Merge remote-tracking branch 'origin/released'

* origin/released:
  mesa pixman git update 18 okt 2013

Conflicts:
	mesalib/include/GL/glext.h
	mesalib/include/GL/glxext.h

9 files changed, 246 insertions, 98 deletions

diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c
index e6cae0034..793c5d382 100644
--- a/mesalib/src/mesa/main/teximage.c
+++ b/mesalib/src/mesa/main/teximage.c
@@ -1112,7 +1112,6 @@ _mesa_get_tex_max_num_levels(GLenum target, GLsizei width, GLsizei height,
    case GL_TEXTURE_CUBE_MAP_ARRAY:
    case GL_PROXY_TEXTURE_CUBE_MAP:
    case GL_PROXY_TEXTURE_CUBE_MAP_ARRAY:
-      ASSERT(width == height);
       size = width;
       break;
    case GL_TEXTURE_2D:
@@ -1447,6 +1446,8 @@ _mesa_legal_texture_dimensions(struct gl_context *ctx, GLenum target,
    case GL_PROXY_TEXTURE_CUBE_MAP_ARB:
       maxSize = 1 << (ctx->Const.MaxCubeTextureLevels - 1);
       maxSize >>= level;
+      if (width != height)
+         return GL_FALSE;
       if (width < 2 * border || width > 2 * border + maxSize)
          return GL_FALSE;
       if (height < 2 * border || height > 2 * border + maxSize)
@@ -1500,7 +1501,9 @@ _mesa_legal_texture_dimensions(struct gl_context *ctx, GLenum target,
          return GL_FALSE;
       if (height < 2 * border || height > 2 * border + maxSize)
          return GL_FALSE;
-      if (depth < 1 || depth > ctx->Const.MaxArrayTextureLayers)
+      if (depth < 1 || depth > ctx->Const.MaxArrayTextureLayers || depth % 6)
+         return GL_FALSE;
+      if (width != height)
          return GL_FALSE;
       if (level >= ctx->Const.MaxCubeTextureLevels)
          return GL_FALSE;
@@ -1991,27 +1994,6 @@ texture_error_check( struct gl_context *ctx,
       }
    }
 
-   if ((target == GL_PROXY_TEXTURE_CUBE_MAP_ARB ||
-        _mesa_is_cube_face(target)) && width != height) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glTexImage2D(cube width != height)");
-      return GL_TRUE;
-   }
-
-   if ((target == GL_PROXY_TEXTURE_CUBE_MAP_ARRAY ||
-        target == GL_TEXTURE_CUBE_MAP_ARRAY) && width != height) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glTexImage3D(cube array width != height)");
-      return GL_TRUE;
-   }
-
-   if ((target == GL_PROXY_TEXTURE_CUBE_MAP_ARRAY ||
-        target == GL_TEXTURE_CUBE_MAP_ARRAY) && (depth % 6)) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glTexImage3D(cube array depth not multiple of 6)");
-      return GL_TRUE;
-   }
-
    /* Check internalFormat */
    if (_mesa_base_tex_format(ctx, internalFormat) < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
@@ -2243,14 +2225,6 @@ compressed_texture_error_check(struct gl_context *ctx, GLint dimensions,
       goto error;
    }
 
-   /* For cube map, width must equal height */
-   if ((target == GL_PROXY_TEXTURE_CUBE_MAP_ARB ||
-        _mesa_is_cube_face(target)) && width != height) {
-      reason = "width != height";
-      error = GL_INVALID_VALUE;
-      goto error;
-   }
-
    /* check image size in bytes */
    if (expectedSize != imageSize) {
       /* Per GL_ARB_texture_compression:  GL_INVALID_VALUE is generated [...]
@@ -2596,13 +2570,6 @@ copytexture_error_check( struct gl_context *ctx, GLuint dimensions,
       }
    }
 
-   if ((target == GL_PROXY_TEXTURE_CUBE_MAP_ARB ||
-        _mesa_is_cube_face(target)) && width != height) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glTexImage2D(cube width != height)");
-      return GL_TRUE;
-   }
-
    if (_mesa_is_compressed_format(ctx, internalFormat)) {
       if (!target_can_be_compressed(ctx, target, internalFormat)) {
          _mesa_error(ctx, GL_INVALID_ENUM,
@@ -4388,7 +4355,7 @@ teximagemultisample(GLuint dims, GLenum target, GLsizei samples,
 
 void GLAPIENTRY
 _mesa_TexImage2DMultisample(GLenum target, GLsizei samples,
-                            GLint internalformat, GLsizei width,
+                            GLenum internalformat, GLsizei width,
                             GLsizei height, GLboolean fixedsamplelocations)
 {
    teximagemultisample(2, target, samples, internalformat,
@@ -4399,7 +4366,7 @@ _mesa_TexImage2DMultisample(GLenum target, GLsizei samples,
 
 void GLAPIENTRY
 _mesa_TexImage3DMultisample(GLenum target, GLsizei samples,
-                            GLint internalformat, GLsizei width,
+                            GLenum internalformat, GLsizei width,
                             GLsizei height, GLsizei depth,
                             GLboolean fixedsamplelocations)
 {
diff --git a/mesalib/src/mesa/main/teximage.h b/mesalib/src/mesa/main/teximage.h
index 889524718..792383d2f 100644
--- a/mesalib/src/mesa/main/teximage.h
+++ b/mesalib/src/mesa/main/teximage.h
@@ -296,12 +296,12 @@ _mesa_TexBufferRange(GLenum target, GLenum internalFormat, GLuint buffer,
 
 extern void GLAPIENTRY
 _mesa_TexImage2DMultisample(GLenum target, GLsizei samples,
-                            GLint internalformat, GLsizei width,
+                            GLenum internalformat, GLsizei width,
                             GLsizei height, GLboolean fixedsamplelocations);
 
 extern void GLAPIENTRY
 _mesa_TexImage3DMultisample(GLenum target, GLsizei samples,
-                            GLint internalformat, GLsizei width,
+                            GLenum internalformat, GLsizei width,
                             GLsizei height, GLsizei depth,
                             GLboolean fixedsamplelocations);
 
diff --git a/mesalib/src/mesa/state_tracker/st_atom.c b/mesalib/src/mesa/state_tracker/st_atom.c
index 1abaf8f3e..682073ed6 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.c
+++ b/mesalib/src/mesa/state_tracker/st_atom.c
@@ -66,6 +66,7 @@ static const struct st_tracked_state *atoms[] =
    &st_update_fs_constants,
    &st_bind_vs_ubos,
    &st_bind_fs_ubos,
+   &st_bind_gs_ubos,
    &st_update_pixel_transfer,
 
    /* this must be done after the vertex program update */
diff --git a/mesalib/src/mesa/state_tracker/st_atom.h b/mesalib/src/mesa/state_tracker/st_atom.h
index 101a3a6bd..ef24face9 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.h
+++ b/mesalib/src/mesa/state_tracker/st_atom.h
@@ -69,6 +69,7 @@ extern const struct st_tracked_state st_update_gs_constants;
 extern const struct st_tracked_state st_update_vs_constants;
 extern const struct st_tracked_state st_bind_fs_ubos;
 extern const struct st_tracked_state st_bind_vs_ubos;
+extern const struct st_tracked_state st_bind_gs_ubos;
 extern const struct st_tracked_state st_update_pixel_transfer;
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_atom_constbuf.c b/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
index 723ab561c..f3c457625 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
@@ -254,4 +254,21 @@ const struct st_tracked_state st_bind_fs_ubos = {
    bind_fs_ubos
 };
 
+static void bind_gs_ubos(struct st_context *st)
+{
+   struct gl_shader_program *prog = st->ctx->Shader.CurrentGeometryProgram;
+
+   if (!prog)
+      return;
+
+   st_bind_ubos(st, prog->_LinkedShaders[MESA_SHADER_GEOMETRY], PIPE_SHADER_GEOMETRY);
+}
 
+const struct st_tracked_state st_bind_gs_ubos = {
+   "st_bind_gs_ubos",
+   {
+      0,
+      ST_NEW_GEOMETRY_PROGRAM | ST_NEW_UNIFORM_BUFFER,
+   },
+   bind_gs_ubos
+};
diff --git a/mesalib/src/mesa/state_tracker/st_context.c b/mesalib/src/mesa/state_tracker/st_context.c
index 80393cff7..4e0d98cc3 100644
--- a/mesalib/src/mesa/state_tracker/st_context.c
+++ b/mesalib/src/mesa/state_tracker/st_context.c
@@ -125,6 +125,9 @@ st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe,
    st->dirty.mesa = ~0;
    st->dirty.st = ~0;
 
+   /* Create upload manager for vertex data for glBitmap, glDrawPixels,
+    * glClear, etc.
+    */
    st->uploader = u_upload_create(st->pipe, 65536, 4, PIPE_BIND_VERTEX_BUFFER);
 
    if (!screen->get_param(screen, PIPE_CAP_USER_INDEX_BUFFERS)) {
@@ -148,7 +151,8 @@ st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe,
    st_init_draw( st );
    st_init_generate_mipmap(st);
 
-   if(pipe->screen->get_param(pipe->screen, PIPE_CAP_NPOT_TEXTURES))
+   /* Choose texture target for glDrawPixels, glBitmap, renderbuffers */
+   if (pipe->screen->get_param(pipe->screen, PIPE_CAP_NPOT_TEXTURES))
       st->internal_target = PIPE_TEXTURE_2D;
    else
       st->internal_target = PIPE_TEXTURE_RECT;
diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index 8d06ac6f7..3e11cce24 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -114,6 +114,8 @@ public:
       this->index2D = 0;
       this->type = type ? type->base_type : GLSL_TYPE_ERROR;
       this->reladdr = NULL;
+      this->reladdr2 = NULL;
+      this->has_index2 = false;
    }
 
    st_src_reg(gl_register_file file, int index, int type)
@@ -125,6 +127,8 @@ public:
       this->swizzle = SWIZZLE_XYZW;
       this->negate = 0;
       this->reladdr = NULL;
+      this->reladdr2 = NULL;
+      this->has_index2 = false;
    }
 
    st_src_reg(gl_register_file file, int index, int type, int index2D)
@@ -136,6 +140,8 @@ public:
       this->swizzle = SWIZZLE_XYZW;
       this->negate = 0;
       this->reladdr = NULL;
+      this->reladdr2 = NULL;
+      this->has_index2 = false;
    }
 
    st_src_reg()
@@ -147,6 +153,8 @@ public:
       this->swizzle = 0;
       this->negate = 0;
       this->reladdr = NULL;
+      this->reladdr2 = NULL;
+      this->has_index2 = false;
    }
 
    explicit st_src_reg(st_dst_reg reg);
@@ -159,10 +167,22 @@ public:
    int type; /** GLSL_TYPE_* from GLSL IR (enum glsl_base_type) */
    /** Register index should be offset by the integer in this reg. */
    st_src_reg *reladdr;
+   st_src_reg *reladdr2;
+   bool has_index2;
 };
 
 class st_dst_reg {
 public:
+   st_dst_reg(gl_register_file file, int writemask, int type, int index)
+   {
+      this->file = file;
+      this->index = index;
+      this->writemask = writemask;
+      this->cond_mask = COND_TR;
+      this->reladdr = NULL;
+      this->type = type;
+   }
+
    st_dst_reg(gl_register_file file, int writemask, int type)
    {
       this->file = file;
@@ -203,6 +223,8 @@ st_src_reg::st_src_reg(st_dst_reg reg)
    this->negate = 0;
    this->reladdr = reg.reladdr;
    this->index2D = 0;
+   this->reladdr2 = NULL;
+   this->has_index2 = false;
 }
 
 st_dst_reg::st_dst_reg(st_src_reg reg)
@@ -449,7 +471,8 @@ static st_src_reg undef_src = st_src_reg(PROGRAM_UNDEFINED, 0, GLSL_TYPE_ERROR);
 
 static st_dst_reg undef_dst = st_dst_reg(PROGRAM_UNDEFINED, SWIZZLE_NOOP, GLSL_TYPE_ERROR);
 
-static st_dst_reg address_reg = st_dst_reg(PROGRAM_ADDRESS, WRITEMASK_X, GLSL_TYPE_FLOAT);
+static st_dst_reg address_reg = st_dst_reg(PROGRAM_ADDRESS, WRITEMASK_X, GLSL_TYPE_FLOAT, 0);
+static st_dst_reg address_reg2 = st_dst_reg(PROGRAM_ADDRESS, WRITEMASK_X, GLSL_TYPE_FLOAT, 1);
 
 static void
 fail_link(struct gl_shader_program *prog, const char *fmt, ...) PRINTFLIKE(2, 3);
@@ -515,9 +538,9 @@ glsl_to_tgsi_visitor::emit(ir_instruction *ir, unsigned op,
     * sources into temps.
     */
    num_reladdr += dst.reladdr != NULL;
-   num_reladdr += src0.reladdr != NULL;
-   num_reladdr += src1.reladdr != NULL;
-   num_reladdr += src2.reladdr != NULL;
+   num_reladdr += src0.reladdr != NULL || src0.reladdr2 != NULL;
+   num_reladdr += src1.reladdr != NULL || src1.reladdr2 != NULL;
+   num_reladdr += src2.reladdr != NULL || src2.reladdr2 != NULL;
 
    reladdr_to_temp(ir, &src2, &num_reladdr);
    reladdr_to_temp(ir, &src1, &num_reladdr);
@@ -539,9 +562,6 @@ glsl_to_tgsi_visitor::emit(ir_instruction *ir, unsigned op,
 
    inst->function = NULL;
    
-   if (op == TGSI_OPCODE_ARL || op == TGSI_OPCODE_UARL)
-      this->num_address_regs = 1;
-   
    /* Update indirect addressing status used by TGSI */
    if (dst.reladdr) {
       switch(dst.file) {
@@ -765,6 +785,10 @@ glsl_to_tgsi_visitor::emit_arl(ir_instruction *ir,
    if (src0.type == GLSL_TYPE_INT || src0.type == GLSL_TYPE_UINT)
       op = TGSI_OPCODE_UARL;
 
+   assert(dst.file == PROGRAM_ADDRESS);
+   if (dst.index >= this->num_address_regs)
+      this->num_address_regs = dst.index + 1;
+
    emit(NULL, op, dst, src0);
 }
 
@@ -1328,10 +1352,11 @@ void
 glsl_to_tgsi_visitor::reladdr_to_temp(ir_instruction *ir,
         			    st_src_reg *reg, int *num_reladdr)
 {
-   if (!reg->reladdr)
+   if (!reg->reladdr && !reg->reladdr2)
       return;
 
-   emit_arl(ir, address_reg, *reg->reladdr);
+   if (reg->reladdr) emit_arl(ir, address_reg, *reg->reladdr);
+   if (reg->reladdr2) emit_arl(ir, address_reg2, *reg->reladdr2);
 
    if (*num_reladdr != 1) {
       st_src_reg temp = get_temp(glsl_type::vec4_type);
@@ -2098,14 +2123,26 @@ glsl_to_tgsi_visitor::visit(ir_dereference_array *ir)
    ir_constant *index;
    st_src_reg src;
    int element_size = type_size(ir->type);
+   bool is_2D_input;
 
    index = ir->array_index->constant_expression_value();
 
    ir->array->accept(this);
    src = this->result;
 
+   is_2D_input = this->prog->Target == GL_GEOMETRY_PROGRAM_NV &&
+                 src.file == PROGRAM_INPUT &&
+                 ir->array->ir_type != ir_type_dereference_array;
+
+   if (is_2D_input)
+      element_size = 1;
+
    if (index) {
-      src.index += index->value.i[0] * element_size;
+      if (is_2D_input) {
+         src.index2D = index->value.i[0];
+         src.has_index2 = true;
+      } else
+         src.index += index->value.i[0] * element_size;
    } else {
       /* Variable index array dereference.  It eats the "vec4" of the
        * base of the array and an index that offsets the TGSI register
@@ -2128,7 +2165,7 @@ glsl_to_tgsi_visitor::visit(ir_dereference_array *ir)
       /* If there was already a relative address register involved, add the
        * new and the old together to get the new offset.
        */
-      if (src.reladdr != NULL) {
+      if (!is_2D_input && src.reladdr != NULL) {
          st_src_reg accum_reg = get_temp(native_integers ?
                                 glsl_type::int_type : glsl_type::float_type);
 
@@ -2138,8 +2175,15 @@ glsl_to_tgsi_visitor::visit(ir_dereference_array *ir)
          index_reg = accum_reg;
       }
 
-      src.reladdr = ralloc(mem_ctx, st_src_reg);
-      memcpy(src.reladdr, &index_reg, sizeof(index_reg));
+      if (is_2D_input) {
+         src.reladdr2 = ralloc(mem_ctx, st_src_reg);
+         memcpy(src.reladdr2, &index_reg, sizeof(index_reg));
+         src.index2D = 0;
+         src.has_index2 = true;
+      } else {
+         src.reladdr = ralloc(mem_ctx, st_src_reg);
+         memcpy(src.reladdr, &index_reg, sizeof(index_reg));
+      }
    }
 
    /* If the type is smaller than a vec4, replicate the last channel out. */
@@ -3018,16 +3062,19 @@ glsl_to_tgsi_visitor::visit(ir_if *ir)
    if_inst = emit(ir->condition, TGSI_OPCODE_ENDIF);
 }
 
+
 void
 glsl_to_tgsi_visitor::visit(ir_emit_vertex *ir)
 {
-   assert(!"Geometry shaders not supported.");
+   assert(this->prog->Target == GL_GEOMETRY_PROGRAM_NV);
+   emit(ir, TGSI_OPCODE_EMIT);
 }
 
 void
 glsl_to_tgsi_visitor::visit(ir_end_primitive *ir)
 {
-   assert(!"Geometry shaders not supported.");
+   assert(this->prog->Target == GL_GEOMETRY_PROGRAM_NV);
+   emit(ir, TGSI_OPCODE_ENDPRIM);
 }
 
 glsl_to_tgsi_visitor::glsl_to_tgsi_visitor()
@@ -3440,7 +3487,8 @@ glsl_to_tgsi_visitor::copy_propagate(void)
          int acp_base = inst->src[r].index * 4;
 
          if (inst->src[r].file != PROGRAM_TEMPORARY ||
-             inst->src[r].reladdr)
+             inst->src[r].reladdr ||
+             inst->src[r].reladdr2)
             continue;
 
          /* See if we can find entries in the ACP consisting of MOVs
@@ -3475,6 +3523,8 @@ glsl_to_tgsi_visitor::copy_propagate(void)
              */
             inst->src[r].file = first->src[0].file;
             inst->src[r].index = first->src[0].index;
+            inst->src[r].index2D = first->src[0].index2D;
+            inst->src[r].has_index2 = first->src[0].has_index2;
 
             int swizzle = 0;
             for (int i = 0; i < 4; i++) {
@@ -3579,6 +3629,7 @@ glsl_to_tgsi_visitor::copy_propagate(void)
           !inst->dst.reladdr &&
           !inst->saturate &&
           !inst->src[0].reladdr &&
+          !inst->src[0].reladdr2 &&
           !inst->src[0].negate) {
          for (int i = 0; i < 4; i++) {
             if (inst->dst.writemask & (1 << i)) {
@@ -4078,7 +4129,7 @@ struct st_translate {
    struct ureg_src *immediates;
    struct ureg_dst outputs[PIPE_MAX_SHADER_OUTPUTS];
    struct ureg_src inputs[PIPE_MAX_SHADER_INPUTS];
-   struct ureg_dst address[1];
+   struct ureg_dst address[2];
    struct ureg_src samplers[PIPE_MAX_SAMPLERS];
    struct ureg_src systemValues[SYSTEM_VALUE_MAX];
 
@@ -4355,6 +4406,15 @@ translate_src(struct st_translate *t, const st_src_reg *src_reg)
 {
    struct ureg_src src = src_register(t, src_reg->file, src_reg->index, src_reg->index2D);
 
+   if (t->procType == TGSI_PROCESSOR_GEOMETRY && src_reg->has_index2) {
+      src = src_register(t, src_reg->file, src_reg->index, src_reg->index2D);
+      if (src_reg->reladdr2)
+         src = ureg_src_dimension_indirect(src, ureg_src(t->address[1]),
+                                           src_reg->index2D);
+      else
+         src = ureg_src_dimension(src, src_reg->index2D);
+   }
+
    src = ureg_swizzle(src,
                       GET_SWZ(src_reg->swizzle, 0) & 0x3,
                       GET_SWZ(src_reg->swizzle, 1) & 0x3,
@@ -4844,8 +4904,10 @@ st_translate_program(
    /* Declare address register.
     */
    if (program->num_address_regs > 0) {
-      assert(program->num_address_regs == 1);
+      assert(program->num_address_regs <= 2);
       t->address[0] = ureg_DECL_address(ureg);
+      if (program->num_address_regs == 2)
+         t->address[1] = ureg_DECL_address(ureg);
    }
 
    /* Declare misc input registers
@@ -5166,6 +5228,9 @@ get_mesa_program(struct gl_context *ctx,
    case GL_GEOMETRY_SHADER:
       stgp = (struct st_geometry_program *)prog;
       stgp->glsl_to_tgsi = v;
+      stgp->Base.InputType = shader_program->Geom.InputType;
+      stgp->Base.OutputType = shader_program->Geom.OutputType;
+      stgp->Base.VerticesOut = shader_program->Geom.VerticesOut;
       break;
    default:
       assert(!"should not be reached");
diff --git a/mesalib/src/mesa/state_tracker/st_program.c b/mesalib/src/mesa/state_tracker/st_program.c
index 60cc37cf8..f010e1889 100644
--- a/mesalib/src/mesa/state_tracker/st_program.c
+++ b/mesalib/src/mesa/state_tracker/st_program.c
@@ -828,7 +828,6 @@ st_translate_geometry_program(struct st_context *st,
    GLuint attr;
    GLbitfield64 inputsRead;
    GLuint vslot = 0;
-   GLuint num_generic = 0;
 
    uint gs_num_inputs = 0;
    uint gs_builtin_inputs = 0;
@@ -848,7 +847,9 @@ st_translate_geometry_program(struct st_context *st,
    if (!gpv)
       return NULL;
 
-   _mesa_remove_output_reads(&stgp->Base.Base, PROGRAM_OUTPUT);
+   if (!stgp->glsl_to_tgsi) {
+      _mesa_remove_output_reads(&stgp->Base.Base, PROGRAM_OUTPUT);
+   }
 
    ureg = ureg_create( TGSI_PROCESSOR_GEOMETRY );
    if (ureg == NULL) {
@@ -910,15 +911,42 @@ st_translate_geometry_program(struct st_context *st,
             stgp->input_semantic_name[slot] = TGSI_SEMANTIC_FOG;
             stgp->input_semantic_index[slot] = 0;
             break;
+         case VARYING_SLOT_CLIP_VERTEX:
+            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPVERTEX;
+            stgp->input_semantic_index[slot] = 0;
+            break;
+         case VARYING_SLOT_CLIP_DIST0:
+            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            stgp->input_semantic_index[slot] = 0;
+            break;
+         case VARYING_SLOT_CLIP_DIST1:
+            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            stgp->input_semantic_index[slot] = 1;
+            break;
+         case VARYING_SLOT_PSIZ:
+            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
+            stgp->input_semantic_index[slot] = 0;
+            break;
          case VARYING_SLOT_TEX0:
+         case VARYING_SLOT_TEX1:
+         case VARYING_SLOT_TEX2:
+         case VARYING_SLOT_TEX3:
+         case VARYING_SLOT_TEX4:
+         case VARYING_SLOT_TEX5:
+         case VARYING_SLOT_TEX6:
+         case VARYING_SLOT_TEX7:
             stgp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stgp->input_semantic_index[slot] = num_generic++;
+            stgp->input_semantic_index[slot] = (attr - VARYING_SLOT_TEX0);
             break;
          case VARYING_SLOT_VAR0:
-            /* fall-through */
          default:
+            assert(attr >= VARYING_SLOT_VAR0 && attr < VARYING_SLOT_MAX);
             stgp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stgp->input_semantic_index[slot] = num_generic++;
+            stgp->input_semantic_index[slot] = (VARYING_SLOT_VAR0 -
+                                                VARYING_SLOT_TEX0 +
+                                                attr -
+                                                VARYING_SLOT_VAR0);
+         break;
          }
       }
    }
@@ -929,7 +957,6 @@ st_translate_geometry_program(struct st_context *st,
       gs_output_semantic_index[i] = 0;
    }
 
-   num_generic = 0;
    /*
     * Determine number of outputs, the (default) output register
     * mapping and the semantic information for each output.
@@ -972,6 +999,26 @@ st_translate_geometry_program(struct st_context *st,
             gs_output_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
             gs_output_semantic_index[slot] = 0;
             break;
+         case VARYING_SLOT_CLIP_VERTEX:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_CLIPVERTEX;
+            gs_output_semantic_index[slot] = 0;
+            break;
+         case VARYING_SLOT_CLIP_DIST0:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            gs_output_semantic_index[slot] = 0;
+            break;
+         case VARYING_SLOT_CLIP_DIST1:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            gs_output_semantic_index[slot] = 1;
+            break;
+         case VARYING_SLOT_LAYER:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_LAYER;
+            gs_output_semantic_index[slot] = 0;
+            break;
+         case VARYING_SLOT_PRIMITIVE_ID:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_PRIMID;
+            gs_output_semantic_index[slot] = 0;
+            break;
          case VARYING_SLOT_TEX0:
          case VARYING_SLOT_TEX1:
          case VARYING_SLOT_TEX2:
@@ -980,20 +1027,22 @@ st_translate_geometry_program(struct st_context *st,
          case VARYING_SLOT_TEX5:
          case VARYING_SLOT_TEX6:
          case VARYING_SLOT_TEX7:
-            /* fall-through */
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
+            gs_output_semantic_index[slot] = (attr - VARYING_SLOT_TEX0);
+            break;
          case VARYING_SLOT_VAR0:
-            /* fall-through */
          default:
             assert(slot < Elements(gs_output_semantic_name));
-            /* use default semantic info */
+            assert(attr >= VARYING_SLOT_VAR0);
             gs_output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            gs_output_semantic_index[slot] = num_generic++;
+            gs_output_semantic_index[slot] = (VARYING_SLOT_VAR0 -
+                                              VARYING_SLOT_TEX0 +
+                                              attr - 
+                                              VARYING_SLOT_VAR0);
          }
       }
    }
 
-   assert(gs_output_semantic_name[0] == TGSI_SEMANTIC_POSITION);
-
    /* find max output slot referenced to compute gs_num_outputs */
    for (attr = 0; attr < VARYING_SLOT_MAX; attr++) {
       if (outputMapping[attr] != ~0 && outputMapping[attr] > maxSlot)
@@ -1031,23 +1080,44 @@ st_translate_geometry_program(struct st_context *st,
    ureg_property_gs_output_prim(ureg, stgp->Base.OutputType);
    ureg_property_gs_max_vertices(ureg, stgp->Base.VerticesOut);
 
-   st_translate_mesa_program(st->ctx,
-                             TGSI_PROCESSOR_GEOMETRY,
-                             ureg,
-                             &stgp->Base.Base,
-                             /* inputs */
-                             gs_num_inputs,
-                             inputMapping,
-                             stgp->input_semantic_name,
-                             stgp->input_semantic_index,
-                             NULL,
-                             /* outputs */
-                             gs_num_outputs,
-                             outputMapping,
-                             gs_output_semantic_name,
-                             gs_output_semantic_index,
-                             FALSE,
-                             FALSE);
+   if (stgp->glsl_to_tgsi)
+      st_translate_program(st->ctx,
+                           TGSI_PROCESSOR_GEOMETRY,
+                           ureg,
+                           stgp->glsl_to_tgsi,
+                           &stgp->Base.Base,
+                           /* inputs */
+                           gs_num_inputs,
+                           inputMapping,
+                           stgp->input_semantic_name,
+                           stgp->input_semantic_index,
+                           NULL,
+                           NULL,
+                           /* outputs */
+                           gs_num_outputs,
+                           outputMapping,
+                           gs_output_semantic_name,
+                           gs_output_semantic_index,
+                           FALSE,
+                           FALSE);
+   else
+      st_translate_mesa_program(st->ctx,
+                                TGSI_PROCESSOR_GEOMETRY,
+                                ureg,
+                                &stgp->Base.Base,
+                                /* inputs */
+                                gs_num_inputs,
+                                inputMapping,
+                                stgp->input_semantic_name,
+                                stgp->input_semantic_index,
+                                NULL,
+                                /* outputs */
+                                gs_num_outputs,
+                                outputMapping,
+                                gs_output_semantic_name,
+                                gs_output_semantic_index,
+                                FALSE,
+                                FALSE);
 
    stgp->num_inputs = gs_num_inputs;
    stgp->tgsi.tokens = ureg_get_tokens( ureg, NULL );
diff --git a/mesalib/src/mesa/vbo/vbo_save_api.c b/mesalib/src/mesa/vbo/vbo_save_api.c
index b5f951787..411c00604 100644
--- a/mesalib/src/mesa/vbo/vbo_save_api.c
+++ b/mesalib/src/mesa/vbo/vbo_save_api.c
@@ -237,16 +237,31 @@ GLfloat *
 vbo_save_map_vertex_store(struct gl_context *ctx,
                           struct vbo_save_vertex_store *vertex_store)
 {
+   const GLbitfield access = (GL_MAP_WRITE_BIT |
+                              GL_MAP_INVALIDATE_RANGE_BIT |
+                              GL_MAP_UNSYNCHRONIZED_BIT |
+                              GL_MAP_FLUSH_EXPLICIT_BIT);
+
    assert(vertex_store->bufferobj);
-   assert(!vertex_store->buffer);
+   assert(!vertex_store->buffer);  /* the buffer should not be mapped */
+
    if (vertex_store->bufferobj->Size > 0) {
-      vertex_store->buffer =
-         (GLfloat *) ctx->Driver.MapBufferRange(ctx, 0,
-                                                vertex_store->bufferobj->Size,
-                                                GL_MAP_WRITE_BIT,  /* not used */
-                                                vertex_store->bufferobj);
-      assert(vertex_store->buffer);
-      return vertex_store->buffer + vertex_store->used;
+      /* Map the remaining free space in the VBO */
+      GLintptr offset = vertex_store->used * sizeof(GLfloat);
+      GLsizeiptr size = vertex_store->bufferobj->Size - offset;
+      GLfloat *range = (GLfloat *)
+         ctx->Driver.MapBufferRange(ctx, offset, size, access,
+                                    vertex_store->bufferobj);
+      if (range) {
+         /* compute address of start of whole buffer (needed elsewhere) */
+         vertex_store->buffer = range - vertex_store->used;
+         assert(vertex_store->buffer);
+         return range;
+      }
+      else {
+         vertex_store->buffer = NULL;
+         return NULL;
+      }
    }
    else {
       /* probably ran out of memory for buffers */
@@ -260,6 +275,14 @@ vbo_save_unmap_vertex_store(struct gl_context *ctx,
                             struct vbo_save_vertex_store *vertex_store)
 {
    if (vertex_store->bufferobj->Size > 0) {
+      GLintptr offset = 0;
+      GLsizeiptr length = vertex_store->used * sizeof(GLfloat)
+         - vertex_store->bufferobj->Offset;
+
+      /* Explicitly flush the region we wrote to */
+      ctx->Driver.FlushMappedBufferRange(ctx, offset, length,
+                                         vertex_store->bufferobj);
+
       ctx->Driver.UnmapBuffer(ctx, vertex_store->bufferobj);
    }
    vertex_store->buffer = NULL;