelKHR", 20, -1 },
+   { "glGetObjectPtrLabelKHR", 20, -1 },
+   { "glObjectLabelKHR", 20, -1 },
+   { "glObjectPtrLabelKHR", 20, -1 },
+
    { NULL, 0, -1 }
 };
 
diff --git a/src/mesa/main/uniform_query.cpp b/src/mesa/main/uniform_query.cpp
index 33c959dc1a5..d48729778ae 100644
--- a/src/mesa/main/uniform_query.cpp
+++ b/src/mesa/main/uniform_query.cpp
@@ -804,11 +804,11 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
 
 	 /* If the shader stage doesn't use the sampler uniform, skip this.
 	  */
-	 if (sh == NULL || !uni->sampler[i].active)
+	 if (sh == NULL || !uni->opaque[i].active)
 	    continue;
 
          for (int j = 0; j < count; j++) {
-            sh->SamplerUnits[uni->sampler[i].index + offset + j] =
+            sh->SamplerUnits[uni->opaque[i].index + offset + j] =
                ((unsigned *) values)[j];
          }
 
@@ -850,11 +850,11 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
     */
    if (uni->type->is_image()) {
       for (int i = 0; i < MESA_SHADER_STAGES; i++) {
-	 if (uni->image[i].active) {
+	 if (uni->opaque[i].active) {
             struct gl_shader *sh = shProg->_LinkedShaders[i];
 
             for (int j = 0; j < count; j++)
-               sh->ImageUnits[uni->image[i].index + offset + j] =
+               sh->ImageUnits[uni->opaque[i].index + offset + j] =
                   ((GLint *) values)[j];
          }
       }
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index 4df57c148c7..887d0c03a50 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@@ -2354,7 +2354,7 @@ _mesa_print_arrays(struct gl_context *ctx)
 void
 _mesa_init_varray(struct gl_context *ctx)
 {
-   ctx->Array.DefaultVAO = ctx->Driver.NewArrayObject(ctx, 0);
+   ctx->Array.DefaultVAO = _mesa_new_vao(ctx, 0);
    _mesa_reference_vao(ctx, &ctx->Array.VAO, ctx->Array.DefaultVAO);
    ctx->Array.ActiveTexture = 0;   /* GL_ARB_multitexture */
 
diff --git a/src/mesa/program/ir_to_mesa.cpp b/src/mesa/program/ir_to_mesa.cpp
index 98032456662..0214b8e684c 100644
--- a/src/mesa/program/ir_to_mesa.cpp
+++ b/src/mesa/program/ir_to_mesa.cpp
@@ -2353,11 +2353,12 @@ add_uniform_to_shader::visit_field(const glsl_type *type, const char *name,
 	 struct gl_uniform_storage *storage =
 	    &this->shader_program->UniformStorage[location];
 
-         assert(storage->sampler[shader_type].active);
+         assert(storage->type->is_sampler() &&
+                storage->opaque[shader_type].active);
 
 	 for (unsigned int j = 0; j < size / 4; j++)
             params->ParameterValues[index + j][0].f =
-               storage->sampler[shader_type].index + j;
+               storage->opaque[shader_type].index + j;
       }
    }
 
diff --git a/src/mesa/program/sampler.cpp b/src/mesa/program/sampler.cpp
index b1168fdade8..1198a3c45f1 100644
--- a/src/mesa/program/sampler.cpp
+++ b/src/mesa/program/sampler.cpp
@@ -119,7 +119,7 @@ _mesa_get_sampler_uniform_value(class ir_dereference *sampler,
       return 0;
    }
 
-   if (!shader_program->UniformStorage[location].sampler[shader].active) {
+   if (!shader_program->UniformStorage[location].opaque[shader].active) {
       assert(0 && "cannot return a sampler");
       linker_error(shader_program,
 		   "cannot return a sampler named %s, because it is not "
@@ -128,7 +128,7 @@ _mesa_get_sampler_uniform_value(class ir_dereference *sampler,
       return 0;
    }
 
-   return shader_program->UniformStorage[location].sampler[shader].index +
+   return shader_program->UniformStorage[location].opaque[shader].index +
           getname.offset;
 }
 
diff --git a/src/mesa/state_tracker/st_atom_rasterizer.c b/src/mesa/state_tracker/st_atom_rasterizer.c
index cceed42c828..0f01e9939de 100644
--- a/src/mesa/state_tracker/st_atom_rasterizer.c
+++ b/src/mesa/state_tracker/st_atom_rasterizer.c
@@ -237,6 +237,14 @@ static void update_raster_state( struct st_context *st )
    /* _NEW_MULTISAMPLE */
    raster->multisample = ctx->Multisample._Enabled;
 
+   /* _NEW_MULTISAMPLE | _NEW_BUFFERS */
+   raster->force_persample_interp =
+         st->can_force_persample_interp &&
+         ctx->Multisample._Enabled &&
+         ctx->Multisample.SampleShading &&
+         ctx->Multisample.MinSampleShadingValue *
+         ctx->DrawBuffer->Visual.samples > 1;
+
    /* _NEW_SCISSOR */
    raster->scissor = ctx->Scissor.EnableFlags;
 
diff --git a/src/mesa/state_tracker/st_atom_shader.c b/src/mesa/state_tracker/st_atom_shader.c
index fee15a980f3..1e880a107c0 100644
--- a/src/mesa/state_tracker/st_atom_shader.c
+++ b/src/mesa/state_tracker/st_atom_shader.c
@@ -70,8 +70,15 @@ update_fp( struct st_context *st )
    key.clamp_color = st->clamp_frag_color_in_shader &&
                      st->ctx->Color._ClampFragmentColor;
 
-   /* Ignore sample qualifier while computing this flag. */
+   /* Don't set it if the driver can force the interpolation by itself.
+    * If SAMPLE_ID or SAMPLE_POS are used, the interpolation is set
+    * automatically.
+    * Ignore sample qualifier while computing this flag.
+    */
    key.persample_shading =
+      !st->can_force_persample_interp &&
+      !(stfp->Base.Base.SystemValuesRead & (SYSTEM_BIT_SAMPLE_ID |
+                                            SYSTEM_BIT_SAMPLE_POS)) &&
       _mesa_get_min_invocations_per_fragment(st->ctx, &stfp->Base, true) > 1;
 
    st->fp_variant = st_get_fp_variant(st, stfp, &key);
diff --git a/src/mesa/state_tracker/st_cb_bufferobjects.c b/src/mesa/state_tracker/st_cb_bufferobjects.c
index db254c2144e..8afd336779f 100644
--- a/src/mesa/state_tracker/st_cb_bufferobjects.c
+++ b/src/mesa/state_tracker/st_cb_bufferobjects.c
@@ -527,8 +527,4 @@ st_init_bufferobject_functions(struct dd_function_table *functions)
    functions->UnmapBuffer = st_bufferobj_unmap;
    functions->CopyBufferSubData = st_copy_buffer_subdata;
    functions->ClearBufferSubData = st_clear_buffer_subdata;
-
-   /* For GL_APPLE_vertex_array_object */
-   functions->NewArrayObject = _mesa_new_vao;
-   functions->DeleteArrayObject = _mesa_delete_vao;
 }
diff --git a/src/mesa/state_tracker/st_cb_fbo.c b/src/mesa/state_tracker/st_cb_fbo.c
index 9d06a232bfa..ff703fa41cb 100644
--- a/src/mesa/state_tracker/st_cb_fbo.c
+++ b/src/mesa/state_tracker/st_cb_fbo.c
@@ -246,17 +246,6 @@ st_renderbuffer_delete(struct gl_context *ctx, struct gl_renderbuffer *rb)
 
 
 /**
- * Called via ctx->Driver.NewFramebuffer()
- */
-static struct gl_framebuffer *
-st_new_framebuffer(struct gl_context *ctx, GLuint name)
-{
-   /* XXX not sure we need to subclass gl_framebuffer for pipe */
-   return _mesa_new_framebuffer(ctx, name);
-}
-
-
-/**
  * Called via ctx->Driver.NewRenderbuffer()
  */
 static struct gl_renderbuffer *
@@ -826,7 +815,7 @@ st_UnmapRenderbuffer(struct gl_context *ctx,
 
 void st_init_fbo_functions(struct dd_function_table *functions)
 {
-   functions->NewFramebuffer = st_new_framebuffer;
+   functions->NewFramebuffer = _mesa_new_framebuffer;
    functions->NewRenderbuffer = st_new_renderbuffer;
    functions->FramebufferRenderbuffer = _mesa_FramebufferRenderbuffer_sw;
    functions->RenderTexture = st_render_texture;
diff --git a/src/mesa/state_tracker/st_context.c b/src/mesa/state_tracker/st_context.c
index 72c23cad4bc..a9ab5edcf49 100644
--- a/src/mesa/state_tracker/st_context.c
+++ b/src/mesa/state_tracker/st_context.c
@@ -237,6 +237,8 @@ st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe,
                                               PIPE_BIND_SAMPLER_VIEW);
    st->prefer_blit_based_texture_transfer = screen->get_param(screen,
                               PIPE_CAP_PREFER_BLIT_BASED_TEXTURE_TRANSFER);
+   st->can_force_persample_interp = screen->get_param(screen,
+                                          PIPE_CAP_FORCE_PERSAMPLE_INTERP);
 
    st->needs_texcoord_semantic =
       screen->get_param(screen, PIPE_CAP_TGSI_TEXCOORD);
@@ -410,8 +412,6 @@ void st_init_driver_functions(struct pipe_screen *screen,
    _mesa_init_shader_object_functions(functions);
    _mesa_init_sampler_object_functions(functions);
 
-   functions->Accum = _mesa_accum;
-
    st_init_blit_functions(functions);
    st_init_bufferobject_functions(functions);
    st_init_clear_functions(functions);
diff --git a/src/mesa/state_tracker/st_context.h b/src/mesa/state_tracker/st_context.h
index 81d5480431a..a4cda29059d 100644
--- a/src/mesa/state_tracker/st_context.h
+++ b/src/mesa/state_tracker/st_context.h
@@ -98,6 +98,7 @@ struct st_context
    boolean has_etc1;
    boolean has_etc2;
    boolean prefer_blit_based_texture_transfer;
+   boolean can_force_persample_interp;
 
    boolean needs_texcoord_semantic;
    boolean apply_texture_swizzle_to_border_color;
diff --git a/src/mesa/state_tracker/st_program.c b/src/mesa/state_tracker/st_program.c
index e62dd7aab80..a07f8fec309 100644
--- a/src/mesa/state_tracker/st_program.c
+++ b/src/mesa/state_tracker/st_program.c
@@ -619,7 +619,9 @@ st_translate_fragment_program(struct st_context *st,
          else
             interpLocation[slot] = TGSI_INTERPOLATE_LOC_CENTER;
 
-         if (key->persample_shading)
+         if (stfp->Base.Base.SystemValuesRead & (SYSTEM_BIT_SAMPLE_ID |
+                                                 SYSTEM_BIT_SAMPLE_POS) ||
+             key->persample_shading)
             interpLocation[slot] = TGSI_INTERPOLATE_LOC_SAMPLE;
 
          switch (attr) {
@@ -1681,6 +1683,26 @@ st_precompile_shader_variant(struct st_context *st,
       break;
    }
 
+   case GL_TESS_CONTROL_PROGRAM_NV: {
+      struct st_tessctrl_program *p = (struct st_tessctrl_program *)prog;
+      struct st_tcp_variant_key key;
+
+      memset(&key, 0, sizeof(key));
+      key.st = st;
+      st_get_tcp_variant(st, p, &key);
+      break;
+   }
+
+   case GL_TESS_EVALUATION_PROGRAM_NV: {
+      struct st_tesseval_program *p = (struct st_tesseval_program *)prog;
+      struct st_tep_variant_key key;
+
+      memset(&key, 0, sizeof(key));
+      key.st = st;
+      st_get_tep_variant(st, p, &key);
+      break;
+   }
+
    case GL_GEOMETRY_PROGRAM_NV: {
       struct st_geometry_program *p = (struct st_geometry_program *)prog;
       struct st_gp_variant_key key;
diff --git a/src/mesa/tnl_dd/t_dd_dmatmp.h b/src/mesa/tnl_dd/t_dd_dmatmp.h
index e7e19a03597..56fa1a382f7 100644
--- a/src/mesa/tnl_dd/t_dd_dmatmp.h
+++ b/src/mesa/tnl_dd/t_dd_dmatmp.h
@@ -85,8 +85,8 @@ static void TAG(render_points_verts)(struct gl_context *ctx,
          currentsz = dmasz;
       }
    } else {
-      fprintf(stderr, "%s - cannot draw primitive\n", __func__);
-      return;
+      unreachable("Cannot draw primitive; validate_render should have "
+                  "prevented this");
    }
 }
 
@@ -316,11 +316,12 @@ static void TAG(render_poly_verts)(struct gl_context *ctx,
       }
 
       FLUSH();
-   } else if (ctx->Light.ShadeModel == GL_SMOOTH) {
+   } else if (ctx->Light.ShadeModel == GL_SMOOTH ||
+              ctx->Light.ProvokingVertex == GL_FIRST_VERTEX_CONVENTION) {
       TAG(render_tri_fan_verts)( ctx, start, count, flags );
    } else {
-      fprintf(stderr, "%s - cannot draw primitive\n", __func__);
-      return;
+      unreachable("Cannot draw primitive; validate_render should have "
+                  "prevented this");
    }
 }
 
@@ -331,14 +332,7 @@ static void TAG(render_quad_strip_verts)(struct gl_context *ctx,
 {
    GLuint j, nr;
 
-   if (ctx->Light.ShadeModel == GL_FLAT &&
-       TNL_CONTEXT(ctx)->vb.AttribPtr[_TNL_ATTRIB_COLOR0]->stride) {
-      /* Vertices won't fit in a single buffer or elts not available - should
-       * never happen.
-       */
-      fprintf(stderr, "%s - cannot draw primitive\n", __func__);
-      return;
-   } else {
+   if (ctx->Light.ShadeModel == GL_SMOOTH) {
       LOCAL_VARS;
       const unsigned dmasz = GET_SUBSEQUENT_VB_MAX_VERTS() & ~1;
       unsigned currentsz;
@@ -364,6 +358,9 @@ static void TAG(render_quad_strip_verts)(struct gl_context *ctx,
       }
 
       FLUSH();
+   } else {
+      unreachable("Cannot draw primitive; validate_render should have "
+                  "prevented this");
    }
 }
 
@@ -373,28 +370,33 @@ static void TAG(render_quads_verts)(struct gl_context *ctx,
                                     GLuint count,
                                     GLuint flags)
 {
-   LOCAL_VARS;
-   GLuint j;
+   if (ctx->Light.ShadeModel == GL_SMOOTH ||
+       ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION) {
+      LOCAL_VARS;
+      GLuint j;
 
-   /* Emit whole number of quads in total. */
-   count -= count & 3;
+      /* Emit whole number of quads in total. */
+      count -= count & 3;
 
-   /* Hardware doesn't have a quad primitive type -- try to simulate it using
-    * triangle primitive.  This is a win for gears, but is it useful in the
-    * broader world?
-    */
-   INIT(GL_TRIANGLES);
-
-   for (j = 0; j + 3 < count; j += 4) {
-      void *tmp = ALLOC_VERTS(6);
-      /* Send v0, v1, v3
-       */
-      tmp = EMIT_VERTS(ctx, start + j,     2, tmp);
-      tmp = EMIT_VERTS(ctx, start + j + 3, 1, tmp);
-      /* Send v1, v2, v3
+      /* Hardware doesn't have a quad primitive type -- try to simulate it using
+       * triangle primitive.  This is a win for gears, but is it useful in the
+       * broader world?
        */
-      tmp = EMIT_VERTS(ctx, start + j + 1, 3, tmp);
-      (void) tmp;
+      INIT(GL_TRIANGLES);
+
+      for (j = 0; j + 3 < count; j += 4) {
+         void *tmp = ALLOC_VERTS(6);
+         /* Send v0, v1, v3
+          */
+         tmp = EMIT_VERTS(ctx, start + j,     2, tmp);
+         tmp = EMIT_VERTS(ctx, start + j + 3, 1, tmp);
+         /* Send v1, v2, v3
+          */
+         tmp = EMIT_VERTS(ctx, start + j + 1, 3, tmp);
+         (void) tmp;
+      }
+   } else {
+      unreachable("Cannot draw primitive");
    }
 }
 
@@ -461,15 +463,15 @@ static bool TAG(validate_render)(struct gl_context *ctx,
          ok = true;
          break;
       case GL_POLYGON:
-         ok = (HAVE_POLYGONS) || ctx->Light.ShadeModel == GL_SMOOTH;
+         ok = (HAVE_POLYGONS) || ctx->Light.ShadeModel == GL_SMOOTH ||
+              ctx->Light.ProvokingVertex == GL_FIRST_VERTEX_CONVENTION;
          break;
       case GL_QUAD_STRIP:
-         ok = VB->Elts ||
-              (ctx->Light.ShadeModel != GL_FLAT ||
-               VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride == 0);
+         ok = VB->Elts || ctx->Light.ShadeModel == GL_SMOOTH;
          break;
       case GL_QUADS:
-         ok = true; /* flatshading is ok. */
+         ok = ctx->Light.ShadeModel == GL_SMOOTH ||
+              ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION;
          break;
       default:
          break;
diff --git a/src/mesa/vbo/vbo.h b/src/mesa/vbo/vbo.h
index 2aaff5df019..00e843c9a0f 100644
--- a/src/mesa/vbo/vbo.h
+++ b/src/mesa/vbo/vbo.h
@@ -88,6 +88,14 @@ void
 vbo_initialize_save_dispatch(const struct gl_context *ctx,
                              struct _glapi_table *exec);
 
+void vbo_exec_FlushVertices(struct gl_context *ctx, GLuint flags);
+void vbo_save_SaveFlushVertices(struct gl_context *ctx);
+GLboolean vbo_save_NotifyBegin(struct gl_context *ctx, GLenum mode);
+void vbo_save_NewList(struct gl_context *ctx, GLuint list, GLenum mode);
+void vbo_save_EndList(struct gl_context *ctx);
+void vbo_save_BeginCallList(struct gl_context *ctx, struct gl_display_list *list);
+void vbo_save_EndCallList(struct gl_context *ctx);
+
 
 typedef void (*vbo_draw_func)( struct gl_context *ctx,
 			       const struct _mesa_prim *prims,
diff --git a/src/mesa/vbo/vbo_exec.c b/src/mesa/vbo/vbo_exec.c
index eb903504328..a301c6c9a22 100644
--- a/src/mesa/vbo/vbo_exec.c
+++ b/src/mesa/vbo/vbo_exec.c
@@ -50,8 +50,6 @@ void vbo_exec_init( struct gl_context *ctx )
 
    ctx->Driver.NeedFlush = 0;
    ctx->Driver.CurrentExecPrimitive = PRIM_OUTSIDE_BEGIN_END;
-   ctx->Driver.BeginVertices = vbo_exec_BeginVertices;
-   ctx->Driver.FlushVertices = vbo_exec_FlushVertices;
 
    vbo_exec_invalidate_state( ctx, ~0 );
 }
diff --git a/src/mesa/vbo/vbo_exec.h b/src/mesa/vbo/vbo_exec.h
index f17fe684cc3..80f3015925d 100644
--- a/src/mesa/vbo/vbo_exec.h
+++ b/src/mesa/vbo/vbo_exec.h
@@ -148,7 +148,6 @@ void vbo_exec_destroy( struct gl_context *ctx );
 void vbo_exec_invalidate_state( struct gl_context *ctx, GLuint new_state );
 
 void vbo_exec_BeginVertices( struct gl_context *ctx );
-void vbo_exec_FlushVertices( struct gl_context *ctx, GLuint flags );
 
 
 /* Internal functions:
diff --git a/src/mesa/vbo/vbo_exec_api.c b/src/mesa/vbo/vbo_exec_api.c
index 138cd60513d..583a2f9b79f 100644
--- a/src/mesa/vbo/vbo_exec_api.c
+++ b/src/mesa/vbo/vbo_exec_api.c
@@ -419,7 +419,7 @@ do {									\
    struct vbo_exec_context *exec = &vbo_context(ctx)->exec;		\
    int sz = (sizeof(C) / sizeof(GLfloat));                              \
    if (unlikely(!(ctx->Driver.NeedFlush & FLUSH_UPDATE_CURRENT)))	\
-      ctx->Driver.BeginVertices( ctx );					\
+      vbo_exec_BeginVertices(ctx);					\
                                                                         \
    if (unlikely(exec->vtx.active_sz[A] != N * sz) ||                    \
        unlikely(exec->vtx.attrtype[A] != T))                            \
@@ -1165,7 +1165,14 @@ void vbo_exec_BeginVertices( struct gl_context *ctx )
 
 
 /**
- * Called via ctx->Driver.FlushVertices()
+ * If inside glBegin()/glEnd(), it should assert(0).  Otherwise, if
+ * FLUSH_STORED_VERTICES bit in \p flags is set flushes any buffered
+ * vertices, if FLUSH_UPDATE_CURRENT bit is set updates
+ * __struct gl_contextRec::Current and gl_light_attrib::Material
+ *
+ * Note that the default T&L engine never clears the
+ * FLUSH_UPDATE_CURRENT bit, even after performing the update.
+ *
  * \param flags  bitmask of FLUSH_STORED_VERTICES, FLUSH_UPDATE_CURRENT
  */
 void vbo_exec_FlushVertices( struct gl_context *ctx, GLuint flags )
@@ -1190,7 +1197,7 @@ void vbo_exec_FlushVertices( struct gl_context *ctx, GLuint flags )
    /* Flush (draw), and make sure VBO is left unmapped when done */
    vbo_exec_FlushVertices_internal(exec, GL_TRUE);
 
-   /* Need to do this to ensure BeginVertices gets called again:
+   /* Need to do this to ensure vbo_exec_BeginVertices gets called again:
     */
    ctx->Driver.NeedFlush &= ~(FLUSH_UPDATE_CURRENT | flags);
 
diff --git a/src/mesa/vbo/vbo_save.c b/src/mesa/vbo/vbo_save.c
index a177660c0f2..79603e9b32e 100644
--- a/src/mesa/vbo/vbo_save.c
+++ b/src/mesa/vbo/vbo_save.c
@@ -33,18 +33,6 @@
 #include "vbo_context.h"
 
 
-static void vbo_save_callback_init( struct gl_context *ctx )
-{
-   ctx->Driver.NewList = vbo_save_NewList;
-   ctx->Driver.EndList = vbo_save_EndList;
-   ctx->Driver.SaveFlushVertices = vbo_save_SaveFlushVertices;
-   ctx->Driver.BeginCallList = vbo_save_BeginCallList;
-   ctx->Driver.EndCallList = vbo_save_EndCallList;
-   ctx->Driver.NotifySaveBegin = vbo_save_NotifyBegin;
-}
-
-
-
 /**
  * Called at context creation time.
  */
@@ -56,7 +44,6 @@ void vbo_save_init( struct gl_context *ctx )
    save->ctx = ctx;
 
    vbo_save_api_init( save );
-   vbo_save_callback_init(ctx);
 
    {
       struct gl_client_array *arrays = save->arrays;
diff --git a/src/mesa/vbo/vbo_save.h b/src/mesa/vbo/vbo_save.h
index 5b1ac81771e..8032db8a9e0 100644
--- a/src/mesa/vbo/vbo_save.h
+++ b/src/mesa/vbo/vbo_save.h
@@ -175,13 +175,6 @@ void vbo_loopback_vertex_list( struct gl_context *ctx,
 
 /* Callbacks:
  */
-void vbo_save_EndList( struct gl_context *ctx );
-void vbo_save_NewList( struct gl_context *ctx, GLuint list, GLenum mode );
-void vbo_save_EndCallList( struct gl_context *ctx );
-void vbo_save_BeginCallList( struct gl_context *ctx, struct gl_display_list *list );
-void vbo_save_SaveFlushVertices( struct gl_context *ctx );
-GLboolean vbo_save_NotifyBegin( struct gl_context *ctx, GLenum mode );
-
 void vbo_save_playback_vertex_list( struct gl_context *ctx, void *data );
 
 void vbo_save_api_init( struct vbo_save_context *save );
diff --git a/src/mesa/vbo/vbo_save_api.c b/src/mesa/vbo/vbo_save_api.c
index 29de3d38aaa..1a70d168c55 100644
--- a/src/mesa/vbo/vbo_save_api.c
+++ b/src/mesa/vbo/vbo_save_api.c
@@ -970,8 +970,7 @@ _save_CallLists(GLsizei n, GLenum type, const GLvoid * v)
 
 
 /**
- * Called via ctx->Driver.NotifySaveBegin() when a glBegin is getting
- * compiled into a display list.
+ * Called when a glBegin is getting compiled into a display list.
  * Updating of ctx->Driver.CurrentSavePrimitive is already taken care of.
  */
 GLboolean
@@ -1001,7 +1000,7 @@ vbo_save_NotifyBegin(struct gl_context *ctx, GLenum mode)
       _mesa_install_save_vtxfmt(ctx, &save->vtxfmt);
    }
 
-   /* We need to call SaveFlushVertices() if there's state change */
+   /* We need to call vbo_save_SaveFlushVertices() if there's state change */
    ctx->Driver.SaveNeedFlush = GL_TRUE;
 
    /* GL_TRUE means we've handled this glBegin here; don't compile a BEGIN
@@ -1604,8 +1603,6 @@ vbo_save_api_init(struct vbo_save_context *save)
                                vbo_destroy_vertex_list,
                                vbo_print_vertex_list);
 
-   ctx->Driver.NotifySaveBegin = vbo_save_NotifyBegin;
-
    _save_vtxfmt_init(ctx);
    _save_current_init(ctx);
    _mesa_noop_vtxfmt_init(&save->vtxfmt_noop);