i965: Eliminate brw->vs.prog_data pointer.

Just say no to:

-   brw->vs.base.prog_data = &brw->vs.prog_data->base.base;

We'll just use the brw_stage_prog_data pointer in brw_stage_state
and downcast it to brw_vs_prog_data as needed.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arcero@collabora.com>

18 files changed, 121 insertions, 96 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index dcda574a70..7bc91a8c5d 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1096,7 +1096,6 @@ struct brw_context
 
    struct {
       struct brw_stage_state base;
-      struct brw_vs_prog_data *prog_data;
    } vs;
 
    struct {
diff --git a/src/mesa/drivers/dri/i965/brw_curbe.c b/src/mesa/drivers/dri/i965/brw_curbe.c
index 02c4e38c15..45bdab19b7 100644
--- a/src/mesa/drivers/dri/i965/brw_curbe.c
+++ b/src/mesa/drivers/dri/i965/brw_curbe.c
@@ -79,7 +79,7 @@ static void calculate_curbe_offsets( struct brw_context *brw )
    const GLuint nr_fp_regs = (brw->wm.prog_data->base.nr_params + 15) / 16;
 
    /* BRW_NEW_VS_PROG_DATA */
-   const GLuint nr_vp_regs = (brw->vs.prog_data->base.base.nr_params + 15) / 16;
+   const GLuint nr_vp_regs = (brw->vs.base.prog_data->nr_params + 15) / 16;
    GLuint nr_clip_regs = 0;
    GLuint total_regs;
 
@@ -260,8 +260,8 @@ brw_upload_constant_buffer(struct brw_context *brw)
       GLuint offset = brw->curbe.vs_start * 16;
 
       /* BRW_NEW_VS_PROG_DATA | _NEW_PROGRAM_CONSTANTS: copy uniform values */
-      for (i = 0; i < brw->vs.prog_data->base.base.nr_params; i++) {
-         buf[offset + i] = *brw->vs.prog_data->base.base.param[i];
+      for (i = 0; i < brw->vs.base.prog_data->nr_params; i++) {
+         buf[offset + i] = *brw->vs.base.prog_data->param[i];
       }
    }
 
diff --git a/src/mesa/drivers/dri/i965/brw_draw.c b/src/mesa/drivers/dri/i965/brw_draw.c
index cab67c94e1..0eaa0f2202 100644
--- a/src/mesa/drivers/dri/i965/brw_draw.c
+++ b/src/mesa/drivers/dri/i965/brw_draw.c
@@ -521,15 +521,17 @@ brw_try_draw_prims(struct gl_context *ctx,
       const int new_basevertex =
          prims[i].indexed ? prims[i].basevertex : prims[i].start;
       const int new_baseinstance = prims[i].base_instance;
+      const struct brw_vs_prog_data *vs_prog_data =
+         brw_vs_prog_data(brw->vs.base.prog_data);
       if (i > 0) {
          const bool uses_draw_parameters =
-            brw->vs.prog_data->uses_basevertex ||
-            brw->vs.prog_data->uses_baseinstance;
+            vs_prog_data->uses_basevertex ||
+            vs_prog_data->uses_baseinstance;
 
          if ((uses_draw_parameters && prims[i].is_indirect) ||
-             (brw->vs.prog_data->uses_basevertex &&
+             (vs_prog_data->uses_basevertex &&
               brw->draw.params.gl_basevertex != new_basevertex) ||
-             (brw->vs.prog_data->uses_baseinstance &&
+             (vs_prog_data->uses_baseinstance &&
               brw->draw.params.gl_baseinstance != new_baseinstance))
             brw->ctx.NewDriverState |= BRW_NEW_VERTICES;
       }
@@ -556,13 +558,13 @@ brw_try_draw_prims(struct gl_context *ctx,
       /* gl_DrawID always needs its own vertex buffer since it's not part of
        * the indirect parameter buffer. If the program uses gl_DrawID we need
        * to flag BRW_NEW_VERTICES. For the first iteration, we don't have
-       * valid brw->vs.prog_data, but we always flag BRW_NEW_VERTICES before
+       * valid vs_prog_data, but we always flag BRW_NEW_VERTICES before
        * the loop.
        */
       brw->draw.gl_drawid = prims[i].draw_id;
       drm_intel_bo_unreference(brw->draw.draw_id_bo);
       brw->draw.draw_id_bo = NULL;
-      if (i > 0 && brw->vs.prog_data->uses_drawid)
+      if (i > 0 && vs_prog_data->uses_drawid)
          brw->ctx.NewDriverState |= BRW_NEW_VERTICES;
 
       if (brw->gen < 6)
diff --git a/src/mesa/drivers/dri/i965/brw_draw_upload.c b/src/mesa/drivers/dri/i965/brw_draw_upload.c
index a44e04be6b..993b21c30e 100644
--- a/src/mesa/drivers/dri/i965/brw_draw_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_draw_upload.c
@@ -439,7 +439,9 @@ brw_prepare_vertices(struct brw_context *brw)
 {
    struct gl_context *ctx = &brw->ctx;
    /* BRW_NEW_VS_PROG_DATA */
-   GLbitfield64 vs_inputs = brw->vs.prog_data->inputs_read;
+   const struct brw_vs_prog_data *vs_prog_data =
+      brw_vs_prog_data(brw->vs.base.prog_data);
+   GLbitfield64 vs_inputs = vs_prog_data->inputs_read;
    const unsigned char *ptr = NULL;
    GLuint interleaved = 0;
    unsigned int min_index = brw->vb.min_index + brw->basevertex;
@@ -665,16 +667,18 @@ brw_prepare_vertices(struct brw_context *brw)
 void
 brw_prepare_shader_draw_parameters(struct brw_context *brw)
 {
+   const struct brw_vs_prog_data *vs_prog_data =
+      brw_vs_prog_data(brw->vs.base.prog_data);
+
    /* For non-indirect draws, upload gl_BaseVertex. */
-   if ((brw->vs.prog_data->uses_basevertex ||
-        brw->vs.prog_data->uses_baseinstance) &&
+   if ((vs_prog_data->uses_basevertex || vs_prog_data->uses_baseinstance) &&
        brw->draw.draw_params_bo == NULL) {
       intel_upload_data(brw, &brw->draw.params, sizeof(brw->draw.params), 4,
 			&brw->draw.draw_params_bo,
                         &brw->draw.draw_params_offset);
    }
 
-   if (brw->vs.prog_data->uses_drawid) {
+   if (vs_prog_data->uses_drawid) {
       intel_upload_data(brw, &brw->draw.gl_drawid, sizeof(brw->draw.gl_drawid), 4,
                         &brw->draw.draw_id_bo,
                         &brw->draw.draw_id_offset);
@@ -770,11 +774,14 @@ brw_emit_vertices(struct brw_context *brw)
 
    brw_emit_query_begin(brw);
 
+   const struct brw_vs_prog_data *vs_prog_data =
+      brw_vs_prog_data(brw->vs.base.prog_data);
+
    unsigned nr_elements = brw->vb.nr_enabled;
-   if (brw->vs.prog_data->uses_vertexid || brw->vs.prog_data->uses_instanceid ||
-       brw->vs.prog_data->uses_basevertex || brw->vs.prog_data->uses_baseinstance)
+   if (vs_prog_data->uses_vertexid || vs_prog_data->uses_instanceid ||
+       vs_prog_data->uses_basevertex || vs_prog_data->uses_baseinstance)
       ++nr_elements;
-   if (brw->vs.prog_data->uses_drawid)
+   if (vs_prog_data->uses_drawid)
       nr_elements++;
 
    /* If the VS doesn't read any inputs (calculating vertex position from
@@ -810,10 +817,10 @@ brw_emit_vertices(struct brw_context *brw)
     */
 
    const bool uses_draw_params =
-      brw->vs.prog_data->uses_basevertex ||
-      brw->vs.prog_data->uses_baseinstance;
+      vs_prog_data->uses_basevertex ||
+      vs_prog_data->uses_baseinstance;
    const unsigned nr_buffers = brw->vb.nr_buffers +
-      uses_draw_params + brw->vs.prog_data->uses_drawid;
+      uses_draw_params + vs_prog_data->uses_drawid;
 
    if (nr_buffers) {
       if (brw->gen >= 6) {
@@ -848,7 +855,7 @@ brw_emit_vertices(struct brw_context *brw)
                                   0); /* step rate */
       }
 
-      if (brw->vs.prog_data->uses_drawid) {
+      if (vs_prog_data->uses_drawid) {
          EMIT_VERTEX_BUFFER_STATE(brw, brw->vb.nr_buffers + 1,
                                   brw->draw.draw_id_bo,
                                   brw->draw.draw_id_offset,
@@ -926,24 +933,24 @@ brw_emit_vertices(struct brw_context *brw)
                     ((i * 4) << BRW_VE1_DST_OFFSET_SHIFT));
    }
 
-   if (brw->vs.prog_data->uses_vertexid || brw->vs.prog_data->uses_instanceid ||
-       brw->vs.prog_data->uses_basevertex || brw->vs.prog_data->uses_baseinstance) {
+   if (vs_prog_data->uses_vertexid || vs_prog_data->uses_instanceid ||
+       vs_prog_data->uses_basevertex || vs_prog_data->uses_baseinstance) {
       uint32_t dw0 = 0, dw1 = 0;
       uint32_t comp0 = BRW_VE1_COMPONENT_STORE_0;
       uint32_t comp1 = BRW_VE1_COMPONENT_STORE_0;
       uint32_t comp2 = BRW_VE1_COMPONENT_STORE_0;
       uint32_t comp3 = BRW_VE1_COMPONENT_STORE_0;
 
-      if (brw->vs.prog_data->uses_basevertex)
+      if (vs_prog_data->uses_basevertex)
          comp0 = BRW_VE1_COMPONENT_STORE_SRC;
 
-      if (brw->vs.prog_data->uses_baseinstance)
+      if (vs_prog_data->uses_baseinstance)
          comp1 = BRW_VE1_COMPONENT_STORE_SRC;
 
-      if (brw->vs.prog_data->uses_vertexid)
+      if (vs_prog_data->uses_vertexid)
          comp2 = BRW_VE1_COMPONENT_STORE_VID;
 
-      if (brw->vs.prog_data->uses_instanceid)
+      if (vs_prog_data->uses_instanceid)
          comp3 = BRW_VE1_COMPONENT_STORE_IID;
 
       dw1 = (comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
@@ -970,7 +977,7 @@ brw_emit_vertices(struct brw_context *brw)
       OUT_BATCH(dw1);
    }
 
-   if (brw->vs.prog_data->uses_drawid) {
+   if (vs_prog_data->uses_drawid) {
       uint32_t dw0 = 0, dw1 = 0;
 
       dw1 = (BRW_VE1_COMPONENT_STORE_SRC << BRW_VE1_COMPONENT_0_SHIFT) |
diff --git a/src/mesa/drivers/dri/i965/brw_ff_gs.c b/src/mesa/drivers/dri/i965/brw_ff_gs.c
index d8eb1464f5..c7129393e5 100644
--- a/src/mesa/drivers/dri/i965/brw_ff_gs.c
+++ b/src/mesa/drivers/dri/i965/brw_ff_gs.c
@@ -56,7 +56,7 @@ brw_codegen_ff_gs_prog(struct brw_context *brw,
    memset(&c, 0, sizeof(c));
 
    c.key = *key;
-   c.vue_map = brw->vs.prog_data->base.vue_map;
+   c.vue_map = brw_vue_prog_data(brw->vs.base.prog_data)->vue_map;
    c.nr_regs = (c.vue_map.num_slots + 1)/2;
 
    mem_ctx = ralloc_context(NULL);
@@ -174,7 +174,7 @@ brw_ff_gs_populate_key(struct brw_context *brw,
    memset(key, 0, sizeof(*key));
 
    /* BRW_NEW_VS_PROG_DATA (part of VUE map) */
-   key->attrs = brw->vs.prog_data->base.vue_map.slots_valid;
+   key->attrs = brw_vue_prog_data(brw->vs.base.prog_data)->vue_map.slots_valid;
 
    /* BRW_NEW_PRIMITIVE */
    key->primitive = brw->primitive;
diff --git a/src/mesa/drivers/dri/i965/brw_state_cache.c b/src/mesa/drivers/dri/i965/brw_state_cache.c
index 43b0a202f2..7fc8aa5c04 100644
--- a/src/mesa/drivers/dri/i965/brw_state_cache.c
+++ b/src/mesa/drivers/dri/i965/brw_state_cache.c
@@ -398,7 +398,6 @@ brw_clear_cache(struct brw_context *brw, struct brw_cache *cache)
    brw->state.pipelines[BRW_COMPUTE_PIPELINE].brw = ~0ull;
 
    /* Also, NULL out any stale program pointers. */
-   brw->vs.prog_data = NULL;
    brw->vs.base.prog_data = NULL;
    brw->tcs.prog_data = NULL;
    brw->tcs.base.prog_data = NULL;
diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c b/src/mesa/drivers/dri/i965/brw_state_upload.c
index 055fed128a..6ace37bdc0 100644
--- a/src/mesa/drivers/dri/i965/brw_state_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
@@ -713,12 +713,15 @@ brw_upload_programs(struct brw_context *brw,
        */
       GLbitfield64 old_slots = brw->vue_map_geom_out.slots_valid;
       bool old_separate = brw->vue_map_geom_out.separate;
+      struct brw_vue_prog_data *vue_prog_data;
       if (brw->geometry_program)
-         brw->vue_map_geom_out = brw->gs.prog_data->base.vue_map;
+         vue_prog_data = brw_vue_prog_data(brw->gs.base.prog_data);
       else if (brw->tess_eval_program)
-         brw->vue_map_geom_out = brw->tes.prog_data->base.vue_map;
+         vue_prog_data = brw_vue_prog_data(brw->tes.base.prog_data);
       else
-         brw->vue_map_geom_out = brw->vs.prog_data->base.vue_map;
+         vue_prog_data = brw_vue_prog_data(brw->vs.base.prog_data);
+
+      brw->vue_map_geom_out = vue_prog_data->vue_map;
 
       /* If the layout has changed, signal BRW_NEW_VUE_MAP_GEOM_OUT. */
       if (old_slots != brw->vue_map_geom_out.slots_valid ||
diff --git a/src/mesa/drivers/dri/i965/brw_urb.c b/src/mesa/drivers/dri/i965/brw_urb.c
index 83851c1205..d61e713800 100644
--- a/src/mesa/drivers/dri/i965/brw_urb.c
+++ b/src/mesa/drivers/dri/i965/brw_urb.c
@@ -115,7 +115,7 @@ static bool check_urb_layout(struct brw_context *brw)
 static void recalculate_urb_fence( struct brw_context *brw )
 {
    GLuint csize = brw->curbe.total_size;
-   GLuint vsize = brw->vs.prog_data->base.urb_entry_size;
+   GLuint vsize = brw_vue_prog_data(brw->vs.base.prog_data)->urb_entry_size;
    GLuint sfsize = brw->sf.prog_data->urb_entry_size;
 
    if (csize < limits[CS].min_entry_size)
diff --git a/src/mesa/drivers/dri/i965/brw_vs.c b/src/mesa/drivers/dri/i965/brw_vs.c
index 86cdbe05ae..bd194ca226 100644
--- a/src/mesa/drivers/dri/i965/brw_vs.c
+++ b/src/mesa/drivers/dri/i965/brw_vs.c
@@ -229,7 +229,7 @@ brw_codegen_vs_prog(struct brw_context *brw,
 		    key, sizeof(struct brw_vs_prog_key),
 		    program, program_size,
 		    &prog_data, sizeof(prog_data),
-		    &brw->vs.base.prog_offset, &brw->vs.prog_data);
+		    &brw->vs.base.prog_offset, &brw->vs.base.prog_data);
    ralloc_free(mem_ctx);
 
    return true;
@@ -372,13 +372,12 @@ brw_upload_vs_prog(struct brw_context *brw)
 
    if (!brw_search_cache(&brw->cache, BRW_CACHE_VS_PROG,
 			 &key, sizeof(key),
-			 &brw->vs.base.prog_offset, &brw->vs.prog_data)) {
+			 &brw->vs.base.prog_offset, &brw->vs.base.prog_data)) {
       bool success = brw_codegen_vs_prog(brw, current[MESA_SHADER_VERTEX],
                                          vp, &key);
       (void) success;
       assert(success);
    }
-   brw->vs.base.prog_data = &brw->vs.prog_data->base.base;
 }
 
 bool
@@ -389,7 +388,7 @@ brw_vs_precompile(struct gl_context *ctx,
    struct brw_context *brw = brw_context(ctx);
    struct brw_vs_prog_key key;
    uint32_t old_prog_offset = brw->vs.base.prog_offset;
-   struct brw_vs_prog_data *old_prog_data = brw->vs.prog_data;
+   struct brw_stage_prog_data *old_prog_data = brw->vs.base.prog_data;
    bool success;
 
    struct gl_vertex_program *vp = (struct gl_vertex_program *) prog;
@@ -406,7 +405,7 @@ brw_vs_precompile(struct gl_context *ctx,
    success = brw_codegen_vs_prog(brw, shader_prog, bvp, &key);
 
    brw->vs.base.prog_offset = old_prog_offset;
-   brw->vs.prog_data = old_prog_data;
+   brw->vs.base.prog_data = old_prog_data;
 
    return success;
 }
diff --git a/src/mesa/drivers/dri/i965/brw_vs_state.c b/src/mesa/drivers/dri/i965/brw_vs_state.c
index 25ecbbd96d..d0d2695a49 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_state.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_state.c
@@ -41,6 +41,9 @@ brw_upload_vs_unit(struct brw_context *brw)
 {
    const struct gen_device_info *devinfo = &brw->screen->devinfo;
    struct brw_stage_state *stage_state = &brw->vs.base;
+   const struct brw_stage_prog_data *prog_data = stage_state->prog_data;
+   const struct brw_vue_prog_data *vue_prog_data =
+      brw_vue_prog_data(stage_state->prog_data);
 
    struct brw_vs_unit_state *vs;
 
@@ -49,8 +52,7 @@ brw_upload_vs_unit(struct brw_context *brw)
    memset(vs, 0, sizeof(*vs));
 
    /* BRW_NEW_PROGRAM_CACHE | BRW_NEW_VS_PROG_DATA */
-   vs->thread0.grf_reg_count =
-      ALIGN(brw->vs.prog_data->base.total_grf, 16) / 16 - 1;
+   vs->thread0.grf_reg_count = ALIGN(vue_prog_data->total_grf, 16) / 16 - 1;
    vs->thread0.kernel_start_pointer =
       brw_program_reloc(brw,
 			stage_state->state_offset +
@@ -58,7 +60,7 @@ brw_upload_vs_unit(struct brw_context *brw)
 			stage_state->prog_offset +
 			(vs->thread0.grf_reg_count << 1)) >> 6;
 
-   if (brw->vs.prog_data->base.base.use_alt_mode)
+   if (prog_data->use_alt_mode)
       vs->thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
    else
       vs->thread1.floating_point_mode = BRW_FLOATING_POINT_IEEE_754;
@@ -78,9 +80,9 @@ brw_upload_vs_unit(struct brw_context *brw)
    vs->thread1.single_program_flow = (brw->gen == 5);
 
    vs->thread1.binding_table_entry_count =
-      brw->vs.prog_data->base.base.binding_table.size_bytes / 4;
+      prog_data->binding_table.size_bytes / 4;
 
-   if (brw->vs.prog_data->base.base.total_scratch != 0) {
+   if (prog_data->total_scratch != 0) {
       vs->thread2.scratch_space_base_pointer =
 	 stage_state->scratch_bo->offset64 >> 10; /* reloc */
       vs->thread2.per_thread_scratch_space =
@@ -90,11 +92,9 @@ brw_upload_vs_unit(struct brw_context *brw)
       vs->thread2.per_thread_scratch_space = 0;
    }
 
-   vs->thread3.urb_entry_read_length = brw->vs.prog_data->base.urb_read_length;
-   vs->thread3.const_urb_entry_read_length
-      = brw->vs.prog_data->base.base.curb_read_length;
-   vs->thread3.dispatch_grf_start_reg =
-      brw->vs.prog_data->base.base.dispatch_grf_start_reg;
+   vs->thread3.urb_entry_read_length = vue_prog_data->urb_read_length;
+   vs->thread3.const_urb_entry_read_length = prog_data->curb_read_length;
+   vs->thread3.dispatch_grf_start_reg = prog_data->dispatch_grf_start_reg;
    vs->thread3.urb_entry_read_offset = 0;
 
    /* BRW_NEW_CURBE_OFFSETS */
@@ -170,7 +170,7 @@ brw_upload_vs_unit(struct brw_context *brw)
    }
 
    /* Emit scratch space relocation */
-   if (brw->vs.prog_data->base.base.total_scratch != 0) {
+   if (prog_data->total_scratch != 0) {
       drm_intel_bo_emit_reloc(brw->batch.bo,
 			      stage_state->state_offset +
 			      offsetof(struct brw_vs_unit_state, thread2),
diff --git a/src/mesa/drivers/dri/i965/brw_vs_surface_state.c b/src/mesa/drivers/dri/i965/brw_vs_surface_state.c
index 387a54bdb0..f8c72180fd 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_surface_state.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_surface_state.c
@@ -117,7 +117,7 @@ brw_upload_vs_pull_constants(struct brw_context *brw)
       (struct brw_vertex_program *) brw->vertex_program;
 
    /* BRW_NEW_VS_PROG_DATA */
-   const struct brw_stage_prog_data *prog_data = &brw->vs.prog_data->base.base;
+   const struct brw_stage_prog_data *prog_data = brw->vs.base.prog_data;
 
    _mesa_shader_write_subroutine_indices(&brw->ctx, MESA_SHADER_VERTEX);
    /* _NEW_PROGRAM_CONSTANTS */
@@ -149,7 +149,7 @@ brw_upload_vs_ubo_surfaces(struct brw_context *brw)
 
    /* BRW_NEW_VS_PROG_DATA */
    brw_upload_ubo_surfaces(brw, prog->_LinkedShaders[MESA_SHADER_VERTEX],
-                           &brw->vs.base, &brw->vs.prog_data->base.base);
+                           &brw->vs.base, brw->vs.base.prog_data);
 }
 
 const struct brw_tracked_state brw_vs_ubo_surfaces = {
@@ -174,7 +174,7 @@ brw_upload_vs_abo_surfaces(struct brw_context *brw)
    if (prog) {
       /* BRW_NEW_VS_PROG_DATA */
       brw_upload_abo_surfaces(brw, prog->_LinkedShaders[MESA_SHADER_VERTEX],
-                              &brw->vs.base, &brw->vs.prog_data->base.base);
+                              &brw->vs.base, brw->vs.base.prog_data);
    }
 }
 
@@ -200,7 +200,7 @@ brw_upload_vs_image_surfaces(struct brw_context *brw)
    if (prog) {
       /* BRW_NEW_VS_PROG_DATA, BRW_NEW_IMAGE_UNITS, _NEW_TEXTURE */
       brw_upload_image_surfaces(brw, prog->_LinkedShaders[MESA_SHADER_VERTEX],
-                                &brw->vs.base, &brw->vs.prog_data->base.base);
+                                &brw->vs.base, brw->vs.base.prog_data);
    }
 }
 
diff --git a/src/mesa/drivers/dri/i965/gen6_clip_state.c b/src/mesa/drivers/dri/i965/gen6_clip_state.c
index 16a96fcaf1..7aff0f59ce 100644
--- a/src/mesa/drivers/dri/i965/gen6_clip_state.c
+++ b/src/mesa/drivers/dri/i965/gen6_clip_state.c
@@ -103,7 +103,7 @@ upload_clip_state(struct brw_context *brw)
    }
 
    /* BRW_NEW_VS_PROG_DATA */
-   dw1 |= brw->vs.prog_data->base.cull_distance_mask;
+   dw1 |= brw_vue_prog_data(brw->vs.base.prog_data)->cull_distance_mask;
 
    if (brw->gen >= 7)
       dw1 |= GEN7_CLIP_EARLY_CULL;
diff --git a/src/mesa/drivers/dri/i965/gen6_urb.c b/src/mesa/drivers/dri/i965/gen6_urb.c
index df12ea66fe..c3dcf7d0a3 100644
--- a/src/mesa/drivers/dri/i965/gen6_urb.c
+++ b/src/mesa/drivers/dri/i965/gen6_urb.c
@@ -109,7 +109,9 @@ static void
 upload_urb(struct brw_context *brw)
 {
    /* BRW_NEW_VS_PROG_DATA */
-   const unsigned vs_size = MAX2(brw->vs.prog_data->base.urb_entry_size, 1);
+   const struct brw_vue_prog_data *vs_vue_prog_data =
+      brw_vue_prog_data(brw->vs.base.prog_data);
+   const unsigned vs_size = MAX2(vs_vue_prog_data->urb_entry_size, 1);
 
    /* BRW_NEW_GEOMETRY_PROGRAM, BRW_NEW_GS_PROG_DATA */
    const bool gs_present = brw->ff_gs.prog_active || brw->geometry_program;
diff --git a/src/mesa/drivers/dri/i965/gen6_vs_state.c b/src/mesa/drivers/dri/i965/gen6_vs_state.c
index 15644b6155..780b1639c7 100644
--- a/src/mesa/drivers/dri/i965/gen6_vs_state.c
+++ b/src/mesa/drivers/dri/i965/gen6_vs_state.c
@@ -43,7 +43,7 @@ gen6_upload_vs_push_constants(struct brw_context *brw)
    const struct brw_vertex_program *vp =
       brw_vertex_program_const(brw->vertex_program);
    /* BRW_NEW_VS_PROG_DATA */
-   const struct brw_stage_prog_data *prog_data = &brw->vs.prog_data->base.base;
+   const struct brw_stage_prog_data *prog_data = brw->vs.base.prog_data;
 
    _mesa_shader_write_subroutine_indices(&brw->ctx, MESA_SHADER_VERTEX);
    gen6_upload_push_constants(brw, &vp->program.Base, prog_data,
@@ -76,6 +76,9 @@ upload_vs_state(struct brw_context *brw)
 {
    const struct gen_device_info *devinfo = &brw->screen->devinfo;
    const struct brw_stage_state *stage_state = &brw->vs.base;
+   const struct brw_stage_prog_data *prog_data = stage_state->prog_data;
+   const struct brw_vue_prog_data *vue_prog_data =
+      brw_vue_prog_data(stage_state->prog_data);
    uint32_t floating_point_mode = 0;
 
    /* From the BSpec, 3D Pipeline > Geometry > Vertex Shader > State,
@@ -115,7 +118,7 @@ upload_vs_state(struct brw_context *brw)
       ADVANCE_BATCH();
    }
 
-   if (brw->vs.prog_data->base.base.use_alt_mode)
+   if (prog_data->use_alt_mode)
       floating_point_mode = GEN6_VS_FLOATING_POINT_MODE_ALT;
 
    BEGIN_BATCH(6);
@@ -123,10 +126,10 @@ upload_vs_state(struct brw_context *brw)
    OUT_BATCH(stage_state->prog_offset);
    OUT_BATCH(floating_point_mode |
 	     ((ALIGN(stage_state->sampler_count, 4)/4) << GEN6_VS_SAMPLER_COUNT_SHIFT) |
-             ((brw->vs.prog_data->base.base.binding_table.size_bytes / 4) <<
+             ((prog_data->binding_table.size_bytes / 4) <<
               GEN6_VS_BINDING_TABLE_ENTRY_COUNT_SHIFT));
 
-   if (brw->vs.prog_data->base.base.total_scratch) {
+   if (prog_data->total_scratch) {
       OUT_RELOC(stage_state->scratch_bo,
 		I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
 		ffs(stage_state->per_thread_scratch) - 11);
@@ -134,9 +137,9 @@ upload_vs_state(struct brw_context *brw)
       OUT_BATCH(0);
    }
 
-   OUT_BATCH((brw->vs.prog_data->base.base.dispatch_grf_start_reg <<
+   OUT_BATCH((prog_data->dispatch_grf_start_reg <<
               GEN6_VS_DISPATCH_START_GRF_SHIFT) |
-	     (brw->vs.prog_data->base.urb_read_length << GEN6_VS_URB_READ_LENGTH_SHIFT) |
+	     (vue_prog_data->urb_read_length << GEN6_VS_URB_READ_LENGTH_SHIFT) |
 	     (0 << GEN6_VS_URB_ENTRY_READ_OFFSET_SHIFT));
 
    OUT_BATCH(((devinfo->max_vs_threads - 1) << GEN6_VS_MAX_THREADS_SHIFT) |
diff --git a/src/mesa/drivers/dri/i965/gen7_urb.c b/src/mesa/drivers/dri/i965/gen7_urb.c
index d38fc26822..2c14dcee43 100644
--- a/src/mesa/drivers/dri/i965/gen7_urb.c
+++ b/src/mesa/drivers/dri/i965/gen7_urb.c
@@ -189,7 +189,9 @@ static void
 upload_urb(struct brw_context *brw)
 {
    /* BRW_NEW_VS_PROG_DATA */
-   const unsigned vs_size = MAX2(brw->vs.prog_data->base.urb_entry_size, 1);
+   const struct brw_vue_prog_data *vs_vue_prog_data =
+      brw_vue_prog_data(brw->vs.base.prog_data);
+   const unsigned vs_size = MAX2(vs_vue_prog_data->urb_entry_size, 1);
    /* BRW_NEW_GEOMETRY_PROGRAM, BRW_NEW_GS_PROG_DATA */
    const bool gs_present = brw->geometry_program;
    /* BRW_NEW_TESS_PROGRAMS */
diff --git a/src/mesa/drivers/dri/i965/gen7_vs_state.c b/src/mesa/drivers/dri/i965/gen7_vs_state.c
index 0b51eb012a..623c78471f 100644
--- a/src/mesa/drivers/dri/i965/gen7_vs_state.c
+++ b/src/mesa/drivers/dri/i965/gen7_vs_state.c
@@ -34,15 +34,17 @@ upload_vs_state(struct brw_context *brw)
 {
    const struct gen_device_info *devinfo = &brw->screen->devinfo;
    const struct brw_stage_state *stage_state = &brw->vs.base;
+   const struct brw_stage_prog_data *prog_data = stage_state->prog_data;
+   const struct brw_vue_prog_data *vue_prog_data =
+      brw_vue_prog_data(stage_state->prog_data);
    uint32_t floating_point_mode = 0;
    const int max_threads_shift = brw->is_haswell ?
       HSW_VS_MAX_THREADS_SHIFT : GEN6_VS_MAX_THREADS_SHIFT;
-   const struct brw_vue_prog_data *prog_data = &brw->vs.prog_data->base;
 
    if (!brw->is_haswell && !brw->is_baytrail)
       gen7_emit_vs_workaround_flush(brw);
 
-   if (prog_data->base.use_alt_mode)
+   if (prog_data->use_alt_mode)
       floating_point_mode = GEN6_VS_FLOATING_POINT_MODE_ALT;
 
    BEGIN_BATCH(6);
@@ -51,10 +53,10 @@ upload_vs_state(struct brw_context *brw)
    OUT_BATCH(floating_point_mode |
 	     ((ALIGN(stage_state->sampler_count, 4)/4) <<
               GEN6_VS_SAMPLER_COUNT_SHIFT) |
-             ((prog_data->base.binding_table.size_bytes / 4) <<
+             ((prog_data->binding_table.size_bytes / 4) <<
               GEN6_VS_BINDING_TABLE_ENTRY_COUNT_SHIFT));
 
-   if (prog_data->base.total_scratch) {
+   if (prog_data->total_scratch) {
       OUT_RELOC(stage_state->scratch_bo,
 		I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
 		ffs(stage_state->per_thread_scratch) - 11);
@@ -62,9 +64,9 @@ upload_vs_state(struct brw_context *brw)
       OUT_BATCH(0);
    }
 
-   OUT_BATCH((prog_data->base.dispatch_grf_start_reg <<
+   OUT_BATCH((prog_data->dispatch_grf_start_reg <<
               GEN6_VS_DISPATCH_START_GRF_SHIFT) |
-	     (prog_data->urb_read_length << GEN6_VS_URB_READ_LENGTH_SHIFT) |
+	     (vue_prog_data->urb_read_length << GEN6_VS_URB_READ_LENGTH_SHIFT) |
 	     (0 << GEN6_VS_URB_ENTRY_READ_OFFSET_SHIFT));
 
    OUT_BATCH(((devinfo->max_vs_threads - 1) << max_threads_shift) |
diff --git a/src/mesa/drivers/dri/i965/gen8_draw_upload.c b/src/mesa/drivers/dri/i965/gen8_draw_upload.c
index 71962664fd..23c7587945 100644
--- a/src/mesa/drivers/dri/i965/gen8_draw_upload.c
+++ b/src/mesa/drivers/dri/i965/gen8_draw_upload.c
@@ -62,7 +62,10 @@ gen8_emit_vertices(struct brw_context *brw)
    uses_edge_flag = (ctx->Polygon.FrontMode != GL_FILL ||
                      ctx->Polygon.BackMode != GL_FILL);
 
-   if (brw->vs.prog_data->uses_vertexid || brw->vs.prog_data->uses_instanceid) {
+   const struct brw_vs_prog_data *vs_prog_data =
+      brw_vs_prog_data(brw->vs.base.prog_data);
+
+   if (vs_prog_data->uses_vertexid || vs_prog_data->uses_instanceid) {
       unsigned vue = brw->vb.nr_enabled;
 
       /* The element for the edge flags must always be last, so we have to
@@ -78,13 +81,13 @@ gen8_emit_vertices(struct brw_context *brw)
                 "need to reorder the vertex attrbutes.");
 
       unsigned dw1 = 0;
-      if (brw->vs.prog_data->uses_vertexid) {
+      if (vs_prog_data->uses_vertexid) {
          dw1 |= GEN8_SGVS_ENABLE_VERTEX_ID |
                 (2 << GEN8_SGVS_VERTEX_ID_COMPONENT_SHIFT) |  /* .z channel */
                 (vue << GEN8_SGVS_VERTEX_ID_ELEMENT_OFFSET_SHIFT);
       }
 
-      if (brw->vs.prog_data->uses_instanceid) {
+      if (vs_prog_data->uses_instanceid) {
          dw1 |= GEN8_SGVS_ENABLE_INSTANCE_ID |
                 (3 << GEN8_SGVS_INSTANCE_ID_COMPONENT_SHIFT) | /* .w channel */
                 (vue << GEN8_SGVS_INSTANCE_ID_ELEMENT_OFFSET_SHIFT);
@@ -131,10 +134,10 @@ gen8_emit_vertices(struct brw_context *brw)
 
    /* Now emit 3DSTATE_VERTEX_BUFFERS and 3DSTATE_VERTEX_ELEMENTS packets. */
    const bool uses_draw_params =
-      brw->vs.prog_data->uses_basevertex ||
-      brw->vs.prog_data->uses_baseinstance;
+      vs_prog_data->uses_basevertex ||
+      vs_prog_data->uses_baseinstance;
    const unsigned nr_buffers = brw->vb.nr_buffers +
-      uses_draw_params + brw->vs.prog_data->uses_drawid;
+      uses_draw_params + vs_prog_data->uses_drawid;
 
    if (nr_buffers) {
       assert(nr_buffers <= 33);
@@ -158,7 +161,7 @@ gen8_emit_vertices(struct brw_context *brw)
                                   0 /* unused */);
       }
 
-      if (brw->vs.prog_data->uses_drawid) {
+      if (vs_prog_data->uses_drawid) {
          EMIT_VERTEX_BUFFER_STATE(brw, brw->vb.nr_buffers + 1,
                                   brw->draw.draw_id_bo,
                                   brw->draw.draw_id_offset,
@@ -177,13 +180,13 @@ gen8_emit_vertices(struct brw_context *brw)
     * can't be inserted past that so we need a dummy element to ensure that
     * the edge flag is the last one.
     */
-   const bool needs_sgvs_element = (brw->vs.prog_data->uses_basevertex ||
-                                    brw->vs.prog_data->uses_baseinstance ||
-                                    ((brw->vs.prog_data->uses_instanceid ||
-                                      brw->vs.prog_data->uses_vertexid) &&
+   const bool needs_sgvs_element = (vs_prog_data->uses_basevertex ||
+                                    vs_prog_data->uses_baseinstance ||
+                                    ((vs_prog_data->uses_instanceid ||
+                                      vs_prog_data->uses_vertexid) &&
                                      uses_edge_flag));
    const unsigned nr_elements =
-      brw->vb.nr_enabled + needs_sgvs_element + brw->vs.prog_data->uses_drawid;
+      brw->vb.nr_enabled + needs_sgvs_element + vs_prog_data->uses_drawid;
 
    /* The hardware allows one more VERTEX_ELEMENTS than VERTEX_BUFFERS,
     * presumably for VertexID/InstanceID.
@@ -279,8 +282,8 @@ gen8_emit_vertices(struct brw_context *brw)
    }
 
    if (needs_sgvs_element) {
-      if (brw->vs.prog_data->uses_basevertex ||
-          brw->vs.prog_data->uses_baseinstance) {
+      if (vs_prog_data->uses_basevertex ||
+          vs_prog_data->uses_baseinstance) {
          OUT_BATCH(GEN6_VE0_VALID |
                    brw->vb.nr_buffers << GEN6_VE0_INDEX_SHIFT |
                    BRW_SURFACEFORMAT_R32G32_UINT << BRW_VE0_FORMAT_SHIFT);
@@ -297,7 +300,7 @@ gen8_emit_vertices(struct brw_context *brw)
       }
    }
 
-   if (brw->vs.prog_data->uses_drawid) {
+   if (vs_prog_data->uses_drawid) {
       OUT_BATCH(GEN6_VE0_VALID |
                 ((brw->vb.nr_buffers + 1) << GEN6_VE0_INDEX_SHIFT) |
                 (BRW_SURFACEFORMAT_R32_UINT << BRW_VE0_FORMAT_SHIFT));
@@ -345,7 +348,7 @@ gen8_emit_vertices(struct brw_context *brw)
       ADVANCE_BATCH();
    }
 
-   if (brw->vs.prog_data->uses_drawid) {
+   if (vs_prog_data->uses_drawid) {
       const unsigned element = brw->vb.nr_enabled + needs_sgvs_element;
       BEGIN_BATCH(3);
       OUT_BATCH(_3DSTATE_VF_INSTANCING << 16 | (3 - 2));
diff --git a/src/mesa/drivers/dri/i965/gen8_vs_state.c b/src/mesa/drivers/dri/i965/gen8_vs_state.c
index 963647cdf3..aed6be08a6 100644
--- a/src/mesa/drivers/dri/i965/gen8_vs_state.c
+++ b/src/mesa/drivers/dri/i965/gen8_vs_state.c
@@ -38,12 +38,14 @@ upload_vs_state(struct brw_context *brw)
    uint32_t floating_point_mode = 0;
 
    /* BRW_NEW_VS_PROG_DATA */
-   const struct brw_vue_prog_data *prog_data = &brw->vs.prog_data->base;
+   const struct brw_stage_prog_data *prog_data = stage_state->prog_data;
+   const struct brw_vue_prog_data *vue_prog_data =
+      brw_vue_prog_data(stage_state->prog_data);
 
-   assert(prog_data->dispatch_mode == DISPATCH_MODE_SIMD8 ||
-          prog_data->dispatch_mode == DISPATCH_MODE_4X2_DUAL_OBJECT);
+   assert(vue_prog_data->dispatch_mode == DISPATCH_MODE_SIMD8 ||
+          vue_prog_data->dispatch_mode == DISPATCH_MODE_4X2_DUAL_OBJECT);
 
-   if (prog_data->base.use_alt_mode)
+   if (prog_data->use_alt_mode)
       floating_point_mode = GEN6_VS_FLOATING_POINT_MODE_ALT;
 
    BEGIN_BATCH(9);
@@ -53,10 +55,10 @@ upload_vs_state(struct brw_context *brw)
    OUT_BATCH(floating_point_mode |
              ((ALIGN(stage_state->sampler_count, 4) / 4) <<
                GEN6_VS_SAMPLER_COUNT_SHIFT) |
-             ((prog_data->base.binding_table.size_bytes / 4) <<
+             ((prog_data->binding_table.size_bytes / 4) <<
                GEN6_VS_BINDING_TABLE_ENTRY_COUNT_SHIFT));
 
-   if (prog_data->base.total_scratch) {
+   if (prog_data->total_scratch) {
       OUT_RELOC64(stage_state->scratch_bo,
                   I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
                   ffs(stage_state->per_thread_scratch) - 11);
@@ -65,12 +67,14 @@ upload_vs_state(struct brw_context *brw)
       OUT_BATCH(0);
    }
 
-   OUT_BATCH((prog_data->base.dispatch_grf_start_reg <<
+   OUT_BATCH((prog_data->dispatch_grf_start_reg <<
               GEN6_VS_DISPATCH_START_GRF_SHIFT) |
-             (prog_data->urb_read_length << GEN6_VS_URB_READ_LENGTH_SHIFT) |
+             (vue_prog_data->urb_read_length <<
+              GEN6_VS_URB_READ_LENGTH_SHIFT) |
              (0 << GEN6_VS_URB_ENTRY_READ_OFFSET_SHIFT));
 
-   uint32_t simd8_enable = prog_data->dispatch_mode == DISPATCH_MODE_SIMD8 ?
+   uint32_t simd8_enable =
+      vue_prog_data->dispatch_mode == DISPATCH_MODE_SIMD8 ?
       GEN8_VS_SIMD8_ENABLE : 0;
    OUT_BATCH(((devinfo->max_vs_threads - 1) << HSW_VS_MAX_THREADS_SHIFT) |
              GEN6_VS_STATISTICS_ENABLE |
@@ -78,7 +82,7 @@ upload_vs_state(struct brw_context *brw)
              GEN6_VS_ENABLE);
 
    /* _NEW_TRANSFORM */
-   OUT_BATCH(prog_data->cull_distance_mask |
+   OUT_BATCH(vue_prog_data->cull_distance_mask |
              (ctx->Transform.ClipPlanesEnabled <<
               GEN8_VS_USER_CLIP_DISTANCE_SHIFT));
    ADVANCE_BATCH();