Merge remote-tracking branch 'origin/released'

Conflicts:
	mesalib/src/mapi/glapi/glapi.h
	mesalib/src/mapi/glapi/glapi_nop.c
	mesalib/src/mesa/main/bufferobj.c

97 files changed, 2850 insertions, 1806 deletions

diff --git a/mesalib/src/gallium/Makefile.am b/mesalib/src/gallium/Makefile.am
index 652cb1310..ede6e2123 100644
--- a/mesalib/src/gallium/Makefile.am
+++ b/mesalib/src/gallium/Makefile.am
@@ -11,7 +11,6 @@ SUBDIRS += auxiliary
 ##
 
 SUBDIRS += \
-	drivers/galahad \
 	drivers/noop \
 	drivers/trace \
 	drivers/rbug
@@ -96,14 +95,6 @@ if HAVE_DRI2
 SUBDIRS += winsys/sw/kms-dri
 endif
 
-if HAVE_EGL_PLATFORM_FBDEV
-SUBDIRS += winsys/sw/fbdev
-endif
-
-if HAVE_EGL_PLATFORM_WAYLAND
-SUBDIRS += winsys/sw/wayland
-endif
-
 SUBDIRS += winsys/sw/wrapper
 
 ##
@@ -114,8 +105,6 @@ EXTRA_DIST = \
 	docs \
 	README.portability \
 	SConscript \
-	state_trackers/gbm \
-	targets/gbm \
 	winsys/sw/gdi \
 	winsys/sw/hgl
 
diff --git a/mesalib/src/gallium/SConscript b/mesalib/src/gallium/SConscript
index 9b9011f7f..680ad925f 100644
--- a/mesalib/src/gallium/SConscript
+++ b/mesalib/src/gallium/SConscript
@@ -12,7 +12,6 @@ SConscript('auxiliary/SConscript')
 
 # These are common and work across all platforms
 SConscript([
-    'drivers/galahad/SConscript',
     'drivers/llvmpipe/SConscript',
     'drivers/rbug/SConscript',
     'drivers/softpipe/SConscript',
diff --git a/mesalib/src/gallium/auxiliary/util/u_slab.h b/mesalib/src/gallium/auxiliary/util/u_slab.h
index 29d0252ba..0df039bcd 100644
--- a/mesalib/src/gallium/auxiliary/util/u_slab.h
+++ b/mesalib/src/gallium/auxiliary/util/u_slab.h
@@ -81,7 +81,16 @@ void util_slab_destroy(struct util_slab_mempool *pool);
 void util_slab_set_thread_safety(struct util_slab_mempool *pool,
                                  enum util_slab_threading threading);
 
-#define util_slab_alloc(pool)     (pool)->alloc(pool)
-#define util_slab_free(pool, ptr) (pool)->free(pool, ptr)
+static inline void *
+util_slab_alloc(struct util_slab_mempool *pool)
+{
+   return pool->alloc(pool);
+}
+
+static inline void
+util_slab_free(struct util_slab_mempool *pool, void *ptr)
+{
+   pool->free(pool, ptr);
+}
 
 #endif
diff --git a/mesalib/src/glsl/SConscript b/mesalib/src/glsl/SConscript
index 26de455f0..284b37584 100644
--- a/mesalib/src/glsl/SConscript
+++ b/mesalib/src/glsl/SConscript
@@ -24,22 +24,26 @@ env.Prepend(LIBS = [mesautil])
 # Make glcpp-parse.h and glsl_parser.h reachable from the include path.
 env.Append(CPPPATH = [Dir('.').abspath, Dir('glcpp').abspath])
 
-env.Append(YACCFLAGS = '-d -p "glcpp_parser_"')
+glcpp_env = env.Clone()
+glcpp_env.Append(YACCFLAGS = [
+    '-d',
+    '-p', 'glcpp_parser_'
+])
 
-parser_env = env.Clone()
-parser_env.Append(YACCFLAGS = [
+glsl_env = env.Clone()
+glsl_env.Append(YACCFLAGS = [
     '--defines=%s' % File('glsl_parser.h').abspath,
     '-p', '_mesa_glsl_',
 ])
 
 # without this line scons will expect "glsl_parser.hpp" instead of
 # "glsl_parser.h", causing glsl_parser.cpp to be regenerated every time
-parser_env['YACCHXXFILESUFFIX'] = '.h'
+glsl_env['YACCHXXFILESUFFIX'] = '.h'
 
-glcpp_lexer = env.CFile('glcpp/glcpp-lex.c', 'glcpp/glcpp-lex.l')
-glcpp_parser = env.CFile('glcpp/glcpp-parse.c', 'glcpp/glcpp-parse.y')
-glsl_lexer = parser_env.CXXFile('glsl_lexer.cpp', 'glsl_lexer.ll')
-glsl_parser = parser_env.CXXFile('glsl_parser.cpp', 'glsl_parser.yy')
+glcpp_lexer = glcpp_env.CFile('glcpp/glcpp-lex.c', 'glcpp/glcpp-lex.l')
+glcpp_parser = glcpp_env.CFile('glcpp/glcpp-parse.c', 'glcpp/glcpp-parse.y')
+glsl_lexer = glsl_env.CXXFile('glsl_lexer.cpp', 'glsl_lexer.ll')
+glsl_parser = glsl_env.CXXFile('glsl_parser.cpp', 'glsl_parser.yy')
 
 # common generated sources
 glsl_sources = [
diff --git a/mesalib/src/glsl/ast_to_hir.cpp b/mesalib/src/glsl/ast_to_hir.cpp
index acb5c763c..d387b2e35 100644
--- a/mesalib/src/glsl/ast_to_hir.cpp
+++ b/mesalib/src/glsl/ast_to_hir.cpp
@@ -1617,6 +1617,12 @@ ast_expression::do_hir(exec_list *instructions,
           && cond_val != NULL) {
          result = cond_val->value.b[0] ? op[1] : op[2];
       } else {
+         /* The copy to conditional_tmp reads the whole array. */
+         if (type->is_array()) {
+            mark_whole_array_access(op[1]);
+            mark_whole_array_access(op[2]);
+         }
+
          ir_variable *const tmp =
             new(ctx) ir_variable(type, "conditional_tmp", ir_var_temporary);
          instructions->push_tail(tmp);
diff --git a/mesalib/src/glsl/builtin_functions.cpp b/mesalib/src/glsl/builtin_functions.cpp
index 28bcfdb3c..e46db8552 100755
--- a/mesalib/src/glsl/builtin_functions.cpp
+++ b/mesalib/src/glsl/builtin_functions.cpp
@@ -201,7 +201,7 @@ static bool
 shader_packing_or_es3(const _mesa_glsl_parse_state *state)
 {
    return state->ARB_shading_language_packing_enable ||
-          state->is_version(400, 300);
+          state->is_version(420, 300);
 }
 
 static bool
diff --git a/mesalib/src/glsl/glsl_parser.yy b/mesalib/src/glsl/glsl_parser.yy
index 4a2ebeb40..d33fcfe6b 100644
--- a/mesalib/src/glsl/glsl_parser.yy
+++ b/mesalib/src/glsl/glsl_parser.yy
@@ -35,6 +35,10 @@
 #include "glsl_types.h"
 #include "main/context.h"
 
+#ifdef _MSC_VER
+#pragma warning( disable : 4065 ) // switch statement contains 'default' but no 'case' labels
+#endif
+
 #undef yyerror
 
 static void yyerror(YYLTYPE *loc, _mesa_glsl_parse_state *st, const char *msg)
diff --git a/mesalib/src/glsl/ir.h b/mesalib/src/glsl/ir.h
index 25f2ecada..fdc22edf1 100644
--- a/mesalib/src/glsl/ir.h
+++ b/mesalib/src/glsl/ir.h
@@ -119,6 +119,7 @@ public:
    /*@{*/
    class ir_rvalue *as_rvalue()
    {
+      assume(this != NULL);
       if (ir_type == ir_type_dereference_array ||
           ir_type == ir_type_dereference_record ||
           ir_type == ir_type_dereference_variable ||
@@ -132,6 +133,7 @@ public:
 
    class ir_dereference *as_dereference()
    {
+      assume(this != NULL);
       if (ir_type == ir_type_dereference_array ||
           ir_type == ir_type_dereference_record ||
           ir_type == ir_type_dereference_variable)
@@ -141,6 +143,7 @@ public:
 
    class ir_jump *as_jump()
    {
+      assume(this != NULL);
       if (ir_type == ir_type_loop_jump ||
           ir_type == ir_type_return ||
           ir_type == ir_type_discard)
@@ -151,6 +154,7 @@ public:
    #define AS_CHILD(TYPE) \
    class ir_##TYPE * as_##TYPE() \
    { \
+      assume(this != NULL);                                         \
       return ir_type == ir_type_##TYPE ? (ir_##TYPE *) this : NULL; \
    }
    AS_CHILD(variable)
diff --git a/mesalib/src/glsl/ir_constant_expression.cpp b/mesalib/src/glsl/ir_constant_expression.cpp
index e24cad98d..d198276c6 100755
--- a/mesalib/src/glsl/ir_constant_expression.cpp
+++ b/mesalib/src/glsl/ir_constant_expression.cpp
@@ -35,6 +35,7 @@
 
 #include <math.h>
 #include "main/core.h" /* for MAX2, MIN2, CLAMP */
+#include "util/rounding.h" /* for _mesa_roundeven */
 #include "ir.h"
 #include "glsl_types.h"
 #include "program/hash_table.h"
@@ -238,8 +239,8 @@ pack_snorm_1x8(float x)
      * We must first cast the float to an int, because casting a negative
      * float to a uint is undefined.
      */
-   return (uint8_t) (int8_t)
-          _mesa_round_to_even(CLAMP(x, -1.0f, +1.0f) * 127.0f);
+   return (uint8_t) (int)
+          _mesa_roundevenf(CLAMP(x, -1.0f, +1.0f) * 127.0f);
 }
 
 /**
@@ -260,8 +261,8 @@ pack_snorm_1x16(float x)
      * We must first cast the float to an int, because casting a negative
      * float to a uint is undefined.
      */
-   return (uint16_t) (int16_t)
-          _mesa_round_to_even(CLAMP(x, -1.0f, +1.0f) * 32767.0f);
+   return (uint16_t) (int)
+          _mesa_roundevenf(CLAMP(x, -1.0f, +1.0f) * 32767.0f);
 }
 
 /**
@@ -315,7 +316,7 @@ pack_unorm_1x8(float x)
      *
      *       packUnorm4x8: round(clamp(c, 0, +1) * 255.0)
      */
-   return (uint8_t) _mesa_round_to_even(CLAMP(x, 0.0f, 1.0f) * 255.0f);
+   return (uint8_t) (int) _mesa_roundevenf(CLAMP(x, 0.0f, 1.0f) * 255.0f);
 }
 
 /**
@@ -333,7 +334,8 @@ pack_unorm_1x16(float x)
      *
      *       packUnorm2x16: round(clamp(c, 0, +1) * 65535.0)
      */
-   return (uint16_t) _mesa_round_to_even(CLAMP(x, 0.0f, 1.0f) * 65535.0f);
+   return (uint16_t) (int)
+          _mesa_roundevenf(CLAMP(x, 0.0f, 1.0f) * 65535.0f);
 }
 
 /**
@@ -726,9 +728,9 @@ ir_expression::constant_expression_value(struct hash_table *variable_context)
    case ir_unop_round_even:
       for (unsigned c = 0; c < op[0]->type->components(); c++) {
          if (op[0]->type->base_type == GLSL_TYPE_DOUBLE)
-            data.d[c] = _mesa_round_to_even(op[0]->value.d[c]);
+            data.d[c] = _mesa_roundeven(op[0]->value.d[c]);
          else
-            data.f[c] = _mesa_round_to_even(op[0]->value.f[c]);
+            data.f[c] = _mesa_roundevenf(op[0]->value.f[c]);
       }
       break;
 
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index 0c4467779..4349f0973 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -2542,8 +2542,9 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
 	 goto done;
       }
 
-      prog->ARB_fragment_coord_conventions_enable |=
-         prog->Shaders[i]->ARB_fragment_coord_conventions_enable;
+      if (prog->Shaders[i]->ARB_fragment_coord_conventions_enable) {
+         prog->ARB_fragment_coord_conventions_enable = true;
+      }
 
       gl_shader_stage shader_type = prog->Shaders[i]->Stage;
       shader_list[shader_type][num_shaders[shader_type]] = prog->Shaders[i];
diff --git a/mesalib/src/glsl/list.h b/mesalib/src/glsl/list.h
index 8554a4652..d855240cb 100644
--- a/mesalib/src/glsl/list.h
+++ b/mesalib/src/glsl/list.h
@@ -692,7 +692,7 @@ inline void exec_node::insert_before(exec_list *before)
            exec_node_data(__type, (__list)->head, __field),                \
                * __next =                                                  \
            exec_node_data(__type, (__node)->__field.next, __field);        \
-        __next != NULL;                                                    \
+        (__node)->__field.next != NULL;                                    \
         __node = __next, __next =                                          \
            exec_node_data(__type, (__next)->__field.next, __field))
 
@@ -701,7 +701,7 @@ inline void exec_node::insert_before(exec_list *before)
            exec_node_data(__type, (__list)->tail_pred, __field),           \
                * __prev =                                                  \
            exec_node_data(__type, (__node)->__field.prev, __field);        \
-        __prev != NULL;                                                    \
+        (__node)->__field.prev != NULL;                                    \
         __node = __prev, __prev =                                          \
            exec_node_data(__type, (__prev)->__field.prev, __field))
 
diff --git a/mesalib/src/glsl/nir/glsl_to_nir.cpp b/mesalib/src/glsl/nir/glsl_to_nir.cpp
index adef19c80..357944da6 100644
--- a/mesalib/src/glsl/nir/glsl_to_nir.cpp
+++ b/mesalib/src/glsl/nir/glsl_to_nir.cpp
@@ -43,7 +43,7 @@ namespace {
 class nir_visitor : public ir_visitor
 {
 public:
-   nir_visitor(nir_shader *shader, bool supports_ints);
+   nir_visitor(nir_shader *shader, gl_shader_stage stage);
    ~nir_visitor();
 
    virtual void visit(ir_variable *);
@@ -83,6 +83,7 @@ private:
    bool supports_ints;
 
    nir_shader *shader;
+   gl_shader_stage stage;
    nir_function_impl *impl;
    exec_list *cf_node_list;
    nir_instr *result; /* result of the expression tree last visited */
@@ -124,59 +125,24 @@ private:
 
 }; /* end of anonymous namespace */
 
-static const nir_shader_compiler_options default_options = {
-};
-
 nir_shader *
-glsl_to_nir(exec_list *ir, _mesa_glsl_parse_state *state,
-            bool native_integers)
+glsl_to_nir(struct gl_shader *sh, const nir_shader_compiler_options *options)
 {
-   const nir_shader_compiler_options *options;
-
-   if (state) {
-      struct gl_context *ctx = state->ctx;
-      struct gl_shader_compiler_options *gl_options =
-         &ctx->Const.ShaderCompilerOptions[state->stage];
-
-      if (!gl_options->NirOptions) {
-         nir_shader_compiler_options *new_options =
-            rzalloc(ctx, nir_shader_compiler_options);
-         options = gl_options->NirOptions = new_options;
-
-         if (gl_options->EmitNoPow)
-            new_options->lower_fpow = true;
-      } else {
-         options = gl_options->NirOptions;
-      }
-   } else {
-      options = &default_options;
-   }
-
    nir_shader *shader = nir_shader_create(NULL, options);
 
-   if (state) {
-      shader->num_user_structures = state->num_user_structures;
-      shader->user_structures = ralloc_array(shader, glsl_type *,
-                                             shader->num_user_structures);
-      memcpy(shader->user_structures, state->user_structures,
-            shader->num_user_structures * sizeof(glsl_type *));
-   } else {
-      shader->num_user_structures = 0;
-      shader->user_structures = NULL;
-   }
-
-   nir_visitor v1(shader, native_integers);
+   nir_visitor v1(shader, sh->Stage);
    nir_function_visitor v2(&v1);
-   v2.run(ir);
-   visit_exec_list(ir, &v1);
+   v2.run(sh->ir);
+   visit_exec_list(sh->ir, &v1);
 
    return shader;
 }
 
-nir_visitor::nir_visitor(nir_shader *shader, bool supports_ints)
+nir_visitor::nir_visitor(nir_shader *shader, gl_shader_stage stage)
 {
-   this->supports_ints = supports_ints;
+   this->supports_ints = shader->options->native_integers;
    this->shader = shader;
+   this->stage = stage;
    this->is_global = true;
    this->var_table = _mesa_hash_table_create(NULL, _mesa_hash_pointer,
                                              _mesa_key_pointer_equal);
@@ -285,7 +251,8 @@ nir_visitor::visit(ir_variable *ir)
       break;
 
    case ir_var_shader_in:
-      if (ir->data.location == VARYING_SLOT_FACE) {
+      if (stage == MESA_SHADER_FRAGMENT &&
+          ir->data.location == VARYING_SLOT_FACE) {
          /* For whatever reason, GLSL IR makes gl_FrontFacing an input */
          var->data.location = SYSTEM_VALUE_FRONT_FACE;
          var->data.mode = nir_var_system_value;
@@ -385,15 +352,15 @@ nir_visitor::visit(ir_variable *ir)
       break;
 
    case nir_var_shader_in:
-      _mesa_hash_table_insert(shader->inputs, var->name, var);
+      exec_list_push_tail(&shader->inputs, &var->node);
       break;
 
    case nir_var_shader_out:
-      _mesa_hash_table_insert(shader->outputs, var->name, var);
+      exec_list_push_tail(&shader->outputs, &var->node);
       break;
 
    case nir_var_uniform:
-      _mesa_hash_table_insert(shader->uniforms, var->name, var);
+      exec_list_push_tail(&shader->uniforms, &var->node);
       break;
 
    case nir_var_system_value:
diff --git a/mesalib/src/glsl/nir/glsl_to_nir.h b/mesalib/src/glsl/nir/glsl_to_nir.h
index 58b2cee6a..3801e8c55 100644
--- a/mesalib/src/glsl/nir/glsl_to_nir.h
+++ b/mesalib/src/glsl/nir/glsl_to_nir.h
@@ -32,8 +32,8 @@
 extern "C" {
 #endif
 
-nir_shader *glsl_to_nir(exec_list * ir, _mesa_glsl_parse_state *state,
-                        bool native_integers);
+nir_shader *glsl_to_nir(struct gl_shader *sh,
+                        const nir_shader_compiler_options *options);
 
 #ifdef __cplusplus
 }
diff --git a/mesalib/src/glsl/nir/nir.c b/mesalib/src/glsl/nir/nir.c
index ab57fd4e2..6459d5108 100644
--- a/mesalib/src/glsl/nir/nir.c
+++ b/mesalib/src/glsl/nir/nir.c
@@ -33,18 +33,12 @@ nir_shader_create(void *mem_ctx, const nir_shader_compiler_options *options)
 {
    nir_shader *shader = ralloc(mem_ctx, nir_shader);
 
-   shader->uniforms = _mesa_hash_table_create(shader, _mesa_key_hash_string,
-                                              _mesa_key_string_equal);
-   shader->inputs = _mesa_hash_table_create(shader, _mesa_key_hash_string,
-                                            _mesa_key_string_equal);
-   shader->outputs = _mesa_hash_table_create(shader, _mesa_key_hash_string,
-                                             _mesa_key_string_equal);
+   exec_list_make_empty(&shader->uniforms);
+   exec_list_make_empty(&shader->inputs);
+   exec_list_make_empty(&shader->outputs);
 
    shader->options = options;
 
-   shader->num_user_structures = 0;
-   shader->user_structures = NULL;
-
    exec_list_make_empty(&shader->functions);
    exec_list_make_empty(&shader->registers);
    exec_list_make_empty(&shader->globals);
diff --git a/mesalib/src/glsl/nir/nir.h b/mesalib/src/glsl/nir/nir.h
index d5df59609..29fe94243 100644
--- a/mesalib/src/glsl/nir/nir.h
+++ b/mesalib/src/glsl/nir/nir.h
@@ -1370,17 +1370,23 @@ typedef struct nir_shader_compiler_options {
    bool lower_fsqrt;
    /** lowers fneg and ineg to fsub and isub. */
    bool lower_negate;
+
+   /**
+    * Does the driver support real 32-bit integers?  (Otherwise, integers
+    * are simulated by floats.)
+    */
+   bool native_integers;
 } nir_shader_compiler_options;
 
 typedef struct nir_shader {
    /** hash table of name -> uniform nir_variable */
-   struct hash_table *uniforms;
+   struct exec_list uniforms;
 
    /** hash table of name -> input nir_variable */
-   struct hash_table *inputs;
+   struct exec_list inputs;
 
    /** hash table of name -> output nir_variable */
-   struct hash_table *outputs;
+   struct exec_list outputs;
 
    /** Set of driver-specific options for the shader.
     *
@@ -1400,10 +1406,6 @@ typedef struct nir_shader {
    /** list of global register in the shader */
    struct exec_list registers;
 
-   /** structures used in this shader */
-   unsigned num_user_structures;
-   struct glsl_type **user_structures;
-
    /** next available global register index */
    unsigned reg_alloc;
 
@@ -1575,6 +1577,13 @@ void nir_lower_global_vars_to_local(nir_shader *shader);
 
 void nir_lower_locals_to_regs(nir_shader *shader);
 
+void nir_assign_var_locations_scalar(struct exec_list *var_list,
+                                     unsigned *size);
+void nir_assign_var_locations_scalar_direct_first(nir_shader *shader,
+                                                  struct exec_list *var_list,
+                                                  unsigned *direct_size,
+                                                  unsigned *size);
+
 void nir_lower_io(nir_shader *shader);
 
 void nir_lower_vars_to_ssa(nir_shader *shader);
diff --git a/mesalib/src/glsl/nir/nir_constant_expressions.py b/mesalib/src/glsl/nir/nir_constant_expressions.py
index 22bc4f095..bf82fe533 100644
--- a/mesalib/src/glsl/nir/nir_constant_expressions.py
+++ b/mesalib/src/glsl/nir/nir_constant_expressions.py
@@ -28,6 +28,7 @@ template = """\
 
 #include <math.h>
 #include "main/core.h"
+#include "util/rounding.h" /* for _mesa_roundeven */
 #include "nir_constant_expressions.h"
 
 #if defined(_MSC_VER) && (_MSC_VER < 1800)
@@ -68,8 +69,8 @@ pack_snorm_1x8(float x)
      * We must first cast the float to an int, because casting a negative
      * float to a uint is undefined.
      */
-   return (uint8_t) (int8_t)
-          _mesa_round_to_even(CLAMP(x, -1.0f, +1.0f) * 127.0f);
+   return (uint8_t) (int)
+          _mesa_roundevenf(CLAMP(x, -1.0f, +1.0f) * 127.0f);
 }
 
 /**
@@ -90,8 +91,8 @@ pack_snorm_1x16(float x)
      * We must first cast the float to an int, because casting a negative
      * float to a uint is undefined.
      */
-   return (uint16_t) (int16_t)
-          _mesa_round_to_even(CLAMP(x, -1.0f, +1.0f) * 32767.0f);
+   return (uint16_t) (int)
+          _mesa_roundevenf(CLAMP(x, -1.0f, +1.0f) * 32767.0f);
 }
 
 /**
@@ -145,7 +146,8 @@ pack_unorm_1x8(float x)
      *
      *       packUnorm4x8: round(clamp(c, 0, +1) * 255.0)
      */
-   return (uint8_t) _mesa_round_to_even(CLAMP(x, 0.0f, 1.0f) * 255.0f);
+   return (uint8_t) (int)
+          _mesa_roundevenf(CLAMP(x, 0.0f, 1.0f) * 255.0f);
 }
 
 /**
@@ -163,7 +165,8 @@ pack_unorm_1x16(float x)
      *
      *       packUnorm2x16: round(clamp(c, 0, +1) * 65535.0)
      */
-   return (uint16_t) _mesa_round_to_even(CLAMP(x, 0.0f, 1.0f) * 65535.0f);
+   return (uint16_t) (int)
+          _mesa_roundevenf(CLAMP(x, 0.0f, 1.0f) * 65535.0f);
 }
 
 /**
diff --git a/mesalib/src/glsl/nir/nir_intrinsics.h b/mesalib/src/glsl/nir/nir_intrinsics.h
index 3bf102fc1..8e28765c1 100644
--- a/mesalib/src/glsl/nir/nir_intrinsics.h
+++ b/mesalib/src/glsl/nir/nir_intrinsics.h
@@ -95,6 +95,8 @@ ATOMIC(read, NIR_INTRINSIC_CAN_ELIMINATE)
 
 SYSTEM_VALUE(front_face, 1)
 SYSTEM_VALUE(vertex_id, 1)
+SYSTEM_VALUE(vertex_id_zero_base, 1)
+SYSTEM_VALUE(base_vertex, 1)
 SYSTEM_VALUE(instance_id, 1)
 SYSTEM_VALUE(sample_id, 1)
 SYSTEM_VALUE(sample_pos, 2)
diff --git a/mesalib/src/glsl/nir/nir_lower_io.c b/mesalib/src/glsl/nir/nir_lower_io.c
index 207f8daa1..03eed04e1 100644
--- a/mesalib/src/glsl/nir/nir_lower_io.c
+++ b/mesalib/src/glsl/nir/nir_lower_io.c
@@ -76,15 +76,12 @@ type_size(const struct glsl_type *type)
    return 0;
 }
 
-static void
-assign_var_locations(struct hash_table *ht, unsigned *size)
+void
+nir_assign_var_locations_scalar(struct exec_list *var_list, unsigned *size)
 {
    unsigned location = 0;
 
-   struct hash_entry *entry;
-   hash_table_foreach(ht, entry) {
-      nir_variable *var = (nir_variable *) entry->data;
-
+   foreach_list_typed(nir_variable, var, node, var_list) {
       /*
        * UBO's have their own address spaces, so don't count them towards the
        * number of global uniforms
@@ -99,14 +96,6 @@ assign_var_locations(struct hash_table *ht, unsigned *size)
    *size = location;
 }
 
-static void
-assign_var_locations_shader(nir_shader *shader)
-{
-   assign_var_locations(shader->inputs, &shader->num_inputs);
-   assign_var_locations(shader->outputs, &shader->num_outputs);
-   assign_var_locations(shader->uniforms, &shader->num_uniforms);
-}
-
 static bool
 deref_has_indirect(nir_deref_var *deref)
 {
@@ -121,6 +110,77 @@ deref_has_indirect(nir_deref_var *deref)
    return false;
 }
 
+static bool
+mark_indirect_uses_block(nir_block *block, void *void_state)
+{
+   struct set *indirect_set = void_state;
+
+   nir_foreach_instr(block, instr) {
+      if (instr->type != nir_instr_type_intrinsic)
+         continue;
+
+      nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
+
+      for (unsigned i = 0;
+           i < nir_intrinsic_infos[intrin->intrinsic].num_variables; i++) {
+         if (deref_has_indirect(intrin->variables[i]))
+            _mesa_set_add(indirect_set, intrin->variables[i]->var);
+      }
+   }
+
+   return true;
+}
+
+/* Identical to nir_assign_var_locations_packed except that it assigns
+ * locations to the variables that are used 100% directly first and then
+ * assigns locations to variables that are used indirectly.
+ */
+void
+nir_assign_var_locations_scalar_direct_first(nir_shader *shader,
+                                             struct exec_list *var_list,
+                                             unsigned *direct_size,
+                                             unsigned *size)
+{
+   struct set *indirect_set = _mesa_set_create(NULL, _mesa_hash_pointer,
+                                               _mesa_key_pointer_equal);
+
+   nir_foreach_overload(shader, overload) {
+      if (overload->impl)
+         nir_foreach_block(overload->impl, mark_indirect_uses_block,
+                           indirect_set);
+   }
+
+   unsigned location = 0;
+
+   foreach_list_typed(nir_variable, var, node, var_list) {
+      if (var->data.mode == nir_var_uniform && var->interface_type != NULL)
+         continue;
+
+      if (_mesa_set_search(indirect_set, var))
+         continue;
+
+      var->data.driver_location = location;
+      location += type_size(var->type);
+   }
+
+   *direct_size = location;
+
+   foreach_list_typed(nir_variable, var, node, var_list) {
+      if (var->data.mode == nir_var_uniform && var->interface_type != NULL)
+         continue;
+
+      if (!_mesa_set_search(indirect_set, var))
+         continue;
+
+      var->data.driver_location = location;
+      location += type_size(var->type);
+   }
+
+   *size = location;
+
+   _mesa_set_destroy(indirect_set, NULL);
+}
+
 static unsigned
 get_io_offset(nir_deref_var *deref, nir_instr *instr, nir_src *indirect,
               struct lower_io_state *state)
@@ -307,8 +367,6 @@ nir_lower_io_impl(nir_function_impl *impl)
 void
 nir_lower_io(nir_shader *shader)
 {
-   assign_var_locations_shader(shader);
-
    nir_foreach_overload(shader, overload) {
       if (overload->impl)
          nir_lower_io_impl(overload->impl);
diff --git a/mesalib/src/glsl/nir/nir_lower_system_values.c b/mesalib/src/glsl/nir/nir_lower_system_values.c
index 328d4f1ab..a6eec653e 100644
--- a/mesalib/src/glsl/nir/nir_lower_system_values.c
+++ b/mesalib/src/glsl/nir/nir_lower_system_values.c
@@ -49,6 +49,12 @@ convert_instr(nir_intrinsic_instr *instr)
    case SYSTEM_VALUE_VERTEX_ID:
       op = nir_intrinsic_load_vertex_id;
       break;
+   case SYSTEM_VALUE_VERTEX_ID_ZERO_BASE:
+      op = nir_intrinsic_load_vertex_id_zero_base;
+      break;
+   case SYSTEM_VALUE_BASE_VERTEX:
+      op = nir_intrinsic_load_base_vertex;
+      break;
    case SYSTEM_VALUE_INSTANCE_ID:
       op = nir_intrinsic_load_instance_id;
       break;
diff --git a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
index 9e9a418e3..86e6ab416 100644
--- a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
+++ b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
@@ -35,6 +35,13 @@ struct deref_node {
 
    bool lower_to_ssa;
 
+   /* Only valid for things that end up in the direct list.
+    * Note that multiple nir_deref_vars may correspond to this node, but they
+    * will all be equivalent, so any is as good as the other.
+    */
+   nir_deref_var *deref;
+   struct exec_node direct_derefs_link;
+
    struct set *loads;
    struct set *stores;
    struct set *copies;
@@ -69,7 +76,7 @@ struct lower_variables_state {
     * wildcards and no indirects, these are precisely the derefs that we
     * can actually consider lowering.
     */
-   struct hash_table *direct_deref_nodes;
+   struct exec_list direct_deref_nodes;
 
    /* Controls whether get_deref_node will add variables to the
     * direct_deref_nodes table.  This is turned on when we are initially
@@ -83,88 +90,6 @@ struct lower_variables_state {
    struct hash_table *phi_table;
 };
 
-/* The following two functions implement a hash and equality check for
- * variable dreferences.  When the hash or equality function encounters an
- * array, all indirects are treated as equal and are never equal to a
- * direct dereference or a wildcard.
- */
-static uint32_t
-hash_deref(const void *void_deref)
-{
-   uint32_t hash = _mesa_fnv32_1a_offset_bias;
-
-   const nir_deref_var *deref_var = void_deref;
-   hash = _mesa_fnv32_1a_accumulate(hash, deref_var->var);
-
-   for (const nir_deref *deref = deref_var->deref.child;
-        deref; deref = deref->child) {
-      switch (deref->deref_type) {
-      case nir_deref_type_array: {
-         nir_deref_array *deref_array = nir_deref_as_array(deref);
-
-         hash = _mesa_fnv32_1a_accumulate(hash, deref_array->deref_array_type);
-
-         if (deref_array->deref_array_type == nir_deref_array_type_direct)
-            hash = _mesa_fnv32_1a_accumulate(hash, deref_array->base_offset);
-         break;
-      }
-      case nir_deref_type_struct: {
-         nir_deref_struct *deref_struct = nir_deref_as_struct(deref);
-         hash = _mesa_fnv32_1a_accumulate(hash, deref_struct->index);
-         break;
-      }
-      default:
-         assert("Invalid deref chain");
-      }
-   }
-
-   return hash;
-}
-
-static bool
-derefs_equal(const void *void_a, const void *void_b)
-{
-   const nir_deref_var *a_var = void_a;
-   const nir_deref_var *b_var = void_b;
-
-   if (a_var->var != b_var->var)
-      return false;
-
-   for (const nir_deref *a = a_var->deref.child, *b = b_var->deref.child;
-        a != NULL; a = a->child, b = b->child) {
-      if (a->deref_type != b->deref_type)
-         return false;
-
-      switch (a->deref_type) {
-      case nir_deref_type_array: {
-         nir_deref_array *a_arr = nir_deref_as_array(a);
-         nir_deref_array *b_arr = nir_deref_as_array(b);
-
-         if (a_arr->deref_array_type != b_arr->deref_array_type)
-            return false;
-
-         if (a_arr->deref_array_type == nir_deref_array_type_direct &&
-             a_arr->base_offset != b_arr->base_offset)
-            return false;
-         break;
-      }
-      case nir_deref_type_struct:
-         if (nir_deref_as_struct(a)->index != nir_deref_as_struct(b)->index)
-            return false;
-         break;
-      default:
-         assert("Invalid deref chain");
-         return false;
-      }
-
-      assert((a->child == NULL) == (b->child == NULL));
-      if((a->child == NULL) != (b->child == NULL))
-         return false;
-   }
-
-   return true;
-}
-
 static int
 type_get_length(const struct glsl_type *type)
 {
@@ -195,6 +120,8 @@ deref_node_create(struct deref_node *parent,
    struct deref_node *node = rzalloc_size(mem_ctx, size);
    node->type = type;
    node->parent = parent;
+   node->deref = NULL;
+   exec_node_init(&node->direct_derefs_link);
 
    return node;
 }
@@ -297,8 +224,14 @@ get_deref_node(nir_deref_var *deref, struct lower_variables_state *state)
 
    assert(node);
 
-   if (is_direct && state->add_to_direct_deref_nodes)
-      _mesa_hash_table_insert(state->direct_deref_nodes, deref, node);
+   /* Only insert if it isn't already in the list. */
+   if (is_direct && state->add_to_direct_deref_nodes &&
+       node->direct_derefs_link.next == NULL) {
+      node->deref = deref;
+      assert(deref->var != NULL);
+      exec_list_push_tail(&state->direct_deref_nodes,
+                          &node->direct_derefs_link);
+   }
 
    return node;
 }
@@ -917,10 +850,8 @@ insert_phi_nodes(struct lower_variables_state *state)
    unsigned w_start, w_end;
    unsigned iter_count = 0;
 
-   struct hash_entry *deref_entry;
-   hash_table_foreach(state->direct_deref_nodes, deref_entry) {
-      struct deref_node *node = deref_entry->data;
-
+   foreach_list_typed(struct deref_node, node, direct_derefs_link,
+                      &state->direct_deref_nodes) {
       if (node->stores == NULL)
          continue;
 
@@ -1014,8 +945,7 @@ nir_lower_vars_to_ssa_impl(nir_function_impl *impl)
    state.deref_var_nodes = _mesa_hash_table_create(state.dead_ctx,
                                                    _mesa_hash_pointer,
                                                    _mesa_key_pointer_equal);
-   state.direct_deref_nodes = _mesa_hash_table_create(state.dead_ctx,
-                                                      hash_deref, derefs_equal);
+   exec_list_make_empty(&state.direct_deref_nodes);
    state.phi_table = _mesa_hash_table_create(state.dead_ctx,
                                              _mesa_hash_pointer,
                                              _mesa_key_pointer_equal);
@@ -1035,18 +965,17 @@ nir_lower_vars_to_ssa_impl(nir_function_impl *impl)
    /* We're about to iterate through direct_deref_nodes.  Don't modify it. */
    state.add_to_direct_deref_nodes = false;
 
-   struct hash_entry *entry;
-   hash_table_foreach(state.direct_deref_nodes, entry) {
-      nir_deref_var *deref = (void *)entry->key;
-      struct deref_node *node = entry->data;
+   foreach_list_typed_safe(struct deref_node, node, direct_derefs_link,
+                           &state.direct_deref_nodes) {
+      nir_deref_var *deref = node->deref;
 
       if (deref->var->data.mode != nir_var_local) {
-         _mesa_hash_table_remove(state.direct_deref_nodes, entry);
+         exec_node_remove(&node->direct_derefs_link);
          continue;
       }
 
       if (deref_may_be_aliased(deref, &state)) {
-         _mesa_hash_table_remove(state.direct_deref_nodes, entry);
+         exec_node_remove(&node->direct_derefs_link);
          continue;
       }
 
diff --git a/mesalib/src/glsl/nir/nir_opcodes.py b/mesalib/src/glsl/nir/nir_opcodes.py
index 77f3bb826..062cd628b 100644
--- a/mesalib/src/glsl/nir/nir_opcodes.py
+++ b/mesalib/src/glsl/nir/nir_opcodes.py
@@ -183,7 +183,7 @@ unop("ftrunc", tfloat, "truncf(src0)")
 unop("fceil", tfloat, "ceilf(src0)")
 unop("ffloor", tfloat, "floorf(src0)")
 unop("ffract", tfloat, "src0 - floorf(src0)")
-unop("fround_even", tfloat, "_mesa_round_to_even(src0)")
+unop("fround_even", tfloat, "_mesa_roundevenf(src0)")
 
 
 # Trigonometric operations.
diff --git a/mesalib/src/glsl/nir/nir_opt_algebraic.py b/mesalib/src/glsl/nir/nir_opt_algebraic.py
index 7bf643134..ef855aa77 100644
--- a/mesalib/src/glsl/nir/nir_opt_algebraic.py
+++ b/mesalib/src/glsl/nir/nir_opt_algebraic.py
@@ -56,6 +56,10 @@ optimizations = [
    (('iabs', ('ineg', a)), ('iabs', a)),
    (('fadd', a, 0.0), a),
    (('iadd', a, 0), a),
+   (('fadd', ('fmul', a, b), ('fmul', a, c)), ('fmul', a, ('fadd', b, c))),
+   (('iadd', ('imul', a, b), ('imul', a, c)), ('imul', a, ('iadd', b, c))),
+   (('fadd', ('fneg', a), a), 0.0),
+   (('iadd', ('ineg', a), a), 0),
    (('fmul', a, 0.0), 0.0),
    (('imul', a, 0), 0),
    (('fmul', a, 1.0), a),
diff --git a/mesalib/src/glsl/nir/nir_opt_peephole_select.c b/mesalib/src/glsl/nir/nir_opt_peephole_select.c
index ab08f286f..b89451b09 100644
--- a/mesalib/src/glsl/nir/nir_opt_peephole_select.c
+++ b/mesalib/src/glsl/nir/nir_opt_peephole_select.c
@@ -52,36 +52,66 @@ struct peephole_select_state {
 };
 
 static bool
-are_all_move_to_phi(nir_block *block)
+block_check_for_allowed_instrs(nir_block *block)
 {
    nir_foreach_instr(block, instr) {
-      if (instr->type != nir_instr_type_alu)
-         return false;
+      switch (instr->type) {
+      case nir_instr_type_intrinsic: {
+         nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
+
+         switch (intrin->intrinsic) {
+         case nir_intrinsic_load_var:
+            switch (intrin->variables[0]->var->data.mode) {
+            case nir_var_shader_in:
+            case nir_var_uniform:
+               break;
+
+            default:
+               return false;
+            }
+            break;
+
+         default:
+            return false;
+         }
 
-      /* It must be a move operation */
-      nir_alu_instr *mov = nir_instr_as_alu(instr);
-      if (mov->op != nir_op_fmov && mov->op != nir_op_imov)
-         return false;
+         break;
+      }
 
-      /* Can't handle saturate */
-      if (mov->dest.saturate)
-         return false;
+      case nir_instr_type_load_const:
+         break;
 
-      /* It must be SSA */
-      if (!mov->dest.dest.is_ssa)
-         return false;
+      case nir_instr_type_alu: {
+         /* It must be a move operation */
+         nir_alu_instr *mov = nir_instr_as_alu(instr);
+         if (mov->op != nir_op_fmov && mov->op != nir_op_imov)
+            return false;
 
-      /* It cannot have any if-uses */
-      if (mov->dest.dest.ssa.if_uses->entries != 0)
-         return false;
+         /* Can't handle saturate */
+         if (mov->dest.saturate)
+            return false;
+
+         /* It must be SSA */
+         if (!mov->dest.dest.is_ssa)
+            return false;
 
-      /* The only uses of this definition must be phi's in the successor */
-      struct set_entry *entry;
-      set_foreach(mov->dest.dest.ssa.uses, entry) {
-         const nir_instr *dest_instr = entry->key;
-         if (dest_instr->type != nir_instr_type_phi ||
-             dest_instr->block != block->successors[0])
+         /* It cannot have any if-uses */
+         if (mov->dest.dest.ssa.if_uses->entries != 0)
             return false;
+
+         /* The only uses of this definition must be phi's in the successor */
+         struct set_entry *entry;
+         set_foreach(mov->dest.dest.ssa.uses, entry) {
+            const nir_instr *dest_instr = entry->key;
+            if (dest_instr->type != nir_instr_type_phi ||
+                dest_instr->block != block->successors[0])
+               return false;
+         }
+         break;
+      }
+
+      default:
+         return false;
       }
    }
 
@@ -119,8 +149,9 @@ nir_opt_peephole_select_block(nir_block *block, void *void_state)
    nir_block *then_block = nir_cf_node_as_block(then_node);
    nir_block *else_block = nir_cf_node_as_block(else_node);
 
-   /* ... and those blocks must only contain move-to-phi. */
-   if (!are_all_move_to_phi(then_block) || !are_all_move_to_phi(else_block))
+   /* ... and those blocks must only contain "allowed" instructions. */
+   if (!block_check_for_allowed_instrs(then_block) ||
+       !block_check_for_allowed_instrs(else_block))
       return true;
 
    /* At this point, we know that the previous CFG node is an if-then
@@ -129,6 +160,25 @@ nir_opt_peephole_select_block(nir_block *block, void *void_state)
     * selects.
     */
 
+   nir_block *prev_block = nir_cf_node_as_block(nir_cf_node_prev(prev_node));
+   assert(prev_block->cf_node.type == nir_cf_node_block);
+
+   /* First, we move the remaining instructions from the blocks to the
+    * block before.  We have already guaranteed that this is safe by
+    * calling block_check_for_allowed_instrs()
+    */
+   nir_foreach_instr_safe(then_block, instr) {
+      exec_node_remove(&instr->node);
+      instr->block = prev_block;
+      exec_list_push_tail(&prev_block->instr_list, &instr->node);
+   }
+
+   nir_foreach_instr_safe(else_block, instr) {
+      exec_node_remove(&instr->node);
+      instr->block = prev_block;
+      exec_list_push_tail(&prev_block->instr_list, &instr->node);
+   }
+
    nir_foreach_instr_safe(block, instr) {
       if (instr->type != nir_instr_type_phi)
          break;
@@ -145,19 +195,7 @@ nir_opt_peephole_select_block(nir_block *block, void *void_state)
          assert(src->src.is_ssa);
 
          unsigned idx = src->pred == then_block ? 1 : 2;
-
-         if (src->src.ssa->parent_instr->block == src->pred) {
-            /* We already know that this instruction must be a move with
-             * this phi's in this block as its only users.
-             */
-            nir_alu_instr *mov = nir_instr_as_alu(src->src.ssa->parent_instr);
-            assert(mov->instr.type == nir_instr_type_alu);
-            assert(mov->op == nir_op_fmov || mov->op == nir_op_imov);
-
-            nir_alu_src_copy(&sel->src[idx], &mov->src[0], state->mem_ctx);
-         } else {
-            nir_src_copy(&sel->src[idx].src, &src->src, state->mem_ctx);
-         }
+         nir_src_copy(&sel->src[idx].src, &src->src, state->mem_ctx);
       }
 
       nir_ssa_dest_init(&sel->instr, &sel->dest.dest,
diff --git a/mesalib/src/glsl/nir/nir_print.c b/mesalib/src/glsl/nir/nir_print.c
index 6a3c6a027..fa11a312e 100644
--- a/mesalib/src/glsl/nir/nir_print.c
+++ b/mesalib/src/glsl/nir/nir_print.c
@@ -228,7 +228,7 @@ print_var_decl(nir_variable *var, print_var_state *state, FILE *fp)
    if (var->data.mode == nir_var_shader_in ||
        var->data.mode == nir_var_shader_out ||
        var->data.mode == nir_var_uniform) {
-      fprintf(fp, " (%u)", var->data.driver_location);
+      fprintf(fp, " (%u, %u)", var->data.location, var->data.driver_location);
    }
 
    fprintf(fp, "\n");
@@ -844,22 +844,16 @@ nir_print_shader(nir_shader *shader, FILE *fp)
    print_var_state state;
    init_print_state(&state);
 
-   for (unsigned i = 0; i < shader->num_user_structures; i++) {
-      glsl_print_struct(shader->user_structures[i], fp);
-   }
-
-   struct hash_entry *entry;
-
-   hash_table_foreach(shader->uniforms, entry) {
-      print_var_decl((nir_variable *) entry->data, &state, fp);
+   foreach_list_typed(nir_variable, var, node, &shader->uniforms) {
+      print_var_decl(var, &state, fp);
    }
 
-   hash_table_foreach(shader->inputs, entry) {
-      print_var_decl((nir_variable *) entry->data, &state, fp);
+   foreach_list_typed(nir_variable, var, node, &shader->inputs) {
+      print_var_decl(var, &state, fp);
    }
 
-   hash_table_foreach(shader->outputs, entry) {
-      print_var_decl((nir_variable *) entry->data, &state, fp);
+   foreach_list_typed(nir_variable, var, node, &shader->outputs) {
+      print_var_decl(var, &state, fp);
    }
 
    foreach_list_typed(nir_variable, var, node, &shader->globals) {
diff --git a/mesalib/src/glsl/nir/nir_validate.c b/mesalib/src/glsl/nir/nir_validate.c
index a3fe9d620..f247ae069 100644
--- a/mesalib/src/glsl/nir/nir_validate.c
+++ b/mesalib/src/glsl/nir/nir_validate.c
@@ -931,17 +931,19 @@ nir_validate_shader(nir_shader *shader)
 
    state.shader = shader;
 
-   struct hash_entry *entry;
-   hash_table_foreach(shader->uniforms, entry) {
-      validate_var_decl((nir_variable *) entry->data, true, &state);
+   exec_list_validate(&shader->uniforms);
+   foreach_list_typed(nir_variable, var, node, &shader->uniforms) {
+      validate_var_decl(var, true, &state);
    }
 
-   hash_table_foreach(shader->inputs, entry) {
-     validate_var_decl((nir_variable *) entry->data, true, &state);
+   exec_list_validate(&shader->inputs);
+   foreach_list_typed(nir_variable, var, node, &shader->inputs) {
+     validate_var_decl(var, true, &state);
    }
 
-   hash_table_foreach(shader->outputs, entry) {
-      validate_var_decl((nir_variable *) entry->data, true, &state);
+   exec_list_validate(&shader->outputs);
+   foreach_list_typed(nir_variable, var, node, &shader->outputs) {
+     validate_var_decl(var, true, &state);
    }
 
    exec_list_validate(&shader->globals);
diff --git a/mesalib/src/glsl/nir/nir_worklist.c b/mesalib/src/glsl/nir/nir_worklist.c
index a8baae937..3087a1d23 100644
--- a/mesalib/src/glsl/nir/nir_worklist.c
+++ b/mesalib/src/glsl/nir/nir_worklist.c
@@ -82,7 +82,7 @@ nir_block_worklist_push_head(nir_block_worklist *w, nir_block *block)
 }
 
 nir_block *
-nir_block_worklist_peek_head(nir_block_worklist *w)
+nir_block_worklist_peek_head(const nir_block_worklist *w)
 {
    assert(w->count > 0);
 
@@ -114,18 +114,18 @@ nir_block_worklist_push_tail(nir_block_worklist *w, nir_block *block)
 
    w->count++;
 
-   unsigned tail = w->start = (w->start + w->count - 1) % w->size;
+   unsigned tail = (w->start + w->count - 1) % w->size;
 
    w->blocks[tail] = block;
    BITSET_SET(w->blocks_present, block->index);
 }
 
 nir_block *
-nir_block_worklist_peek_tail(nir_block_worklist *w)
+nir_block_worklist_peek_tail(const nir_block_worklist *w)
 {
    assert(w->count > 0);
 
-   unsigned tail = w->start = (w->start + w->count - 1) % w->size;
+   unsigned tail = (w->start + w->count - 1) % w->size;
 
    return w->blocks[tail];
 }
@@ -135,7 +135,7 @@ nir_block_worklist_pop_tail(nir_block_worklist *w)
 {
    assert(w->count > 0);
 
-   unsigned tail = w->start = (w->start + w->count - 1) % w->size;
+   unsigned tail = (w->start + w->count - 1) % w->size;
 
    w->count--;
 
diff --git a/mesalib/src/glsl/nir/nir_worklist.h b/mesalib/src/glsl/nir/nir_worklist.h
index d5a8568e4..829bff24a 100644
--- a/mesalib/src/glsl/nir/nir_worklist.h
+++ b/mesalib/src/glsl/nir/nir_worklist.h
@@ -74,13 +74,13 @@ nir_block_worklist_is_empty(const nir_block_worklist *w)
 
 void nir_block_worklist_push_head(nir_block_worklist *w, nir_block *block);
 
-nir_block *nir_block_worklist_peek_head(nir_block_worklist *w);
+nir_block *nir_block_worklist_peek_head(const nir_block_worklist *w);
 
 nir_block *nir_block_worklist_pop_head(nir_block_worklist *w);
 
 void nir_block_worklist_push_tail(nir_block_worklist *w, nir_block *block);
 
-nir_block *nir_block_worklist_peek_tail(nir_block_worklist *w);
+nir_block *nir_block_worklist_peek_tail(const nir_block_worklist *w);
 
 nir_block *nir_block_worklist_pop_tail(nir_block_worklist *w);
 
diff --git a/mesalib/src/glsl/opt_algebraic.cpp b/mesalib/src/glsl/opt_algebraic.cpp
index c6040bff8..69c03ea8b 100644
--- a/mesalib/src/glsl/opt_algebraic.cpp
+++ b/mesalib/src/glsl/opt_algebraic.cpp
@@ -626,9 +626,18 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
          if (!is_vec_zero(zero))
             continue;
 
-         return new(mem_ctx) ir_expression(ir->operation,
-                                           add->operands[0],
-                                           neg(add->operands[1]));
+         /* Depending of the zero position we want to optimize
+          * (0 cmp x+y) into (-x cmp y) or (x+y cmp 0) into (x cmp -y)
+          */
+         if (add_pos == 1) {
+            return new(mem_ctx) ir_expression(ir->operation,
+                                              neg(add->operands[0]),
+                                              add->operands[1]);
+         } else {
+            return new(mem_ctx) ir_expression(ir->operation,
+                                              add->operands[0],
+                                              neg(add->operands[1]));
+         }
       }
       break;
 
diff --git a/mesalib/src/loader/loader.c b/mesalib/src/loader/loader.c
index 9ff511522..17bf13360 100644
--- a/mesalib/src/loader/loader.c
+++ b/mesalib/src/loader/loader.c
@@ -64,6 +64,7 @@
  *    Rob Clark <robclark@freedesktop.org>
  */
 
+#include <sys/stat.h>
 #include <stdarg.h>
 #include <stdio.h>
 #include <string.h>
@@ -80,7 +81,6 @@
 #endif
 #endif
 #ifdef HAVE_SYSFS
-#include <sys/stat.h>
 #include <sys/types.h>
 #endif
 #include "loader.h"
diff --git a/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml b/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
index 2fe1638fd..641e68f71 100644
--- a/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
+++ b/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
@@ -7,6 +7,110 @@
    <enum name="QUERY_TARGET"    value="0x82EA"/>
    <enum name="TEXTURE_BINDING" value="0x82EB"/>
 
+   <!-- Buffer object functions -->
+
+   <function name="CreateBuffers" offset="assign">
+      <param name="n" type="GLsizei" />
+      <param name="buffers" type="GLuint *" />
+   </function>
+
+   <function name="NamedBufferStorage" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="size" type="GLsizeiptr" />
+      <param name="data" type="const GLvoid *" />
+      <param name="flags" type="GLbitfield" />
+   </function>
+
+   <function name="NamedBufferData" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="size" type="GLsizeiptr" />
+      <param name="data" type="const GLvoid *" />
+      <param name="usage" type="GLenum" />
+   </function>
+
+   <function name="NamedBufferSubData" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+      <param name="data" type="const GLvoid *" />
+   </function>
+
+   <function name="CopyNamedBufferSubData" offset="assign">
+      <param name="readBuffer" type="GLuint" />
+      <param name="writeBuffer" type="GLuint" />
+      <param name="readOffset" type="GLintptr" />
+      <param name="writeOffset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+   </function>
+
+   <function name="ClearNamedBufferData" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="internalformat" type="GLenum" />
+      <param name="format" type="GLenum" />
+      <param name="type" type="GLenum" />
+      <param name="data" type="const GLvoid *" />
+   </function>
+
+   <function name="ClearNamedBufferSubData" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="internalformat" type="GLenum" />
+      <param name="offset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+      <param name="format" type="GLenum" />
+      <param name="type" type="GLenum" />
+      <param name="data" type="const GLvoid *" />
+   </function>
+
+   <function name="MapNamedBuffer" offset="assign">
+      <return type="GLvoid *" />
+      <param name="buffer" type="GLuint" />
+      <param name="access" type="GLenum" />
+   </function>
+
+   <function name="MapNamedBufferRange" offset="assign">
+      <return type="GLvoid *" />
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="length" type="GLsizeiptr" />
+      <param name="access" type="GLbitfield" />
+   </function>
+
+   <function name="UnmapNamedBuffer" offset="assign">
+      <return type="GLboolean" />
+      <param name="buffer" type="GLuint" />
+   </function>
+
+   <function name="FlushMappedNamedBufferRange" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="length" type="GLsizeiptr" />
+   </function>
+
+   <function name="GetNamedBufferParameteriv" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="params" type="GLint *" />
+   </function>
+
+   <function name="GetNamedBufferParameteri64v" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="params" type="GLint64 *" />
+   </function>
+
+   <function name="GetNamedBufferPointerv" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="params" type="GLvoid **" />
+   </function>
+
+   <function name="GetNamedBufferSubData" offset="assign">
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+      <param name="data" type="GLvoid *" />
+   </function>
+
    <!-- Texture object functions -->
 
    <function name="CreateTextures" offset="assign">
@@ -21,6 +125,14 @@
       <param name="buffer" type="GLuint" />
    </function>
 
+   <function name="TextureBufferRange" offset="assign">
+      <param name="texture" type="GLuint" />
+      <param name="internalformat" type="GLenum" />
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+   </function>
+
    <function name="TextureStorage1D" offset="assign">
       <param name="texture" type="GLuint" />
       <param name="levels" type="GLsizei" />
diff --git a/mesalib/src/mapi/glapi/gen/gl_enums.py b/mesalib/src/mapi/glapi/gen/gl_enums.py
index a756ad971..7c55ee443 100755
--- a/mesalib/src/mapi/glapi/gen/gl_enums.py
+++ b/mesalib/src/mapi/glapi/gen/gl_enums.py
@@ -157,6 +157,12 @@ _mesa_lookup_prim_by_nr(GLuint nr)
 
         string_offsets = {}
         i = 0;
+        print '#if defined(__GNUC__)'
+        print '# define LONGSTRING __extension__'
+        print '#else'
+        print '# define LONGSTRING'
+        print '#endif'
+        print ''
         print 'LONGSTRING static const char enum_string_table[] = '
         for enum, name in enum_table:
             print '   "%s\\0"' % (name)
diff --git a/mesalib/src/mapi/glapi/glapi.h b/mesalib/src/mapi/glapi/glapi.h
index 89fb1dceb..8084e5757 100755
--- a/mesalib/src/mapi/glapi/glapi.h
+++ b/mesalib/src/mapi/glapi/glapi.h
@@ -76,6 +76,8 @@ struct _glapi_table;
 
 typedef void (*_glapi_proc)(void); /* generic function pointer */
 
+typedef void (*_glapi_nop_handler_proc)(const char *name);
+
 typedef void (*_glapi_warning_func)(void *ctx, const char *str, ...);
 
 
@@ -179,6 +181,14 @@ extern struct _glapi_table *
 _glapi_create_table_from_handle(void *handle, const char *symbol_prefix);
 
 
+void
+_glapi_set_nop_handler(_glapi_nop_handler_proc func);
+
+/** Return pointer to new dispatch table filled with no-op functions */
+struct _glapi_table *
+_glapi_new_nop_table(unsigned num_entries);
+
+
 
 #ifdef __cplusplus
 }
diff --git a/mesalib/src/mapi/glapi/glapi_getproc.c b/mesalib/src/mapi/glapi/glapi_getproc.c
index d52b26450..7f4dc7f4d 100644
--- a/mesalib/src/mapi/glapi/glapi_getproc.c
+++ b/mesalib/src/mapi/glapi/glapi_getproc.c
@@ -30,6 +30,9 @@
  */
 
 
+#include <assert.h>
+#include <string.h>
+#include <stdlib.h>
 #include "glapi/glapi_priv.h"
 #include "glapi/glapitable.h"
 
@@ -37,6 +40,7 @@
 #define FIRST_DYNAMIC_OFFSET (sizeof(struct _glapi_table) / sizeof(void *))
 
 
+
 /**********************************************************************
  * Static function management.
  */
diff --git a/mesalib/src/mapi/glapi/glapi_nop.c b/mesalib/src/mapi/glapi/glapi_nop.c
index cbfbd2433..e006e0259 100644..100755
--- a/mesalib/src/mapi/glapi/glapi_nop.c
+++ b/mesalib/src/mapi/glapi/glapi_nop.c
@@ -30,14 +30,22 @@
  * This file defines a special dispatch table which is loaded with no-op
  * functions.
  *
- * When there's no current rendering context, calling a GL function like
- * glBegin() is a no-op.  Apps should never normally do this.  So as a
- * debugging aid, each of the no-op functions will emit a warning to
- * stderr if the MESA_DEBUG or LIBGL_DEBUG env var is set.
+ * Mesa can register a "no-op handler function" which will be called in
+ * the event that a no-op function is called.
+ *
+ * In the past, the dispatch table was loaded with pointers to a single
+ * no-op function.  But that broke on Windows because the GL entrypoints
+ * use __stdcall convention.  __stdcall means the callee cleans up the
+ * stack.  So one no-op function can't properly clean up the stack.  This
+ * would lead to crashes.
+ *
+ * Another benefit of unique no-op functions is we can accurately report
+ * the function's name in an error message.
  */
 
 
-
+#include <stdlib.h>
+#include <string.h>
 #include "glapi/glapi_priv.h"
 
 #undef _GLAPI_EXPORT
@@ -53,25 +61,32 @@ _glapi_set_warning_func(_glapi_proc func)
 {
 }
 
-/*
- * When GLAPIENTRY is __stdcall (i.e. Windows), the stack is popped by the
- * callee making the number/type of arguments significant.
+
+/**
+ * We'll jump though this function pointer whenever a no-op function
+ * is called.
+ */
+static _glapi_nop_handler_proc nop_handler = NULL;
+
+
+/**
+ * Register the no-op handler call-back function.
  */
-#if defined(_WIN32) || defined(DEBUG)
+_GLAPI_EXPORT void
+_glapi_set_nop_handler(_glapi_nop_handler_proc func)
+{
+   nop_handler = func;
+}
+
 
 /**
  * Called by each of the no-op GL entrypoints.
  */
-static int
-Warn(const char *func)
+static void
+nop(const char *func)
 {
-#if defined(DEBUG)
-   if (getenv("MESA_DEBUG") || getenv("LIBGL_DEBUG")) {
-      fprintf(stderr, "GL User Error: gl%s called without a rendering context\n",
-              func);
-   }
-#endif
-   return 0;
+   if (nop_handler)
+      nop_handler(func);
 }
 
 
@@ -81,7 +96,8 @@ Warn(const char *func)
 static GLint
 NoOpUnused(void)
 {
-   return Warn(" function");
+   nop("unused GL entry point");
+   return 0;
 }
 
 /*
@@ -91,31 +107,28 @@ NoOpUnused(void)
 #define KEYWORD1_ALT static
 #define KEYWORD2 GLAPIENTRY
 #define NAME(func)  NoOp##func
-#define DISPATCH(func, args, msg)  Warn(#func);
-#define RETURN_DISPATCH(func, args, msg)  Warn(#func); return 0
+#define DISPATCH(func, args, msg)  nop(#func);
+#define RETURN_DISPATCH(func, args, msg)  nop(#func); return 0
 
 
 /*
  * Defines for the table of no-op entry points.
  */
 #define TABLE_ENTRY(name) (_glapi_proc) NoOp##name
+#define DISPATCH_TABLE_NAME __glapi_noop_table
+#define UNUSED_TABLE_NAME __unused_noop_functions
+
+#include "glapi/gen/glapitemp.h"
 
-#else
 
-static int
-NoOpGeneric(void)
+/** Return pointer to new dispatch table filled with no-op functions */
+struct _glapi_table * _GLAPI_EXPORT
+_glapi_new_nop_table(unsigned num_entries)
 {
-   if (getenv("MESA_DEBUG") || getenv("LIBGL_DEBUG")) {
-      fprintf(stderr, "GL User Error: calling GL function without a rendering context\n");
+   struct _glapi_table *table = malloc(num_entries * sizeof(_glapi_proc));
+   if (table) {
+      memcpy(table, __glapi_noop_table,
+             num_entries * sizeof(_glapi_proc));
    }
-   return 0;
+   return table;
 }
-
-#define TABLE_ENTRY(name) (_glapi_proc) NoOpGeneric
-
-#endif
-
-#define DISPATCH_TABLE_NAME __glapi_noop_table
-#define UNUSED_TABLE_NAME __unused_noop_functions
-
-#include "glapi/gen/glapitemp.h"
diff --git a/mesalib/src/mapi/glapi/glapi_priv.h b/mesalib/src/mapi/glapi/glapi_priv.h
index 0a998b475..cb4cb5b97 100644
--- a/mesalib/src/mapi/glapi/glapi_priv.h
+++ b/mesalib/src/mapi/glapi/glapi_priv.h
@@ -26,9 +26,6 @@
 #ifndef _GLAPI_PRIV_H
 #define _GLAPI_PRIV_H
 
-#include <stdlib.h>
-#include <string.h>
-#include <stdio.h>
 
 #ifdef HAVE_DIX_CONFIG_H
 #include <dix-config.h>
diff --git a/mesalib/src/mapi/mapi_glapi.c b/mesalib/src/mapi/mapi_glapi.c
index 127dfafac..70605f3df 100644
--- a/mesalib/src/mapi/mapi_glapi.c
+++ b/mesalib/src/mapi/mapi_glapi.c
@@ -27,6 +27,7 @@
  */
 
 #include <string.h>
+#include <stdlib.h>
 #include "glapi/glapi.h"
 #include "u_current.h"
 #include "table.h" /* for MAPI_TABLE_NUM_SLOTS */
@@ -222,6 +223,28 @@ _glapi_get_proc_name(unsigned int offset)
    return stub ? stub_get_name(stub) : NULL;
 }
 
+/** Return pointer to new dispatch table filled with no-op functions */
+struct _glapi_table *
+_glapi_new_nop_table(unsigned num_entries)
+{
+   struct _glapi_table *table;
+
+   if (num_entries > MAPI_TABLE_NUM_SLOTS)
+      num_entries = MAPI_TABLE_NUM_SLOTS;
+
+   table = malloc(num_entries * sizeof(mapi_func));
+   if (table) {
+      memcpy(table, table_noop_array, num_entries * sizeof(mapi_func));
+   }
+   return table;
+}
+
+void
+_glapi_set_nop_handler(_glapi_nop_handler_proc func)
+{
+   table_set_noop_handler(func);
+}
+
 /**
  * This is a deprecated function which should not be used anymore.
  * It's only present to satisfy linking with older versions of libGL.
diff --git a/mesalib/src/mapi/table.c b/mesalib/src/mapi/table.c
index 0d2866618..748750197 100644
--- a/mesalib/src/mapi/table.c
+++ b/mesalib/src/mapi/table.c
@@ -30,16 +30,29 @@
 
 #include "table.h"
 
+static nop_handler_proc nop_handler = NULL;
+
+void
+table_set_noop_handler(nop_handler_proc func)
+{
+   nop_handler = func;
+}
+
 static void
 noop_warn(const char *name)
 {
-   static int debug = -1;
+   if (nop_handler) {
+      nop_handler(name);
+   }
+   else {
+      static int debug = -1;
    
-   if (debug < 0)
-      debug = (getenv("MESA_DEBUG") || getenv("LIBGL_DEBUG"));
+      if (debug < 0)
+         debug = (getenv("MESA_DEBUG") || getenv("LIBGL_DEBUG"));
 
-   if (debug)
-      fprintf(stderr, "%s is no-op\n", name);
+      if (debug)
+         fprintf(stderr, "%s is no-op\n", name);
+   }
 }
 
 static int
diff --git a/mesalib/src/mapi/table.h b/mesalib/src/mapi/table.h
index e2d6ef03d..a1af40c6f 100644
--- a/mesalib/src/mapi/table.h
+++ b/mesalib/src/mapi/table.h
@@ -41,6 +41,14 @@ struct mapi_table;
 
 extern const mapi_func table_noop_array[];
 
+
+typedef void (*nop_handler_proc)(const char *name);
+
+
+void
+table_set_noop_handler(nop_handler_proc func);
+
+
 /**
  * Get the no-op dispatch table.
  */
diff --git a/mesalib/src/mesa/SConscript b/mesalib/src/mesa/SConscript
index 81939f981..a563fd2ff 100644
--- a/mesalib/src/mesa/SConscript
+++ b/mesalib/src/mesa/SConscript
@@ -39,18 +39,13 @@ else:
 # parse Makefile.sources
 source_lists = env.ParseSourceList('Makefile.sources')
 
-env.Append(YACCFLAGS = '-d -p "_mesa_program_"')
-program_lex = env.CFile('program/lex.yy.c', 'program/program_lexer.l')
-program_parse = env.CFile('program/program_parse.tab.c',
-                          'program/program_parse.y')
-program_sources = source_lists['PROGRAM_FILES'] + [
-    program_lex,
-    program_parse[0],
-]
+env.Append(YACCFLAGS = ['-d', '-p', '_mesa_program_'])
+env.CFile('program/lex.yy.c', 'program/program_lexer.l')
+env.CFile('program/program_parse.tab.c', 'program/program_parse.y')
 
 mesa_sources = (
     source_lists['MESA_FILES'] +
-    program_sources +
+    source_lists['PROGRAM_FILES'] +
     source_lists['STATETRACKER_FILES']
 )
 
@@ -135,7 +130,9 @@ def write_git_sha1_h_file(filename):
         (commit, foo) = subprocess.Popen(args, stdout=subprocess.PIPE).communicate()
     except:
         # git log command didn't work
-        if not os.path.exists(filename):
+        dirname = os.path.dirname(filename)
+        if not os.path.exists(dirname):
+            os.makedirs(dirname)
             # create an empty file if none already exists
             f = open(filename, "w")
             f.close()
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index fdc4cf1e9..cf99d9563 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -247,9 +247,9 @@ _mesa_meta_setup_blit_shader(struct gl_context *ctx,
                              struct blit_shader_table *table)
 {
    char *vs_source, *fs_source;
-   void *const mem_ctx = ralloc_context(NULL);
    struct blit_shader *shader = choose_blit_shader(target, table);
    const char *vs_input, *vs_output, *fs_input, *vs_preprocess, *fs_preprocess;
+   void *mem_ctx;
 
    if (ctx->Const.GLSLVersion < 130) {
       vs_preprocess = "";
@@ -273,6 +273,8 @@ _mesa_meta_setup_blit_shader(struct gl_context *ctx,
       return;
    }
 
+   mem_ctx = ralloc_context(NULL);
+
    vs_source = ralloc_asprintf(mem_ctx,
                 "%s\n"
                 "%s vec2 position;\n"
diff --git a/mesalib/src/mesa/drivers/common/meta_tex_subimage.c b/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
index e29addb73..ad6e7873e 100644
--- a/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
+++ b/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
@@ -150,9 +150,6 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
    bool success = false;
    int z;
 
-   /* XXX: This should probably be passed in from somewhere */
-   const char *where = "_mesa_meta_pbo_TexSubImage";
-
    if (!_mesa_is_bufferobj(packing->BufferObj) && !create_pbo)
       return false;
 
@@ -165,19 +162,6 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
    if (ctx->_ImageTransferState)
       return false;
 
-   if (!_mesa_validate_pbo_access(dims, packing, width, height, depth,
-                                  format, type, INT_MAX, pixels)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "%s(out of bounds PBO access)", where);
-      return true;
-   }
-
-   if (_mesa_check_disallowed_mapping(packing->BufferObj)) {
-      /* buffer is mapped - that's an error */
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(PBO is mapped)", where);
-      return true;
-   }
-
    /* For arrays, use a tall (height * depth) 2D texture but taking into
     * account the inter-image padding specified with the image height packing
     * property.
@@ -277,9 +261,6 @@ _mesa_meta_pbo_GetTexSubImage(struct gl_context *ctx, GLuint dims,
    bool success = false;
    int z;
 
-   /* XXX: This should probably be passed in from somewhere */
-   const char *where = "_mesa_meta_pbo_GetTexSubImage";
-
    if (!_mesa_is_bufferobj(packing->BufferObj))
       return false;
 
@@ -292,19 +273,6 @@ _mesa_meta_pbo_GetTexSubImage(struct gl_context *ctx, GLuint dims,
    if (ctx->_ImageTransferState)
       return false;
 
-   if (!_mesa_validate_pbo_access(dims, packing, width, height, depth,
-                                  format, type, INT_MAX, pixels)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "%s(out of bounds PBO access)", where);
-      return true;
-   }
-
-   if (_mesa_check_disallowed_mapping(packing->BufferObj)) {
-      /* buffer is mapped - that's an error */
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(PBO is mapped)", where);
-      return true;
-   }
-
    /* For arrays, use a tall (height * depth) 2D texture but taking into
     * account the inter-image padding specified with the image height packing
     * property.
diff --git a/mesalib/src/mesa/drivers/dri/common/dri_test.c b/mesalib/src/mesa/drivers/dri/common/dri_test.c
index 7ab50d923..57bfa5b93 100644
--- a/mesalib/src/mesa/drivers/dri/common/dri_test.c
+++ b/mesalib/src/mesa/drivers/dri/common/dri_test.c
@@ -76,6 +76,17 @@ _glapi_get_dispatch_table_size(void)
 	return 0;
 }
 
+PUBLIC void
+_glapi_set_nop_handler(_glapi_nop_handler_proc func)
+{
+}
+
+PUBLIC struct _glapi_table *
+_glapi_new_nop_table(unsigned num_entries)
+{
+	return NULL;
+}
+
 #ifndef NO_MAIN
 int main(int argc, char** argv)
 {
diff --git a/mesalib/src/mesa/main/api_exec.h b/mesalib/src/mesa/main/api_exec.h
index 1e4a9d61a..12249fec2 100644
--- a/mesalib/src/mesa/main/api_exec.h
+++ b/mesalib/src/mesa/main/api_exec.h
@@ -30,12 +30,8 @@
 extern "C" {
 #endif
 
-struct _glapi_table;
 struct gl_context;
 
-extern struct _glapi_table *
-_mesa_alloc_dispatch_table(void);
-
 extern void
 _mesa_initialize_exec_table(struct gl_context *ctx);
 
diff --git a/mesalib/src/mesa/main/bufferobj.c b/mesalib/src/mesa/main/bufferobj.c
index 008d1aa5f..25110c33b 100755
--- a/mesalib/src/mesa/main/bufferobj.c
+++ b/mesalib/src/mesa/main/bufferobj.c
@@ -232,67 +232,62 @@ bufferobj_range_mapped(const struct gl_buffer_object *obj,
  * \c glClearBufferSubData.
  *
  * \param ctx     GL context.
- * \param target  Buffer object target on which to operate.
+ * \param bufObj  The buffer object.
  * \param offset  Offset of the first byte of the subdata range.
  * \param size    Size, in bytes, of the subdata range.
  * \param mappedRange  If true, checks if an overlapping range is mapped.
  *                     If false, checks if buffer is mapped.
- * \param errorNoBuffer  Error code if no buffer is bound to target.
  * \param caller  Name of calling function for recording errors.
- * \return   A pointer to the buffer object bound to \c target in the
- *           specified context or \c NULL if any of the parameter or state
- *           conditions are invalid.
+ * \return   false if error, true otherwise
  *
  * \sa glBufferSubDataARB, glGetBufferSubDataARB, glClearBufferSubData
  */
-static struct gl_buffer_object *
-buffer_object_subdata_range_good(struct gl_context * ctx, GLenum target,
-                                 GLintptrARB offset, GLsizeiptrARB size,
-                                 bool mappedRange, GLenum errorNoBuffer,
-                                 const char *caller)
+static bool
+buffer_object_subdata_range_good(struct gl_context *ctx,
+                                 struct gl_buffer_object *bufObj,
+                                 GLintptr offset, GLsizeiptr size,
+                                 bool mappedRange, const char *caller)
 {
-   struct gl_buffer_object *bufObj;
-
    if (size < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE, "%s(size < 0)", caller);
-      return NULL;
+      return false;
    }
 
    if (offset < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE, "%s(offset < 0)", caller);
-      return NULL;
+      return false;
    }
 
-   bufObj = get_buffer(ctx, caller, target, errorNoBuffer);
-   if (!bufObj)
-      return NULL;
-
    if (offset + size > bufObj->Size) {
       _mesa_error(ctx, GL_INVALID_VALUE,
                   "%s(offset %lu + size %lu > buffer size %lu)", caller,
                   (unsigned long) offset,
                   (unsigned long) size,
                   (unsigned long) bufObj->Size);
-      return NULL;
+      return false;
    }
 
    if (bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_PERSISTENT_BIT)
-      return bufObj;
+      return true;
 
    if (mappedRange) {
       if (bufferobj_range_mapped(bufObj, offset, size)) {
-         _mesa_error(ctx, GL_INVALID_OPERATION, "%s", caller);
-         return NULL;
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "%s(range is mapped without persistent bit)",
+                     caller);
+         return false;
       }
    }
    else {
       if (_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
-         _mesa_error(ctx, GL_INVALID_OPERATION, "%s", caller);
-         return NULL;
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "%s(buffer is mapped without persistent bit)",
+                     caller);
+         return false;
       }
    }
 
-   return bufObj;
+   return true;
 }
 
 
@@ -561,9 +556,9 @@ _mesa_total_buffer_object_memory(struct gl_context *ctx)
  * \sa glBufferDataARB, dd_function_table::BufferData.
  */
 GLboolean
-_mesa_buffer_data( struct gl_context *ctx, GLenum target, GLsizeiptrARB size,
-		   const GLvoid * data, GLenum usage, GLenum storageFlags,
-		   struct gl_buffer_object * bufObj )
+buffer_data_fallback(struct gl_context *ctx, GLenum target, GLsizeiptr size,
+                     const GLvoid *data, GLenum usage, GLenum storageFlags,
+                     struct gl_buffer_object *bufObj)
 {
    void * new_data;
 
@@ -607,9 +602,9 @@ _mesa_buffer_data( struct gl_context *ctx, GLenum target, GLsizeiptrARB size,
  * \sa glBufferSubDataARB, dd_function_table::BufferSubData.
  */
 void
-_mesa_buffer_subdata( struct gl_context *ctx, GLintptrARB offset,
-		      GLsizeiptrARB size, const GLvoid * data,
-		      struct gl_buffer_object * bufObj )
+buffer_sub_data_fallback(struct gl_context *ctx, GLintptr offset,
+                         GLsizeiptr size, const GLvoid *data,
+                         struct gl_buffer_object *bufObj)
 {
    (void) ctx;
 
@@ -670,11 +665,11 @@ _mesa_buffer_get_subdata( struct gl_context *ctx, GLintptrARB offset,
  * dd_function_table::ClearBufferSubData.
  */
 void
-_mesa_buffer_clear_subdata(struct gl_context *ctx,
-                           GLintptr offset, GLsizeiptr size,
-                           const GLvoid *clearValue,
-                           GLsizeiptr clearValueSize,
-                           struct gl_buffer_object *bufObj)
+_mesa_ClearBufferSubData_sw(struct gl_context *ctx,
+                            GLintptr offset, GLsizeiptr size,
+                            const GLvoid *clearValue,
+                            GLsizeiptr clearValueSize,
+                            struct gl_buffer_object *bufObj)
 {
    GLsizeiptr i;
    GLubyte *dest;
@@ -711,10 +706,10 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
  * Called via glMapBufferRange().
  */
 static void *
-_mesa_buffer_map_range( struct gl_context *ctx, GLintptr offset,
-                        GLsizeiptr length, GLbitfield access,
-                        struct gl_buffer_object *bufObj,
-                        gl_map_buffer_index index)
+map_buffer_range_fallback(struct gl_context *ctx, GLintptr offset,
+                          GLsizeiptr length, GLbitfield access,
+                          struct gl_buffer_object *bufObj,
+                          gl_map_buffer_index index)
 {
    (void) ctx;
    assert(!_mesa_bufferobj_mapped(bufObj, index));
@@ -732,10 +727,10 @@ _mesa_buffer_map_range( struct gl_context *ctx, GLintptr offset,
  * Called via glFlushMappedBufferRange().
  */
 static void
-_mesa_buffer_flush_mapped_range( struct gl_context *ctx,
-                                 GLintptr offset, GLsizeiptr length,
-                                 struct gl_buffer_object *obj,
-                                 gl_map_buffer_index index)
+flush_mapped_buffer_range_fallback(struct gl_context *ctx,
+                                   GLintptr offset, GLsizeiptr length,
+                                   struct gl_buffer_object *obj,
+                                   gl_map_buffer_index index)
 {
    (void) ctx;
    (void) offset;
@@ -746,15 +741,15 @@ _mesa_buffer_flush_mapped_range( struct gl_context *ctx,
 
 
 /**
- * Default callback for \c dd_function_table::MapBuffer().
+ * Default callback for \c dd_function_table::UnmapBuffer().
  *
  * The input parameters will have been already tested for errors.
  *
  * \sa glUnmapBufferARB, dd_function_table::UnmapBuffer
  */
 GLboolean
-_mesa_buffer_unmap(struct gl_context *ctx, struct gl_buffer_object *bufObj,
-                   gl_map_buffer_index index)
+unmap_buffer_fallback(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                      gl_map_buffer_index index)
 {
    (void) ctx;
    /* XXX we might assert here that bufObj->Pointer is non-null */
@@ -771,11 +766,11 @@ _mesa_buffer_unmap(struct gl_context *ctx, struct gl_buffer_object *bufObj,
  * Called via glCopyBufferSubData().
  */
 static void
-_mesa_copy_buffer_subdata(struct gl_context *ctx,
-                          struct gl_buffer_object *src,
-                          struct gl_buffer_object *dst,
-                          GLintptr readOffset, GLintptr writeOffset,
-                          GLsizeiptr size)
+copy_buffer_sub_data_fallback(struct gl_context *ctx,
+                              struct gl_buffer_object *src,
+                              struct gl_buffer_object *dst,
+                              GLintptr readOffset, GLintptr writeOffset,
+                              GLsizeiptr size)
 {
    GLubyte *srcPtr, *dstPtr;
 
@@ -1006,6 +1001,27 @@ _mesa_lookup_bufferobj_locked(struct gl_context *ctx, GLuint buffer)
       _mesa_HashLookupLocked(ctx->Shared->BufferObjects, buffer);
 }
 
+/**
+ * A convenience function for direct state access functions that throws
+ * GL_INVALID_OPERATION if buffer is not the name of an existing
+ * buffer object.
+ */
+struct gl_buffer_object *
+_mesa_lookup_bufferobj_err(struct gl_context *ctx, GLuint buffer,
+                           const char *caller)
+{
+   struct gl_buffer_object *bufObj;
+
+   bufObj = _mesa_lookup_bufferobj(ctx, buffer);
+   if (!bufObj || bufObj == &DummyBufferObject) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "%s(non-existent buffer object %u)", caller, buffer);
+      return NULL;
+   }
+
+   return bufObj;
+}
+
 
 void
 _mesa_begin_bufferobj_lookups(struct gl_context *ctx)
@@ -1098,20 +1114,20 @@ _mesa_init_buffer_object_functions(struct dd_function_table *driver)
    /* GL_ARB_vertex/pixel_buffer_object */
    driver->NewBufferObject = _mesa_new_buffer_object;
    driver->DeleteBuffer = _mesa_delete_buffer_object;
-   driver->BufferData = _mesa_buffer_data;
-   driver->BufferSubData = _mesa_buffer_subdata;
+   driver->BufferData = buffer_data_fallback;
+   driver->BufferSubData = buffer_sub_data_fallback;
    driver->GetBufferSubData = _mesa_buffer_get_subdata;
-   driver->UnmapBuffer = _mesa_buffer_unmap;
+   driver->UnmapBuffer = unmap_buffer_fallback;
 
    /* GL_ARB_clear_buffer_object */
-   driver->ClearBufferSubData = _mesa_buffer_clear_subdata;
+   driver->ClearBufferSubData = _mesa_ClearBufferSubData_sw;
 
    /* GL_ARB_map_buffer_range */
-   driver->MapBufferRange = _mesa_buffer_map_range;
-   driver->FlushMappedBufferRange = _mesa_buffer_flush_mapped_range;
+   driver->MapBufferRange = map_buffer_range_fallback;
+   driver->FlushMappedBufferRange = flush_mapped_buffer_range_fallback;
 
    /* GL_ARB_copy_buffer */
-   driver->CopyBufferSubData = _mesa_copy_buffer_subdata;
+   driver->CopyBufferSubData = copy_buffer_sub_data_fallback;
 }
 
 
@@ -1273,27 +1289,29 @@ _mesa_DeleteBuffers(GLsizei n, const GLuint *ids)
 
 
 /**
- * Generate a set of unique buffer object IDs and store them in \c buffer.
- * 
- * \param n       Number of IDs to generate.
- * \param buffer  Array of \c n locations to store the IDs.
+ * This is the implementation for glGenBuffers and glCreateBuffers. It is not
+ * exposed to the rest of Mesa to encourage the use of nameless buffers in
+ * driver internals.
  */
-void GLAPIENTRY
-_mesa_GenBuffers(GLsizei n, GLuint *buffer)
+static void
+create_buffers(GLsizei n, GLuint *buffers, bool dsa)
 {
    GET_CURRENT_CONTEXT(ctx);
    GLuint first;
    GLint i;
+   struct gl_buffer_object *buf;
+
+   const char *func = dsa ? "glCreateBuffers" : "glGenBuffers";
 
    if (MESA_VERBOSE & VERBOSE_API)
-      _mesa_debug(ctx, "glGenBuffers(%d)\n", n);
+      _mesa_debug(ctx, "%s(%d)\n", func, n);
 
    if (n < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenBuffersARB");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(n %d < 0)", func, n);
       return;
    }
 
-   if (!buffer) {
+   if (!buffers) {
       return;
    }
 
@@ -1304,16 +1322,53 @@ _mesa_GenBuffers(GLsizei n, GLuint *buffer)
 
    first = _mesa_HashFindFreeKeyBlock(ctx->Shared->BufferObjects, n);
 
-   /* Insert the ID and pointer to dummy buffer object into hash table */
+   /* Insert the ID and pointer into the hash table. If non-DSA, insert a
+    * DummyBufferObject.  Otherwise, create a new buffer object and insert
+    * it.
+    */
    for (i = 0; i < n; i++) {
-      _mesa_HashInsert(ctx->Shared->BufferObjects, first + i,
-                       &DummyBufferObject);
-      buffer[i] = first + i;
+      buffers[i] = first + i;
+      if (dsa) {
+         assert(ctx->Driver.NewBufferObject);
+         buf = ctx->Driver.NewBufferObject(ctx, buffers[i]);
+         if (!buf) {
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
+            return;
+         }
+      }
+      else
+         buf = &DummyBufferObject;
+
+      _mesa_HashInsert(ctx->Shared->BufferObjects, buffers[i], buf);
    }
 
    mtx_unlock(&ctx->Shared->Mutex);
 }
 
+/**
+ * Generate a set of unique buffer object IDs and store them in \c buffers.
+ *
+ * \param n        Number of IDs to generate.
+ * \param buffers  Array of \c n locations to store the IDs.
+ */
+void GLAPIENTRY
+_mesa_GenBuffers(GLsizei n, GLuint *buffers)
+{
+   create_buffers(n, buffers, false);
+}
+
+/**
+ * Create a set of buffer objects and store their unique IDs in \c buffers.
+ *
+ * \param n        Number of IDs to generate.
+ * \param buffers  Array of \c n locations to store the IDs.
+ */
+void GLAPIENTRY
+_mesa_CreateBuffers(GLsizei n, GLuint *buffers)
+{
+   create_buffers(n, buffers, true);
+}
+
 
 /**
  * Determine if ID is the name of a buffer object.
@@ -1337,15 +1392,13 @@ _mesa_IsBuffer(GLuint id)
 }
 
 
-void GLAPIENTRY
-_mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
-                    GLbitfield flags)
+void
+_mesa_buffer_storage(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                     GLenum target, GLsizeiptr size, const GLvoid *data,
+                     GLbitfield flags, const char *func)
 {
-   GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object *bufObj;
-
    if (size <= 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(size <= 0)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(size <= 0)", func);
       return;
    }
 
@@ -1355,27 +1408,25 @@ _mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
                  GL_MAP_COHERENT_BIT |
                  GL_DYNAMIC_STORAGE_BIT |
                  GL_CLIENT_STORAGE_BIT)) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(flags)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(invalid flag bits set)", func);
       return;
    }
 
    if (flags & GL_MAP_PERSISTENT_BIT &&
        !(flags & (GL_MAP_READ_BIT | GL_MAP_WRITE_BIT))) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(flags!=READ/WRITE)");
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "%s(PERSISTENT and flags!=READ/WRITE)", func);
       return;
    }
 
    if (flags & GL_MAP_COHERENT_BIT && !(flags & GL_MAP_PERSISTENT_BIT)) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(flags!=PERSISTENT)");
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "%s(COHERENT and flags!=PERSISTENT)", func);
       return;
    }
 
-   bufObj = get_buffer(ctx, "glBufferStorage", target, GL_INVALID_OPERATION);
-   if (!bufObj)
-      return;
-
    if (bufObj->Immutable) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferStorage(immutable)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(immutable)", func);
       return;
    }
 
@@ -1395,31 +1446,65 @@ _mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
           * glBufferStorage is not described in the spec, Graham Sellers
           * said that it should behave the same as glBufferData.
           */
-         _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferStorage()");
+         _mesa_error(ctx, GL_INVALID_OPERATION, "%s", func);
       }
       else {
-         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBufferStorage()");
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
       }
    }
 }
 
+void GLAPIENTRY
+_mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
+                    GLbitfield flags)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   bufObj = get_buffer(ctx, "glBufferStorage", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   _mesa_buffer_storage(ctx, bufObj, target, size, data, flags,
+                        "glBufferStorage");
+}
 
 void GLAPIENTRY
-_mesa_BufferData(GLenum target, GLsizeiptrARB size,
-                    const GLvoid * data, GLenum usage)
+_mesa_NamedBufferStorage(GLuint buffer, GLsizeiptr size, const GLvoid *data,
+                         GLbitfield flags)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glNamedBufferStorage");
+   if (!bufObj)
+      return;
+
+   /*
+    * In direct state access, buffer objects have an unspecified target since
+    * they are not required to be bound.
+    */
+   _mesa_buffer_storage(ctx, bufObj, GL_NONE, size, data, flags,
+                        "glNamedBufferStorage");
+}
+
+
+void
+_mesa_buffer_data(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                  GLenum target, GLsizeiptr size, const GLvoid *data,
+                  GLenum usage, const char *func)
+{
    bool valid_usage;
 
    if (MESA_VERBOSE & VERBOSE_API)
-      _mesa_debug(ctx, "glBufferData(%s, %ld, %p, %s)\n",
+      _mesa_debug(ctx, "%s(%s, %ld, %p, %s)\n",
+                  func,
                   _mesa_lookup_enum_by_nr(target),
                   (long int) size, data,
                   _mesa_lookup_enum_by_nr(usage));
 
    if (size < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferDataARB(size < 0)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(size < 0)", func);
       return;
    }
 
@@ -1448,16 +1533,13 @@ _mesa_BufferData(GLenum target, GLsizeiptrARB size,
    }
 
    if (!valid_usage) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glBufferData(usage)");
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid usage: %s)", func,
+                  _mesa_lookup_enum_by_nr(usage));
       return;
    }
 
-   bufObj = get_buffer(ctx, "glBufferDataARB", target, GL_INVALID_OPERATION);
-   if (!bufObj)
-      return;
-
    if (bufObj->Immutable) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferData(immutable)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(immutable)", func);
       return;
    }
 
@@ -1490,33 +1572,73 @@ _mesa_BufferData(GLenum target, GLsizeiptrARB size,
           *   EXTERNAL_VIRTUAL_MEMORY_BUFFER_AMD, and the store cannot be
           *   mapped to the GPU address space.
           */
-         _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferData()");
+         _mesa_error(ctx, GL_INVALID_OPERATION, "%s", func);
       }
       else {
-         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBufferData()");
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
       }
    }
 }
 
+void GLAPIENTRY
+_mesa_BufferData(GLenum target, GLsizeiptr size,
+                 const GLvoid *data, GLenum usage)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   bufObj = get_buffer(ctx, "glBufferData", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   _mesa_buffer_data(ctx, bufObj, target, size, data, usage,
+                     "glBufferData");
+}
 
 void GLAPIENTRY
-_mesa_BufferSubData(GLenum target, GLintptrARB offset,
-                       GLsizeiptrARB size, const GLvoid * data)
+_mesa_NamedBufferData(GLuint buffer, GLsizeiptr size, const GLvoid *data,
+                      GLenum usage)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
 
-   bufObj = buffer_object_subdata_range_good( ctx, target, offset, size,
-                                              false, GL_INVALID_OPERATION,
-                                              "glBufferSubDataARB" );
-   if (!bufObj) {
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glNamedBufferData");
+   if (!bufObj)
+      return;
+
+   /* In direct state access, buffer objects have an unspecified target since
+    * they are not required to be bound.
+    */
+   _mesa_buffer_data(ctx, bufObj, GL_NONE, size, data, usage,
+                     "glNamedBufferData");
+}
+
+
+/**
+ * Implementation for glBufferSubData and glNamedBufferSubData.
+ *
+ * \param ctx     GL context.
+ * \param bufObj  The buffer object.
+ * \param offset  Offset of the first byte of the subdata range.
+ * \param size    Size, in bytes, of the subdata range.
+ * \param data    The data store.
+ * \param func  Name of calling function for recording errors.
+ *
+ */
+void
+_mesa_buffer_sub_data(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                      GLintptr offset, GLsizeiptr size, const GLvoid *data,
+                      const char *func)
+{
+   if (!buffer_object_subdata_range_good(ctx, bufObj, offset, size,
+                                         false, func)) {
       /* error already recorded */
       return;
    }
 
    if (bufObj->Immutable &&
        !(bufObj->StorageFlags & GL_DYNAMIC_STORAGE_BIT)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferSubData");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s", func);
       return;
    }
 
@@ -1526,105 +1648,107 @@ _mesa_BufferSubData(GLenum target, GLintptrARB offset,
    bufObj->Written = GL_TRUE;
 
    assert(ctx->Driver.BufferSubData);
-   ctx->Driver.BufferSubData( ctx, offset, size, data, bufObj );
+   ctx->Driver.BufferSubData(ctx, offset, size, data, bufObj);
 }
 
+void GLAPIENTRY
+_mesa_BufferSubData(GLenum target, GLintptr offset,
+                    GLsizeiptr size, const GLvoid *data)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   bufObj = get_buffer(ctx, "glBufferSubData", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   _mesa_buffer_sub_data(ctx, bufObj, offset, size, data, "glBufferSubData");
+}
 
 void GLAPIENTRY
-_mesa_GetBufferSubData(GLenum target, GLintptrARB offset,
-                          GLsizeiptrARB size, void * data)
+_mesa_NamedBufferSubData(GLuint buffer, GLintptr offset,
+                         GLsizeiptr size, const GLvoid *data)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
 
-   bufObj = buffer_object_subdata_range_good(ctx, target, offset, size,
-                                             false, GL_INVALID_OPERATION,
-                                             "glGetBufferSubDataARB");
-   if (!bufObj) {
-      /* error already recorded */
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glNamedBufferSubData");
+   if (!bufObj)
       return;
-   }
 
-   assert(ctx->Driver.GetBufferSubData);
-   ctx->Driver.GetBufferSubData( ctx, offset, size, data, bufObj );
+   _mesa_buffer_sub_data(ctx, bufObj, offset, size, data,
+                         "glNamedBufferSubData");
 }
 
 
 void GLAPIENTRY
-_mesa_ClearBufferData(GLenum target, GLenum internalformat, GLenum format,
-                      GLenum type, const GLvoid* data)
+_mesa_GetBufferSubData(GLenum target, GLintptr offset,
+                       GLsizeiptr size, GLvoid *data)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object* bufObj;
-   mesa_format mesaFormat;
-   GLubyte clearValue[MAX_PIXEL_BYTES];
-   GLsizeiptr clearValueSize;
+   struct gl_buffer_object *bufObj;
 
-   bufObj = get_buffer(ctx, "glClearBufferData", target, GL_INVALID_VALUE);
-   if (!bufObj) {
+   bufObj = get_buffer(ctx, "glGetBufferSubData", target,
+                       GL_INVALID_OPERATION);
+   if (!bufObj)
       return;
-   }
 
-   if (_mesa_check_disallowed_mapping(bufObj)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glClearBufferData(buffer currently mapped)");
+   if (!buffer_object_subdata_range_good(ctx, bufObj, offset, size, false,
+                                         "glGetBufferSubData")) {
       return;
    }
 
-   mesaFormat = validate_clear_buffer_format(ctx, internalformat,
-                                             format, type,
-                                             "glClearBufferData");
-   if (mesaFormat == MESA_FORMAT_NONE) {
-      return;
-   }
+   assert(ctx->Driver.GetBufferSubData);
+   ctx->Driver.GetBufferSubData(ctx, offset, size, data, bufObj);
+}
 
-   clearValueSize = _mesa_get_format_bytes(mesaFormat);
-   if (bufObj->Size % clearValueSize != 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glClearBufferData(size is not a multiple of "
-                  "internalformat size)");
-      return;
-   }
+void GLAPIENTRY
+_mesa_GetNamedBufferSubData(GLuint buffer, GLintptr offset,
+                            GLsizeiptr size, GLvoid *data)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
 
-   if (data == NULL) {
-      /* clear to zeros, per the spec */
-      ctx->Driver.ClearBufferSubData(ctx, 0, bufObj->Size,
-                                     NULL, clearValueSize, bufObj);
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glGetNamedBufferSubData");
+   if (!bufObj)
       return;
-   }
 
-   if (!convert_clear_buffer_data(ctx, mesaFormat, clearValue,
-                                  format, type, data, "glClearBufferData")) {
+   if (!buffer_object_subdata_range_good(ctx, bufObj, offset, size, false,
+                                         "glGetNamedBufferSubData")) {
       return;
    }
 
-   ctx->Driver.ClearBufferSubData(ctx, 0, bufObj->Size,
-                                  clearValue, clearValueSize, bufObj);
+   assert(ctx->Driver.GetBufferSubData);
+   ctx->Driver.GetBufferSubData(ctx, offset, size, data, bufObj);
 }
 
 
-void GLAPIENTRY
-_mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
-                         GLintptr offset, GLsizeiptr size,
-                         GLenum format, GLenum type,
-                         const GLvoid* data)
+/**
+ * \param subdata   true if caller is *SubData, false if *Data
+ */
+void
+_mesa_clear_buffer_sub_data(struct gl_context *ctx,
+                            struct gl_buffer_object *bufObj,
+                            GLenum internalformat,
+                            GLintptr offset, GLsizeiptr size,
+                            GLenum format, GLenum type,
+                            const GLvoid *data,
+                            const char *func, bool subdata)
 {
-   GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object* bufObj;
    mesa_format mesaFormat;
    GLubyte clearValue[MAX_PIXEL_BYTES];
    GLsizeiptr clearValueSize;
 
-   bufObj = buffer_object_subdata_range_good(ctx, target, offset, size,
-                                             true, GL_INVALID_VALUE,
-                                             "glClearBufferSubData");
-   if (!bufObj) {
+   /* This checks for disallowed mappings. */
+   if (!buffer_object_subdata_range_good(ctx, bufObj, offset, size,
+                                         subdata, func)) {
       return;
    }
 
    mesaFormat = validate_clear_buffer_format(ctx, internalformat,
-                                             format, type,
-                                             "glClearBufferSubData");
+                                             format, type, func);
+
    if (mesaFormat == MESA_FORMAT_NONE) {
       return;
    }
@@ -1632,8 +1756,8 @@ _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
    clearValueSize = _mesa_get_format_bytes(mesaFormat);
    if (offset % clearValueSize != 0 || size % clearValueSize != 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glClearBufferSubData(offset or size is not a multiple of "
-                  "internalformat size)");
+                  "%s(offset or size is not a multiple of "
+                  "internalformat size)", func);
       return;
    }
 
@@ -1647,8 +1771,7 @@ _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
    }
 
    if (!convert_clear_buffer_data(ctx, mesaFormat, clearValue,
-                                  format, type, data,
-                                  "glClearBufferSubData")) {
+                                  format, type, data, func)) {
       return;
    }
 
@@ -1658,131 +1781,87 @@ _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
    }
 }
 
-
-void * GLAPIENTRY
-_mesa_MapBuffer(GLenum target, GLenum access)
+void GLAPIENTRY
+_mesa_ClearBufferData(GLenum target, GLenum internalformat, GLenum format,
+                      GLenum type, const GLvoid *data)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object * bufObj;
-   GLbitfield accessFlags;
-   void *map;
-   bool valid_access;
+   struct gl_buffer_object *bufObj;
 
-   ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, NULL);
+   bufObj = get_buffer(ctx, "glClearBufferData", target, GL_INVALID_VALUE);
+   if (!bufObj)
+      return;
 
-   switch (access) {
-   case GL_READ_ONLY_ARB:
-      accessFlags = GL_MAP_READ_BIT;
-      valid_access = _mesa_is_desktop_gl(ctx);
-      break;
-   case GL_WRITE_ONLY_ARB:
-      accessFlags = GL_MAP_WRITE_BIT;
-      valid_access = true;
-      break;
-   case GL_READ_WRITE_ARB:
-      accessFlags = GL_MAP_READ_BIT | GL_MAP_WRITE_BIT;
-      valid_access = _mesa_is_desktop_gl(ctx);
-      break;
-   default:
-      valid_access = false;
-      break;
-   }
+   _mesa_clear_buffer_sub_data(ctx, bufObj, internalformat, 0, bufObj->Size,
+                               format, type, data,
+                               "glClearBufferData", false);
+}
 
-   if (!valid_access) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glMapBufferARB(access)");
-      return NULL;
-   }
+void GLAPIENTRY
+_mesa_ClearNamedBufferData(GLuint buffer, GLenum internalformat,
+                           GLenum format, GLenum type, const GLvoid *data)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
 
-   bufObj = get_buffer(ctx, "glMapBufferARB", target, GL_INVALID_OPERATION);
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glClearNamedBufferData");
    if (!bufObj)
-      return NULL;
+      return;
 
-   if (accessFlags & GL_MAP_READ_BIT &&
-       !(bufObj->StorageFlags & GL_MAP_READ_BIT)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBuffer(invalid read flag)");
-      return NULL;
-   }
+   _mesa_clear_buffer_sub_data(ctx, bufObj, internalformat, 0, bufObj->Size,
+                               format, type, data,
+                               "glClearNamedBufferData", false);
+}
 
-   if (accessFlags & GL_MAP_WRITE_BIT &&
-       !(bufObj->StorageFlags & GL_MAP_WRITE_BIT)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBuffer(invalid write flag)");
-      return NULL;
-   }
 
-   if (_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glMapBufferARB(already mapped)");
-      return NULL;
-   }
-
-   if (!bufObj->Size) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY,
-                  "glMapBuffer(buffer size = 0)");
-      return NULL;
-   }
+void GLAPIENTRY
+_mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
+                         GLintptr offset, GLsizeiptr size,
+                         GLenum format, GLenum type,
+                         const GLvoid *data)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
 
-   assert(ctx->Driver.MapBufferRange);
-   map = ctx->Driver.MapBufferRange(ctx, 0, bufObj->Size, accessFlags, bufObj,
-                                    MAP_USER);
-   if (!map) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glMapBufferARB(map failed)");
-      return NULL;
-   }
-   else {
-      /* The driver callback should have set these fields.
-       * This is important because other modules (like VBO) might call
-       * the driver function directly.
-       */
-      assert(bufObj->Mappings[MAP_USER].Pointer == map);
-      assert(bufObj->Mappings[MAP_USER].Length == bufObj->Size);
-      assert(bufObj->Mappings[MAP_USER].Offset == 0);
-      bufObj->Mappings[MAP_USER].AccessFlags = accessFlags;
-   }
+   bufObj = get_buffer(ctx, "glClearBufferSubData", target, GL_INVALID_VALUE);
+   if (!bufObj)
+      return;
 
-   if (access == GL_WRITE_ONLY_ARB || access == GL_READ_WRITE_ARB)
-      bufObj->Written = GL_TRUE;
+   _mesa_clear_buffer_sub_data(ctx, bufObj, internalformat, offset, size,
+                               format, type, data,
+                               "glClearBufferSubData", true);
+}
 
-#ifdef VBO_DEBUG
-   printf("glMapBufferARB(%u, sz %ld, access 0x%x)\n",
-	  bufObj->Name, bufObj->Size, access);
-   if (access == GL_WRITE_ONLY_ARB) {
-      GLuint i;
-      GLubyte *b = (GLubyte *) bufObj->Pointer;
-      for (i = 0; i < bufObj->Size; i++)
-         b[i] = i & 0xff;
-   }
-#endif
+void GLAPIENTRY
+_mesa_ClearNamedBufferSubData(GLuint buffer, GLenum internalformat,
+                              GLintptr offset, GLsizeiptr size,
+                              GLenum format, GLenum type,
+                              const GLvoid *data)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
 
-#ifdef BOUNDS_CHECK
-   {
-      GLubyte *buf = (GLubyte *) bufObj->Pointer;
-      GLuint i;
-      /* buffer is 100 bytes larger than requested, fill with magic value */
-      for (i = 0; i < 100; i++) {
-         buf[bufObj->Size - i - 1] = 123;
-      }
-   }
-#endif
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glClearNamedBufferSubData");
+   if (!bufObj)
+      return;
 
-   return bufObj->Mappings[MAP_USER].Pointer;
+   _mesa_clear_buffer_sub_data(ctx, bufObj, internalformat, offset, size,
+                               format, type, data,
+                               "glClearNamedBufferSubData", true);
 }
 
 
-GLboolean GLAPIENTRY
-_mesa_UnmapBuffer(GLenum target)
+GLboolean
+_mesa_unmap_buffer(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                   const char *func)
 {
-   GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object *bufObj;
    GLboolean status = GL_TRUE;
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
 
-   bufObj = get_buffer(ctx, "glUnmapBufferARB", target, GL_INVALID_OPERATION);
-   if (!bufObj)
-      return GL_FALSE;
-
    if (!_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glUnmapBufferARB");
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "%s(buffer is not mapped)", func);
       return GL_FALSE;
    }
 
@@ -1831,129 +1910,164 @@ _mesa_UnmapBuffer(GLenum target)
    return status;
 }
 
+GLboolean GLAPIENTRY
+_mesa_UnmapBuffer(GLenum target)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
 
-void GLAPIENTRY
-_mesa_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params)
+   bufObj = get_buffer(ctx, "glUnmapBuffer", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return GL_FALSE;
+
+   return _mesa_unmap_buffer(ctx, bufObj, "glUnmapBuffer");
+}
+
+GLboolean GLAPIENTRY
+_mesa_UnmapNamedBuffer(GLuint buffer)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
 
-   bufObj = get_buffer(ctx, "glGetBufferParameterivARB", target,
-                       GL_INVALID_OPERATION);
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glUnmapNamedBuffer");
    if (!bufObj)
-      return;
+      return GL_FALSE;
 
+   return _mesa_unmap_buffer(ctx, bufObj, "glUnmapNamedBuffer");
+}
+
+
+static bool
+get_buffer_parameter(struct gl_context *ctx,
+                     struct gl_buffer_object *bufObj, GLenum pname,
+                     GLint64 *params, const char *func)
+{
    switch (pname) {
    case GL_BUFFER_SIZE_ARB:
-      *params = (GLint) bufObj->Size;
-      return;
+      *params = bufObj->Size;
+      break;
    case GL_BUFFER_USAGE_ARB:
       *params = bufObj->Usage;
-      return;
+      break;
    case GL_BUFFER_ACCESS_ARB:
       *params = simplified_access_mode(ctx,
                             bufObj->Mappings[MAP_USER].AccessFlags);
-      return;
+      break;
    case GL_BUFFER_MAPPED_ARB:
       *params = _mesa_bufferobj_mapped(bufObj, MAP_USER);
-      return;
+      break;
    case GL_BUFFER_ACCESS_FLAGS:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
       *params = bufObj->Mappings[MAP_USER].AccessFlags;
-      return;
+      break;
    case GL_BUFFER_MAP_OFFSET:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = (GLint) bufObj->Mappings[MAP_USER].Offset;
-      return;
+      *params = bufObj->Mappings[MAP_USER].Offset;
+      break;
    case GL_BUFFER_MAP_LENGTH:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = (GLint) bufObj->Mappings[MAP_USER].Length;
-      return;
+      *params = bufObj->Mappings[MAP_USER].Length;
+      break;
    case GL_BUFFER_IMMUTABLE_STORAGE:
       if (!ctx->Extensions.ARB_buffer_storage)
          goto invalid_pname;
       *params = bufObj->Immutable;
-      return;
+      break;
    case GL_BUFFER_STORAGE_FLAGS:
       if (!ctx->Extensions.ARB_buffer_storage)
          goto invalid_pname;
       *params = bufObj->StorageFlags;
-      return;
+      break;
    default:
-      ; /* fall-through */
+      goto invalid_pname;
    }
 
+   return true;
+
 invalid_pname:
-   _mesa_error(ctx, GL_INVALID_ENUM, "glGetBufferParameterivARB(pname=%s)",
+   _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid pname: %s)", func,
                _mesa_lookup_enum_by_nr(pname));
+   return false;
 }
 
+void GLAPIENTRY
+_mesa_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+   GLint64 parameter;
+
+   bufObj = get_buffer(ctx, "glGetBufferParameteriv", target,
+                       GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   if (!get_buffer_parameter(ctx, bufObj, pname, &parameter,
+                             "glGetBufferParameteriv"))
+      return; /* Error already recorded. */
+
+   *params = (GLint) parameter;
+}
 
-/**
- * New in GL 3.2
- * This is pretty much a duplicate of GetBufferParameteriv() but the
- * GL_BUFFER_SIZE_ARB attribute will be 64-bits on a 64-bit system.
- */
 void GLAPIENTRY
 _mesa_GetBufferParameteri64v(GLenum target, GLenum pname, GLint64 *params)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
+   GLint64 parameter;
 
    bufObj = get_buffer(ctx, "glGetBufferParameteri64v", target,
                        GL_INVALID_OPERATION);
    if (!bufObj)
       return;
 
-   switch (pname) {
-   case GL_BUFFER_SIZE_ARB:
-      *params = bufObj->Size;
-      return;
-   case GL_BUFFER_USAGE_ARB:
-      *params = bufObj->Usage;
-      return;
-   case GL_BUFFER_ACCESS_ARB:
-      *params = simplified_access_mode(ctx,
-                             bufObj->Mappings[MAP_USER].AccessFlags);
-      return;
-   case GL_BUFFER_ACCESS_FLAGS:
-      if (!ctx->Extensions.ARB_map_buffer_range)
-         goto invalid_pname;
-      *params = bufObj->Mappings[MAP_USER].AccessFlags;
-      return;
-   case GL_BUFFER_MAPPED_ARB:
-      *params = _mesa_bufferobj_mapped(bufObj, MAP_USER);
-      return;
-   case GL_BUFFER_MAP_OFFSET:
-      if (!ctx->Extensions.ARB_map_buffer_range)
-         goto invalid_pname;
-      *params = bufObj->Mappings[MAP_USER].Offset;
-      return;
-   case GL_BUFFER_MAP_LENGTH:
-      if (!ctx->Extensions.ARB_map_buffer_range)
-         goto invalid_pname;
-      *params = bufObj->Mappings[MAP_USER].Length;
-      return;
-   case GL_BUFFER_IMMUTABLE_STORAGE:
-      if (!ctx->Extensions.ARB_buffer_storage)
-         goto invalid_pname;
-      *params = bufObj->Immutable;
+   if (!get_buffer_parameter(ctx, bufObj, pname, &parameter,
+                             "glGetBufferParameteri64v"))
+      return; /* Error already recorded. */
+
+   *params = parameter;
+}
+
+void GLAPIENTRY
+_mesa_GetNamedBufferParameteriv(GLuint buffer, GLenum pname, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+   GLint64 parameter;
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glGetNamedBufferParameteriv");
+   if (!bufObj)
       return;
-   case GL_BUFFER_STORAGE_FLAGS:
-      if (!ctx->Extensions.ARB_buffer_storage)
-         goto invalid_pname;
-      *params = bufObj->StorageFlags;
+
+   if (!get_buffer_parameter(ctx, bufObj, pname, &parameter,
+                             "glGetNamedBufferParameteriv"))
+      return; /* Error already recorded. */
+
+   *params = (GLint) parameter;
+}
+
+void GLAPIENTRY
+_mesa_GetNamedBufferParameteri64v(GLuint buffer, GLenum pname,
+                                  GLint64 *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+   GLint64 parameter;
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glGetNamedBufferParameteri64v");
+   if (!bufObj)
       return;
-   default:
-      ; /* fall-through */
-   }
 
-invalid_pname:
-   _mesa_error(ctx, GL_INVALID_ENUM, "glGetBufferParameteri64v(pname=%s)",
-               _mesa_lookup_enum_by_nr(pname));
+   if (!get_buffer_parameter(ctx, bufObj, pname, &parameter,
+                             "glGetNamedBufferParameteri64v"))
+      return; /* Error already recorded. */
+
+   *params = parameter;
 }
 
 
@@ -1961,14 +2075,15 @@ void GLAPIENTRY
 _mesa_GetBufferPointerv(GLenum target, GLenum pname, GLvoid **params)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object * bufObj;
+   struct gl_buffer_object *bufObj;
 
-   if (pname != GL_BUFFER_MAP_POINTER_ARB) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glGetBufferPointervARB(pname)");
+   if (pname != GL_BUFFER_MAP_POINTER) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetBufferPointerv(pname != "
+                  "GL_BUFFER_MAP_POINTER)");
       return;
    }
 
-   bufObj = get_buffer(ctx, "glGetBufferPointervARB", target,
+   bufObj = get_buffer(ctx, "glGetBufferPointerv", target,
                        GL_INVALID_OPERATION);
    if (!bufObj)
       return;
@@ -1976,66 +2091,75 @@ _mesa_GetBufferPointerv(GLenum target, GLenum pname, GLvoid **params)
    *params = bufObj->Mappings[MAP_USER].Pointer;
 }
 
-
 void GLAPIENTRY
-_mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
-                        GLintptr readOffset, GLintptr writeOffset,
-                        GLsizeiptr size)
+_mesa_GetNamedBufferPointerv(GLuint buffer, GLenum pname, GLvoid **params)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object *src, *dst;
+   struct gl_buffer_object *bufObj;
 
-   src = get_buffer(ctx, "glCopyBufferSubData", readTarget,
-                    GL_INVALID_OPERATION);
-   if (!src)
+   if (pname != GL_BUFFER_MAP_POINTER) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetNamedBufferPointerv(pname != "
+                  "GL_BUFFER_MAP_POINTER)");
       return;
+   }
 
-   dst = get_buffer(ctx, "glCopyBufferSubData", writeTarget,
-                    GL_INVALID_OPERATION);
-   if (!dst)
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glGetNamedBufferPointerv");
+   if (!bufObj)
       return;
 
+   *params = bufObj->Mappings[MAP_USER].Pointer;
+}
+
+
+void
+_mesa_copy_buffer_sub_data(struct gl_context *ctx,
+                           struct gl_buffer_object *src,
+                           struct gl_buffer_object *dst,
+                           GLintptr readOffset, GLintptr writeOffset,
+                           GLsizeiptr size, const char *func)
+{
    if (_mesa_check_disallowed_mapping(src)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCopyBufferSubData(readBuffer is mapped)");
+                  "%s(readBuffer is mapped)", func);
       return;
    }
 
    if (_mesa_check_disallowed_mapping(dst)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCopyBufferSubData(writeBuffer is mapped)");
+                  "%s(writeBuffer is mapped)", func);
       return;
    }
 
    if (readOffset < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyBufferSubData(readOffset = %d)", (int) readOffset);
+                  "%s(readOffset %d < 0)", func, (int) readOffset);
       return;
    }
 
    if (writeOffset < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyBufferSubData(writeOffset = %d)", (int) writeOffset);
+                  "%s(writeOffset %d < 0)", func, (int) writeOffset);
       return;
    }
 
    if (size < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyBufferSubData(writeOffset = %d)", (int) size);
+                  "%s(size %d < 0)", func, (int) size);
       return;
    }
 
    if (readOffset + size > src->Size) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyBufferSubData(readOffset + size = %d)",
-                  (int) (readOffset + size));
+                  "%s(readOffset %d + size %d > src_buffer_size %d)", func,
+                  (int) readOffset, (int) size, (int) src->Size);
       return;
    }
 
    if (writeOffset + size > dst->Size) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyBufferSubData(writeOffset + size = %d)",
-                  (int) (writeOffset + size));
+                  "%s(writeOffset %d + size %d > dst_buffer_size %d)", func,
+                  (int) writeOffset, (int) size, (int) dst->Size);
       return;
    }
 
@@ -2049,7 +2173,7 @@ _mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
       else {
          /* overlapping src/dst is illegal */
          _mesa_error(ctx, GL_INVALID_VALUE,
-                     "glCopyBufferSubData(overlapping src/dst)");
+                     "%s(overlapping src/dst)", func);
          return;
       }
    }
@@ -2057,36 +2181,71 @@ _mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
    ctx->Driver.CopyBufferSubData(ctx, src, dst, readOffset, writeOffset, size);
 }
 
+void GLAPIENTRY
+_mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
+                        GLintptr readOffset, GLintptr writeOffset,
+                        GLsizeiptr size)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *src, *dst;
+
+   src = get_buffer(ctx, "glCopyBufferSubData", readTarget,
+                    GL_INVALID_OPERATION);
+   if (!src)
+      return;
 
-/**
- * See GL_ARB_map_buffer_range spec
- */
-void * GLAPIENTRY
-_mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
-                     GLbitfield access)
+   dst = get_buffer(ctx, "glCopyBufferSubData", writeTarget,
+                    GL_INVALID_OPERATION);
+   if (!dst)
+      return;
+
+   _mesa_copy_buffer_sub_data(ctx, src, dst, readOffset, writeOffset, size,
+                              "glCopyBufferSubData");
+}
+
+void GLAPIENTRY
+_mesa_CopyNamedBufferSubData(GLuint readBuffer, GLuint writeBuffer,
+                             GLintptr readOffset, GLintptr writeOffset,
+                             GLsizeiptr size)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_buffer_object *bufObj;
+   struct gl_buffer_object *src, *dst;
+
+   src = _mesa_lookup_bufferobj_err(ctx, readBuffer,
+                                    "glCopyNamedBufferSubData");
+   if (!src)
+      return;
+
+   dst = _mesa_lookup_bufferobj_err(ctx, writeBuffer,
+                                    "glCopyNamedBufferSubData");
+   if (!dst)
+      return;
+
+   _mesa_copy_buffer_sub_data(ctx, src, dst, readOffset, writeOffset, size,
+                              "glCopyNamedBufferSubData");
+}
+
+
+void *
+_mesa_map_buffer_range(struct gl_context *ctx,
+                       struct gl_buffer_object *bufObj,
+                       GLintptr offset, GLsizeiptr length,
+                       GLbitfield access, const char *func)
+{
    void *map;
    GLbitfield allowed_access;
 
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, NULL);
 
-   if (!ctx->Extensions.ARB_map_buffer_range) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(extension not supported)");
-      return NULL;
-   }
-
    if (offset < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glMapBufferRange(offset = %ld)", (long)offset);
+                  "%s(offset %ld < 0)", func, (long) offset);
       return NULL;
    }
 
    if (length < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glMapBufferRange(length = %ld)", (long)length);
+                  "%s(length %ld < 0)", func, (long) length);
       return NULL;
    }
 
@@ -2096,10 +2255,13 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
     *     conditions:
     *
     *     * <length> is zero."
+    *
+    * Additionally, page 94 of the PDF of the OpenGL 4.5 core spec
+    * (30.10.2014) also says this, so it's no longer allowed for desktop GL,
+    * either.
     */
-   if (_mesa_is_gles(ctx) && length == 0) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(length = 0)");
+   if (length == 0) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(length = 0)", func);
       return NULL;
    }
 
@@ -2116,14 +2278,15 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
    }
 
    if (access & ~allowed_access) {
-      /* generate an error if any other than allowed bit is set */
-      _mesa_error(ctx, GL_INVALID_VALUE, "glMapBufferRange(access)");
+      /* generate an error if any bits other than those allowed are set */
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "%s(access has undefined bits set)", func);
       return NULL;
    }
 
    if ((access & (GL_MAP_READ_BIT | GL_MAP_WRITE_BIT)) == 0) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(access indicates neither read or write)");
+                  "%s(access indicates neither read or write)", func);
       return NULL;
    }
 
@@ -2132,82 +2295,69 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
                   GL_MAP_INVALIDATE_BUFFER_BIT |
                   GL_MAP_UNSYNCHRONIZED_BIT))) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid access flags)");
+                  "%s(read access with disallowed bits)", func);
       return NULL;
    }
 
    if ((access & GL_MAP_FLUSH_EXPLICIT_BIT) &&
        ((access & GL_MAP_WRITE_BIT) == 0)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid access flags)");
+                  "%s(access has flush explicit without write)", func);
       return NULL;
    }
 
-   bufObj = get_buffer(ctx, "glMapBufferRange", target, GL_INVALID_OPERATION);
-   if (!bufObj)
-      return NULL;
-
    if (access & GL_MAP_READ_BIT &&
        !(bufObj->StorageFlags & GL_MAP_READ_BIT)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid read flag)");
+                  "%s(buffer does not allow read access)", func);
       return NULL;
    }
 
    if (access & GL_MAP_WRITE_BIT &&
        !(bufObj->StorageFlags & GL_MAP_WRITE_BIT)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid write flag)");
+                  "%s(buffer does not allow write access)", func);
       return NULL;
    }
 
    if (access & GL_MAP_COHERENT_BIT &&
        !(bufObj->StorageFlags & GL_MAP_COHERENT_BIT)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid coherent flag)");
+                  "%s(buffer does not allow coherent access)", func);
       return NULL;
    }
 
    if (access & GL_MAP_PERSISTENT_BIT &&
        !(bufObj->StorageFlags & GL_MAP_PERSISTENT_BIT)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(invalid persistent flag)");
+                  "%s(buffer does not allow persistent access)", func);
       return NULL;
    }
 
    if (offset + length > bufObj->Size) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glMapBufferRange(offset + length > size)");
+                  "%s(offset %ld + length %ld > buffer_size %ld)", func,
+                  offset, length, bufObj->Size);
       return NULL;
    }
 
    if (_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glMapBufferRange(buffer already mapped)");
+                  "%s(buffer already mapped)", func);
       return NULL;
    }
 
    if (!bufObj->Size) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY,
-                  "glMapBufferRange(buffer size = 0)");
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s(buffer size = 0)", func);
       return NULL;
    }
 
-   /* Mapping zero bytes should return a non-null pointer. */
-   if (!length) {
-      static long dummy = 0;
-      bufObj->Mappings[MAP_USER].Pointer = &dummy;
-      bufObj->Mappings[MAP_USER].Length = length;
-      bufObj->Mappings[MAP_USER].Offset = offset;
-      bufObj->Mappings[MAP_USER].AccessFlags = access;
-      return bufObj->Mappings[MAP_USER].Pointer;
-   }
 
    assert(ctx->Driver.MapBufferRange);
    map = ctx->Driver.MapBufferRange(ctx, offset, length, access, bufObj,
                                     MAP_USER);
    if (!map) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glMapBufferARB(map failed)");
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s(map failed)", func);
    }
    else {
       /* The driver callback should have set all these fields.
@@ -2220,61 +2370,189 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
       assert(bufObj->Mappings[MAP_USER].AccessFlags == access);
    }
 
+   if (access & GL_MAP_WRITE_BIT)
+      bufObj->Written = GL_TRUE;
+
+#ifdef VBO_DEBUG
+   if (strstr(func, "Range") == NULL) { /* If not MapRange */
+      printf("glMapBuffer(%u, sz %ld, access 0x%x)\n",
+            bufObj->Name, bufObj->Size, access);
+      /* Access must be write only */
+      if ((access & GL_MAP_WRITE_BIT) && (!(access & ~GL_MAP_WRITE_BIT))) {
+         GLuint i;
+         GLubyte *b = (GLubyte *) bufObj->Pointer;
+         for (i = 0; i < bufObj->Size; i++)
+            b[i] = i & 0xff;
+      }
+   }
+#endif
+
+#ifdef BOUNDS_CHECK
+   if (strstr(func, "Range") == NULL) { /* If not MapRange */
+      GLubyte *buf = (GLubyte *) bufObj->Pointer;
+      GLuint i;
+      /* buffer is 100 bytes larger than requested, fill with magic value */
+      for (i = 0; i < 100; i++) {
+         buf[bufObj->Size - i - 1] = 123;
+      }
+   }
+#endif
+
    return map;
 }
 
+void * GLAPIENTRY
+_mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
+                     GLbitfield access)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   if (!ctx->Extensions.ARB_map_buffer_range) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapBufferRange(ARB_map_buffer_range not supported)");
+      return NULL;
+   }
+
+   bufObj = get_buffer(ctx, "glMapBufferRange", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return NULL;
+
+   return _mesa_map_buffer_range(ctx, bufObj, offset, length, access,
+                                 "glMapBufferRange");
+}
+
+void * GLAPIENTRY
+_mesa_MapNamedBufferRange(GLuint buffer, GLintptr offset, GLsizeiptr length,
+                          GLbitfield access)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   if (!ctx->Extensions.ARB_map_buffer_range) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapNamedBufferRange("
+                  "ARB_map_buffer_range not supported)");
+      return NULL;
+   }
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glMapNamedBufferRange");
+   if (!bufObj)
+      return NULL;
+
+   return _mesa_map_buffer_range(ctx, bufObj, offset, length, access,
+                                 "glMapNamedBufferRange");
+}
 
 /**
- * See GL_ARB_map_buffer_range spec
+ * Converts GLenum access from MapBuffer and MapNamedBuffer into
+ * flags for input to _mesa_map_buffer_range.
+ *
+ * \return true if the type of requested access is permissible.
  */
-void GLAPIENTRY
-_mesa_FlushMappedBufferRange(GLenum target, GLintptr offset, GLsizeiptr length)
+static bool
+get_map_buffer_access_flags(struct gl_context *ctx, GLenum access,
+                            GLbitfield *flags)
+{
+   switch (access) {
+   case GL_READ_ONLY_ARB:
+      *flags = GL_MAP_READ_BIT;
+      return _mesa_is_desktop_gl(ctx);
+   case GL_WRITE_ONLY_ARB:
+      *flags = GL_MAP_WRITE_BIT;
+      return true;
+   case GL_READ_WRITE_ARB:
+      *flags = GL_MAP_READ_BIT | GL_MAP_WRITE_BIT;
+      return _mesa_is_desktop_gl(ctx);
+   default:
+      return false;
+   }
+}
+
+void * GLAPIENTRY
+_mesa_MapBuffer(GLenum target, GLenum access)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+   GLbitfield accessFlags;
+
+   if (!get_map_buffer_access_flags(ctx, access, &accessFlags)) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glMapBuffer(invalid access)");
+      return NULL;
+   }
+
+   bufObj = get_buffer(ctx, "glMapBuffer", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return NULL;
+
+   return _mesa_map_buffer_range(ctx, bufObj, 0, bufObj->Size, accessFlags,
+                                 "glMapBuffer");
+}
+
+void * GLAPIENTRY
+_mesa_MapNamedBuffer(GLuint buffer, GLenum access)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
+   GLbitfield accessFlags;
 
+   if (!get_map_buffer_access_flags(ctx, access, &accessFlags)) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glMapNamedBuffer(invalid access)");
+      return NULL;
+   }
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glMapNamedBuffer");
+   if (!bufObj)
+      return NULL;
+
+   return _mesa_map_buffer_range(ctx, bufObj, 0, bufObj->Size, accessFlags,
+                                 "glMapNamedBuffer");
+}
+
+
+void
+_mesa_flush_mapped_buffer_range(struct gl_context *ctx,
+                                struct gl_buffer_object *bufObj,
+                                GLintptr offset, GLsizeiptr length,
+                                const char *func)
+{
    if (!ctx->Extensions.ARB_map_buffer_range) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glFlushMappedBufferRange(extension not supported)");
+                  "%s(ARB_map_buffer_range not supported)", func);
       return;
    }
 
    if (offset < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glFlushMappedBufferRange(offset = %ld)", (long)offset);
+                  "%s(offset %ld < 0)", func, (long) offset);
       return;
    }
 
    if (length < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glFlushMappedBufferRange(length = %ld)", (long)length);
+                  "%s(length %ld < 0)", func, (long) length);
       return;
    }
 
-   bufObj = get_buffer(ctx, "glFlushMappedBufferRange", target,
-                       GL_INVALID_OPERATION);
-   if (!bufObj)
-      return;
-
    if (!_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
       /* buffer is not mapped */
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glFlushMappedBufferRange(buffer is not mapped)");
+                  "%s(buffer is not mapped)", func);
       return;
    }
 
    if ((bufObj->Mappings[MAP_USER].AccessFlags &
         GL_MAP_FLUSH_EXPLICIT_BIT) == 0) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glFlushMappedBufferRange(GL_MAP_FLUSH_EXPLICIT_BIT not set)");
+                  "%s(GL_MAP_FLUSH_EXPLICIT_BIT not set)", func);
       return;
    }
 
    if (offset + length > bufObj->Mappings[MAP_USER].Length) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-		  "glFlushMappedBufferRange(offset %ld + length %ld > mapped length %ld)",
-		  (long)offset, (long)length,
-                  (long)bufObj->Mappings[MAP_USER].Length);
+                  "%s(offset %ld + length %ld > mapped length %ld)", func,
+                  (long) offset, (long) length,
+                  (long) bufObj->Mappings[MAP_USER].Length);
       return;
    }
 
@@ -2285,6 +2563,38 @@ _mesa_FlushMappedBufferRange(GLenum target, GLintptr offset, GLsizeiptr length)
                                          MAP_USER);
 }
 
+void GLAPIENTRY
+_mesa_FlushMappedBufferRange(GLenum target, GLintptr offset,
+                             GLsizeiptr length)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   bufObj = get_buffer(ctx, "glFlushMappedBufferRange", target,
+                       GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   _mesa_flush_mapped_buffer_range(ctx, bufObj, offset, length,
+                                   "glFlushMappedBufferRange");
+}
+
+void GLAPIENTRY
+_mesa_FlushMappedNamedBufferRange(GLuint buffer, GLintptr offset,
+                                  GLsizeiptr length)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                       "glFlushMappedNamedBufferRange");
+   if (!bufObj)
+      return;
+
+   _mesa_flush_mapped_buffer_range(ctx, bufObj, offset, length,
+                                   "glFlushMappedNamedBufferRange");
+}
+
 
 static GLenum
 buffer_object_purgeable(struct gl_context *ctx, GLuint name, GLenum option)
diff --git a/mesalib/src/mesa/main/bufferobj.h b/mesalib/src/mesa/main/bufferobj.h
index fe294fc0b..b5d73aec0 100644
--- a/mesalib/src/mesa/main/bufferobj.h
+++ b/mesalib/src/mesa/main/bufferobj.h
@@ -89,6 +89,10 @@ _mesa_lookup_bufferobj(struct gl_context *ctx, GLuint buffer);
 extern struct gl_buffer_object *
 _mesa_lookup_bufferobj_locked(struct gl_context *ctx, GLuint buffer);
 
+extern struct gl_buffer_object *
+_mesa_lookup_bufferobj_err(struct gl_context *ctx, GLuint buffer,
+                           const char *caller);
+
 extern void
 _mesa_begin_bufferobj_lookups(struct gl_context *ctx);
 
@@ -126,15 +130,62 @@ extern void
 _mesa_init_buffer_object_functions(struct dd_function_table *driver);
 
 extern void
+_mesa_buffer_storage(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                     GLenum target, GLsizeiptr size, const GLvoid *data,
+                     GLbitfield flags, const char *func);
+
+extern void
+_mesa_buffer_data(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                  GLenum target, GLsizeiptr size, const GLvoid *data,
+                  GLenum usage, const char *func);
+
+extern void
+_mesa_buffer_sub_data(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                      GLintptr offset, GLsizeiptr size, const GLvoid *data,
+                      const char *func);
+
+extern void
 _mesa_buffer_unmap_all_mappings(struct gl_context *ctx,
                                 struct gl_buffer_object *bufObj);
 
 extern void
-_mesa_buffer_clear_subdata(struct gl_context *ctx,
-                           GLintptr offset, GLsizeiptr size,
-                           const GLvoid *clearValue,
-                           GLsizeiptr clearValueSize,
-                           struct gl_buffer_object *bufObj);
+_mesa_copy_buffer_sub_data(struct gl_context *ctx,
+                           struct gl_buffer_object *src,
+                           struct gl_buffer_object *dst,
+                           GLintptr readOffset, GLintptr writeOffset,
+                           GLsizeiptr size, const char *func);
+
+extern void *
+_mesa_map_buffer_range(struct gl_context *ctx,
+                       struct gl_buffer_object *bufObj,
+                       GLintptr offset, GLsizeiptr length,
+                       GLbitfield access, const char *func);
+
+extern void
+_mesa_flush_mapped_buffer_range(struct gl_context *ctx,
+                                struct gl_buffer_object *bufObj,
+                                GLintptr offset, GLsizeiptr length,
+                                const char *func);
+
+extern void
+_mesa_ClearBufferSubData_sw(struct gl_context *ctx,
+                            GLintptr offset, GLsizeiptr size,
+                            const GLvoid *clearValue,
+                            GLsizeiptr clearValueSize,
+                            struct gl_buffer_object *bufObj);
+
+extern void
+_mesa_clear_buffer_sub_data(struct gl_context *ctx,
+                            struct gl_buffer_object *bufObj,
+                            GLenum internalformat,
+                            GLintptr offset, GLsizeiptr size,
+                            GLenum format, GLenum type,
+                            const GLvoid *data,
+                            const char *func, bool subdata);
+
+extern GLboolean
+_mesa_unmap_buffer(struct gl_context *ctx, struct gl_buffer_object *bufObj,
+                   const char *func);
 
 /*
  * API functions
@@ -146,7 +197,10 @@ void GLAPIENTRY
 _mesa_DeleteBuffers(GLsizei n, const GLuint * buffer);
 
 void GLAPIENTRY
-_mesa_GenBuffers(GLsizei n, GLuint * buffer);
+_mesa_GenBuffers(GLsizei n, GLuint *buffers);
+
+void GLAPIENTRY
+_mesa_CreateBuffers(GLsizei n, GLuint *buffers);
 
 GLboolean GLAPIENTRY
 _mesa_IsBuffer(GLuint buffer);
@@ -156,34 +210,61 @@ _mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
                     GLbitfield flags);
 
 void GLAPIENTRY
-_mesa_BufferData(GLenum target, GLsizeiptrARB size,
-                 const GLvoid * data, GLenum usage);
+_mesa_NamedBufferStorage(GLuint buffer, GLsizeiptr size, const GLvoid *data,
+                         GLbitfield flags);
+
+void GLAPIENTRY
+_mesa_BufferData(GLenum target, GLsizeiptr size,
+                 const GLvoid *data, GLenum usage);
+
+void GLAPIENTRY
+_mesa_NamedBufferData(GLuint buffer, GLsizeiptr size,
+                      const GLvoid *data, GLenum usage);
+
+void GLAPIENTRY
+_mesa_BufferSubData(GLenum target, GLintptr offset,
+                    GLsizeiptr size, const GLvoid *data);
 
 void GLAPIENTRY
-_mesa_BufferSubData(GLenum target, GLintptrARB offset,
-                    GLsizeiptrARB size, const GLvoid * data);
+_mesa_NamedBufferSubData(GLuint buffer, GLintptr offset,
+                         GLsizeiptr size, const GLvoid *data);
 
 void GLAPIENTRY
-_mesa_GetBufferSubData(GLenum target, GLintptrARB offset,
-                       GLsizeiptrARB size, void * data);
+_mesa_GetBufferSubData(GLenum target, GLintptr offset,
+                       GLsizeiptr size, GLvoid *data);
+
+void GLAPIENTRY
+_mesa_GetNamedBufferSubData(GLuint buffer, GLintptr offset,
+                            GLsizeiptr size, GLvoid *data);
 
 void GLAPIENTRY
 _mesa_ClearBufferData(GLenum target, GLenum internalformat,
                       GLenum format, GLenum type,
-                      const GLvoid * data);
+                      const GLvoid *data);
+
+void GLAPIENTRY
+_mesa_ClearNamedBufferData(GLuint buffer, GLenum internalformat,
+                           GLenum format, GLenum type,
+                           const GLvoid *data);
 
 void GLAPIENTRY
 _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
                          GLintptr offset, GLsizeiptr size,
                          GLenum format, GLenum type,
-                         const GLvoid * data);
+                         const GLvoid *data);
 
-void * GLAPIENTRY
-_mesa_MapBuffer(GLenum target, GLenum access);
+void GLAPIENTRY
+_mesa_ClearNamedBufferSubData(GLuint buffer, GLenum internalformat,
+                              GLintptr offset, GLsizeiptr size,
+                              GLenum format, GLenum type,
+                              const GLvoid *data);
 
 GLboolean GLAPIENTRY
 _mesa_UnmapBuffer(GLenum target);
 
+GLboolean GLAPIENTRY
+_mesa_UnmapNamedBuffer(GLuint buffer);
+
 void GLAPIENTRY
 _mesa_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params);
 
@@ -191,21 +272,52 @@ void GLAPIENTRY
 _mesa_GetBufferParameteri64v(GLenum target, GLenum pname, GLint64 *params);
 
 void GLAPIENTRY
+_mesa_GetNamedBufferParameteriv(GLuint buffer, GLenum pname, GLint *params);
+
+void GLAPIENTRY
+_mesa_GetNamedBufferParameteri64v(GLuint buffer, GLenum pname,
+                                  GLint64 *params);
+
+void GLAPIENTRY
 _mesa_GetBufferPointerv(GLenum target, GLenum pname, GLvoid **params);
 
 void GLAPIENTRY
+_mesa_GetNamedBufferPointerv(GLuint buffer, GLenum pname, GLvoid **params);
+
+
+void GLAPIENTRY
 _mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
                         GLintptr readOffset, GLintptr writeOffset,
                         GLsizeiptr size);
 
+void GLAPIENTRY
+_mesa_CopyNamedBufferSubData(GLuint readBuffer, GLuint writeBuffer,
+                             GLintptr readOffset, GLintptr writeOffset,
+                             GLsizeiptr size);
+
 void * GLAPIENTRY
 _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
                      GLbitfield access);
 
+void * GLAPIENTRY
+_mesa_MapNamedBufferRange(GLuint buffer, GLintptr offset, GLsizeiptr length,
+                          GLbitfield access);
+
+void * GLAPIENTRY
+_mesa_MapBuffer(GLenum target, GLenum access);
+
+void * GLAPIENTRY
+_mesa_MapNamedBuffer(GLuint buffer, GLenum access);
+
+
 void GLAPIENTRY
 _mesa_FlushMappedBufferRange(GLenum target,
                              GLintptr offset, GLsizeiptr length);
 
+void GLAPIENTRY
+_mesa_FlushMappedNamedBufferRange(GLuint buffer, GLintptr offset,
+                                  GLsizeiptr length);
+
 GLenum GLAPIENTRY
 _mesa_ObjectPurgeableAPPLE(GLenum objectType, GLuint name, GLenum option);
 
diff --git a/mesalib/src/mesa/main/compiler.h b/mesalib/src/mesa/main/compiler.h
index 95581fb49..55152fdef 100644
--- a/mesalib/src/mesa/main/compiler.h
+++ b/mesalib/src/mesa/main/compiler.h
@@ -57,29 +57,6 @@ extern "C" {
 # elif !defined(__sparc__) && defined(__sparc)
 #  define __sparc__
 # endif
-# if !defined(__volatile)
-#  define __volatile volatile
-# endif
-#endif
-
-
-/**
- * Disable assorted warnings
- */
-#if defined(_WIN32) && !defined(__CYGWIN__)
-#  if !defined(__GNUC__) /* mingw environment */
-#    pragma warning( disable : 4068 ) /* unknown pragma */
-#    pragma warning( disable : 4710 ) /* function 'foo' not inlined */
-#    pragma warning( disable : 4711 ) /* function 'foo' selected for automatic inline expansion */
-#    pragma warning( disable : 4127 ) /* conditional expression is constant */
-#    if defined(MESA_MINWARN)
-#      pragma warning( disable : 4244 ) /* '=' : conversion from 'const double ' to 'float ', possible loss of data */
-#      pragma warning( disable : 4018 ) /* '<' : signed/unsigned mismatch */
-#      pragma warning( disable : 4305 ) /* '=' : truncation from 'const double ' to 'float ' */
-#      pragma warning( disable : 4550 ) /* 'function' undefined; assuming extern returning int */
-#      pragma warning( disable : 4761 ) /* integral size mismatch in argument; conversion supplied */
-#    endif
-#  endif
 #endif
 
 
@@ -101,14 +78,6 @@ extern "C" {
 #elif defined(__APPLE__)
 #include <CoreFoundation/CFByteOrder.h>
 #define CPU_TO_LE32( x )	CFSwapInt32HostToLittle( x )
-#elif (defined(_AIX))
-static inline GLuint CPU_TO_LE32(GLuint x)
-{
-   return (((x & 0x000000ff) << 24) |
-           ((x & 0x0000ff00) <<  8) |
-           ((x & 0x00ff0000) >>  8) |
-           ((x & 0xff000000) >> 24));
-}
 #elif defined(__OpenBSD__)
 #include <sys/types.h>
 #define CPU_TO_LE32( x )	htole32( x )
@@ -125,34 +94,6 @@ static inline GLuint CPU_TO_LE32(GLuint x)
 
 
 
-/**
- * Create a macro so that asm functions can be linked into compilers other
- * than GNU C
- */
-#ifndef _ASMAPI
-#if defined(_WIN32)
-#define _ASMAPI __cdecl
-#else
-#define _ASMAPI
-#endif
-#ifdef	PTR_DECL_IN_FRONT
-#define	_ASMAPIP * _ASMAPI
-#else
-#define	_ASMAPIP _ASMAPI *
-#endif
-#endif
-
-
-/**
- * LONGSTRING macro
- * gcc -pedantic warns about long string literals, LONGSTRING silences that.
- */
-#if !defined(__GNUC__)
-# define LONGSTRING
-#else
-# define LONGSTRING __extension__
-#endif
-
 #define IEEE_ONE 0x3f800000
 
 
diff --git a/mesalib/src/mesa/main/context.c b/mesalib/src/mesa/main/context.c
index 22c2341d6..c1acda980 100644
--- a/mesalib/src/mesa/main/context.c
+++ b/mesalib/src/mesa/main/context.c
@@ -882,18 +882,35 @@ update_default_objects(struct gl_context *ctx)
 
 
 /**
- * This is the default function we plug into all dispatch table slots
- * This helps prevents a segfault when someone calls a GL function without
- * first checking if the extension's supported.
+ * This function is called by the glapi no-op functions.  For each OpenGL
+ * function/entrypoint there's a simple no-op function.  These "no-op"
+ * functions call this function.
+ *
+ * If there's a current OpenGL context for the calling thread, we record a
+ * GL_INVALID_OPERATION error.  This can happen either because the app's
+ * calling an unsupported extension function, or calling an illegal function
+ * (such as glClear between glBegin/glEnd).
+ *
+ * If there's no current OpenGL context for the calling thread, we can
+ * print a message to stderr.
+ *
+ * \param name  the name of the OpenGL function, without the "gl" prefix
  */
-int
-_mesa_generic_nop(void)
+static void
+nop_handler(const char *name)
 {
    GET_CURRENT_CONTEXT(ctx);
-   _mesa_error(ctx, GL_INVALID_OPERATION,
-               "unsupported function called "
-               "(unsupported extension or deprecated function?)");
-   return 0;
+   if (ctx) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "gl%s(invalid call)", name);
+   }
+#if defined(DEBUG)
+   else if (getenv("MESA_DEBUG") || getenv("LIBGL_DEBUG")) {
+      fprintf(stderr,
+              "GL User Error: gl%s called without a rendering context\n",
+              name);
+      fflush(stderr);
+   }
+#endif
 }
 
 
@@ -909,16 +926,13 @@ nop_glFlush(void)
 #endif
 
 
-extern void (*__glapi_noop_table[])(void);
-
-
 /**
- * Allocate and initialize a new dispatch table.  All the dispatch
- * function pointers will point at the _mesa_generic_nop() function
- * which raises GL_INVALID_OPERATION.
+ * Allocate and initialize a new dispatch table.  The table will be
+ * populated with pointers to "no-op" functions.  In turn, the no-op
+ * functions will call nop_handler() above.
  */
-struct _glapi_table *
-_mesa_alloc_dispatch_table(void)
+static struct _glapi_table *
+alloc_dispatch_table(void)
 {
    /* Find the larger of Mesa's dispatch table and libGL's dispatch table.
     * In practice, this'll be the same for stand-alone Mesa.  But for DRI
@@ -926,23 +940,10 @@ _mesa_alloc_dispatch_table(void)
     * DRI drivers.
     */
    GLint numEntries = MAX2(_glapi_get_dispatch_table_size(), _gloffset_COUNT);
-   struct _glapi_table *table;
-
-   table = malloc(numEntries * sizeof(_glapi_proc));
-   if (table) {
-      _glapi_proc *entry = (_glapi_proc *) table;
-      GLint i;
-      for (i = 0; i < numEntries; i++) {
-#if defined(_WIN32)
-         /* FIXME: This will not generate an error, but at least it won't
-          * corrupt the stack like _mesa_generic_nop does. */
-         entry[i] = __glapi_noop_table[i];
-#else
-         entry[i] = (_glapi_proc) _mesa_generic_nop;
-#endif
-      }
+   struct _glapi_table *table = _glapi_new_nop_table(numEntries);
 
 #if defined(_WIN32)
+   if (table) {
       /* This is a special case for Windows in the event that
        * wglGetProcAddress is called between glBegin/End().
        *
@@ -960,8 +961,11 @@ _mesa_alloc_dispatch_table(void)
        * assertion passes and the test continues.
        */
       SET_Flush(table, nop_glFlush);
-#endif
    }
+#endif
+
+   _glapi_set_nop_handler(nop_handler);
+
    return table;
 }
 
@@ -997,7 +1001,7 @@ create_beginend_table(const struct gl_context *ctx)
 {
    struct _glapi_table *table;
 
-   table = _mesa_alloc_dispatch_table();
+   table = alloc_dispatch_table();
    if (!table)
       return NULL;
 
@@ -1136,7 +1140,7 @@ _mesa_initialize_context(struct gl_context *ctx,
       goto fail;
 
    /* setup the API dispatch tables with all nop functions */
-   ctx->OutsideBeginEnd = _mesa_alloc_dispatch_table();
+   ctx->OutsideBeginEnd = alloc_dispatch_table();
    if (!ctx->OutsideBeginEnd)
       goto fail;
    ctx->Exec = ctx->OutsideBeginEnd;
@@ -1163,7 +1167,7 @@ _mesa_initialize_context(struct gl_context *ctx,
    switch (ctx->API) {
    case API_OPENGL_COMPAT:
       ctx->BeginEnd = create_beginend_table(ctx);
-      ctx->Save = _mesa_alloc_dispatch_table();
+      ctx->Save = alloc_dispatch_table();
       if (!ctx->BeginEnd || !ctx->Save)
          goto fail;
 
diff --git a/mesalib/src/mesa/main/context.h b/mesalib/src/mesa/main/context.h
index d5650877e..1cd89a84a 100644
--- a/mesalib/src/mesa/main/context.h
+++ b/mesalib/src/mesa/main/context.h
@@ -175,9 +175,6 @@ _mesa_finish(struct gl_context *ctx);
 extern void
 _mesa_flush(struct gl_context *ctx);
 
-extern int
-_mesa_generic_nop(void);
-
 extern void GLAPIENTRY
 _mesa_Finish( void );
 
diff --git a/mesalib/src/mesa/main/copyimage.c b/mesalib/src/mesa/main/copyimage.c
index 455929dc2..fd22f2889 100644
--- a/mesalib/src/mesa/main/copyimage.c
+++ b/mesalib/src/mesa/main/copyimage.c
@@ -33,6 +33,12 @@
 #include "texobj.h"
 #include "fbobject.h"
 #include "textureview.h"
+#include "glformats.h"
+
+enum mesa_block_class {
+   BLOCK_CLASS_128_BITS,
+   BLOCK_CLASS_64_BITS
+};
 
 static bool
 prepare_target(struct gl_context *ctx, GLuint name, GLenum *target, int level,
@@ -253,6 +259,124 @@ check_region_bounds(struct gl_context *ctx, struct gl_texture_image *tex_image,
    return true;
 }
 
+static bool
+compressed_format_compatible(struct gl_context *ctx,
+                             GLenum compressedFormat, GLenum otherFormat)
+{
+   enum mesa_block_class compressedClass, otherClass;
+
+   /* Two view-incompatible compressed formats are never compatible. */
+   if (_mesa_is_compressed_format(ctx, otherFormat)) {
+      return false;
+   }
+
+   /*
+    * From ARB_copy_image spec:
+    *    Table 4.X.1 (Compatible internal formats for copying between
+    *                 compressed and uncompressed internal formats)
+    *    ---------------------------------------------------------------------
+    *    | Texel / | Uncompressed      |                                     |
+    *    | Block   | internal format   | Compressed internal format          |
+    *    | size    |                   |                                     |
+    *    ---------------------------------------------------------------------
+    *    | 128-bit | RGBA32UI,         | COMPRESSED_RGBA_S3TC_DXT3_EXT,      |
+    *    |         | RGBA32I,          | COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT,|
+    *    |         | RGBA32F           | COMPRESSED_RGBA_S3TC_DXT5_EXT,      |
+    *    |         |                   | COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT,|
+    *    |         |                   | COMPRESSED_RG_RGTC2,                |
+    *    |         |                   | COMPRESSED_SIGNED_RG_RGTC2,         |
+    *    |         |                   | COMPRESSED_RGBA_BPTC_UNORM,         |
+    *    |         |                   | COMPRESSED_SRGB_ALPHA_BPTC_UNORM,   |
+    *    |         |                   | COMPRESSED_RGB_BPTC_SIGNED_FLOAT,   |
+    *    |         |                   | COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT  |
+    *    ---------------------------------------------------------------------
+    *    | 64-bit  | RGBA16F, RG32F,   | COMPRESSED_RGB_S3TC_DXT1_EXT,       |
+    *    |         | RGBA16UI, RG32UI, | COMPRESSED_SRGB_S3TC_DXT1_EXT,      |
+    *    |         | RGBA16I, RG32I,   | COMPRESSED_RGBA_S3TC_DXT1_EXT,      |
+    *    |         | RGBA16,           | COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT,|
+    *    |         | RGBA16_SNORM      | COMPRESSED_RED_RGTC1,               |
+    *    |         |                   | COMPRESSED_SIGNED_RED_RGTC1         |
+    *    ---------------------------------------------------------------------
+    */
+
+   switch (compressedFormat) {
+      case GL_COMPRESSED_RGBA_S3TC_DXT3_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:
+      case GL_COMPRESSED_RGBA_S3TC_DXT5_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:
+      case GL_COMPRESSED_RG_RGTC2:
+      case GL_COMPRESSED_SIGNED_RG_RGTC2:
+      case GL_COMPRESSED_RGBA_BPTC_UNORM:
+      case GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM:
+      case GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT:
+      case GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT:
+         compressedClass = BLOCK_CLASS_128_BITS;
+         break;
+      case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_SRGB_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_RED_RGTC1:
+      case GL_COMPRESSED_SIGNED_RED_RGTC1:
+         compressedClass = BLOCK_CLASS_64_BITS;
+         break;
+      default:
+         return false;
+   }
+
+   switch (otherFormat) {
+      case GL_RGBA32UI:
+      case GL_RGBA32I:
+      case GL_RGBA32F:
+         otherClass = BLOCK_CLASS_128_BITS;
+         break;
+      case GL_RGBA16F:
+      case GL_RG32F:
+      case GL_RGBA16UI:
+      case GL_RG32UI:
+      case GL_RGBA16I:
+      case GL_RG32I:
+      case GL_RGBA16:
+      case GL_RGBA16_SNORM:
+         otherClass = BLOCK_CLASS_64_BITS;
+         break;
+      default:
+         return false;
+   }
+
+   return compressedClass == otherClass;
+}
+
+static bool
+copy_format_compatible(struct gl_context *ctx,
+                                GLenum srcFormat, GLenum dstFormat)
+{
+   /*
+    * From ARB_copy_image spec:
+    *    For the purposes of CopyImageSubData, two internal formats
+    *    are considered compatible if any of the following conditions are
+    *    met:
+    *    * the formats are the same,
+    *    * the formats are considered compatible according to the
+    *      compatibility rules used for texture views as defined in
+    *      section 3.9.X. In particular, if both internal formats are listed
+    *      in the same entry of Table 3.X.2, they are considered compatible, or
+    *    * one format is compressed and the other is uncompressed and
+    *      Table 4.X.1 lists the two formats in the same row.
+    */
+
+   if (_mesa_texture_view_compatible_format(ctx, srcFormat, dstFormat)) {
+      /* Also checks if formats are equal. */
+      return true;
+   } else if (_mesa_is_compressed_format(ctx, srcFormat)) {
+      return compressed_format_compatible(ctx, srcFormat, dstFormat);
+   } else if (_mesa_is_compressed_format(ctx, dstFormat)) {
+      return compressed_format_compatible(ctx, dstFormat, srcFormat);
+   }
+
+   return false;
+}
+
 void GLAPIENTRY
 _mesa_CopyImageSubData(GLuint srcName, GLenum srcTarget, GLint srcLevel,
                        GLint srcX, GLint srcY, GLint srcZ,
@@ -265,7 +389,7 @@ _mesa_CopyImageSubData(GLuint srcName, GLenum srcTarget, GLint srcLevel,
    struct gl_texture_object *srcTexObj, *dstTexObj;
    struct gl_texture_image *srcTexImage, *dstTexImage;
    GLuint src_bw, src_bh, dst_bw, dst_bh;
-   int i, srcNewZ, dstNewZ, Bpt;
+   int i, srcNewZ, dstNewZ;
 
    if (MESA_VERBOSE & VERBOSE_API)
       _mesa_debug(ctx, "glCopyImageSubData(%u, %s, %d, %d, %d, %d, "
@@ -306,15 +430,6 @@ _mesa_CopyImageSubData(GLuint srcName, GLenum srcTarget, GLint srcLevel,
       goto cleanup;
    }
 
-   /* Very simple sanity check.  This is sufficient if one of the textures
-    * is compressed. */
-   Bpt = _mesa_get_format_bytes(srcTexImage->TexFormat);
-   if (_mesa_get_format_bytes(dstTexImage->TexFormat) != Bpt) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyImageSubData(internalFormat mismatch)");
-      goto cleanup;
-   }
-
    if (!check_region_bounds(ctx, srcTexImage, srcX, srcY, srcZ,
                             srcWidth, srcHeight, srcDepth, "src"))
       goto cleanup;
@@ -324,17 +439,11 @@ _mesa_CopyImageSubData(GLuint srcName, GLenum srcTarget, GLint srcLevel,
                             (srcHeight / src_bh) * dst_bh, srcDepth, "dst"))
       goto cleanup;
 
-   if (_mesa_is_format_compressed(srcTexImage->TexFormat)) {
-      /* XXX: Technically, we should probaby do some more specific checking
-       * here.  However, this should be sufficient for all compressed
-       * formats that mesa supports since it is a direct memory copy.
-       */
-   } else if (_mesa_is_format_compressed(dstTexImage->TexFormat)) {
-   } else if (_mesa_texture_view_compatible_format(ctx,
-                                                   srcTexImage->InternalFormat,
-                                                   dstTexImage->InternalFormat)) {
-   } else {
-      return; /* Error logged by _mesa_texture_view_compatible_format */
+   if (!copy_format_compatible(ctx, srcTexImage->InternalFormat,
+                               dstTexImage->InternalFormat)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glCopyImageSubData(internalFormat mismatch)");
+      goto cleanup;
    }
 
    for (i = 0; i < srcDepth; ++i) {
diff --git a/mesalib/src/mesa/main/formats.c b/mesalib/src/mesa/main/formats.c
index 422c9dc46..2bc8bcad9 100644
--- a/mesalib/src/mesa/main/formats.c
+++ b/mesalib/src/mesa/main/formats.c
@@ -388,7 +388,7 @@ array_formats_equal(const void *a, const void *b)
 }
 
 static void
-format_array_format_table_init()
+format_array_format_table_init(void)
 {
    const struct gl_format_info *info;
    mesa_array_format array_format;
diff --git a/mesalib/src/mesa/main/imports.c b/mesalib/src/mesa/main/imports.c
index a7ffe2296..68c731657 100644
--- a/mesalib/src/mesa/main/imports.c
+++ b/mesalib/src/mesa/main/imports.c
@@ -45,6 +45,7 @@
 #include <stdio.h>
 #include <stdarg.h>
 #include "c99_math.h"
+#include "util/rounding.h" /* for _mesa_roundeven */
 #include "imports.h"
 #include "context.h"
 #include "mtypes.h"
@@ -307,26 +308,6 @@ _mesa_bitcount_64(uint64_t n)
 #endif
 
 
-/* Using C99 rounding functions for roundToEven() implementation is
- * difficult, because round(), rint, and nearbyint() are affected by
- * fesetenv(), which the application may have done for its own
- * purposes.  Mesa's IROUND macro is close to what we want, but it
- * rounds away from 0 on n + 0.5.
- */
-int
-_mesa_round_to_even(float val)
-{
-   int rounded = IROUND(val);
-
-   if (val - floor(val) == 0.5) {
-      if (rounded % 2 != 0)
-         rounded += val > 0 ? -1 : 1;
-   }
-
-   return rounded;
-}
-
-
 /**
  * Convert a 4-byte float to a 2-byte half float.
  *
@@ -388,7 +369,7 @@ _mesa_float_to_half(float val)
           * or normal.
           */
          e = 0;
-         m = _mesa_round_to_even((1 << 24) * fabsf(fi.f));
+         m = (int) _mesa_roundevenf((1 << 24) * fabsf(fi.f));
       }
       else if (new_exp > 15) {
          /* map this value to infinity */
@@ -402,7 +383,7 @@ _mesa_float_to_half(float val)
           * either normal or infinite.
           */
          e = new_exp + 15;
-         m = _mesa_round_to_even(flt_m / (float) (1 << 13));
+         m = (int) _mesa_roundevenf(flt_m / (float) (1 << 13));
       }
    }
 
@@ -481,24 +462,6 @@ _mesa_half_to_float(GLhalfARB val)
 /** \name String */
 /*@{*/
 
-/**
- * Implemented using malloc() and strcpy.
- * Note that NULL is handled accordingly.
- */
-char *
-_mesa_strdup( const char *s )
-{
-   if (s) {
-      size_t l = strlen(s);
-      char *s2 = malloc(l + 1);
-      if (s2)
-         strcpy(s2, s);
-      return s2;
-   }
-   else {
-      return NULL;
-   }
-}
 
 /** Compute simple checksum/hash for a string */
 unsigned int
diff --git a/mesalib/src/mesa/main/imports.h b/mesalib/src/mesa/main/imports.h
index c211b1555..8877966ec 100755
--- a/mesalib/src/mesa/main/imports.h
+++ b/mesalib/src/mesa/main/imports.h
@@ -432,9 +432,6 @@ _mesa_fls(unsigned int n)
 #endif
 }
 
-extern int
-_mesa_round_to_even(float val);
-
 extern GLhalfARB
 _mesa_float_to_half(float f);
 
@@ -447,9 +444,6 @@ _mesa_half_is_negative(GLhalfARB h)
    return h & 0x8000;
 }
 
-extern char *
-_mesa_strdup( const char *s );
-
 extern unsigned int
 _mesa_str_checksum(const char *str);
 
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index efeee8bff..8e1dba6f0 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -693,15 +693,16 @@ struct gl_colorbuffer_attrib
     * \name Logic op
     */
    /*@{*/
-   GLenum LogicOp;			/**< Logic operator */
    GLboolean IndexLogicOpEnabled;	/**< Color index logic op enabled flag */
    GLboolean ColorLogicOpEnabled;	/**< RGBA logic op enabled flag */
+   GLenum LogicOp;			/**< Logic operator */
+
    /*@}*/
 
    GLboolean DitherFlag;		/**< Dither enable flag */
 
-   GLenum ClampFragmentColor; /**< GL_TRUE, GL_FALSE or GL_FIXED_ONLY_ARB */
    GLboolean _ClampFragmentColor; /** < with GL_FIXED_ONLY_ARB resolved */
+   GLenum ClampFragmentColor; /**< GL_TRUE, GL_FALSE or GL_FIXED_ONLY_ARB */
    GLenum ClampReadColor;     /**< GL_TRUE, GL_FALSE or GL_FIXED_ONLY_ARB */
 
    GLboolean sRGBEnabled;    /**< Framebuffer sRGB blending/updating requested */
@@ -799,6 +800,7 @@ struct gl_eval_attrib
 struct gl_fog_attrib
 {
    GLboolean Enabled;		/**< Fog enabled flag */
+   GLboolean ColorSumEnabled;
    GLfloat ColorUnclamped[4];            /**< Fog color */
    GLfloat Color[4];		/**< Fog color */
    GLfloat Density;		/**< Density >= 0.0 */
@@ -806,7 +808,6 @@ struct gl_fog_attrib
    GLfloat End;			/**< End distance in eye coords */
    GLfloat Index;		/**< Fog index */
    GLenum Mode;			/**< Fog mode */
-   GLboolean ColorSumEnabled;
    GLenum FogCoordinateSource;  /**< GL_EXT_fog_coord */
    GLfloat _Scale;		/**< (End == Start) ? 1.0 : 1.0 / (End - Start) */
    GLenum FogDistanceMode;     /**< GL_NV_fog_distance */
@@ -846,16 +847,17 @@ struct gl_light_attrib
    struct gl_material Material;
 
    GLboolean Enabled;			/**< Lighting enabled flag */
+   GLboolean ColorMaterialEnabled;
+
    GLenum ShadeModel;			/**< GL_FLAT or GL_SMOOTH */
    GLenum ProvokingVertex;              /**< GL_EXT_provoking_vertex */
    GLenum ColorMaterialFace;		/**< GL_FRONT, BACK or FRONT_AND_BACK */
    GLenum ColorMaterialMode;		/**< GL_AMBIENT, GL_DIFFUSE, etc */
    GLbitfield _ColorMaterialBitmask;	/**< bitmask formed from Face and Mode */
-   GLboolean ColorMaterialEnabled;
-   GLenum ClampVertexColor;             /**< GL_TRUE, GL_FALSE, GL_FIXED_ONLY */
-   GLboolean _ClampVertexColor;
 
-   struct gl_light EnabledList;         /**< List sentinel */
+
+   GLboolean _ClampVertexColor;
+   GLenum ClampVertexColor;             /**< GL_TRUE, GL_FALSE, GL_FIXED_ONLY */
 
    /** 
     * Derived state for optimizations: 
@@ -863,6 +865,8 @@ struct gl_light_attrib
    /*@{*/
    GLboolean _NeedEyeCoords;		
    GLboolean _NeedVertices;		/**< Use fast shader? */
+   struct gl_light EnabledList;         /**< List sentinel */
+
    GLfloat _BaseColor[2][3];
    /*@}*/
 };
@@ -900,13 +904,15 @@ struct gl_multisample_attrib
    GLboolean SampleAlphaToCoverage;
    GLboolean SampleAlphaToOne;
    GLboolean SampleCoverage;
-   GLfloat SampleCoverageValue;
    GLboolean SampleCoverageInvert;
    GLboolean SampleShading;
-   GLfloat MinSampleShadingValue;
 
    /* ARB_texture_multisample / GL3.2 additions */
    GLboolean SampleMask;
+
+   GLfloat SampleCoverageValue;
+   GLfloat MinSampleShadingValue;
+
    /** The GL spec defines this as an array but >32x MSAA is madness */
    GLbitfield SampleMaskValue;
 };
@@ -977,11 +983,11 @@ struct gl_pixel_attrib
  */
 struct gl_point_attrib
 {
-   GLboolean SmoothFlag;	/**< True if GL_POINT_SMOOTH is enabled */
    GLfloat Size;		/**< User-specified point size */
    GLfloat Params[3];		/**< GL_EXT_point_parameters */
    GLfloat MinSize, MaxSize;	/**< GL_EXT_point_parameters */
    GLfloat Threshold;		/**< GL_EXT_point_parameters */
+   GLboolean SmoothFlag;	/**< True if GL_POINT_SMOOTH is enabled */
    GLboolean _Attenuated;	/**< True if Params != [1, 0, 0] */
    GLboolean PointSprite;	/**< GL_NV/ARB_point_sprite */
    GLboolean CoordReplace[MAX_TEXTURE_COORD_UNITS]; /**< GL_ARB_point_sprite*/
@@ -1347,6 +1353,9 @@ struct gl_texture_unit
 
    GLfloat LodBias;		/**< for biasing mipmap levels */
 
+   /** Texture targets that have a non-default texture bound */
+   GLbitfield _BoundTextures;
+
    /** Current sampler object (GL_ARB_sampler_objects) */
    struct gl_sampler_object *Sampler;
 
@@ -1373,8 +1382,6 @@ struct gl_texture_unit
    /** Points to highest priority, complete and enabled texture object */
    struct gl_texture_object *_Current;
 
-   /** Texture targets that have a non-default texture bound */
-   GLbitfield _BoundTextures;
 };
 
 
@@ -1384,16 +1391,15 @@ struct gl_texture_unit
 struct gl_texture_attrib
 {
    GLuint CurrentUnit;   /**< GL_ACTIVE_TEXTURE */
-   struct gl_texture_unit Unit[MAX_COMBINED_TEXTURE_IMAGE_UNITS];
+
+   /** GL_ARB_seamless_cubemap */
+   GLboolean CubeMapSeamless;
 
    struct gl_texture_object *ProxyTex[NUM_TEXTURE_TARGETS];
 
    /** GL_ARB_texture_buffer_object */
    struct gl_buffer_object *BufferObject;
 
-   /** GL_ARB_seamless_cubemap */
-   GLboolean CubeMapSeamless;
-
    /** Texture coord units/sets used for fragment texturing */
    GLbitfield _EnabledCoordUnits;
 
@@ -1411,6 +1417,8 @@ struct gl_texture_attrib
 
    /** Largest index + 1 of texture units that have had any CurrentTex set. */
    GLint NumCurrentTexUsed;
+
+   struct gl_texture_unit Unit[MAX_COMBINED_TEXTURE_IMAGE_UNITS];
 };
 
 
@@ -1536,12 +1544,12 @@ struct gl_client_array
    GLenum Format;               /**< default: GL_RGBA, but may be GL_BGRA */
    GLsizei Stride;		/**< user-specified stride */
    GLsizei StrideB;		/**< actual stride in bytes */
+   GLuint _ElementSize;         /**< size of each element in bytes */
    const GLubyte *Ptr;          /**< Points to array data */
    GLboolean Enabled;		/**< Enabled flag is a boolean */
    GLboolean Normalized;        /**< GL_ARB_vertex_program */
    GLboolean Integer;           /**< Integer-valued? */
    GLuint InstanceDivisor;      /**< GL_ARB_instanced_arrays */
-   GLuint _ElementSize;         /**< size of each element in bytes */
 
    struct gl_buffer_object *BufferObj;/**< GL_ARB_vertex_buffer_object */
 };
@@ -1600,9 +1608,11 @@ struct gl_vertex_array_object
 {
    /** Name of the VAO as received from glGenVertexArray. */
    GLuint Name;
-   GLchar *Label;       /**< GL_KHR_debug */
 
    GLint RefCount;
+
+   GLchar *Label;       /**< GL_KHR_debug */
+
    mtx_t Mutex;
 
    /**
@@ -1707,6 +1717,9 @@ struct gl_array_attrib
    GLuint RestartIndex;
    /*@}*/
 
+   /** One of the DRAW_xxx flags, not consumed by drivers */
+   gl_draw_method DrawMethod;
+
    /* GL_ARB_vertex_buffer_object */
    struct gl_buffer_object *ArrayBufferObj;
 
@@ -1716,9 +1729,6 @@ struct gl_array_attrib
     */
    const struct gl_client_array **_DrawArrays; /**< 0..VERT_ATTRIB_MAX-1 */
 
-   /** One of the DRAW_xxx flags, not consumed by drivers */
-   gl_draw_method DrawMethod;
-
    /** Legal array datatypes and the API for which they have been computed */
    GLbitfield LegalTypesMask;
    gl_api LegalTypesMaskAPI;
@@ -1880,8 +1890,8 @@ struct gl_transform_feedback_info
 struct gl_transform_feedback_object
 {
    GLuint Name;  /**< AKA the object ID */
-   GLchar *Label;     /**< GL_KHR_debug */
    GLint RefCount;
+   GLchar *Label;     /**< GL_KHR_debug */
    GLboolean Active;  /**< Is transform feedback enabled? */
    GLboolean Paused;  /**< Is transform feedback paused? */
    GLboolean EndedAnytime; /**< Has EndTransformFeedback been called
@@ -1889,14 +1899,6 @@ struct gl_transform_feedback_object
    GLboolean EverBound; /**< Has this object been bound? */
 
    /**
-    * The shader program active when BeginTransformFeedback() was called.
-    * When active and unpaused, this equals ctx->Shader.CurrentProgram[stage],
-    * where stage is the pipeline stage that is the source of data for
-    * transform feedback.
-    */
-   struct gl_shader_program *shader_program;
-
-   /**
     * GLES: if Active is true, remaining number of primitives which can be
     * rendered without overflow.  This is necessary to track because GLES
     * requires us to generate INVALID_OPERATION if a call to glDrawArrays or
@@ -1907,6 +1909,14 @@ struct gl_transform_feedback_object
     */
    unsigned GlesRemainingPrims;
 
+   /**
+    * The shader program active when BeginTransformFeedback() was called.
+    * When active and unpaused, this equals ctx->Shader.CurrentProgram[stage],
+    * where stage is the pipeline stage that is the source of data for
+    * transform feedback.
+    */
+   struct gl_shader_program *shader_program;
+
    /** The feedback buffers */
    GLuint BufferNames[MAX_FEEDBACK_BUFFERS];
    struct gl_buffer_object *Buffers[MAX_FEEDBACK_BUFFERS];
@@ -2229,8 +2239,9 @@ enum gl_frag_depth_layout
 struct gl_program
 {
    GLuint Id;
-   GLubyte *String;  /**< Null-terminated program text */
    GLint RefCount;
+   GLubyte *String;  /**< Null-terminated program text */
+
    GLenum Target;    /**< GL_VERTEX/FRAGMENT_PROGRAM_ARB, GL_GEOMETRY_PROGRAM_NV */
    GLenum Format;    /**< String encoding format */
 
@@ -2533,18 +2544,20 @@ struct gl_shader
    GLenum Type;
    gl_shader_stage Stage;
    GLuint Name;  /**< AKA the handle */
-   GLchar *Label;   /**< GL_KHR_debug */
    GLint RefCount;  /**< Reference count */
+   GLchar *Label;   /**< GL_KHR_debug */
    GLboolean DeletePending;
    GLboolean CompileStatus;
-   const GLchar *Source;  /**< Source code string */
+   bool IsES;              /**< True if this shader uses GLSL ES */
+
    GLuint SourceChecksum;       /**< for debug/logging purposes */
+   const GLchar *Source;  /**< Source code string */
+
    struct gl_program *Program;  /**< Post-compile assembly code */
    GLchar *InfoLog;
    struct gl_sl_pragmas Pragmas;
 
    unsigned Version;       /**< GLSL version used for linking */
-   GLboolean IsES;         /**< True if this shader uses GLSL ES */
 
    /**
     * \name Sampler tracking
@@ -2588,8 +2601,8 @@ struct gl_shader
     *
     * These fields are only set post-linking.
     */
-   struct gl_uniform_block *UniformBlocks;
    unsigned NumUniformBlocks;
+   struct gl_uniform_block *UniformBlocks;
 
    struct exec_list *ir;
    struct glsl_symbol_table *symbols;
@@ -2879,8 +2892,8 @@ struct gl_shader_program
     */
    unsigned LastClipDistanceArraySize;
 
-   struct gl_uniform_block *UniformBlocks;
    unsigned NumUniformBlocks;
+   struct gl_uniform_block *UniformBlocks;
 
    /**
     * Indices into the _LinkedShaders's UniformBlocks[] array for each stage
@@ -2911,7 +2924,7 @@ struct gl_shader_program
    GLchar *InfoLog;
 
    unsigned Version;       /**< GLSL version used for linking */
-   GLboolean IsES;         /**< True if this program uses GLSL ES */
+   bool IsES;              /**< True if this program uses GLSL ES */
 
    /**
     * Per-stage shaders resulting from the first stage of linking.
@@ -3036,7 +3049,7 @@ struct gl_shader_compiler_options
 
    struct gl_sl_pragmas DefaultPragmas; /**< Default #pragma settings */
 
-   struct nir_shader_compiler_options *NirOptions;
+   const struct nir_shader_compiler_options *NirOptions;
 };
 
 
@@ -3265,9 +3278,10 @@ struct gl_framebuffer
     * polygon face orientation, and polygon stipple will have to be inverted.
     */
    GLuint Name;
+   GLint RefCount;
+
    GLchar *Label;       /**< GL_KHR_debug */
 
-   GLint RefCount;
    GLboolean DeletePending;
 
    /**
@@ -3301,6 +3315,13 @@ struct gl_framebuffer
    GLboolean _AllColorBuffersFixedPoint; /* no integer, no float */
    GLboolean _HasSNormOrFloatColorBuffer;
 
+   /**
+    * The maximum number of layers in the framebuffer, or 0 if the framebuffer
+    * is not layered.  For cube maps and cube map arrays, each cube face
+    * counts as a layer.
+    */
+   GLuint MaxNumLayers;
+
    /** Array of all renderbuffer attachments, indexed by BUFFER_* tokens. */
    struct gl_renderbuffer_attachment Attachment[BUFFER_COUNT];
 
@@ -3317,13 +3338,6 @@ struct gl_framebuffer
    struct gl_renderbuffer *_ColorDrawBuffers[MAX_DRAW_BUFFERS];
    struct gl_renderbuffer *_ColorReadBuffer;
 
-   /**
-    * The maximum number of layers in the framebuffer, or 0 if the framebuffer
-    * is not layered.  For cube maps and cube map arrays, each cube face
-    * counts as a layer.
-    */
-   GLuint MaxNumLayers;
-
    /** Delete this framebuffer */
    void (*Delete)(struct gl_framebuffer *fb);
 };
@@ -4161,6 +4175,13 @@ struct gl_image_unit
    GLboolean Layered;
 
    /**
+    * GL_TRUE if the state of this image unit is valid and access from
+    * the shader is allowed.  Otherwise loads from this unit should
+    * return zero and stores should have no effect.
+    */
+   GLboolean _Valid;
+
+   /**
     * Layer of the texture object bound to this unit, or zero if the
     * whole level is bound.
     */
@@ -4184,12 +4205,6 @@ struct gl_image_unit
     */
    mesa_format _ActualFormat;
 
-   /**
-    * GL_TRUE if the state of this image unit is valid and access from
-    * the shader is allowed.  Otherwise loads from this unit should
-    * return zero and stores should have no effect.
-    */
-   GLboolean _Valid;
 };
 
 /**
diff --git a/mesalib/src/mesa/main/objectlabel.c b/mesalib/src/mesa/main/objectlabel.c
index 78df96b9b..aecb5b1fa 100644
--- a/mesalib/src/mesa/main/objectlabel.c
+++ b/mesalib/src/mesa/main/objectlabel.c
@@ -76,7 +76,7 @@ set_label(struct gl_context *ctx, char **labelPtr, const char *label,
                 MAX_LABEL_LENGTH);
 
          /* null-terminated string */
-         *labelPtr = _mesa_strdup(label);
+         *labelPtr = strdup(label);
       }
    }
 }
diff --git a/mesalib/src/mesa/main/pbo.c b/mesalib/src/mesa/main/pbo.c
index 5c906ed74..0c1602532 100644
--- a/mesalib/src/mesa/main/pbo.c
+++ b/mesalib/src/mesa/main/pbo.c
@@ -80,7 +80,7 @@ _mesa_validate_pbo_access(GLuint dimensions,
     */
    if (!_mesa_is_bufferobj(pack->BufferObj)) {
       offset = 0;
-      size = clientMemSize;
+      size = (clientMemSize == INT_MAX) ? UINTPTR_MAX : clientMemSize;
    } else {
       offset = (uintptr_t)ptr;
       size = pack->BufferObj->Size;
@@ -164,23 +164,18 @@ _mesa_map_pbo_source(struct gl_context *ctx,
    return buf;
 }
 
-
 /**
- * Combine PBO-read validation and mapping.
- * If any GL errors are detected, they'll be recorded and NULL returned.
+ * Perform PBO validation for read operations with uncompressed textures.
+ * If any GL errors are detected, false is returned, otherwise returns true.
  * \sa _mesa_validate_pbo_access
- * \sa _mesa_map_pbo_source
- * A call to this function should have a matching call to
- * _mesa_unmap_pbo_source().
  */
-const GLvoid *
-_mesa_map_validate_pbo_source(struct gl_context *ctx,
-                              GLuint dimensions,
-                              const struct gl_pixelstore_attrib *unpack,
-                              GLsizei width, GLsizei height, GLsizei depth,
-                              GLenum format, GLenum type,
-                              GLsizei clientMemSize,
-                              const GLvoid *ptr, const char *where)
+bool
+_mesa_validate_pbo_source(struct gl_context *ctx, GLuint dimensions,
+                          const struct gl_pixelstore_attrib *unpack,
+                          GLsizei width, GLsizei height, GLsizei depth,
+                          GLenum format, GLenum type,
+                          GLsizei clientMemSize,
+                          const GLvoid *ptr, const char *where)
 {
    assert(dimensions == 1 || dimensions == 2 || dimensions == 3);
 
@@ -188,24 +183,85 @@ _mesa_map_validate_pbo_source(struct gl_context *ctx,
                                   format, type, clientMemSize, ptr)) {
       if (_mesa_is_bufferobj(unpack->BufferObj)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s(out of bounds PBO access)", where);
+                     "%s(out of bounds PBO access)",
+                     where);
       } else {
          _mesa_error(ctx, GL_INVALID_OPERATION,
                      "%s(out of bounds access: bufSize (%d) is too small)",
                      where, clientMemSize);
       }
-      return NULL;
+      return false;
    }
 
    if (!_mesa_is_bufferobj(unpack->BufferObj)) {
       /* non-PBO access: no further validation to be done */
-      return ptr;
+      return true;
    }
 
    if (_mesa_check_disallowed_mapping(unpack->BufferObj)) {
       /* buffer is already mapped - that's an error */
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(PBO is mapped)", where);
-      return NULL;
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(PBO is mapped)",
+                  where);
+      return false;
+   }
+
+   return true;
+}
+
+/**
+ * Perform PBO validation for read operations with compressed textures.
+ * If any GL errors are detected, false is returned, otherwise returns true.
+ */
+bool
+_mesa_validate_pbo_source_compressed(struct gl_context *ctx, GLuint dimensions,
+                                     const struct gl_pixelstore_attrib *unpack,
+                                     GLsizei imageSize, const GLvoid *pixels,
+                                     const char *where)
+{
+   if (!_mesa_is_bufferobj(unpack->BufferObj)) {
+      /* not using a PBO */
+      return true;
+   }
+
+   if ((const GLubyte *) pixels + imageSize >
+       ((const GLubyte *) 0) + unpack->BufferObj->Size) {
+      /* out of bounds read! */
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(invalid PBO access)",
+                  where);
+      return false;
+   }
+
+   if (_mesa_check_disallowed_mapping(unpack->BufferObj)) {
+      /* buffer is already mapped - that's an error */
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(PBO is mapped)",
+                  where);
+      return false;
+   }
+
+   return true;
+}
+
+/**
+ * Perform PBO-read mapping.
+ * If any GL errors are detected, they'll be recorded and NULL returned.
+ * \sa _mesa_validate_pbo_source
+ * \sa _mesa_map_pbo_source
+ * A call to this function should have a matching call to
+ * _mesa_unmap_pbo_source().
+ */
+const GLvoid *
+_mesa_map_validate_pbo_source(struct gl_context *ctx,
+                              GLuint dimensions,
+                              const struct gl_pixelstore_attrib *unpack,
+                              GLsizei width, GLsizei height, GLsizei depth,
+                              GLenum format, GLenum type,
+                              GLsizei clientMemSize,
+                              const GLvoid *ptr, const char *where)
+{
+   if (!_mesa_validate_pbo_source(ctx, dimensions, unpack,
+                                  width, height, depth, format, type,
+                                  clientMemSize, ptr, where)) {
+     return NULL;
    }
 
    ptr = _mesa_map_pbo_source(ctx, unpack, ptr);
@@ -381,28 +437,27 @@ _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
 {
    GLubyte *buf;
 
+   if (!_mesa_validate_pbo_source_compressed(ctx, dimensions, packing,
+                                             imageSize, pixels, funcName)) {
+     /* error is already set during validation */
+      return NULL;
+   }
+
    if (!_mesa_is_bufferobj(packing->BufferObj)) {
       /* not using a PBO - return pointer unchanged */
       return pixels;
    }
-   if ((const GLubyte *) pixels + imageSize >
-       ((const GLubyte *) 0) + packing->BufferObj->Size) {
-      /* out of bounds read! */
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(invalid PBO access)",
-                  funcName, dimensions);
-      return NULL;
-   }
 
    buf = (GLubyte*) ctx->Driver.MapBufferRange(ctx, 0,
 					       packing->BufferObj->Size,
 					       GL_MAP_READ_BIT,
 					       packing->BufferObj,
                                                MAP_INTERNAL);
-   if (!buf) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(PBO is mapped)", funcName,
-                  dimensions);
-      return NULL;
-   }
+
+   /* Validation above already checked that PBO is not mapped, so buffer
+    * should not be null.
+    */
+   assert(buf);
 
    return ADD_POINTERS(buf, pixels);
 }
diff --git a/mesalib/src/mesa/main/pbo.h b/mesalib/src/mesa/main/pbo.h
index 9851ef1a1..b3f24e62b 100644
--- a/mesalib/src/mesa/main/pbo.h
+++ b/mesalib/src/mesa/main/pbo.h
@@ -92,4 +92,18 @@ _mesa_unmap_teximage_pbo(struct gl_context *ctx,
                          const struct gl_pixelstore_attrib *unpack);
 
 
+extern bool
+_mesa_validate_pbo_source(struct gl_context *ctx, GLuint dimensions,
+                          const struct gl_pixelstore_attrib *unpack,
+                          GLsizei width, GLsizei height, GLsizei depth,
+                          GLenum format, GLenum type,
+                          GLsizei clientMemSize,
+                          const GLvoid *ptr, const char *where);
+
+extern bool
+_mesa_validate_pbo_source_compressed(struct gl_context *ctx, GLuint dimensions,
+                                     const struct gl_pixelstore_attrib *unpack,
+                                     GLsizei imageSize, const GLvoid *ptr,
+                                     const char *where);
+
 #endif
diff --git a/mesalib/src/mesa/main/querymatrix.c b/mesalib/src/mesa/main/querymatrix.c
index ef8517571..18361c929 100644
--- a/mesalib/src/mesa/main/querymatrix.c
+++ b/mesalib/src/mesa/main/querymatrix.c
@@ -13,7 +13,7 @@
 
 
 #include <stdlib.h>
-#include <math.h>
+#include "c99_math.h"
 #include "glheader.h"
 #include "querymatrix.h"
 #include "main/get.h"
@@ -37,169 +37,120 @@
 #define INT_TO_FIXED(x) ((GLfixed) ((x) << 16))
 #define FLOAT_TO_FIXED(x) ((GLfixed) ((x) * 65536.0))
 
-#if defined(fpclassify)
-/* ISO C99 says that fpclassify is a macro.  Assume that any implementation
- * of fpclassify, whether it's in a C99 compiler or not, will be a macro.
- */
-#elif defined(_MSC_VER)
-/* Not required on VS2013 and above. */
-/* Oddly, the fpclassify() function doesn't exist in such a form
- * on MSVC.  This is an implementation using slightly different
- * lower-level Windows functions.
- */
-#include <float.h>
-
-enum {FP_NAN, FP_INFINITE, FP_ZERO, FP_SUBNORMAL, FP_NORMAL}
-fpclassify(double x)
-{
-    switch(_fpclass(x)) {
-        case _FPCLASS_SNAN: /* signaling NaN */
-        case _FPCLASS_QNAN: /* quiet NaN */
-            return FP_NAN;
-        case _FPCLASS_NINF: /* negative infinity */
-        case _FPCLASS_PINF: /* positive infinity */
-            return FP_INFINITE;
-        case _FPCLASS_NN:   /* negative normal */
-        case _FPCLASS_PN:   /* positive normal */
-            return FP_NORMAL;
-        case _FPCLASS_ND:   /* negative denormalized */
-        case _FPCLASS_PD:   /* positive denormalized */
-            return FP_SUBNORMAL;
-        case _FPCLASS_NZ:   /* negative zero */
-        case _FPCLASS_PZ:   /* positive zero */
-            return FP_ZERO;
-        default:
-            /* Should never get here; but if we do, this will guarantee
-             * that the pattern is not treated like a number.
-             */
-            return FP_NAN;
-    }
-}
-
-#else
-
-enum {FP_NAN, FP_INFINITE, FP_ZERO, FP_SUBNORMAL, FP_NORMAL}
-fpclassify(double x)
-{
-   /* XXX do something better someday */
-   return FP_NORMAL;
-}
-
-#endif
 
-GLbitfield GLAPIENTRY _mesa_QueryMatrixxOES(GLfixed mantissa[16], GLint exponent[16])
+GLbitfield GLAPIENTRY
+_mesa_QueryMatrixxOES(GLfixed mantissa[16], GLint exponent[16])
 {
-    GLfloat matrix[16];
-    GLint tmp;
-    GLenum currentMode = GL_FALSE;
-    GLenum desiredMatrix = GL_FALSE;
-    /* The bitfield returns 1 for each component that is invalid (i.e.
-     * NaN or Inf).  In case of error, everything is invalid.
-     */
-    GLbitfield rv;
-    register unsigned int i;
-    unsigned int bit;
-
-    /* This data structure defines the mapping between the current matrix
-     * mode and the desired matrix identifier.
-     */
-    static struct {
-        GLenum currentMode;
-        GLenum desiredMatrix;
-    } modes[] = {
-        {GL_MODELVIEW, GL_MODELVIEW_MATRIX},
-        {GL_PROJECTION, GL_PROJECTION_MATRIX},
-        {GL_TEXTURE, GL_TEXTURE_MATRIX},
-    };
-
-    /* Call Mesa to get the current matrix in floating-point form.  First,
-     * we have to figure out what the current matrix mode is.
-     */
-    _mesa_GetIntegerv(GL_MATRIX_MODE, &tmp);
-    currentMode = (GLenum) tmp;
-
-    /* The mode is either GL_FALSE, if for some reason we failed to query
-     * the mode, or a given mode from the above table.  Search for the
-     * returned mode to get the desired matrix; if we don't find it,
-     * we can return immediately, as _mesa_GetInteger() will have
-     * logged the necessary error already.
-     */
-    for (i = 0; i < sizeof(modes)/sizeof(modes[0]); i++) {
-        if (modes[i].currentMode == currentMode) {
-            desiredMatrix = modes[i].desiredMatrix;
-            break;
-        }
-    }
-    if (desiredMatrix == GL_FALSE) {
-        /* Early error means all values are invalid. */
-        return 0xffff;
-    }
-
-    /* Now pull the matrix itself. */
-    _mesa_GetFloatv(desiredMatrix, matrix);
-
-    rv = 0;
-    for (i = 0, bit = 1; i < 16; i++, bit<<=1) {
-        float normalizedFraction;
-        int exp;
-
-        switch (fpclassify(matrix[i])) {
-            /* A "subnormal" or denormalized number is too small to be
-             * represented in normal format; but despite that it's a
-             * valid floating point number.  FP_ZERO and FP_NORMAL
-             * are both valid as well.  We should be fine treating
-             * these three cases as legitimate floating-point numbers.
-             */
-            case FP_SUBNORMAL:
-            case FP_NORMAL:
-            case FP_ZERO:
-                normalizedFraction = (GLfloat)frexp(matrix[i], &exp);
-                mantissa[i] = FLOAT_TO_FIXED(normalizedFraction);
-                exponent[i] = (GLint) exp;
-                break;
-
-            /* If the entry is not-a-number or an infinity, then the
-             * matrix component is invalid.  The invalid flag for
-             * the component is already set; might as well set the
-             * other return values to known values.  We'll set
-             * distinct values so that a savvy end user could determine
-             * whether the matrix component was a NaN or an infinity,
-             * but this is more useful for debugging than anything else
-             * since the standard doesn't specify any such magic
-             * values to return.
-             */
-            case FP_NAN:
-                mantissa[i] = INT_TO_FIXED(0);
-                exponent[i] = (GLint) 0;
-                rv |= bit;
-                break;
-
-            case FP_INFINITE:
-                /* Return +/- 1 based on whether it's a positive or
-                 * negative infinity.
-                 */
-                if (matrix[i] > 0) {
-                    mantissa[i] = INT_TO_FIXED(1);
-                }
-                else {
-                    mantissa[i] = -INT_TO_FIXED(1);
-                }
-                exponent[i] = (GLint) 0;
-                rv |= bit;
-                break;
-
-            /* We should never get here; but here's a catching case
-             * in case fpclassify() is returnings something unexpected.
-             */
-            default:
-                mantissa[i] = INT_TO_FIXED(2);
-                exponent[i] = (GLint) 0;
-                rv |= bit;
-                break;
-        }
-
-    } /* for each component */
-
-    /* All done */
-    return rv;
+   GLfloat matrix[16];
+   GLint tmp;
+   GLenum currentMode = GL_FALSE;
+   GLenum desiredMatrix = GL_FALSE;
+   /* The bitfield returns 1 for each component that is invalid (i.e.
+    * NaN or Inf).  In case of error, everything is invalid.
+    */
+   GLbitfield rv;
+   unsigned i, bit;
+
+   /* This data structure defines the mapping between the current matrix
+    * mode and the desired matrix identifier.
+    */
+   static const struct {
+      GLenum currentMode;
+      GLenum desiredMatrix;
+   } modes[] = {
+      {GL_MODELVIEW, GL_MODELVIEW_MATRIX},
+      {GL_PROJECTION, GL_PROJECTION_MATRIX},
+      {GL_TEXTURE, GL_TEXTURE_MATRIX},
+   };
+
+   /* Call Mesa to get the current matrix in floating-point form.  First,
+    * we have to figure out what the current matrix mode is.
+    */
+   _mesa_GetIntegerv(GL_MATRIX_MODE, &tmp);
+   currentMode = (GLenum) tmp;
+
+   /* The mode is either GL_FALSE, if for some reason we failed to query
+    * the mode, or a given mode from the above table.  Search for the
+    * returned mode to get the desired matrix; if we don't find it,
+    * we can return immediately, as _mesa_GetInteger() will have
+    * logged the necessary error already.
+    */
+   for (i = 0; i < ARRAY_SIZE(modes); i++) {
+      if (modes[i].currentMode == currentMode) {
+         desiredMatrix = modes[i].desiredMatrix;
+         break;
+      }
+   }
+   if (desiredMatrix == GL_FALSE) {
+      /* Early error means all values are invalid. */
+      return 0xffff;
+   }
+
+   /* Now pull the matrix itself. */
+   _mesa_GetFloatv(desiredMatrix, matrix);
+
+   rv = 0;
+   for (i = 0, bit = 1; i < 16; i++, bit<<=1) {
+      float normalizedFraction;
+      int exp;
+
+      switch (fpclassify(matrix[i])) {
+      case FP_SUBNORMAL:
+      case FP_NORMAL:
+      case FP_ZERO:
+         /* A "subnormal" or denormalized number is too small to be
+          * represented in normal format; but despite that it's a
+          * valid floating point number.  FP_ZERO and FP_NORMAL
+          * are both valid as well.  We should be fine treating
+          * these three cases as legitimate floating-point numbers.
+          */
+         normalizedFraction = (GLfloat)frexp(matrix[i], &exp);
+         mantissa[i] = FLOAT_TO_FIXED(normalizedFraction);
+         exponent[i] = (GLint) exp;
+         break;
+
+      case FP_NAN:
+         /* If the entry is not-a-number or an infinity, then the
+          * matrix component is invalid.  The invalid flag for
+          * the component is already set; might as well set the
+          * other return values to known values.  We'll set
+          * distinct values so that a savvy end user could determine
+          * whether the matrix component was a NaN or an infinity,
+          * but this is more useful for debugging than anything else
+          * since the standard doesn't specify any such magic
+          * values to return.
+          */
+         mantissa[i] = INT_TO_FIXED(0);
+         exponent[i] = (GLint) 0;
+         rv |= bit;
+         break;
+
+      case FP_INFINITE:
+         /* Return +/- 1 based on whether it's a positive or
+          * negative infinity.
+          */
+         if (matrix[i] > 0) {
+            mantissa[i] = INT_TO_FIXED(1);
+         }
+         else {
+            mantissa[i] = -INT_TO_FIXED(1);
+         }
+         exponent[i] = (GLint) 0;
+         rv |= bit;
+         break;
+
+      default:
+         /* We should never get here; but here's a catching case
+          * in case fpclassify() is returnings something unexpected.
+          */
+         mantissa[i] = INT_TO_FIXED(2);
+         exponent[i] = (GLint) 0;
+         rv |= bit;
+         break;
+      }
+
+   } /* for each component */
+
+   /* All done */
+   return rv;
 }
diff --git a/mesalib/src/mesa/main/scissor.c b/mesalib/src/mesa/main/scissor.c
index 83f39e2a0..bc8224c7d 100644
--- a/mesalib/src/mesa/main/scissor.c
+++ b/mesalib/src/mesa/main/scissor.c
@@ -201,13 +201,13 @@ void GLAPIENTRY
 _mesa_ScissorIndexed(GLuint index, GLint left, GLint bottom,
                      GLsizei width, GLsizei height)
 {
-   ScissorIndexed(index, left, bottom, width, height, "glScissorIndexd");
+   ScissorIndexed(index, left, bottom, width, height, "glScissorIndexed");
 }
 
 void GLAPIENTRY
 _mesa_ScissorIndexedv(GLuint index, const GLint *v)
 {
-   ScissorIndexed(index, v[0], v[1], v[2], v[3], "glScissorIndexdv");
+   ScissorIndexed(index, v[0], v[1], v[2], v[3], "glScissorIndexedv");
 }
 
 /**
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index 5731d581a..30716f5e3 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -1027,15 +1027,14 @@ _mesa_active_program(struct gl_context *ctx, struct gl_shader_program *shProg,
 
 
 static void
-use_shader_program(struct gl_context *ctx, GLenum type,
+use_shader_program(struct gl_context *ctx, gl_shader_stage stage,
                    struct gl_shader_program *shProg,
                    struct gl_pipeline_object *shTarget)
 {
    struct gl_shader_program **target;
-   gl_shader_stage stage = _mesa_shader_enum_to_shader_stage(type);
 
    target = &shTarget->CurrentProgram[stage];
-   if ((shProg == NULL) || (shProg->_LinkedShaders[stage] == NULL))
+   if ((shProg != NULL) && (shProg->_LinkedShaders[stage] == NULL))
       shProg = NULL;
 
    if (*target != shProg) {
@@ -1048,17 +1047,17 @@ use_shader_program(struct gl_context *ctx, GLenum type,
        * it from that binding point as well.  This ensures that the correct
        * semantics of glDeleteProgram are maintained.
        */
-      switch (type) {
-      case GL_VERTEX_SHADER:
+      switch (stage) {
+      case MESA_SHADER_VERTEX:
 	 /* Empty for now. */
 	 break;
-      case GL_GEOMETRY_SHADER_ARB:
+      case MESA_SHADER_GEOMETRY:
 	 /* Empty for now. */
 	 break;
-      case GL_COMPUTE_SHADER:
+      case MESA_SHADER_COMPUTE:
          /* Empty for now. */
          break;
-      case GL_FRAGMENT_SHADER:
+      case MESA_SHADER_FRAGMENT:
          if (*target == ctx->_Shader->_CurrentFragmentProgram) {
 	    _mesa_reference_shader_program(ctx,
                                            &ctx->_Shader->_CurrentFragmentProgram,
@@ -1079,10 +1078,9 @@ use_shader_program(struct gl_context *ctx, GLenum type,
 void
 _mesa_use_program(struct gl_context *ctx, struct gl_shader_program *shProg)
 {
-   use_shader_program(ctx, GL_VERTEX_SHADER, shProg, &ctx->Shader);
-   use_shader_program(ctx, GL_GEOMETRY_SHADER_ARB, shProg, &ctx->Shader);
-   use_shader_program(ctx, GL_FRAGMENT_SHADER, shProg, &ctx->Shader);
-   use_shader_program(ctx, GL_COMPUTE_SHADER, shProg, &ctx->Shader);
+   int i;
+   for (i = 0; i < MESA_SHADER_STAGES; i++)
+      use_shader_program(ctx, i, shProg, &ctx->Shader);
    _mesa_active_program(ctx, shProg, "glUseProgram");
 
    if (ctx->Driver.UseProgram)
@@ -1460,7 +1458,7 @@ read_shader(const char *fname)
 
    fclose(f);
 
-   shader = _mesa_strdup(buffer);
+   shader = strdup(buffer);
    free(buffer);
 
    return shader;
@@ -1889,7 +1887,8 @@ _mesa_use_shader_program(struct gl_context *ctx, GLenum type,
                          struct gl_shader_program *shProg,
                          struct gl_pipeline_object *shTarget)
 {
-   use_shader_program(ctx, type, shProg, shTarget);
+   gl_shader_stage stage = _mesa_shader_enum_to_shader_stage(type);
+   use_shader_program(ctx, stage, shProg, shTarget);
 
    if (ctx->Driver.UseProgram)
       ctx->Driver.UseProgram(ctx, shProg);
diff --git a/mesalib/src/mesa/main/texgetimage.c b/mesalib/src/mesa/main/texgetimage.c
index f975c160e..255d36559 100644
--- a/mesalib/src/mesa/main/texgetimage.c
+++ b/mesalib/src/mesa/main/texgetimage.c
@@ -1088,48 +1088,9 @@ _mesa_GetTextureImage(GLuint texture, GLint level, GLenum format,
    /* Must handle special case GL_TEXTURE_CUBE_MAP. */
    if (texObj->Target == GL_TEXTURE_CUBE_MAP) {
 
-      /* Error checking */
-      if (texObj->NumLayers < 6) {
-         /* Not enough image planes for a cube map.  The spec does not say
-          * what should happen in this case because the user has always
-          * specified each cube face separately (using
-          * GL_TEXTURE_CUBE_MAP_POSITIVE_X+i) in previous GL versions.
-          * This is addressed in Khronos Bug 13223.
-          */
-         _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "glGetTextureImage(insufficient cube map storage)");
-         return;
-      }
-
-      /*
-       * What do we do if the user created a texture with the following code
-       * and then called this function with its handle?
-       *
-       *    GLuint tex;
-       *    glCreateTextures(GL_TEXTURE_CUBE_MAP, 1, &tex);
-       *    glBindTexture(GL_TEXTURE_CUBE_MAP, tex);
-       *    glTexImage2D(GL_TEXTURE_CUBE_MAP_POSITIVE_X, 0, ...);
-       *    glTexImage2D(GL_TEXTURE_CUBE_MAP_NEGATIVE_X, 0, ...);
-       *    glTexImage2D(GL_TEXTURE_CUBE_MAP_POSITIVE_Y, 0, ...);
-       *    // Note: GL_TEXTURE_CUBE_MAP_NEGATIVE_Y not set, or given the
-       *    // wrong format, or given the wrong size, etc.
-       *    glTexImage2D(GL_TEXTURE_CUBE_MAP_POSITIVE_Z, 0, ...);
-       *    glTexImage2D(GL_TEXTURE_CUBE_MAP_NEGATIVE_Z, 0, ...);
-       *
-       * A bug has been filed against the spec for this case.  In the
-       * meantime, we will check for cube completeness.
-       *
-       * According to Section 8.17 Texture Completeness in the OpenGL 4.5
-       * Core Profile spec (30.10.2014):
-       *    "[A] cube map texture is cube complete if the
-       *    following conditions all hold true: The [base level] texture
-       *    images of each of the six cube map faces have identical, positive,
-       *    and square dimensions. The [base level] images were each specified
-       *    with the same internal format."
-       *
-       * It seems reasonable to check for cube completeness of an arbitrary
-       * level here so that the returned data has a consistent format and size
-       * and therefore fits in the user's buffer.
+      /* Make sure the texture object is a proper cube.
+       * (See texturesubimage in teximage.c for details on why this check is
+       * performed.)
        */
       if (!_mesa_cube_level_complete(texObj, level)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
@@ -1140,6 +1101,8 @@ _mesa_GetTextureImage(GLuint texture, GLint level, GLenum format,
       /* Copy each face. */
       for (i = 0; i < 6; ++i) {
          texImage = texObj->Image[i][level];
+         assert(texImage);
+
          _mesa_get_texture_image(ctx, texObj, texImage, texObj->Target, level,
                                  format, type, bufSize, pixels, true);
 
@@ -1340,13 +1303,21 @@ _mesa_GetCompressedTextureImage(GLuint texture, GLint level,
 
    /* Must handle special case GL_TEXTURE_CUBE_MAP. */
    if (texObj->Target == GL_TEXTURE_CUBE_MAP) {
-      assert(texObj->NumLayers >= 6);
+
+      /* Make sure the texture object is a proper cube.
+       * (See texturesubimage in teximage.c for details on why this check is
+       * performed.)
+       */
+      if (!_mesa_cube_level_complete(texObj, level)) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glGetCompressedTextureImage(cube map incomplete)");
+         return;
+      }
 
       /* Copy each face. */
       for (i = 0; i < 6; ++i) {
          texImage = texObj->Image[i][level];
-         if (!texImage)
-            return;
+         assert(texImage);
 
          _mesa_get_compressed_texture_image(ctx, texObj, texImage,
                                             texObj->Target, level,
diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c
index 611d664b6..8d9d7cfc1 100644
--- a/mesalib/src/mesa/main/teximage.c
+++ b/mesalib/src/mesa/main/teximage.c
@@ -53,6 +53,7 @@
 #include "mtypes.h"
 #include "glformats.h"
 #include "texstore.h"
+#include "pbo.h"
 
 
 /**
@@ -1619,32 +1620,30 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
    /* Check size */
    if (subWidth < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "%s%dD(width=%d)", func, dims, subWidth);
+                  "%s(width=%d)", func, subWidth);
       return GL_TRUE;
    }
 
    if (dims > 1 && subHeight < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "%s%dD(height=%d)", func, dims, subHeight);
+                  "%s(height=%d)", func, subHeight);
       return GL_TRUE;
    }
 
    if (dims > 2 && subDepth < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "%s%dD(depth=%d)", func, dims, subDepth);
+                  "%s(depth=%d)", func, subDepth);
       return GL_TRUE;
    }
 
    /* check xoffset and width */
    if (xoffset < - (GLint) destImage->Border) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "%s%dD(xoffset)",
-                  func, dims);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(xoffset)", func);
       return GL_TRUE;
    }
 
    if (xoffset + subWidth > (GLint) destImage->Width) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "%s%dD(xoffset+width)",
-                  func, dims);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(xoffset+width)", func);
       return GL_TRUE;
    }
 
@@ -1652,13 +1651,11 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
    if (dims > 1) {
       GLint yBorder = (target == GL_TEXTURE_1D_ARRAY) ? 0 : destImage->Border;
       if (yoffset < -yBorder) {
-         _mesa_error(ctx, GL_INVALID_VALUE, "%s%dD(yoffset)",
-                     func, dims);
+         _mesa_error(ctx, GL_INVALID_VALUE, "%s(yoffset)", func);
          return GL_TRUE;
       }
       if (yoffset + subHeight > (GLint) destImage->Height) {
-         _mesa_error(ctx, GL_INVALID_VALUE, "%s%dD(yoffset+height)",
-                     func, dims);
+         _mesa_error(ctx, GL_INVALID_VALUE, "%s(yoffset+height)", func);
          return GL_TRUE;
       }
    }
@@ -1671,7 +1668,7 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
                          0 : destImage->Border;
 
       if (zoffset < -zBorder) {
-         _mesa_error(ctx, GL_INVALID_VALUE, "%s3D(zoffset)", func);
+         _mesa_error(ctx, GL_INVALID_VALUE, "%s(zoffset)", func);
          return GL_TRUE;
       }
 
@@ -1679,7 +1676,7 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
       if (target == GL_TEXTURE_CUBE_MAP)
          depth = 6;
       if (zoffset + subDepth  > depth) {
-         _mesa_error(ctx, GL_INVALID_VALUE, "%s3D(zoffset+depth)", func);
+         _mesa_error(ctx, GL_INVALID_VALUE, "%s(zoffset+depth)", func);
          return GL_TRUE;
       }
    }
@@ -1697,8 +1694,8 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
       /* offset must be multiple of block size */
       if ((xoffset % bw != 0) || (yoffset % bh != 0)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s%dD(xoffset = %d, yoffset = %d)",
-                     func, dims, xoffset, yoffset);
+                     "%s(xoffset = %d, yoffset = %d)",
+                     func, xoffset, yoffset);
          return GL_TRUE;
       }
 
@@ -1710,14 +1707,14 @@ error_check_subtexture_dimensions(struct gl_context *ctx, GLuint dims,
       if ((subWidth % bw != 0) &&
           (xoffset + subWidth != (GLint) destImage->Width)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s%dD(width = %d)", func, dims, subWidth);
+                     "%s(width = %d)", func, subWidth);
          return GL_TRUE;
       }
 
       if ((subHeight % bh != 0) &&
           (yoffset + subHeight != (GLint) destImage->Height)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s%dD(height = %d)", func, dims, subHeight);
+                     "%s(height = %d)", func, subHeight);
          return GL_TRUE;
       }
    }
@@ -2113,7 +2110,8 @@ texture_error_check( struct gl_context *ctx,
                      GLint level, GLint internalFormat,
                      GLenum format, GLenum type,
                      GLint width, GLint height,
-                     GLint depth, GLint border )
+                     GLint depth, GLint border,
+                     const GLvoid *pixels )
 {
    GLenum err;
 
@@ -2198,6 +2196,13 @@ texture_error_check( struct gl_context *ctx,
       return GL_TRUE;
    }
 
+   /* validate the bound PBO, if any */
+   if (!_mesa_validate_pbo_source(ctx, dimensions, &ctx->Unpack,
+                                  width, height, depth, format, type,
+                                  INT_MAX, pixels, "glTexImage")) {
+      return GL_TRUE;
+   }
+
    /* make sure internal format and format basically agree */
    if (!texture_formats_agree(internalFormat, format)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
@@ -2294,7 +2299,7 @@ compressed_texture_error_check(struct gl_context *ctx, GLint dimensions,
                                GLenum target, GLint level,
                                GLenum internalFormat, GLsizei width,
                                GLsizei height, GLsizei depth, GLint border,
-                               GLsizei imageSize)
+                               GLsizei imageSize, const GLvoid *data)
 {
    const GLint maxLevels = _mesa_max_texture_levels(ctx, target);
    GLint expectedSize;
@@ -2322,6 +2327,13 @@ compressed_texture_error_check(struct gl_context *ctx, GLint dimensions,
       return GL_TRUE;
    }
 
+   /* validate the bound PBO, if any */
+   if (!_mesa_validate_pbo_source_compressed(ctx, dimensions, &ctx->Unpack,
+                                             imageSize, data,
+                                             "glCompressedTexImage")) {
+      return GL_TRUE;
+   }
+
    switch (internalFormat) {
    case GL_PALETTE4_RGB8_OES:
    case GL_PALETTE4_RGBA8_OES:
@@ -2454,30 +2466,28 @@ texsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
                         GLenum target, GLint level,
                         GLint xoffset, GLint yoffset, GLint zoffset,
                         GLint width, GLint height, GLint depth,
-                        GLenum format, GLenum type, bool dsa)
+                        GLenum format, GLenum type, const GLvoid *pixels,
+                        bool dsa, const char *callerName)
 {
    struct gl_texture_image *texImage;
    GLenum err;
-   const char* suffix = dsa ? "ture" : "";
 
    if (!texObj) {
       /* must be out of memory */
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glTex%sSubImage%dD()",
-                  suffix, dimensions);
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s()", callerName);
       return GL_TRUE;
    }
 
    /* check target (proxies not allowed) */
    if (!legal_texsubimage_target(ctx, dimensions, target, dsa)) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glTex%sSubImage%uD(target=%s)",
-                  suffix, dimensions, _mesa_lookup_enum_by_nr(target));
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(target=%s)",
+                  callerName, _mesa_lookup_enum_by_nr(target));
       return GL_TRUE;
    }
 
    /* level check */
    if (level < 0 || level >= _mesa_max_texture_levels(ctx, target)) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glTex%sSubImage%uD(level=%d)",
-                  suffix, dimensions, level);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(level=%d)", callerName, level);
       return GL_TRUE;
    }
 
@@ -2489,9 +2499,8 @@ texsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
    if (_mesa_is_gles(ctx) && !_mesa_is_gles3(ctx)) {
       err = _mesa_es_error_check_format_and_type(format, type, dimensions);
       if (err != GL_NO_ERROR) {
-         _mesa_error(ctx, err,
-                     "glTex%sSubImage%dD(format = %s, type = %s)",
-                     suffix, dimensions, _mesa_lookup_enum_by_nr(format),
+         _mesa_error(ctx, err, "%s(format = %s, type = %s)",
+                     callerName, _mesa_lookup_enum_by_nr(format),
                      _mesa_lookup_enum_by_nr(type));
          return GL_TRUE;
       }
@@ -2500,34 +2509,37 @@ texsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
    err = _mesa_error_check_format_and_type(ctx, format, type);
    if (err != GL_NO_ERROR) {
       _mesa_error(ctx, err,
-                  "glTex%sSubImage%dD(incompatible format = %s, type = %s)",
-                  suffix, dimensions, _mesa_lookup_enum_by_nr(format),
+                  "%s(incompatible format = %s, type = %s)",
+                  callerName, _mesa_lookup_enum_by_nr(format),
                   _mesa_lookup_enum_by_nr(type));
       return GL_TRUE;
    }
 
+   /* validate the bound PBO, if any */
+   if (!_mesa_validate_pbo_source(ctx, dimensions, &ctx->Unpack,
+                                  width, height, depth, format, type,
+                                  INT_MAX, pixels, callerName)) {
+      return GL_TRUE;
+   }
+
    texImage = _mesa_select_tex_image(texObj, target, level);
    if (!texImage) {
       /* non-existant texture level */
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glTex%sSubImage%dD(invalid texture image)", suffix,
-                  dimensions);
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(invalid texture image)",
+                  callerName);
       return GL_TRUE;
    }
 
    if (error_check_subtexture_dimensions(ctx, dimensions,
                                          texImage, xoffset, yoffset, zoffset,
-                                         width, height, depth,
-                                         dsa ? "glTextureSubImage" :
-                                         "glTexSubImage")) {
+                                         width, height, depth, callerName)) {
       return GL_TRUE;
    }
 
    if (_mesa_is_format_compressed(texImage->TexFormat)) {
       if (compressedteximage_only_format(ctx, texImage->InternalFormat)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-               "glTex%sSubImage%dD(no compression for format)",
-               suffix, dimensions);
+               "%s(no compression for format)", callerName);
          return GL_TRUE;
       }
    }
@@ -2537,8 +2549,7 @@ texsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
       if (_mesa_is_format_integer_color(texImage->TexFormat) !=
           _mesa_is_enum_format_integer(format)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "glTex%sSubImage%dD(integer/non-integer format mismatch)",
-                     suffix, dimensions);
+                     "%s(integer/non-integer format mismatch)", callerName);
          return GL_TRUE;
       }
    }
@@ -2815,10 +2826,9 @@ copytexsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
                             const struct gl_texture_object *texObj,
                             GLenum target, GLint level,
                             GLint xoffset, GLint yoffset, GLint zoffset,
-                            GLint width, GLint height, bool dsa)
+                            GLint width, GLint height, const char *caller)
 {
    struct gl_texture_image *texImage;
-   const char *suffix = dsa ? "ture" : "";
 
    /* Check that the source buffer is complete */
    if (_mesa_is_user_fbo(ctx->ReadBuffer)) {
@@ -2827,31 +2837,26 @@ copytexsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
       }
       if (ctx->ReadBuffer->_Status != GL_FRAMEBUFFER_COMPLETE_EXT) {
          _mesa_error(ctx, GL_INVALID_FRAMEBUFFER_OPERATION_EXT,
-                     "glCopyTex%sSubImage%dD(invalid readbuffer)",
-                     suffix, dimensions);
+                     "%s(invalid readbuffer)", caller);
          return GL_TRUE;
       }
 
       if (ctx->ReadBuffer->Visual.samples > 0) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                "glCopyTex%sSubImage%dD(multisample FBO)", suffix,
-                dimensions);
+                "%s(multisample FBO)", caller);
          return GL_TRUE;
       }
    }
 
    /* Check level */
    if (level < 0 || level >= _mesa_max_texture_levels(ctx, target)) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCopyTex%sSubImage%dD(level=%d)", suffix,
-                  dimensions, level);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(level=%d)", caller, level);
       return GL_TRUE;
    }
 
    /* Get dest image pointers */
    if (!texObj) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCopyTex%sSubImage%dD()",
-                  suffix, dimensions);
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s()", caller);
       return GL_TRUE;
    }
 
@@ -2859,37 +2864,33 @@ copytexsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
    if (!texImage) {
       /* destination image does not exist */
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCopyTex%sSubImage%dD(invalid texture image)",
-                  suffix, dimensions);
+                  "%s(invalid texture image)", caller);
       return GL_TRUE;
    }
 
    if (error_check_subtexture_dimensions(ctx, dimensions, texImage,
                                          xoffset, yoffset, zoffset,
-                                         width, height, 1, dsa ?
-                                         "glCompressedTextureSubImage" :
-                                         "glCompressedTexSubImage")) {
+                                         width, height, 1, caller)) {
       return GL_TRUE;
    }
 
    if (_mesa_is_format_compressed(texImage->TexFormat)) {
       if (compressedteximage_only_format(ctx, texImage->InternalFormat)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-               "glCopyTex%sSubImage%dD(no compression for format)",
-               suffix, dimensions);
+               "%s(no compression for format)", caller);
          return GL_TRUE;
       }
    }
 
    if (texImage->InternalFormat == GL_YCBCR_MESA) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glCopyTex%sSubImage2D", suffix);
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s()", caller);
       return GL_TRUE;
    }
 
    if (!_mesa_source_buffer_exists(ctx, texImage->_BaseFormat)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCopyTex%sSubImage%dD(missing readbuffer, format=0x%x)",
-                  suffix, dimensions, texImage->_BaseFormat);
+                  "%s(missing readbuffer, format=0x%x)", caller,
+                  texImage->_BaseFormat);
       return GL_TRUE;
    }
 
@@ -2907,8 +2908,7 @@ copytexsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
       if (_mesa_is_format_integer_color(rb->Format) !=
           _mesa_is_format_integer_color(texImage->TexFormat)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "glCopyTex%sSubImage%dD(integer vs non-integer)",
-                     suffix, dimensions);
+                     "%s(integer vs non-integer)", caller);
          return GL_TRUE;
       }
    }
@@ -3218,12 +3218,13 @@ teximage(struct gl_context *ctx, GLboolean compressed, GLuint dims,
       if (compressed_texture_error_check(ctx, dims, target, level,
                                          internalFormat,
                                          width, height, depth,
-                                         border, imageSize))
+                                         border, imageSize, pixels))
          return;
    }
    else {
       if (texture_error_check(ctx, dims, target, level, internalFormat,
-                              format, type, width, height, depth, border))
+                              format, type, width, height, depth, border,
+                              pixels))
          return;
    }
 
@@ -3562,7 +3563,8 @@ static void
 texsubimage(struct gl_context *ctx, GLuint dims, GLenum target, GLint level,
             GLint xoffset, GLint yoffset, GLint zoffset,
             GLsizei width, GLsizei height, GLsizei depth,
-            GLenum format, GLenum type, const GLvoid *pixels)
+            GLenum format, GLenum type, const GLvoid *pixels,
+            const char *callerName)
 {
    struct gl_texture_object *texObj;
    struct gl_texture_image *texImage;
@@ -3573,7 +3575,8 @@ texsubimage(struct gl_context *ctx, GLuint dims, GLenum target, GLint level,
 
    if (texsubimage_error_check(ctx, dims, texObj, target, level,
                                xoffset, yoffset, zoffset,
-                               width, height, depth, format, type, false)) {
+                               width, height, depth, format, type,
+                               pixels, false, callerName)) {
       return;   /* error was detected */
    }
 
@@ -3603,7 +3606,8 @@ texturesubimage(struct gl_context *ctx, GLuint dims,
                 GLuint texture, GLint level,
                 GLint xoffset, GLint yoffset, GLint zoffset,
                 GLsizei width, GLsizei height, GLsizei depth,
-                GLenum format, GLenum type, const GLvoid *pixels)
+                GLenum format, GLenum type, const GLvoid *pixels,
+                const char *callerName)
 {
    struct gl_texture_object *texObj;
    struct gl_texture_image *texImage;
@@ -3627,7 +3631,8 @@ texturesubimage(struct gl_context *ctx, GLuint dims,
 
    if (texsubimage_error_check(ctx, dims, texObj, texObj->Target, level,
                                xoffset, yoffset, zoffset,
-                               width, height, depth, format, type, true)) {
+                               width, height, depth, format, type,
+                               pixels, true, callerName)) {
       return;   /* error was detected */
    }
 
@@ -3636,20 +3641,6 @@ texturesubimage(struct gl_context *ctx, GLuint dims,
    if (texObj->Target == GL_TEXTURE_CUBE_MAP) {
       GLint rowStride;
 
-      /* Error checking */
-      if (texObj->NumLayers < 6) {
-         /* Not enough image planes for a cube map.  The spec does not say
-          * what should happen in this case because the user has always
-          * specified each cube face separately (using
-          * GL_TEXTURE_CUBE_MAP_POSITIVE_X+i) in previous GL versions.
-          * This is addressed in Khronos Bug 13223.
-          */
-         _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "glTextureSubImage%uD(insufficient cube map storage)",
-                     dims);
-         return;
-      }
-
       /*
        * What do we do if the user created a texture with the following code
        * and then called this function with its handle?
@@ -3691,6 +3682,8 @@ texturesubimage(struct gl_context *ctx, GLuint dims,
       /* Copy in each face. */
       for (i = 0; i < 6; ++i) {
          texImage = texObj->Image[i][level];
+         assert(texImage);
+
          _mesa_texture_sub_image(ctx, 3, texObj, texImage, texObj->Target,
                                  level, xoffset, yoffset, zoffset,
                                  width, height, 1, format,
@@ -3700,8 +3693,7 @@ texturesubimage(struct gl_context *ctx, GLuint dims,
    }
    else {
       texImage = _mesa_select_tex_image(texObj, texObj->Target, level);
-      if (!texImage)
-         return;
+      assert(texImage);
 
       _mesa_texture_sub_image(ctx, dims, texObj, texImage, texObj->Target,
                               level, xoffset, yoffset, zoffset,
@@ -3721,7 +3713,7 @@ _mesa_TexSubImage1D( GLenum target, GLint level,
    texsubimage(ctx, 1, target, level,
                xoffset, 0, 0,
                width, 1, 1,
-               format, type, pixels);
+               format, type, pixels, "glTexSubImage1D");
 }
 
 
@@ -3736,7 +3728,7 @@ _mesa_TexSubImage2D( GLenum target, GLint level,
    texsubimage(ctx, 2, target, level,
                xoffset, yoffset, 0,
                width, height, 1,
-               format, type, pixels);
+               format, type, pixels, "glTexSubImage2D");
 }
 
 
@@ -3752,7 +3744,7 @@ _mesa_TexSubImage3D( GLenum target, GLint level,
    texsubimage(ctx, 3, target, level,
                xoffset, yoffset, zoffset,
                width, height, depth,
-               format, type, pixels);
+               format, type, pixels, "glTexSubImage3D");
 }
 
 void GLAPIENTRY
@@ -3765,7 +3757,7 @@ _mesa_TextureSubImage1D(GLuint texture, GLint level,
    texturesubimage(ctx, 1, texture, level,
                    xoffset, 0, 0,
                    width, 1, 1,
-                   format, type, pixels);
+                   format, type, pixels, "glTextureSubImage1D");
 }
 
 
@@ -3780,7 +3772,7 @@ _mesa_TextureSubImage2D(GLuint texture, GLint level,
    texturesubimage(ctx, 2, texture, level,
                    xoffset, yoffset, 0,
                    width, height, 1,
-                   format, type, pixels);
+                   format, type, pixels, "glTextureSubImage2D");
 }
 
 
@@ -3795,7 +3787,7 @@ _mesa_TextureSubImage3D(GLuint texture, GLint level,
    texturesubimage(ctx, 3, texture, level,
                    xoffset, yoffset, zoffset,
                    width, height, depth,
-                   format, type, pixels);
+                   format, type, pixels, "glTextureSubImage3D");
 }
 
 
@@ -4040,15 +4032,14 @@ _mesa_copy_texture_sub_image(struct gl_context *ctx, GLuint dims,
                              GLint xoffset, GLint yoffset, GLint zoffset,
                              GLint x, GLint y,
                              GLsizei width, GLsizei height,
-                             bool dsa)
+                             const char *caller)
 {
    struct gl_texture_image *texImage;
 
    FLUSH_VERTICES(ctx, 0);
 
    if (MESA_VERBOSE & (VERBOSE_API|VERBOSE_TEXTURE))
-      _mesa_debug(ctx, "glCopyTex%sSubImage%uD %s %d %d %d %d %d %d %d %d\n",
-                  dsa ? "ture" : "", dims,
+      _mesa_debug(ctx, "%s %s %d %d %d %d %d %d %d %d\n", caller,
                   _mesa_lookup_enum_by_nr(target),
                   level, xoffset, yoffset, zoffset, x, y, width, height);
 
@@ -4057,7 +4048,7 @@ _mesa_copy_texture_sub_image(struct gl_context *ctx, GLuint dims,
 
    if (copytexsubimage_error_check(ctx, dims, texObj, target, level,
                                    xoffset, yoffset, zoffset,
-                                   width, height, dsa)) {
+                                   width, height, caller)) {
       return;
    }
 
@@ -4103,14 +4094,14 @@ _mesa_CopyTexSubImage1D( GLenum target, GLint level,
                          GLint xoffset, GLint x, GLint y, GLsizei width )
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTexSubImage1D";
    GET_CURRENT_CONTEXT(ctx);
 
    /* Check target (proxies not allowed). Target must be checked prior to
     * calling _mesa_get_current_tex_object.
     */
    if (!legal_texsubimage_target(ctx, 1, target, false)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTexSubImage1D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(target));
       return;
    }
@@ -4120,7 +4111,7 @@ _mesa_CopyTexSubImage1D( GLenum target, GLint level,
       return;
 
    _mesa_copy_texture_sub_image(ctx, 1, texObj, target, level, xoffset, 0, 0,
-                                x, y, width, 1, false);
+                                x, y, width, 1, self);
 }
 
 
@@ -4131,14 +4122,14 @@ _mesa_CopyTexSubImage2D( GLenum target, GLint level,
                          GLint x, GLint y, GLsizei width, GLsizei height )
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTexSubImage2D";
    GET_CURRENT_CONTEXT(ctx);
 
    /* Check target (proxies not allowed). Target must be checked prior to
     * calling _mesa_get_current_tex_object.
     */
    if (!legal_texsubimage_target(ctx, 2, target, false)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTexSubImage2D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(target));
       return;
    }
@@ -4149,7 +4140,7 @@ _mesa_CopyTexSubImage2D( GLenum target, GLint level,
 
    _mesa_copy_texture_sub_image(ctx, 2, texObj, target, level,
                                 xoffset, yoffset, 0,
-                                x, y, width, height, false);
+                                x, y, width, height, self);
 }
 
 
@@ -4160,14 +4151,14 @@ _mesa_CopyTexSubImage3D( GLenum target, GLint level,
                          GLint x, GLint y, GLsizei width, GLsizei height )
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTexSubImage3D";
    GET_CURRENT_CONTEXT(ctx);
 
    /* Check target (proxies not allowed). Target must be checked prior to
     * calling _mesa_get_current_tex_object.
     */
    if (!legal_texsubimage_target(ctx, 3, target, false)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTexSubImage3D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(target));
       return;
    }
@@ -4178,7 +4169,7 @@ _mesa_CopyTexSubImage3D( GLenum target, GLint level,
 
    _mesa_copy_texture_sub_image(ctx, 3, texObj, target, level,
                                 xoffset, yoffset, zoffset,
-                                x, y, width, height, false);
+                                x, y, width, height, self);
 }
 
 void GLAPIENTRY
@@ -4186,22 +4177,22 @@ _mesa_CopyTextureSubImage1D(GLuint texture, GLint level,
                             GLint xoffset, GLint x, GLint y, GLsizei width)
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTextureSubImage1D";
    GET_CURRENT_CONTEXT(ctx);
 
-   texObj = _mesa_lookup_texture_err(ctx, texture, "glCopyTextureSubImage1D");
+   texObj = _mesa_lookup_texture_err(ctx, texture, self);
    if (!texObj)
       return;
 
    /* Check target (proxies not allowed). */
    if (!legal_texsubimage_target(ctx, 1, texObj->Target, true)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTextureSubImage1D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(texObj->Target));
       return;
    }
 
    _mesa_copy_texture_sub_image(ctx, 1, texObj, texObj->Target, level,
-                                xoffset, 0, 0, x, y, width, 1, true);
+                                xoffset, 0, 0, x, y, width, 1, self);
 }
 
 void GLAPIENTRY
@@ -4210,23 +4201,23 @@ _mesa_CopyTextureSubImage2D(GLuint texture, GLint level,
                             GLint x, GLint y, GLsizei width, GLsizei height)
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTextureSubImage2D";
    GET_CURRENT_CONTEXT(ctx);
 
-   texObj = _mesa_lookup_texture_err(ctx, texture, "glCopyTextureSubImage2D");
+   texObj = _mesa_lookup_texture_err(ctx, texture, self);
    if (!texObj)
       return;
 
    /* Check target (proxies not allowed). */
    if (!legal_texsubimage_target(ctx, 2, texObj->Target, true)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTextureSubImage2D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(texObj->Target));
       return;
    }
 
    _mesa_copy_texture_sub_image(ctx, 2, texObj, texObj->Target, level,
                                 xoffset, yoffset, 0,
-                                x, y, width, height, true);
+                                x, y, width, height, self);
 }
 
 
@@ -4237,23 +4228,31 @@ _mesa_CopyTextureSubImage3D(GLuint texture, GLint level,
                             GLint x, GLint y, GLsizei width, GLsizei height)
 {
    struct gl_texture_object* texObj;
+   const char *self = "glCopyTextureSubImage3D";
    GET_CURRENT_CONTEXT(ctx);
 
-   texObj = _mesa_lookup_texture_err(ctx, texture, "glCopyTextureSubImage3D");
+   texObj = _mesa_lookup_texture_err(ctx, texture, self);
    if (!texObj)
       return;
 
    /* Check target (proxies not allowed). */
    if (!legal_texsubimage_target(ctx, 3, texObj->Target, true)) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCopyTextureSubImage3D(invalid target %s)",
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", self,
                   _mesa_lookup_enum_by_nr(texObj->Target));
       return;
    }
 
-   _mesa_copy_texture_sub_image(ctx, 3, texObj, texObj->Target, level,
-                                xoffset, yoffset, zoffset,
-                                x, y, width, height, true);
+   if (texObj->Target == GL_TEXTURE_CUBE_MAP) {
+      /* Act like CopyTexSubImage2D */
+      _mesa_copy_texture_sub_image(ctx, 2, texObj,
+                                   GL_TEXTURE_CUBE_MAP_POSITIVE_X + zoffset,
+                                   level, xoffset, yoffset, 0,
+                                   x, y, width, height, self);
+   }
+   else
+      _mesa_copy_texture_sub_image(ctx, 3, texObj, texObj->Target, level,
+                                   xoffset, yoffset, zoffset,
+                                   x, y, width, height, self);
 }
 
 static bool
@@ -4636,68 +4635,72 @@ compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
                                   GLenum target, GLint level,
                                   GLint xoffset, GLint yoffset, GLint zoffset,
                                   GLsizei width, GLsizei height, GLsizei depth,
-                                  GLenum format, GLsizei imageSize, bool dsa)
+                                  GLenum format, GLsizei imageSize,
+                                  const GLvoid *data, const char *callerName)
 {
    struct gl_texture_image *texImage;
    GLint expectedSize;
-   const char *suffix = dsa ? "ture" : "";
 
    /* this will catch any invalid compressed format token */
    if (!_mesa_is_compressed_format(ctx, format)) {
       _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCompressedTex%sSubImage%uD(format)", suffix, dims);
+                  "%s(format)", callerName);
       return GL_TRUE;
    }
 
    if (level < 0 || level >= _mesa_max_texture_levels(ctx, target)) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCompressedTex%sSubImage%uD(level=%d)",
-                  suffix, dims, level);
+                  "%s(level=%d)",
+                  callerName, level);
+      return GL_TRUE;
+   }
+
+   /* validate the bound PBO, if any */
+   if (!_mesa_validate_pbo_source_compressed(ctx, dims, &ctx->Unpack,
+                                     imageSize, data, callerName)) {
       return GL_TRUE;
    }
 
    /* Check for invalid pixel storage modes */
    if (!_mesa_compressed_pixel_storage_error_check(ctx, dims,
-               &ctx->Unpack,
-               dsa ? "glCompressedTextureSubImage" :
-               "glCompressedTexSubImage")) {
+                                                   &ctx->Unpack, callerName)) {
       return GL_TRUE;
    }
 
    expectedSize = compressed_tex_size(width, height, depth, format);
    if (expectedSize != imageSize) {
       _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glCompressedTex%sSubImage%uD(size=%d)",
-                  suffix, dims, imageSize);
+                  "%s(size=%d)",
+                  callerName, imageSize);
       return GL_TRUE;
    }
 
    texImage = _mesa_select_tex_image(texObj, target, level);
    if (!texImage) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCompressedTex%sSubImage%uD(invalid texture image)",
-                  suffix, dims);
+                  "%s(invalid texture image)",
+                  callerName);
       return GL_TRUE;
    }
 
    if ((GLint) format != texImage->InternalFormat) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCompressedTex%sSubImage%uD(format=0x%x)",
-                  suffix, dims, format);
+                  "%s(format=0x%x)",
+                  callerName, format);
       return GL_TRUE;
    }
 
    if (compressedteximage_only_format(ctx, format)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-               "glCompressedTex%sSubImage%uD(format=0x%x cannot be updated)",
-               suffix, dims, format);
+               "%s(format=0x%x cannot be updated)",
+               callerName, format);
       return GL_TRUE;
    }
 
    if (error_check_subtexture_dimensions(ctx, dims,
                                          texImage, xoffset, yoffset, zoffset,
                                          width, height, depth,
-                                         "glCompressedTexSubImage")) {
+                                         callerName)) {
       return GL_TRUE;
    }
 
@@ -4748,30 +4751,19 @@ _mesa_CompressedTexImage3D(GLenum target, GLint level,
 void
 _mesa_compressed_texture_sub_image(struct gl_context *ctx, GLuint dims,
                                    struct gl_texture_object *texObj,
+                                   struct gl_texture_image *texImage,
                                    GLenum target, GLint level,
                                    GLint xoffset, GLint yoffset,
                                    GLint zoffset,
                                    GLsizei width, GLsizei height,
                                    GLsizei depth,
                                    GLenum format, GLsizei imageSize,
-                                   const GLvoid *data, bool dsa)
+                                   const GLvoid *data)
 {
-   struct gl_texture_image *texImage;
-
-   if (compressed_subtexture_error_check(ctx, dims, texObj, target,
-                                         level, xoffset, yoffset, zoffset,
-                                         width, height, depth,
-                                         format, imageSize, dsa)) {
-      return;
-   }
-
    FLUSH_VERTICES(ctx, 0);
 
    _mesa_lock_texture(ctx, texObj);
    {
-      texImage = _mesa_select_tex_image(texObj, target, level);
-      assert(texImage);
-
       if (width > 0 && height > 0 && depth > 0) {
          ctx->Driver.CompressedTexSubImage(ctx, dims, texImage,
                                            xoffset, yoffset, zoffset,
@@ -4795,6 +4787,8 @@ _mesa_CompressedTexSubImage1D(GLenum target, GLint level, GLint xoffset,
                               GLsizei imageSize, const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    if (compressed_subtexture_target_check(ctx, target, 1, format, false,
@@ -4806,9 +4800,20 @@ _mesa_CompressedTexSubImage1D(GLenum target, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
-   _mesa_compressed_texture_sub_image(ctx, 1, texObj, target, level,
+   if (compressed_subtexture_error_check(ctx, 1, texObj, target,
+                                         level, xoffset, 0, 0,
+                                         width, 1, 1,
+                                         format, imageSize, data,
+                                         "glCompressedTexSubImage1D")) {
+      return;
+   }
+
+   texImage = _mesa_select_tex_image(texObj, target, level);
+   assert(texImage);
+
+   _mesa_compressed_texture_sub_image(ctx, 1, texObj, texImage, target, level,
                                       xoffset, 0, 0, width, 1, 1,
-                                      format, imageSize, data, false);
+                                      format, imageSize, data);
 }
 
 void GLAPIENTRY
@@ -4817,6 +4822,8 @@ _mesa_CompressedTextureSubImage1D(GLuint texture, GLint level, GLint xoffset,
                                   GLsizei imageSize, const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    texObj = _mesa_lookup_texture_err(ctx, texture,
@@ -4830,9 +4837,21 @@ _mesa_CompressedTextureSubImage1D(GLuint texture, GLint level, GLint xoffset,
       return;
    }
 
-   _mesa_compressed_texture_sub_image(ctx, 1, texObj, texObj->Target, level,
+   if (compressed_subtexture_error_check(ctx, 1, texObj, texObj->Target,
+                                         level, xoffset, 0, 0,
+                                         width, 1, 1,
+                                         format, imageSize, data,
+                                         "glCompressedTextureSubImage1D")) {
+      return;
+   }
+
+   texImage = _mesa_select_tex_image(texObj, texObj->Target, level);
+   assert(texImage);
+
+   _mesa_compressed_texture_sub_image(ctx, 1, texObj, texImage,
+                                      texObj->Target, level,
                                       xoffset, 0, 0, width, 1, 1,
-                                      format, imageSize, data, true);
+                                      format, imageSize, data);
 }
 
 
@@ -4843,6 +4862,8 @@ _mesa_CompressedTexSubImage2D(GLenum target, GLint level, GLint xoffset,
                               const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    if (compressed_subtexture_target_check(ctx, target, 2, format, false,
@@ -4854,9 +4875,21 @@ _mesa_CompressedTexSubImage2D(GLenum target, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
-   _mesa_compressed_texture_sub_image(ctx, 2, texObj, target, level,
+   if (compressed_subtexture_error_check(ctx, 2, texObj, target,
+                                         level, xoffset, yoffset, 0,
+                                         width, height, 1,
+                                         format, imageSize, data,
+                                         "glCompressedTexSubImage2D")) {
+      return;
+   }
+
+
+   texImage = _mesa_select_tex_image(texObj, target, level);
+   assert(texImage);
+
+   _mesa_compressed_texture_sub_image(ctx, 2, texObj, texImage, target, level,
                                       xoffset, yoffset, 0, width, height, 1,
-                                      format, imageSize, data, false);
+                                      format, imageSize, data);
 }
 
 void GLAPIENTRY
@@ -4867,6 +4900,8 @@ _mesa_CompressedTextureSubImage2D(GLuint texture, GLint level, GLint xoffset,
                                   const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    texObj = _mesa_lookup_texture_err(ctx, texture,
@@ -4880,9 +4915,21 @@ _mesa_CompressedTextureSubImage2D(GLuint texture, GLint level, GLint xoffset,
       return;
    }
 
-   _mesa_compressed_texture_sub_image(ctx, 2, texObj, texObj->Target, level,
+   if (compressed_subtexture_error_check(ctx, 2, texObj, texObj->Target,
+                                         level, xoffset, yoffset, 0,
+                                         width, height, 1,
+                                         format, imageSize, data,
+                                         "glCompressedTextureSubImage2D")) {
+      return;
+   }
+
+   texImage = _mesa_select_tex_image(texObj, texObj->Target, level);
+   assert(texImage);
+
+   _mesa_compressed_texture_sub_image(ctx, 2, texObj, texImage,
+                                      texObj->Target, level,
                                       xoffset, yoffset, 0, width, height, 1,
-                                      format, imageSize, data, true);
+                                      format, imageSize, data);
 }
 
 void GLAPIENTRY
@@ -4892,6 +4939,8 @@ _mesa_CompressedTexSubImage3D(GLenum target, GLint level, GLint xoffset,
                               GLsizei imageSize, const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    if (compressed_subtexture_target_check(ctx, target, 3, format, false,
@@ -4903,10 +4952,22 @@ _mesa_CompressedTexSubImage3D(GLenum target, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
-   _mesa_compressed_texture_sub_image(ctx, 3, texObj, target, level,
+   if (compressed_subtexture_error_check(ctx, 3, texObj, target,
+                                         level, xoffset, yoffset, zoffset,
+                                         width, height, depth,
+                                         format, imageSize, data,
+                                         "glCompressedTexSubImage3D")) {
+      return;
+   }
+
+
+   texImage = _mesa_select_tex_image(texObj, target, level);
+   assert(texImage);
+
+   _mesa_compressed_texture_sub_image(ctx, 3, texObj, texImage, target, level,
                                       xoffset, yoffset, zoffset,
                                       width, height, depth,
-                                      format, imageSize, data, false);
+                                      format, imageSize, data);
 }
 
 void GLAPIENTRY
@@ -4917,6 +4978,8 @@ _mesa_CompressedTextureSubImage3D(GLuint texture, GLint level, GLint xoffset,
                                   const GLvoid *data)
 {
    struct gl_texture_object *texObj;
+   struct gl_texture_image *texImage;
+
    GET_CURRENT_CONTEXT(ctx);
 
    texObj = _mesa_lookup_texture_err(ctx, texture,
@@ -4930,10 +4993,60 @@ _mesa_CompressedTextureSubImage3D(GLuint texture, GLint level, GLint xoffset,
       return;
    }
 
-   _mesa_compressed_texture_sub_image(ctx, 3, texObj, texObj->Target, level,
-                                      xoffset, yoffset, zoffset,
-                                      width, height, depth,
-                                      format, imageSize, data, true);
+   if (compressed_subtexture_error_check(ctx, 3, texObj, texObj->Target,
+                                         level, xoffset, yoffset, zoffset,
+                                         width, height, depth,
+                                         format, imageSize, data,
+                                         "glCompressedTextureSubImage3D")) {
+      return;
+   }
+
+   /* Must handle special case GL_TEXTURE_CUBE_MAP. */
+   if (texObj->Target == GL_TEXTURE_CUBE_MAP) {
+      const char *pixels = data;
+      int i;
+      GLint image_stride;
+
+      /* Make sure the texture object is a proper cube.
+       * (See texturesubimage in teximage.c for details on why this check is
+       * performed.)
+       */
+      if (!_mesa_cube_level_complete(texObj, level)) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glCompressedTextureSubImage3D(cube map incomplete)");
+         return;
+      }
+
+      /* Copy in each face. */
+      for (i = 0; i < 6; ++i) {
+         texImage = texObj->Image[i][level];
+         assert(texImage);
+
+         _mesa_compressed_texture_sub_image(ctx, 3, texObj, texImage,
+                                            texObj->Target, level,
+                                            xoffset, yoffset, zoffset,
+                                            width, height, 1,
+                                            format, imageSize, pixels);
+
+         /* Compressed images don't have a client format */
+         image_stride = _mesa_format_image_size(texImage->TexFormat,
+                                                texImage->Width,
+                                                texImage->Height, 1);
+
+         pixels += image_stride;
+         imageSize -= image_stride;
+      }
+   }
+   else {
+      texImage = _mesa_select_tex_image(texObj, texObj->Target, level);
+      assert(texImage);
+
+      _mesa_compressed_texture_sub_image(ctx, 3, texObj, texImage,
+                                         texObj->Target, level,
+                                         xoffset, yoffset, zoffset,
+                                         width, height, depth,
+                                         format, imageSize, data);
+   }
 }
 
 static mesa_format
@@ -5152,24 +5265,34 @@ _mesa_validate_texbuffer_format(const struct gl_context *ctx,
 
 void
 _mesa_texture_buffer_range(struct gl_context *ctx,
-                           struct gl_texture_object *texObj, GLenum target,
+                           struct gl_texture_object *texObj,
                            GLenum internalFormat,
                            struct gl_buffer_object *bufObj,
-                           GLintptr offset, GLsizeiptr size, bool range,
-                           bool dsa)
+                           GLintptr offset, GLsizeiptr size,
+                           const char *caller)
 {
    mesa_format format;
 
-   FLUSH_VERTICES(ctx, 0);
+   /* NOTE: ARB_texture_buffer_object has interactions with
+    * the compatibility profile that are not implemented.
+    */
+   if (!(ctx->API == API_OPENGL_CORE &&
+         ctx->Extensions.ARB_texture_buffer_object)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "%s(ARB_texture_buffer_object is not"
+                  " implemented for the compatibility profile)", caller);
+      return;
+   }
 
    format = _mesa_validate_texbuffer_format(ctx, internalFormat);
    if (format == MESA_FORMAT_NONE) {
       _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glTex%sBuffer%s(internalFormat 0x%x)", dsa ? "ture" : "",
-                  range ? "Range" : "", internalFormat);
+                  "%s(internalFormat 0x%x)", caller, internalFormat);
       return;
    }
 
+   FLUSH_VERTICES(ctx, 0);
+
    _mesa_lock_texture(ctx, texObj);
    {
       _mesa_reference_buffer_object(ctx, &texObj->BufferObject, bufObj);
@@ -5188,6 +5311,75 @@ _mesa_texture_buffer_range(struct gl_context *ctx,
 }
 
 
+/**
+ * Make sure the texture buffer target is GL_TEXTURE_BUFFER.
+ * Return true if it is, and return false if it is not
+ * (and throw INVALID ENUM as dictated in the OpenGL 4.5
+ * core spec, 02.02.2015, PDF page 245).
+ */
+static bool
+check_texture_buffer_target(struct gl_context *ctx, GLenum target,
+                            const char *caller)
+{
+   if (target != GL_TEXTURE_BUFFER_ARB) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "%s(texture target is not GL_TEXTURE_BUFFER)", caller);
+      return false;
+   }
+   else
+      return true;
+}
+
+/**
+ * Check for errors related to the texture buffer range.
+ * Return false if errors are found, true if none are found.
+ */
+static bool
+check_texture_buffer_range(struct gl_context *ctx,
+                           struct gl_buffer_object *bufObj,
+                           GLintptr offset, GLsizeiptr size,
+                           const char *caller)
+{
+   /* OpenGL 4.5 core spec (02.02.2015) says in Section 8.9 Buffer
+    * Textures (PDF page 245):
+    *    "An INVALID_VALUE error is generated if offset is negative, if
+    *    size is less than or equal to zero, or if offset + size is greater
+    *    than the value of BUFFER_SIZE for the buffer bound to target."
+    */
+   if (offset < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(offset=%d < 0)", caller,
+                  (int) offset);
+      return false;
+   }
+
+   if (size <= 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(size=%d <= 0)", caller,
+                  (int) size);
+      return false;
+   }
+
+   if (offset + size > bufObj->Size) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "%s(offset=%d + size=%d > buffer_size=%d)", caller,
+                  (int) offset, (int) size, (int) bufObj->Size);
+      return false;
+   }
+
+   /* OpenGL 4.5 core spec (02.02.2015) says in Section 8.9 Buffer
+    * Textures (PDF page 245):
+    *    "An INVALID_VALUE error is generated if offset is not an integer
+    *    multiple of the value of TEXTURE_BUFFER_OFFSET_ALIGNMENT."
+    */
+   if (offset % ctx->Const.TextureBufferOffsetAlignment) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "%s(invalid offset alignment)", caller);
+      return false;
+   }
+
+   return true;
+}
+
+
 /** GL_ARB_texture_buffer_object */
 void GLAPIENTRY
 _mesa_TexBuffer(GLenum target, GLenum internalFormat, GLuint buffer)
@@ -5197,33 +5389,25 @@ _mesa_TexBuffer(GLenum target, GLenum internalFormat, GLuint buffer)
 
    GET_CURRENT_CONTEXT(ctx);
 
-   /* Need to catch this before it gets to _mesa_get_current_tex_object */
-   if (target != GL_TEXTURE_BUFFER_ARB) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glTexBuffer(target)");
-      return;
-   }
-
-   /* NOTE: ARB_texture_buffer_object has interactions with
-    * the compatibility profile that are not implemented.
+   /* Need to catch a bad target before it gets to
+    * _mesa_get_current_tex_object.
     */
-   if (!(ctx->API == API_OPENGL_CORE &&
-         ctx->Extensions.ARB_texture_buffer_object)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTexBuffer");
+   if (!check_texture_buffer_target(ctx, target, "glTexBuffer"))
       return;
-   }
 
-   bufObj = _mesa_lookup_bufferobj(ctx, buffer);
-   if (!bufObj && buffer) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTexBuffer(buffer %u)", buffer);
-      return;
-   }
+   if (buffer) {
+      bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glTexBuffer");
+      if (!bufObj)
+         return;
+   } else
+      bufObj = NULL;
 
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
 
-   _mesa_texture_buffer_range(ctx, texObj, target, internalFormat, bufObj, 0,
-                              buffer ? -1 : 0, false, false);
+   _mesa_texture_buffer_range(ctx, texObj, internalFormat, bufObj, 0,
+                              buffer ? -1 : 0, "glTexBuffer");
 }
 
 
@@ -5237,46 +5421,41 @@ _mesa_TexBufferRange(GLenum target, GLenum internalFormat, GLuint buffer,
 
    GET_CURRENT_CONTEXT(ctx);
 
-   /* Need to catch this before it gets to _mesa_get_current_tex_object */
-   if (target != GL_TEXTURE_BUFFER_ARB) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glTexBufferRange(target)");
-      return;
-   }
-
-   if (!(ctx->API == API_OPENGL_CORE &&
-         ctx->Extensions.ARB_texture_buffer_range)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTexBufferRange");
+   /* Need to catch a bad target before it gets to
+    * _mesa_get_current_tex_object.
+    */
+   if (!check_texture_buffer_target(ctx, target, "glTexBufferRange"))
       return;
-   }
 
-   bufObj = _mesa_lookup_bufferobj(ctx, buffer);
-   if (bufObj) {
-      if (offset < 0 ||
-          size <= 0 ||
-          (offset + size) > bufObj->Size) {
-         _mesa_error(ctx, GL_INVALID_VALUE, "glTexBufferRange");
+   if (buffer) {
+      bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glTexBufferRange");
+      if (!bufObj)
          return;
-      }
-      if (offset % ctx->Const.TextureBufferOffsetAlignment) {
-         _mesa_error(ctx, GL_INVALID_VALUE,
-                     "glTexBufferRange(invalid offset alignment)");
+
+      if (!check_texture_buffer_range(ctx, bufObj, offset, size,
+          "glTexBufferRange"))
          return;
-      }
-   } else if (buffer) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTexBufferRange(buffer %u)",
-                  buffer);
-      return;
+
    } else {
+
+      /* OpenGL 4.5 core spec (02.02.2015) says in Section 8.9 Buffer
+       * Textures (PDF page 254):
+       *    "If buffer is zero, then any buffer object attached to the buffer
+       *    texture is detached, the values offset and size are ignored and
+       *    the state for offset and size for the buffer texture are reset to
+       *    zero."
+       */
       offset = 0;
       size = 0;
+      bufObj = NULL;
    }
 
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
 
-   _mesa_texture_buffer_range(ctx, texObj, target, internalFormat, bufObj,
-                              offset, size, true, false);
+   _mesa_texture_buffer_range(ctx, texObj, internalFormat, bufObj,
+                              offset, size, "glTexBufferRange");
 }
 
 void GLAPIENTRY
@@ -5287,35 +5466,69 @@ _mesa_TextureBuffer(GLuint texture, GLenum internalFormat, GLuint buffer)
 
    GET_CURRENT_CONTEXT(ctx);
 
-   /* NOTE: ARB_texture_buffer_object has interactions with
-    * the compatibility profile that are not implemented.
-    */
-   if (!(ctx->API == API_OPENGL_CORE &&
-         ctx->Extensions.ARB_texture_buffer_object)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTextureBuffer");
+   if (buffer) {
+      bufObj = _mesa_lookup_bufferobj_err(ctx, buffer, "glTextureBuffer");
+      if (!bufObj)
+         return;
+   } else
+      bufObj = NULL;
+
+   /* Get the texture object by Name. */
+   texObj = _mesa_lookup_texture_err(ctx, texture, "glTextureBuffer");
+   if (!texObj)
       return;
-   }
 
-   bufObj = _mesa_lookup_bufferobj(ctx, buffer);
-   if (!bufObj && buffer) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glTextureBuffer(buffer %u)",
-                  buffer);
+   if (!check_texture_buffer_target(ctx, texObj->Target, "glTextureBuffer"))
       return;
+
+   _mesa_texture_buffer_range(ctx, texObj, internalFormat,
+                              bufObj, 0, buffer ? -1 : 0, "glTextureBuffer");
+}
+
+void GLAPIENTRY
+_mesa_TextureBufferRange(GLuint texture, GLenum internalFormat, GLuint buffer,
+                         GLintptr offset, GLsizeiptr size)
+{
+   struct gl_texture_object *texObj;
+   struct gl_buffer_object *bufObj;
+
+   GET_CURRENT_CONTEXT(ctx);
+
+   if (buffer) {
+      bufObj = _mesa_lookup_bufferobj_err(ctx, buffer,
+                                          "glTextureBufferRange");
+      if (!bufObj)
+         return;
+
+      if (!check_texture_buffer_range(ctx, bufObj, offset, size,
+          "glTextureBufferRange"))
+         return;
+
+   } else {
+
+      /* OpenGL 4.5 core spec (02.02.2015) says in Section 8.9 Buffer
+       * Textures (PDF page 254):
+       *    "If buffer is zero, then any buffer object attached to the buffer
+       *    texture is detached, the values offset and size are ignored and
+       *    the state for offset and size for the buffer texture are reset to
+       *    zero."
+       */
+      offset = 0;
+      size = 0;
+      bufObj = NULL;
    }
 
    /* Get the texture object by Name. */
-   texObj = _mesa_lookup_texture_err(ctx, texture,
-                                     "glTextureBuffer(texture)");
+   texObj = _mesa_lookup_texture_err(ctx, texture, "glTextureBufferRange");
    if (!texObj)
       return;
 
-   if (texObj->Target != GL_TEXTURE_BUFFER_ARB) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glTextureBuffer(target)");
+   if (!check_texture_buffer_target(ctx, texObj->Target,
+       "glTextureBufferRange"))
       return;
-   }
 
-   _mesa_texture_buffer_range(ctx, texObj, texObj->Target, internalFormat,
-                              bufObj, 0, buffer ? -1 : 0, false, true);
+   _mesa_texture_buffer_range(ctx, texObj, internalFormat,
+                              bufObj, offset, size, "glTextureBufferRange");
 }
 
 static GLboolean
diff --git a/mesalib/src/mesa/main/teximage.h b/mesalib/src/mesa/main/teximage.h
index b7336bc6c..1eebaa8b6 100644
--- a/mesalib/src/mesa/main/teximage.h
+++ b/mesalib/src/mesa/main/teximage.h
@@ -181,13 +181,14 @@ _mesa_texture_sub_image(struct gl_context *ctx, GLuint dims,
 extern void
 _mesa_compressed_texture_sub_image(struct gl_context *ctx, GLuint dims, 
                                    struct gl_texture_object *texObj, 
+                                   struct gl_texture_image *texImage,
                                    GLenum target, GLint level,
                                    GLint xoffset, GLint yoffset,
                                    GLint zoffset,
                                    GLsizei width, GLsizei height,
                                    GLsizei depth,
                                    GLenum format, GLsizei imageSize,
-                                   const GLvoid *data, bool dsa);
+                                   const GLvoid *data);
 
 extern void
 _mesa_copy_texture_sub_image(struct gl_context *ctx, GLuint dims,
@@ -195,7 +196,8 @@ _mesa_copy_texture_sub_image(struct gl_context *ctx, GLuint dims,
                              GLenum target, GLint level,
                              GLint xoffset, GLint yoffset, GLint zoffset,
                              GLint x, GLint y,
-                             GLsizei width, GLsizei height, bool dsa);
+                             GLsizei width, GLsizei height,
+                             const char *caller);
 
 extern void
 _mesa_texture_image_multisample(struct gl_context *ctx, GLuint dims,
@@ -208,11 +210,11 @@ _mesa_texture_image_multisample(struct gl_context *ctx, GLuint dims,
 
 extern void
 _mesa_texture_buffer_range(struct gl_context *ctx,
-                           struct gl_texture_object *texObj, GLenum target, 
+                           struct gl_texture_object *texObj,
                            GLenum internalFormat,
                            struct gl_buffer_object *bufObj,
-                           GLintptr offset, GLsizeiptr size, bool range,
-                           bool dsa);
+                           GLintptr offset, GLsizeiptr size,
+                           const char *caller);
 /*@}*/
 
 
@@ -408,6 +410,10 @@ _mesa_TexBufferRange(GLenum target, GLenum internalFormat, GLuint buffer,
 extern void GLAPIENTRY
 _mesa_TextureBuffer(GLuint texture, GLenum internalFormat, GLuint buffer);
 
+extern void GLAPIENTRY
+_mesa_TextureBufferRange(GLuint texture, GLenum internalFormat, GLuint buffer,
+                         GLintptr offset, GLsizeiptr size);
+
 
 extern void GLAPIENTRY
 _mesa_TexImage2DMultisample(GLenum target, GLsizei samples,
diff --git a/mesalib/src/mesa/main/transformfeedback.c b/mesalib/src/mesa/main/transformfeedback.c
index a3e23ced5..ce678c864 100644
--- a/mesalib/src/mesa/main/transformfeedback.c
+++ b/mesalib/src/mesa/main/transformfeedback.c
@@ -762,7 +762,7 @@ _mesa_TransformFeedbackVaryings(GLuint program, GLsizei count,
 
    /* Save the new names and the count */
    for (i = 0; i < count; i++) {
-      shProg->TransformFeedback.VaryingNames[i] = _mesa_strdup(varyings[i]);
+      shProg->TransformFeedback.VaryingNames[i] = strdup(varyings[i]);
    }
    shProg->TransformFeedback.NumVarying = count;
 
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index 9f82de952..2ab5528c3 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -260,8 +260,8 @@ validate_uniform_parameters(struct gl_context *ctx,
    if (uni->array_elements == 0) {
       if (count > 1) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s(count > 1 for non-array, location=%d)",
-                     caller, location);
+                     "%s(count = %u for non-array \"%s\"@%d)",
+                     caller, count, uni->name, location);
          return NULL;
       }
 
@@ -601,6 +601,46 @@ _mesa_propagate_uniforms_to_driver_storage(struct gl_uniform_storage *uni,
    }
 }
 
+
+/**
+ * Return printable string for a given GLSL_TYPE_x
+ */
+static const char *
+glsl_type_name(enum glsl_base_type type)
+{
+   switch (type) {
+   case GLSL_TYPE_UINT:
+      return "uint";
+   case GLSL_TYPE_INT:
+      return "int";
+   case GLSL_TYPE_FLOAT:
+      return "float";
+   case GLSL_TYPE_DOUBLE:
+      return "double";
+   case GLSL_TYPE_BOOL:
+      return "bool";
+   case GLSL_TYPE_SAMPLER:
+      return "sampler";
+   case GLSL_TYPE_IMAGE:
+      return "image";
+   case GLSL_TYPE_ATOMIC_UINT:
+      return "atomic_uint";
+   case GLSL_TYPE_STRUCT:
+      return "struct";
+   case GLSL_TYPE_INTERFACE:
+      return "interface";
+   case GLSL_TYPE_ARRAY:
+      return "array";
+   case GLSL_TYPE_VOID:
+      return "void";
+   case GLSL_TYPE_ERROR:
+      return "error";
+   default:
+      return "other";
+   }
+}
+
+
 /**
  * Called via glUniform*() functions.
  */
@@ -620,11 +660,28 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
    if (uni == NULL)
       return;
 
+   if (uni->type->is_matrix()) {
+      /* Can't set matrix uniforms (like mat4) with glUniform */
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glUniform%u(uniform \"%s\"@%d is matrix)",
+                  src_components, uni->name, location);
+      return;
+   }
+
    /* Verify that the types are compatible.
     */
    const unsigned components = uni->type->is_sampler()
       ? 1 : uni->type->vector_elements;
 
+   if (components != src_components) {
+      /* glUniformN() must match float/vecN type */
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glUniform%u(\"%s\"@%u has %u components, not %u)",
+                  src_components, uni->name, location,
+                  components, src_components);
+      return;
+   }
+
    bool match;
    switch (uni->type->base_type) {
    case GLSL_TYPE_BOOL:
@@ -639,8 +696,12 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
       break;
    }
 
-   if (uni->type->is_matrix() || components != src_components || !match) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glUniform(type mismatch)");
+   if (!match) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glUniform%u(\"%s\"@%d is %s, not %s)",
+                  src_components, uni->name, location,
+                  glsl_type_name(uni->type->base_type),
+                  glsl_type_name(basicType));
       return;
    }
 
diff --git a/mesalib/src/mesa/math/m_clip_tmp.h b/mesalib/src/mesa/math/m_clip_tmp.h
index 45dec47f0..e289be7b3 100644
--- a/mesalib/src/mesa/math/m_clip_tmp.h
+++ b/mesalib/src/mesa/math/m_clip_tmp.h
@@ -40,7 +40,7 @@
  * \param andMask bitwise-AND of clipMask values
  * \return proj_vec pointer
  */
-static GLvector4f * _XFORMAPI TAG(cliptest_points4)( GLvector4f *clip_vec,
+static GLvector4f * TAG(cliptest_points4)( GLvector4f *clip_vec,
                                                      GLvector4f *proj_vec,
                                                      GLubyte clipMask[],
                                                      GLubyte *orMask,
@@ -120,7 +120,7 @@ static GLvector4f * _XFORMAPI TAG(cliptest_points4)( GLvector4f *clip_vec,
  * \param andMask bitwise-AND of clipMask values
  * \return clip_vec pointer
  */
-static GLvector4f * _XFORMAPI TAG(cliptest_np_points4)( GLvector4f *clip_vec,
+static GLvector4f * TAG(cliptest_np_points4)( GLvector4f *clip_vec,
 							GLvector4f *proj_vec,
 							GLubyte clipMask[],
 							GLubyte *orMask,
@@ -177,7 +177,7 @@ static GLvector4f * _XFORMAPI TAG(cliptest_np_points4)( GLvector4f *clip_vec,
 }
 
 
-static GLvector4f * _XFORMAPI TAG(cliptest_points3)( GLvector4f *clip_vec,
+static GLvector4f * TAG(cliptest_points3)( GLvector4f *clip_vec,
                                                      GLvector4f *proj_vec,
                                                      GLubyte clipMask[],
                                                      GLubyte *orMask,
@@ -213,7 +213,7 @@ static GLvector4f * _XFORMAPI TAG(cliptest_points3)( GLvector4f *clip_vec,
 }
 
 
-static GLvector4f * _XFORMAPI TAG(cliptest_points2)( GLvector4f *clip_vec,
+static GLvector4f * TAG(cliptest_points2)( GLvector4f *clip_vec,
                                                      GLvector4f *proj_vec,
                                                      GLubyte clipMask[],
                                                      GLubyte *orMask,
diff --git a/mesalib/src/mesa/math/m_norm_tmp.h b/mesalib/src/mesa/math/m_norm_tmp.h
index c8fab0ed3..d3ec1c22e 100644
--- a/mesalib/src/mesa/math/m_norm_tmp.h
+++ b/mesalib/src/mesa/math/m_norm_tmp.h
@@ -39,7 +39,7 @@
  *           optimization)
  * dest - the destination vector of normals
  */
-static void _XFORMAPI
+static void
 TAG(transform_normalize_normals)( const GLmatrix *mat,
                                   GLfloat scale,
                                   const GLvector4f *in,
@@ -106,7 +106,7 @@ TAG(transform_normalize_normals)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_normalize_normals_no_rot)( const GLmatrix *mat,
                                          GLfloat scale,
                                          const GLvector4f *in,
@@ -171,7 +171,7 @@ TAG(transform_normalize_normals_no_rot)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_rescale_normals_no_rot)( const GLmatrix *mat,
                                        GLfloat scale,
                                        const GLvector4f *in,
@@ -200,7 +200,7 @@ TAG(transform_rescale_normals_no_rot)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_rescale_normals)( const GLmatrix *mat,
                                 GLfloat scale,
                                 const GLvector4f *in,
@@ -232,7 +232,7 @@ TAG(transform_rescale_normals)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_normals_no_rot)( const GLmatrix *mat,
 			       GLfloat scale,
 			       const GLvector4f *in,
@@ -262,7 +262,7 @@ TAG(transform_normals_no_rot)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_normals)( const GLmatrix *mat,
                         GLfloat scale,
                         const GLvector4f *in,
@@ -292,7 +292,7 @@ TAG(transform_normals)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(normalize_normals)( const GLmatrix *mat,
                         GLfloat scale,
                         const GLvector4f *in,
@@ -338,7 +338,7 @@ TAG(normalize_normals)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(rescale_normals)( const GLmatrix *mat,
                       GLfloat scale,
                       const GLvector4f *in,
@@ -361,7 +361,7 @@ TAG(rescale_normals)( const GLmatrix *mat,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(init_c_norm_transform)( void )
 {
    _mesa_normal_tab[NORM_TRANSFORM_NO_ROT] =
diff --git a/mesalib/src/mesa/math/m_xform.h b/mesalib/src/mesa/math/m_xform.h
index 2ed62e711..0bb8e9bd8 100644
--- a/mesalib/src/mesa/math/m_xform.h
+++ b/mesalib/src/mesa/math/m_xform.h
@@ -32,14 +32,6 @@
 #include "math/m_matrix.h"
 #include "math/m_vector.h"
 
-#ifdef USE_X86_ASM
-#define _XFORMAPI _ASMAPI
-#define _XFORMAPIP _ASMAPIP
-#else
-#define _XFORMAPI
-#define _XFORMAPIP *
-#endif
-
 
 extern void
 _math_init_transformation(void);
@@ -99,12 +91,12 @@ init_c_cliptest(void);
 #define CLIP_FRUSTUM_BITS    0x3f
 
 
-typedef GLvector4f * (_XFORMAPIP clip_func)( GLvector4f *vClip,
-					     GLvector4f *vProj,
-					     GLubyte clipMask[],
-					     GLubyte *orMask,
-					     GLubyte *andMask,
-					     GLboolean viewport_z_clip );
+typedef GLvector4f * (*clip_func)(GLvector4f *vClip,
+                                  GLvector4f *vProj,
+                                  GLubyte clipMask[],
+                                  GLubyte *orMask,
+                                  GLubyte *andMask,
+                                  GLboolean viewport_z_clip);
 
 typedef void (*dotprod_func)( GLfloat *out,
 			      GLuint out_stride,
@@ -119,11 +111,11 @@ typedef void (*vec_copy_func)( GLvector4f *to,
 /*
  * Functions for transformation of normals in the VB.
  */
-typedef void (_XFORMAPIP normal_func)( const GLmatrix *mat,
-				      GLfloat scale,
-				      const GLvector4f *in,
-				      const GLfloat lengths[],
-				      GLvector4f *dest );
+typedef void (*normal_func)(const GLmatrix *mat,
+                            GLfloat scale,
+                            const GLvector4f *in,
+                            const GLfloat lengths[],
+                            GLvector4f *dest);
 
 
 /* Flags for selecting a normal transformation function.
@@ -141,9 +133,9 @@ typedef void (_XFORMAPIP normal_func)( const GLmatrix *mat,
  *     when the mask byte is zero.  This is always present as a
  *     parameter, to allow a unified interface.
  */
-typedef void (_XFORMAPIP transform_func)( GLvector4f *to_vec,
-					  const GLfloat m[16],
-					  const GLvector4f *from_vec );
+typedef void (*transform_func)(GLvector4f *to_vec,
+                               const GLfloat m[16],
+                               const GLvector4f *from_vec);
 
 
 extern dotprod_func  _mesa_dotprod_tab[5];
diff --git a/mesalib/src/mesa/math/m_xform_tmp.h b/mesalib/src/mesa/math/m_xform_tmp.h
index 8886c9646..af85de3d4 100644
--- a/mesalib/src/mesa/math/m_xform_tmp.h
+++ b/mesalib/src/mesa/math/m_xform_tmp.h
@@ -70,7 +70,7 @@
  *     driver-specific vertex format.
  */
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_general)( GLvector4f *to_vec,
 				const GLfloat m[16],
 				const GLvector4f *from_vec )
@@ -96,7 +96,7 @@ TAG(transform_points1_general)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_identity)( GLvector4f *to_vec,
 				 const GLfloat m[16],
 				 const GLvector4f *from_vec )
@@ -116,7 +116,7 @@ TAG(transform_points1_identity)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_2d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -138,7 +138,7 @@ TAG(transform_points1_2d)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_2d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -159,7 +159,7 @@ TAG(transform_points1_2d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_3d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -183,7 +183,7 @@ TAG(transform_points1_3d)( GLvector4f *to_vec,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_3d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -206,7 +206,7 @@ TAG(transform_points1_3d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points1_perspective)( GLvector4f *to_vec,
 				    const GLfloat m[16],
 				    const GLvector4f *from_vec )
@@ -236,7 +236,7 @@ TAG(transform_points1_perspective)( GLvector4f *to_vec,
  * present early in the geometry pipeline and throughout the
  * texture pipeline.
  */
-static void _XFORMAPI
+static void
 TAG(transform_points2_general)( GLvector4f *to_vec,
 				const GLfloat m[16],
 				const GLvector4f *from_vec )
@@ -262,7 +262,7 @@ TAG(transform_points2_general)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points2_identity)( GLvector4f *to_vec,
 				 const GLfloat m[16],
 				 const GLvector4f *from_vec )
@@ -283,7 +283,7 @@ TAG(transform_points2_identity)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points2_2d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -305,7 +305,7 @@ TAG(transform_points2_2d)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points2_2d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -326,7 +326,7 @@ TAG(transform_points2_2d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points2_3d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -353,7 +353,7 @@ TAG(transform_points2_3d)( GLvector4f *to_vec,
 /* I would actually say this was a fairly important function, from
  * a texture transformation point of view.
  */
-static void _XFORMAPI
+static void
 TAG(transform_points2_3d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -382,7 +382,7 @@ TAG(transform_points2_3d_no_rot)( GLvector4f *to_vec,
 }
 
 
-static void _XFORMAPI
+static void
 TAG(transform_points2_perspective)( GLvector4f *to_vec,
 				    const GLfloat m[16],
 				    const GLvector4f *from_vec )
@@ -407,7 +407,7 @@ TAG(transform_points2_perspective)( GLvector4f *to_vec,
 
 
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_general)( GLvector4f *to_vec,
 				const GLfloat m[16],
 				const GLvector4f *from_vec )
@@ -433,7 +433,7 @@ TAG(transform_points3_general)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_identity)( GLvector4f *to_vec,
 				 const GLfloat m[16],
 				 const GLvector4f *from_vec )
@@ -455,7 +455,7 @@ TAG(transform_points3_identity)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_2d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -478,7 +478,7 @@ TAG(transform_points3_2d)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_2d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -500,7 +500,7 @@ TAG(transform_points3_2d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_3d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -526,7 +526,7 @@ TAG(transform_points3_3d)( GLvector4f *to_vec,
 
 /* previously known as ortho...
  */
-static void _XFORMAPI
+static void
 TAG(transform_points3_3d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -549,7 +549,7 @@ TAG(transform_points3_3d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points3_perspective)( GLvector4f *to_vec,
 				    const GLfloat m[16],
 				    const GLvector4f *from_vec )
@@ -575,7 +575,7 @@ TAG(transform_points3_perspective)( GLvector4f *to_vec,
 
 
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_general)( GLvector4f *to_vec,
 				const GLfloat m[16],
 				const GLvector4f *from_vec )
@@ -601,7 +601,7 @@ TAG(transform_points4_general)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_identity)( GLvector4f *to_vec,
 				 const GLfloat m[16],
 				 const GLvector4f *from_vec )
@@ -624,7 +624,7 @@ TAG(transform_points4_identity)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_2d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -648,7 +648,7 @@ TAG(transform_points4_2d)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_2d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -671,7 +671,7 @@ TAG(transform_points4_2d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_3d)( GLvector4f *to_vec,
 			   const GLfloat m[16],
 			   const GLvector4f *from_vec )
@@ -696,7 +696,7 @@ TAG(transform_points4_3d)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_3d_no_rot)( GLvector4f *to_vec,
 				  const GLfloat m[16],
 				  const GLvector4f *from_vec )
@@ -720,7 +720,7 @@ TAG(transform_points4_3d_no_rot)( GLvector4f *to_vec,
    to_vec->count = from_vec->count;
 }
 
-static void _XFORMAPI
+static void
 TAG(transform_points4_perspective)( GLvector4f *to_vec,
 				    const GLfloat m[16],
 				    const GLvector4f *from_vec )
@@ -753,7 +753,7 @@ static transform_func TAG(transform_tab_4)[7];
  * optimized routines overwriting the arrays.  This only occurs during
  * startup.
  */
-static void _XFORMAPI TAG(init_c_transformations)( void )
+static void TAG(init_c_transformations)( void )
 {
 #define TAG_TAB   _mesa_transform_tab
 #define TAG_TAB_1 TAG(transform_tab_1)
diff --git a/mesalib/src/mesa/program/prog_instruction.c b/mesalib/src/mesa/program/prog_instruction.c
index 6a9bcb7b5..f9ebe4e8f 100644
--- a/mesalib/src/mesa/program/prog_instruction.c
+++ b/mesalib/src/mesa/program/prog_instruction.c
@@ -89,7 +89,7 @@ _mesa_copy_instructions(struct prog_instruction *dest,
    memcpy(dest, src, n * sizeof(struct prog_instruction));
    for (i = 0; i < n; i++) {
       if (src[i].Comment)
-         dest[i].Comment = _mesa_strdup(src[i].Comment);
+         dest[i].Comment = strdup(src[i].Comment);
    }
    return dest;
 }
diff --git a/mesalib/src/mesa/program/prog_instruction.h b/mesalib/src/mesa/program/prog_instruction.h
index 0957bd9d7..ab3acbc02 100644
--- a/mesalib/src/mesa/program/prog_instruction.h
+++ b/mesalib/src/mesa/program/prog_instruction.h
@@ -366,11 +366,11 @@ struct prog_instruction
     */
    GLint BranchTarget;
 
-   /** for debugging purposes */
-   const char *Comment;
-
    /** for driver use (try to remove someday) */
    GLint Aux;
+
+   /** for debugging purposes */
+   const char *Comment;
 };
 
 
diff --git a/mesalib/src/mesa/program/prog_parameter.c b/mesalib/src/mesa/program/prog_parameter.c
index 5939f6f72..cdfe25145 100644
--- a/mesalib/src/mesa/program/prog_parameter.c
+++ b/mesalib/src/mesa/program/prog_parameter.c
@@ -148,7 +148,7 @@ _mesa_add_parameter(struct gl_program_parameter_list *paramList,
 
       for (i = 0; i < sz4; i++) {
          struct gl_program_parameter *p = paramList->Parameters + oldNum + i;
-         p->Name = name ? _mesa_strdup(name) : NULL;
+         p->Name = name ? strdup(name) : NULL;
          p->Type = type;
          p->Size = size;
          p->DataType = datatype;
diff --git a/mesalib/src/mesa/program/prog_statevars.c b/mesalib/src/mesa/program/prog_statevars.c
index 57b25a7e3..0c0c87faa 100644
--- a/mesalib/src/mesa/program/prog_statevars.c
+++ b/mesalib/src/mesa/program/prog_statevars.c
@@ -1045,7 +1045,7 @@ _mesa_program_state_string(const gl_state_index state[STATE_LENGTH])
       break;
    }
 
-   return _mesa_strdup(str);
+   return strdup(str);
 }
 
 
diff --git a/mesalib/src/mesa/program/program.c b/mesalib/src/mesa/program/program.c
index 61a9e97c9..3c214d5e3 100644
--- a/mesalib/src/mesa/program/program.c
+++ b/mesalib/src/mesa/program/program.c
@@ -79,7 +79,7 @@ _mesa_init_program(struct gl_context *ctx)
    STATIC_ASSERT(NUM_TEXTURE_TARGETS <= (1 << 4));
 
    ctx->Program.ErrorPos = -1;
-   ctx->Program.ErrorString = _mesa_strdup("");
+   ctx->Program.ErrorString = strdup("");
 
    ctx->VertexProgram.Enabled = GL_FALSE;
    ctx->VertexProgram.PointSizeEnabled =
@@ -176,7 +176,7 @@ _mesa_set_program_error(struct gl_context *ctx, GLint pos, const char *string)
    free((void *) ctx->Program.ErrorString);
    if (!string)
       string = "";
-   ctx->Program.ErrorString = _mesa_strdup(string);
+   ctx->Program.ErrorString = strdup(string);
 }
 
 
@@ -483,7 +483,7 @@ _mesa_clone_program(struct gl_context *ctx, const struct gl_program *prog)
    assert(clone->Target == prog->Target);
    assert(clone->RefCount == 1);
 
-   clone->String = (GLubyte *) _mesa_strdup((char *) prog->String);
+   clone->String = (GLubyte *) strdup((char *) prog->String);
    clone->Format = prog->Format;
    clone->Instructions = _mesa_alloc_instructions(prog->NumInstructions);
    if (!clone->Instructions) {
diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
index 1dbc4b905..db254c214 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
@@ -485,8 +485,8 @@ st_clear_buffer_subdata(struct gl_context *ctx,
    static const char zeros[16] = {0};
 
    if (!pipe->clear_buffer) {
-      _mesa_buffer_clear_subdata(ctx, offset, size,
-                                 clearValue, clearValueSize, bufObj);
+      _mesa_ClearBufferSubData_sw(ctx, offset, size,
+                                  clearValue, clearValueSize, bufObj);
       return;
    }
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_texture.c b/mesalib/src/mesa/state_tracker/st_cb_texture.c
index a8b19a1f3..5c520b44f 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_texture.c
@@ -178,9 +178,6 @@ st_FreeTextureImageBuffer(struct gl_context *ctx,
       pipe_resource_reference(&stImage->pt, NULL);
    }
 
-   _mesa_align_free(stImage->TexData);
-   stImage->TexData = NULL;
-
    free(stImage->transfer);
    stImage->transfer = NULL;
    stImage->num_transfers = 0;
@@ -501,7 +498,6 @@ st_AllocTextureImageBuffer(struct gl_context *ctx,
 
    DBG("%s\n", __FUNCTION__);
 
-   assert(!stImage->TexData);
    assert(!stImage->pt); /* xxx this might be wrong */
 
    /* Look if the parent texture object has space for this image */
@@ -1521,23 +1517,6 @@ copy_image_data_to_texture(struct st_context *st,
 
       pipe_resource_reference(&stImage->pt, NULL);
    }
-   else if (stImage->TexData) {
-      /* Copy from malloc'd memory */
-      /* XXX this should be re-examined/tested with a compressed format */
-      GLuint blockSize = util_format_get_blocksize(stObj->pt->format);
-      GLuint srcRowStride = stImage->base.Width * blockSize;
-      GLuint srcSliceStride = stImage->base.Height * srcRowStride;
-      st_texture_image_data(st,
-                            stObj->pt,
-                            stImage->base.Face,
-                            dstLevel,
-                            stImage->TexData,
-                            srcRowStride,
-                            srcSliceStride);
-      _mesa_align_free(stImage->TexData);
-      stImage->TexData = NULL;
-   }
-
    pipe_resource_reference(&stImage->pt, stObj->pt);
 }
 
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 bd191d864..efee4b258 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -332,6 +332,7 @@ public:
    int glsl_version;
    bool native_integers;
    bool have_sqrt;
+   bool have_fma;
 
    variable_storage *find_variable_storage(ir_variable *var);
 
@@ -836,6 +837,7 @@ glsl_to_tgsi_visitor::get_opcode(ir_instruction *ir, unsigned op,
       case3fid(ADD, UADD, DADD);
       case3fid(MUL, UMUL, DMUL);
       case3fid(MAD, UMAD, DMAD);
+      case3fid(FMA, UMAD, DFMA);
       case3(DIV, IDIV, UDIV);
       case4d(MAX, IMAX, UMAX, DMAX);
       case4d(MIN, IMIN, UMIN, DMIN);
@@ -2222,10 +2224,11 @@ glsl_to_tgsi_visitor::visit(ir_expression *ir)
       emit(ir, TGSI_OPCODE_IMUL_HI, result_dst, op[0], op[1]);
       break;
    case ir_triop_fma:
-      /* NOTE: Perhaps there should be a special opcode that enforces fused
-       * mul-add. Just use MAD for now.
-       */
-      emit(ir, TGSI_OPCODE_MAD, result_dst, op[0], op[1], op[2]);
+      /* In theory, MAD is incorrect here. */
+      if (have_fma)
+         emit(ir, TGSI_OPCODE_FMA, result_dst, op[0], op[1], op[2]);
+      else
+         emit(ir, TGSI_OPCODE_MAD, result_dst, op[0], op[1], op[2]);
       break;
    case ir_unop_interpolate_at_centroid:
       emit(ir, TGSI_OPCODE_INTERP_CENTROID, result_dst, op[0]);
@@ -5564,6 +5567,8 @@ get_mesa_program(struct gl_context *ctx,
 
    v->have_sqrt = pscreen->get_shader_param(pscreen, ptarget,
                                             PIPE_SHADER_CAP_TGSI_SQRT_SUPPORTED);
+   v->have_fma = pscreen->get_shader_param(pscreen, ptarget,
+                                           PIPE_SHADER_CAP_TGSI_FMA_SUPPORTED);
 
    _mesa_copy_linked_program_data(shader->Stage, shader_program, prog);
    _mesa_generate_parameters_list_for_uniforms(shader_program, shader,
diff --git a/mesalib/src/mesa/state_tracker/st_texture.c b/mesalib/src/mesa/state_tracker/st_texture.c
index ada984189..ca7c83c21 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_texture.c
@@ -310,46 +310,6 @@ st_texture_image_unmap(struct st_context *st,
    *transfer = NULL;
 }
 
-
-/* Upload data for a particular image.
- */
-void
-st_texture_image_data(struct st_context *st,
-                      struct pipe_resource *dst,
-                      GLuint face,
-                      GLuint level,
-                      void *src,
-                      GLuint src_row_stride, GLuint src_image_stride)
-{
-   struct pipe_context *pipe = st->pipe;
-   GLuint i;
-   const GLubyte *srcUB = src;
-   GLuint layers;
-
-   if (dst->target == PIPE_TEXTURE_1D_ARRAY ||
-       dst->target == PIPE_TEXTURE_2D_ARRAY ||
-       dst->target == PIPE_TEXTURE_CUBE_ARRAY)
-      layers = dst->array_size;
-   else
-      layers = u_minify(dst->depth0, level);
-
-   DBG("%s\n", __FUNCTION__);
-
-   for (i = 0; i < layers; i++) {
-      struct pipe_box box;
-      u_box_2d_zslice(0, 0, face + i,
-                      u_minify(dst->width0, level),
-                      u_minify(dst->height0, level),
-                      &box);
-
-      pipe->transfer_inline_write(pipe, dst, level, PIPE_TRANSFER_WRITE,
-                                  &box, srcUB, src_row_stride, 0);
-
-      srcUB += src_image_stride;
-   }
-}
-
-
 /**
  * For debug only: get/print center pixel in the src resource.
  */
diff --git a/mesalib/src/mesa/state_tracker/st_texture.h b/mesalib/src/mesa/state_tracker/st_texture.h
index 6b7f8c750..d8cd7c7b5 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.h
+++ b/mesalib/src/mesa/state_tracker/st_texture.h
@@ -55,13 +55,7 @@ struct st_texture_image
 {
    struct gl_texture_image base;
 
-   /** Used to store texture data that doesn't fit in the parent
-    * object's mipmap buffer.
-    */
-   GLubyte *TexData;
-
    /* If stImage->pt != NULL, image data is stored here.
-    * Else if stImage->TexData != NULL, image is stored there.
     * Else there is no image data.
     */
    struct pipe_resource *pt;
@@ -230,16 +224,6 @@ st_texture_image_unmap(struct st_context *st,
 extern const GLuint *
 st_texture_depth_offsets(struct pipe_resource *pt, GLuint level);
 
-
-/* Upload an image into a texture
- */
-extern void
-st_texture_image_data(struct st_context *st,
-                      struct pipe_resource *dst,
-                      GLuint face, GLuint level, void *src,
-                      GLuint src_row_pitch, GLuint src_image_pitch);
-
-
 /* Copy an image between two textures
  */
 extern void
diff --git a/mesalib/src/mesa/state_tracker/st_vdpau.c b/mesalib/src/mesa/state_tracker/st_vdpau.c
index 6ccaf3e3f..63af1196a 100644
--- a/mesalib/src/mesa/state_tracker/st_vdpau.c
+++ b/mesalib/src/mesa/state_tracker/st_vdpau.c
@@ -66,7 +66,7 @@ st_vdpau_map_surface(struct gl_context *ctx, GLenum target, GLenum access,
    struct pipe_sampler_view templ, **sampler_view;
    mesa_format texFormat;
 
-   getProcAddr = ctx->vdpGetProcAddress;
+   getProcAddr = (void *)ctx->vdpGetProcAddress;
    if (output) {
       VdpOutputSurfaceGallium *f;
       
diff --git a/mesalib/src/mesa/swrast/s_blend.c b/mesalib/src/mesa/swrast/s_blend.c
index 7cb119407..8479b0bc0 100644
--- a/mesalib/src/mesa/swrast/s_blend.c
+++ b/mesalib/src/mesa/swrast/s_blend.c
@@ -48,9 +48,6 @@
 #if defined(USE_MMX_ASM)
 #include "x86/mmx.h"
 #include "x86/common_x86_asm.h"
-#define _BLENDAPI _ASMAPI
-#else
-#define _BLENDAPI
 #endif
 
 
@@ -69,7 +66,7 @@
  * No-op means the framebuffer values remain unchanged.
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_noop(struct gl_context *ctx, GLuint n, const GLubyte mask[],
            GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -97,7 +94,7 @@ blend_noop(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Special case for glBlendFunc(GL_ONE, GL_ZERO)
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_replace(struct gl_context *ctx, GLuint n, const GLubyte mask[],
               GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -117,7 +114,7 @@ blend_replace(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Common transparency blending mode:
  * glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA).
  */
-static void _BLENDAPI
+static void
 blend_transparency_ubyte(struct gl_context *ctx, GLuint n, const GLubyte mask[],
                          GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -162,7 +159,7 @@ blend_transparency_ubyte(struct gl_context *ctx, GLuint n, const GLubyte mask[],
 }
 
 
-static void _BLENDAPI
+static void
 blend_transparency_ushort(struct gl_context *ctx, GLuint n, const GLubyte mask[],
                           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -200,7 +197,7 @@ blend_transparency_ushort(struct gl_context *ctx, GLuint n, const GLubyte mask[]
 }
 
 
-static void _BLENDAPI
+static void
 blend_transparency_float(struct gl_context *ctx, GLuint n, const GLubyte mask[],
                          GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -242,7 +239,7 @@ blend_transparency_float(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Add src and dest: glBlendFunc(GL_ONE, GL_ONE).
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_add(struct gl_context *ctx, GLuint n, const GLubyte mask[],
           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -308,7 +305,7 @@ blend_add(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Blend min function.
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_min(struct gl_context *ctx, GLuint n, const GLubyte mask[],
           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -361,7 +358,7 @@ blend_min(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Blend max function.
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_max(struct gl_context *ctx, GLuint n, const GLubyte mask[],
           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
@@ -415,7 +412,7 @@ blend_max(struct gl_context *ctx, GLuint n, const GLubyte mask[],
  * Modulate:  result = src * dest
  * Any chanType ok.
  */
-static void _BLENDAPI
+static void
 blend_modulate(struct gl_context *ctx, GLuint n, const GLubyte mask[],
                GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
diff --git a/mesalib/src/mesa/swrast/s_context.c b/mesalib/src/mesa/swrast/s_context.c
index 51cc22760..ecde292e3 100644
--- a/mesalib/src/mesa/swrast/s_context.c
+++ b/mesalib/src/mesa/swrast/s_context.c
@@ -409,7 +409,7 @@ _swrast_validate_point( struct gl_context *ctx, const SWvertex *v0 )
  * Called via swrast->BlendFunc.  Examine GL state to choose a blending
  * function, then call it.
  */
-static void _ASMAPI
+static void
 _swrast_validate_blend_func(struct gl_context *ctx, GLuint n, const GLubyte mask[],
                             GLvoid *src, const GLvoid *dst,
                             GLenum chanType )
diff --git a/mesalib/src/mesa/swrast/s_context.h b/mesalib/src/mesa/swrast/s_context.h
index d6fbc5d54..7cf0e30dc 100644
--- a/mesalib/src/mesa/swrast/s_context.h
+++ b/mesalib/src/mesa/swrast/s_context.h
@@ -58,10 +58,10 @@ typedef void (*texture_sample_func)(struct gl_context *ctx,
                                     GLuint n, const GLfloat texcoords[][4],
                                     const GLfloat lambda[], GLfloat rgba[][4]);
 
-typedef void (_ASMAPIP blend_func)( struct gl_context *ctx, GLuint n,
-                                    const GLubyte mask[],
-                                    GLvoid *src, const GLvoid *dst,
-                                    GLenum chanType);
+typedef void (*blend_func)(struct gl_context *ctx, GLuint n,
+                           const GLubyte mask[],
+                           GLvoid *src, const GLvoid *dst,
+                           GLenum chanType);
 
 typedef void (*swrast_point_func)( struct gl_context *ctx, const SWvertex *);
 
diff --git a/mesalib/src/mesa/swrast/s_span.c b/mesalib/src/mesa/swrast/s_span.c
index 5d618f048..e304b6b5a 100644
--- a/mesalib/src/mesa/swrast/s_span.c
+++ b/mesalib/src/mesa/swrast/s_span.c
@@ -788,7 +788,7 @@ clip_span( struct gl_context *ctx, SWspan *span )
          memmove(ARRAY, ARRAY + (SHIFT), (LEN) * sizeof(ARRAY[0]))
 
          for (i = 0; i < VARYING_SLOT_MAX; i++) {
-            if (span->arrayAttribs & (1 << i)) {
+            if (span->arrayAttribs & BITFIELD64_BIT(i)) {
                /* shift array elements left by 'leftClip' */
                SHIFT_ARRAY(span->array->attribs[i], leftClip, n - leftClip);
             }
diff --git a/mesalib/src/mesa/vbo/vbo_attrib_tmp.h b/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
index b1c3d9842..17e057836 100644
--- a/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
+++ b/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
@@ -30,35 +30,30 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 
 /* ATTR */
-#define ATTR( A, N, T, V0, V1, V2, V3 ) \
-        ATTR_##T((A), (N), (T), (V0), (V1), (V2), (V3))
-
-#define ATTR_GL_UNSIGNED_INT( A, N, T, V0, V1, V2, V3 ) \
-    ATTR_UNION(A, N, T, UINT_AS_UNION(V0), UINT_AS_UNION(V1), \
-        UINT_AS_UNION(V2), UINT_AS_UNION(V3))
-#define ATTR_GL_INT( A, N, T, V0, V1, V2, V3 ) \
-    ATTR_UNION(A, N, T, INT_AS_UNION(V0), INT_AS_UNION(V1), \
+#define ATTRI( A, N, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, GL_INT, INT_AS_UNION(V0), INT_AS_UNION(V1), \
         INT_AS_UNION(V2), INT_AS_UNION(V3))
-#define ATTR_GL_FLOAT( A, N, T, V0, V1, V2, V3 ) \
-    ATTR_UNION(A, N, T, FLOAT_AS_UNION(V0), FLOAT_AS_UNION(V1),\
+#define ATTRUI( A, N, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, GL_UNSIGNED_INT, UINT_AS_UNION(V0), UINT_AS_UNION(V1), \
+        UINT_AS_UNION(V2), UINT_AS_UNION(V3))
+#define ATTRF( A, N, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, GL_FLOAT, FLOAT_AS_UNION(V0), FLOAT_AS_UNION(V1),\
         FLOAT_AS_UNION(V2), FLOAT_AS_UNION(V3))
 
 
 /* float */
-#define ATTR1FV( A, V ) ATTR( A, 1, GL_FLOAT, (V)[0], 0, 0, 1 )
-#define ATTR2FV( A, V ) ATTR( A, 2, GL_FLOAT, (V)[0], (V)[1], 0, 1 )
-#define ATTR3FV( A, V ) ATTR( A, 3, GL_FLOAT, (V)[0], (V)[1], (V)[2], 1 )
-#define ATTR4FV( A, V ) ATTR( A, 4, GL_FLOAT, (V)[0], (V)[1], (V)[2], (V)[3] )
+#define ATTR1FV( A, V ) ATTRF( A, 1, (V)[0], 0, 0, 1 )
+#define ATTR2FV( A, V ) ATTRF( A, 2, (V)[0], (V)[1], 0, 1 )
+#define ATTR3FV( A, V ) ATTRF( A, 3, (V)[0], (V)[1], (V)[2], 1 )
+#define ATTR4FV( A, V ) ATTRF( A, 4, (V)[0], (V)[1], (V)[2], (V)[3] )
 
-#define ATTR1F( A, X )          ATTR( A, 1, GL_FLOAT, X, 0, 0, 1 )
-#define ATTR2F( A, X, Y )       ATTR( A, 2, GL_FLOAT, X, Y, 0, 1 )
-#define ATTR3F( A, X, Y, Z )    ATTR( A, 3, GL_FLOAT, X, Y, Z, 1 )
-#define ATTR4F( A, X, Y, Z, W ) ATTR( A, 4, GL_FLOAT, X, Y, Z, W )
+#define ATTR1F( A, X )          ATTRF( A, 1, X, 0, 0, 1 )
+#define ATTR2F( A, X, Y )       ATTRF( A, 2, X, Y, 0, 1 )
+#define ATTR3F( A, X, Y, Z )    ATTRF( A, 3, X, Y, Z, 1 )
+#define ATTR4F( A, X, Y, Z, W ) ATTRF( A, 4, X, Y, Z, W )
 
-/* int */
-#define ATTRI( A, N, X, Y, Z, W) ATTR( A, N, GL_INT, \
-                                       X, Y, Z, W )
 
+/* int */
 #define ATTR2IV( A, V ) ATTRI( A, 2, (V)[0], (V)[1], 0, 1 )
 #define ATTR3IV( A, V ) ATTRI( A, 3, (V)[0], (V)[1], (V)[2], 1 )
 #define ATTR4IV( A, V ) ATTRI( A, 4, (V)[0], (V)[1], (V)[2], (V)[3] )
@@ -70,9 +65,6 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 
 /* uint */
-#define ATTRUI( A, N, X, Y, Z, W) ATTR( A, N, GL_UNSIGNED_INT, \
-                                        X, Y, Z, W )
-
 #define ATTR2UIV( A, V ) ATTRUI( A, 2, (V)[0], (V)[1], 0, 1 )
 #define ATTR3UIV( A, V ) ATTRUI( A, 3, (V)[0], (V)[1], (V)[2], 1 )
 #define ATTR4UIV( A, V ) ATTRUI( A, 4, (V)[0], (V)[1], (V)[2], (V)[3] )
@@ -82,7 +74,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define ATTR3UI( A, X, Y, Z )    ATTRUI( A, 3, X, Y, Z, 1 )
 #define ATTR4UI( A, X, Y, Z, W ) ATTRUI( A, 4, X, Y, Z, W )
 
-#define MAT_ATTR( A, N, V ) ATTR( A, N, GL_FLOAT, (V)[0], (V)[1], (V)[2], (V)[3] )
+#define MAT_ATTR( A, N, V ) ATTRF( A, N, (V)[0], (V)[1], (V)[2], (V)[3] )
 
 static inline float conv_ui10_to_norm_float(unsigned ui10)
 {
@@ -94,20 +86,20 @@ static inline float conv_ui2_to_norm_float(unsigned ui2)
    return ui2 / 3.0f;
 }
 
-#define ATTRUI10_1( A, UI ) ATTR( A, 1, GL_FLOAT, (UI) & 0x3ff, 0, 0, 1 )
-#define ATTRUI10_2( A, UI ) ATTR( A, 2, GL_FLOAT, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, 0, 1 )
-#define ATTRUI10_3( A, UI ) ATTR( A, 3, GL_FLOAT, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, ((UI) >> 20) & 0x3ff, 1 )
-#define ATTRUI10_4( A, UI ) ATTR( A, 4, GL_FLOAT, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, ((UI) >> 20) & 0x3ff, ((UI) >> 30) & 0x3 )
+#define ATTRUI10_1( A, UI ) ATTRF( A, 1, (UI) & 0x3ff, 0, 0, 1 )
+#define ATTRUI10_2( A, UI ) ATTRF( A, 2, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, 0, 1 )
+#define ATTRUI10_3( A, UI ) ATTRF( A, 3, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, ((UI) >> 20) & 0x3ff, 1 )
+#define ATTRUI10_4( A, UI ) ATTRF( A, 4, (UI) & 0x3ff, ((UI) >> 10) & 0x3ff, ((UI) >> 20) & 0x3ff, ((UI) >> 30) & 0x3 )
 
-#define ATTRUI10N_1( A, UI ) ATTR( A, 1, GL_FLOAT, conv_ui10_to_norm_float((UI) & 0x3ff), 0, 0, 1 )
-#define ATTRUI10N_2( A, UI ) ATTR( A, 2, GL_FLOAT, \
+#define ATTRUI10N_1( A, UI ) ATTRF( A, 1, conv_ui10_to_norm_float((UI) & 0x3ff), 0, 0, 1 )
+#define ATTRUI10N_2( A, UI ) ATTRF( A, 2, \
 				   conv_ui10_to_norm_float((UI) & 0x3ff), \
 				   conv_ui10_to_norm_float(((UI) >> 10) & 0x3ff), 0, 1 )
-#define ATTRUI10N_3( A, UI ) ATTR( A, 3, GL_FLOAT, \
+#define ATTRUI10N_3( A, UI ) ATTRF( A, 3, \
 				   conv_ui10_to_norm_float((UI) & 0x3ff), \
 				   conv_ui10_to_norm_float(((UI) >> 10) & 0x3ff), \
 				   conv_ui10_to_norm_float(((UI) >> 20) & 0x3ff), 1 )
-#define ATTRUI10N_4( A, UI ) ATTR( A, 4, GL_FLOAT, \
+#define ATTRUI10N_4( A, UI ) ATTRF( A, 4, \
 				   conv_ui10_to_norm_float((UI) & 0x3ff), \
 				   conv_ui10_to_norm_float(((UI) >> 10) & 0x3ff), \
 				   conv_ui10_to_norm_float(((UI) >> 20) & 0x3ff), \
@@ -180,30 +172,30 @@ static inline float conv_i2_to_norm_float(const struct gl_context *ctx, int i2)
    }
 }
 
-#define ATTRI10_1( A, I10 ) ATTR( A, 1, GL_FLOAT, conv_i10_to_i((I10) & 0x3ff), 0, 0, 1 )
-#define ATTRI10_2( A, I10 ) ATTR( A, 2, GL_FLOAT, \
+#define ATTRI10_1( A, I10 ) ATTRF( A, 1, conv_i10_to_i((I10) & 0x3ff), 0, 0, 1 )
+#define ATTRI10_2( A, I10 ) ATTRF( A, 2, \
 				conv_i10_to_i((I10) & 0x3ff),		\
 				conv_i10_to_i(((I10) >> 10) & 0x3ff), 0, 1 )
-#define ATTRI10_3( A, I10 ) ATTR( A, 3, GL_FLOAT, \
+#define ATTRI10_3( A, I10 ) ATTRF( A, 3, \
 				conv_i10_to_i((I10) & 0x3ff),	    \
 				conv_i10_to_i(((I10) >> 10) & 0x3ff), \
 				conv_i10_to_i(((I10) >> 20) & 0x3ff), 1 )
-#define ATTRI10_4( A, I10 ) ATTR( A, 4, GL_FLOAT, \
+#define ATTRI10_4( A, I10 ) ATTRF( A, 4, \
 				conv_i10_to_i((I10) & 0x3ff),		\
 				conv_i10_to_i(((I10) >> 10) & 0x3ff), \
 				conv_i10_to_i(((I10) >> 20) & 0x3ff), \
 				conv_i2_to_i(((I10) >> 30) & 0x3))
 
 
-#define ATTRI10N_1(ctx, A, I10) ATTR(A, 1, GL_FLOAT, conv_i10_to_norm_float(ctx, (I10) & 0x3ff), 0, 0, 1 )
-#define ATTRI10N_2(ctx, A, I10) ATTR(A, 2, GL_FLOAT, \
+#define ATTRI10N_1(ctx, A, I10) ATTRF(A, 1, conv_i10_to_norm_float(ctx, (I10) & 0x3ff), 0, 0, 1 )
+#define ATTRI10N_2(ctx, A, I10) ATTRF(A, 2, \
 				conv_i10_to_norm_float(ctx, (I10) & 0x3ff),		\
 				conv_i10_to_norm_float(ctx, ((I10) >> 10) & 0x3ff), 0, 1 )
-#define ATTRI10N_3(ctx, A, I10) ATTR(A, 3, GL_FLOAT, \
+#define ATTRI10N_3(ctx, A, I10) ATTRF(A, 3, \
 				conv_i10_to_norm_float(ctx, (I10) & 0x3ff),	    \
 				conv_i10_to_norm_float(ctx, ((I10) >> 10) & 0x3ff), \
 				conv_i10_to_norm_float(ctx, ((I10) >> 20) & 0x3ff), 1 )
-#define ATTRI10N_4(ctx, A, I10) ATTR(A, 4, GL_FLOAT, \
+#define ATTRI10N_4(ctx, A, I10) ATTRF(A, 4, \
 				conv_i10_to_norm_float(ctx, (I10) & 0x3ff),		\
 				conv_i10_to_norm_float(ctx, ((I10) >> 10) & 0x3ff), \
 				conv_i10_to_norm_float(ctx, ((I10) >> 20) & 0x3ff), \
diff --git a/mesalib/src/mesa/vbo/vbo_exec_api.c b/mesalib/src/mesa/vbo/vbo_exec_api.c
index 9669abe7d..02741c2bc 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_api.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_api.c
@@ -1239,7 +1239,7 @@ VertexAttrib4f_nopos(GLuint index, GLfloat x, GLfloat y, GLfloat z, GLfloat w)
 {
    GET_CURRENT_CONTEXT(ctx);
    if (index < MAX_VERTEX_GENERIC_ATTRIBS)
-      ATTR(VBO_ATTRIB_GENERIC0 + index, 4, GL_FLOAT, x, y, z, w);
+      ATTRF(VBO_ATTRIB_GENERIC0 + index, 4, x, y, z, w);
    else
       ERROR(GL_INVALID_VALUE);
 }
diff --git a/mesalib/src/mesa/x86/3dnow.c b/mesalib/src/mesa/x86/3dnow.c
index c46cfbc57..df7c64fcb 100644
--- a/mesalib/src/mesa/x86/3dnow.c
+++ b/mesalib/src/mesa/x86/3dnow.c
@@ -47,20 +47,20 @@ DECLARE_XFORM_GROUP( 3dnow, 3 )
 DECLARE_XFORM_GROUP( 3dnow, 4 )
 
 
-extern void _ASMAPI
+extern void
 _mesa_v16_3dnow_general_xform( GLfloat *first_vert,
 			       const GLfloat *m,
 			       const GLfloat *src,
 			       GLuint src_stride,
 			       GLuint count );
 
-extern void _ASMAPI
+extern void
 _mesa_3dnow_project_vertices( GLfloat *first,
 			      GLfloat *last,
 			      const GLfloat *m,
 			      GLuint stride );
 
-extern void _ASMAPI
+extern void
 _mesa_3dnow_project_clipped_vertices( GLfloat *first,
 				      GLfloat *last,
 				      const GLfloat *m,
diff --git a/mesalib/src/mesa/x86/common_x86.c b/mesalib/src/mesa/x86/common_x86.c
index 14b497d06..86fbca91e 100644
--- a/mesalib/src/mesa/x86/common_x86.c
+++ b/mesalib/src/mesa/x86/common_x86.c
@@ -68,12 +68,12 @@ static int detection_debug = GL_FALSE;
 
 /* No reason for this to be public.
  */
-extern GLuint	_ASMAPI _mesa_x86_has_cpuid(void);
-extern void	_ASMAPI _mesa_x86_cpuid(GLuint op, GLuint *reg_eax, GLuint *reg_ebx, GLuint *reg_ecx, GLuint *reg_edx);
-extern GLuint	_ASMAPI _mesa_x86_cpuid_eax(GLuint op);
-extern GLuint	_ASMAPI _mesa_x86_cpuid_ebx(GLuint op);
-extern GLuint	_ASMAPI _mesa_x86_cpuid_ecx(GLuint op);
-extern GLuint	_ASMAPI _mesa_x86_cpuid_edx(GLuint op);
+extern GLuint _mesa_x86_has_cpuid(void);
+extern void _mesa_x86_cpuid(GLuint op, GLuint *reg_eax, GLuint *reg_ebx, GLuint *reg_ecx, GLuint *reg_edx);
+extern GLuint _mesa_x86_cpuid_eax(GLuint op);
+extern GLuint _mesa_x86_cpuid_ebx(GLuint op);
+extern GLuint _mesa_x86_cpuid_ecx(GLuint op);
+extern GLuint _mesa_x86_cpuid_edx(GLuint op);
 
 
 #if defined(USE_SSE_ASM)
diff --git a/mesalib/src/mesa/x86/mmx.h b/mesalib/src/mesa/x86/mmx.h
index 8101cf889..0c354907f 100644
--- a/mesalib/src/mesa/x86/mmx.h
+++ b/mesalib/src/mesa/x86/mmx.h
@@ -31,27 +31,27 @@
 
 struct gl_context;
 
-extern void _ASMAPI
+extern void
 _mesa_mmx_blend_transparency( struct gl_context *ctx, GLuint n, const GLubyte mask[],
                               GLvoid *rgba, const GLvoid *dest,
                               GLenum chanType );
 
-extern void _ASMAPI
+extern void
 _mesa_mmx_blend_add( struct gl_context *ctx, GLuint n, const GLubyte mask[],
                      GLvoid *rgba, const GLvoid *dest,
                      GLenum chanType );
 
-extern void _ASMAPI
+extern void
 _mesa_mmx_blend_min( struct gl_context *ctx, GLuint n, const GLubyte mask[],
                      GLvoid *rgba, const GLvoid *dest,
                      GLenum chanType );
 
-extern void _ASMAPI
+extern void
 _mesa_mmx_blend_max( struct gl_context *ctx, GLuint n, const GLubyte mask[],
                      GLvoid *rgba, const GLvoid *dest,
                      GLenum chanType );
 
-extern void _ASMAPI
+extern void
 _mesa_mmx_blend_modulate( struct gl_context *ctx, GLuint n, const GLubyte mask[],
                           GLvoid *rgba, const GLvoid *dest,
                           GLenum chanType );
diff --git a/mesalib/src/mesa/x86/sse.c b/mesalib/src/mesa/x86/sse.c
index 846184fa3..02fe18fa1 100644
--- a/mesalib/src/mesa/x86/sse.c
+++ b/mesalib/src/mesa/x86/sse.c
@@ -46,35 +46,35 @@ DECLARE_XFORM_GROUP( sse, 3 )
 
 #if 1
 /* Some functions are not written in SSE-assembly, because the fpu ones are faster */
-extern void _ASMAPI _mesa_sse_transform_normals_no_rot( NORM_ARGS );
-extern void _ASMAPI _mesa_sse_transform_rescale_normals( NORM_ARGS );
-extern void _ASMAPI _mesa_sse_transform_rescale_normals_no_rot( NORM_ARGS );
+extern void _mesa_sse_transform_normals_no_rot( NORM_ARGS );
+extern void _mesa_sse_transform_rescale_normals( NORM_ARGS );
+extern void _mesa_sse_transform_rescale_normals_no_rot( NORM_ARGS );
 
-extern void _ASMAPI _mesa_sse_transform_points4_general( XFORM_ARGS );
-extern void _ASMAPI _mesa_sse_transform_points4_3d( XFORM_ARGS );
+extern void _mesa_sse_transform_points4_general( XFORM_ARGS );
+extern void _mesa_sse_transform_points4_3d( XFORM_ARGS );
 /* XXX this function segfaults, see below */
-extern void _ASMAPI _mesa_sse_transform_points4_identity( XFORM_ARGS );
+extern void _mesa_sse_transform_points4_identity( XFORM_ARGS );
 /* XXX this one works, see below */
-extern void _ASMAPI _mesa_x86_transform_points4_identity( XFORM_ARGS );
+extern void _mesa_x86_transform_points4_identity( XFORM_ARGS );
 #else
 DECLARE_NORM_GROUP( sse )
 #endif
 
 
-extern void _ASMAPI
+extern void
 _mesa_v16_sse_general_xform( GLfloat *first_vert,
 			     const GLfloat *m,
 			     const GLfloat *src,
 			     GLuint src_stride,
 			     GLuint count );
 
-extern void _ASMAPI
+extern void
 _mesa_sse_project_vertices( GLfloat *first,
 			    GLfloat *last,
 			    const GLfloat *m,
 			    GLuint stride );
 
-extern void _ASMAPI
+extern void
 _mesa_sse_project_clipped_vertices( GLfloat *first,
 				    GLfloat *last,
 				    const GLfloat *m,
diff --git a/mesalib/src/mesa/x86/x86_xform.c b/mesalib/src/mesa/x86/x86_xform.c
index 007cd704d..8b5cf9139 100644
--- a/mesalib/src/mesa/x86/x86_xform.c
+++ b/mesalib/src/mesa/x86/x86_xform.c
@@ -54,7 +54,7 @@ DECLARE_XFORM_GROUP( x86, 3 )
 DECLARE_XFORM_GROUP( x86, 4 )
 
 
-extern GLvector4f * _ASMAPI
+extern GLvector4f *
 _mesa_x86_cliptest_points4( GLvector4f *clip_vec,
 			    GLvector4f *proj_vec,
 			    GLubyte clipMask[],
@@ -62,7 +62,7 @@ _mesa_x86_cliptest_points4( GLvector4f *clip_vec,
 			    GLubyte *andMask,
 			    GLboolean viewport_z_clip );
 
-extern GLvector4f * _ASMAPI
+extern GLvector4f *
 _mesa_x86_cliptest_points4_np( GLvector4f *clip_vec,
 			       GLvector4f *proj_vec,
 			       GLubyte clipMask[],
@@ -70,7 +70,7 @@ _mesa_x86_cliptest_points4_np( GLvector4f *clip_vec,
 			       GLubyte *andMask,
 			       GLboolean viewport_z_clip );
 
-extern void _ASMAPI
+extern void
 _mesa_v16_x86_cliptest_points4( GLfloat *first_vert,
 				GLfloat *last_vert,
 				GLubyte *or_mask,
@@ -78,7 +78,7 @@ _mesa_v16_x86_cliptest_points4( GLfloat *first_vert,
 				GLubyte *clip_mask,
 				GLboolean viewport_z_clip );
 
-extern void _ASMAPI
+extern void
 _mesa_v16_x86_general_xform( GLfloat *dest,
 			     const GLfloat *m,
 			     const GLfloat *src,
diff --git a/mesalib/src/mesa/x86/x86_xform.h b/mesalib/src/mesa/x86/x86_xform.h
index 12b9fb8ff..e398f64a4 100644
--- a/mesalib/src/mesa/x86/x86_xform.h
+++ b/mesalib/src/mesa/x86/x86_xform.h
@@ -39,13 +39,13 @@
 			const GLvector4f *from_vec
 
 #define DECLARE_XFORM_GROUP( pfx, sz ) \
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_general( XFORM_ARGS );		\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_identity( XFORM_ARGS );	\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d_no_rot( XFORM_ARGS );	\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_perspective( XFORM_ARGS );	\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d( XFORM_ARGS );		\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d_no_rot( XFORM_ARGS );	\
-extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d( XFORM_ARGS );
+extern void _mesa_##pfx##_transform_points##sz##_general( XFORM_ARGS );		\
+extern void _mesa_##pfx##_transform_points##sz##_identity( XFORM_ARGS );	\
+extern void _mesa_##pfx##_transform_points##sz##_3d_no_rot( XFORM_ARGS );	\
+extern void _mesa_##pfx##_transform_points##sz##_perspective( XFORM_ARGS );	\
+extern void _mesa_##pfx##_transform_points##sz##_2d( XFORM_ARGS );		\
+extern void _mesa_##pfx##_transform_points##sz##_2d_no_rot( XFORM_ARGS );	\
+extern void _mesa_##pfx##_transform_points##sz##_3d( XFORM_ARGS );
 
 #define ASSIGN_XFORM_GROUP( pfx, sz )					\
    _mesa_transform_tab[sz][MATRIX_GENERAL] =				\
@@ -75,14 +75,14 @@ extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d( XFORM_ARGS );
 			GLvector4f *dest
 
 #define DECLARE_NORM_GROUP( pfx ) \
-extern void _ASMAPI _mesa_##pfx##_rescale_normals( NORM_ARGS );				\
-extern void _ASMAPI _mesa_##pfx##_normalize_normals( NORM_ARGS );			\
-extern void _ASMAPI _mesa_##pfx##_transform_normals( NORM_ARGS );			\
-extern void _ASMAPI _mesa_##pfx##_transform_normals_no_rot( NORM_ARGS );		\
-extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals( NORM_ARGS );		\
-extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals_no_rot( NORM_ARGS );	\
-extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals( NORM_ARGS );		\
-extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals_no_rot( NORM_ARGS );
+extern void _mesa_##pfx##_rescale_normals( NORM_ARGS );				\
+extern void _mesa_##pfx##_normalize_normals( NORM_ARGS );			\
+extern void _mesa_##pfx##_transform_normals( NORM_ARGS );			\
+extern void _mesa_##pfx##_transform_normals_no_rot( NORM_ARGS );		\
+extern void _mesa_##pfx##_transform_rescale_normals( NORM_ARGS );		\
+extern void _mesa_##pfx##_transform_rescale_normals_no_rot( NORM_ARGS );	\
+extern void _mesa_##pfx##_transform_normalize_normals( NORM_ARGS );		\
+extern void _mesa_##pfx##_transform_normalize_normals_no_rot( NORM_ARGS );
 
 #define ASSIGN_NORM_GROUP( pfx )					\
    _mesa_normal_tab[NORM_RESCALE] =					\
diff --git a/mesalib/src/util/Makefile.am b/mesalib/src/util/Makefile.am
index ec49dc6cf..2e7542e42 100644
--- a/mesalib/src/util/Makefile.am
+++ b/mesalib/src/util/Makefile.am
@@ -50,7 +50,9 @@ endif
 
 libmesautil_la_LIBADD = $(SHA1_LIBS)
 
-check_PROGRAMS = u_atomic_test
+roundeven_test_LDADD = -lm
+
+check_PROGRAMS = u_atomic_test roundeven_test
 TESTS = $(check_PROGRAMS)
 
 BUILT_SOURCES = $(MESA_UTIL_GENERATED_FILES)
diff --git a/mesalib/src/util/Makefile.sources b/mesalib/src/util/Makefile.sources
index 560ea836a..3e0d02bad 100644
--- a/mesalib/src/util/Makefile.sources
+++ b/mesalib/src/util/Makefile.sources
@@ -14,6 +14,7 @@ MESA_UTIL_FILES :=	\
 	register_allocate.h \
 	rgtc.c \
 	rgtc.h \
+	rounding.h \
 	set.c \
 	set.h \
 	simple_list.h \
diff --git a/mesalib/src/util/macros.h b/mesalib/src/util/macros.h
index b862bfd5f..6c7bda7ae 100644
--- a/mesalib/src/util/macros.h
+++ b/mesalib/src/util/macros.h
@@ -73,15 +73,13 @@ do {                        \
    assert(!str);            \
    __builtin_unreachable(); \
 } while (0)
-#elif _MSC_VER >= 1200
+#elif defined (_MSC_VER)
 #define unreachable(str)    \
 do {                        \
    assert(!str);            \
    __assume(0);             \
 } while (0)
-#endif
-
-#ifndef unreachable
+#else
 #define unreachable(str) assert(!str)
 #endif
 
@@ -99,7 +97,7 @@ do {                       \
 #define assume(expr) ((expr) ? ((void) 0) \
                              : (assert(!"assumption failed"), \
                                 __builtin_unreachable()))
-#elif _MSC_VER >= 1200
+#elif defined (_MSC_VER)
 #define assume(expr) __assume(expr)
 #else
 #define assume(expr) assert(expr)
@@ -178,5 +176,11 @@ do {                       \
 #  endif
 #endif
 
+#ifdef HAVE_FUNC_ATTRIBUTE_UNUSED
+#define UNUSED __attribute__((unused))
+#else
+#define UNUSED
+#endif
+
 
 #endif /* UTIL_MACROS_H */
diff --git a/mesalib/src/util/roundeven_test.c b/mesalib/src/util/roundeven_test.c
new file mode 100644
index 000000000..7526db1f3
--- /dev/null
+++ b/mesalib/src/util/roundeven_test.c
@@ -0,0 +1,140 @@
+/*
+ * 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 <stdio.h>
+#include <stdbool.h>
+#include <string.h>
+#include <math.h>
+
+#include "macros.h"
+#include "rounding.h"
+
+int main(int argc, char *argv[])
+{
+   const struct {
+      float input, expected;
+   } float_data[] = {
+      { 0.0,                  0.0 },
+      { nextafterf(0.5, 0.0), 0.0 },
+      { 0.5,                  0.0 },
+      { nextafterf(0.5, 1.0), 1.0 },
+      { 1.0,                  1.0 },
+      { nextafterf(1.5, 1.0), 1.0 },
+      { 1.5,                  2.0 },
+      { nextafterf(1.5, 2.0), 2.0 },
+      { 2.0,                  2.0 },
+      { nextafterf(2.5, 2.0), 2.0 },
+      { 2.5,                  2.0 },
+      { nextafterf(2.5, 3.0), 3.0 },
+   };
+
+   const struct {
+      double input, expected;
+   } double_data[] = {
+      { 0.0,                 0.0 },
+      { nextafter(0.5, 0.0), 0.0 },
+      { 0.5,                 0.0 },
+      { nextafter(0.5, 1.0), 1.0 },
+      { 1.0,                 1.0 },
+      { nextafter(1.5, 1.0), 1.0 },
+      { 1.5,                 2.0 },
+      { nextafter(1.5, 2.0), 2.0 },
+      { 2.0,                 2.0 },
+      { nextafter(2.5, 2.0), 2.0 },
+      { 2.5,                 2.0 },
+      { nextafter(2.5, 3.0), 3.0 },
+   };
+
+   bool failed = false;
+   int i;
+
+   for (i = 0; i < ARRAY_SIZE(float_data); i++) {
+      float output = _mesa_roundevenf(float_data[i].input);
+      if (memcmp(&float_data[i].expected, &output, sizeof(float))) {
+         fprintf(stderr, "%d float: expected %f (%a) from "
+                         "_mesa_roundevenf(%f (%a)) but got %f (%a)\n",
+                 i,
+                 float_data[i].expected,
+                 float_data[i].expected,
+                 float_data[i].input,
+                 float_data[i].input,
+                 output,
+                 output);
+         failed = true;
+      }
+   }
+
+   /* Test negated values */
+   for (i = 0; i < ARRAY_SIZE(float_data); i++) {
+      float output = _mesa_roundevenf(-float_data[i].input);
+      float negated_expected = -float_data[i].expected;
+      if (memcmp(&negated_expected, &output, sizeof(float))) {
+         fprintf(stderr, "%d float: expected %f (%a) from "
+                         "_mesa_roundevenf(%f (%a)) but got %f (%a)\n",
+                 i,
+                 negated_expected,
+                 negated_expected,
+                 -float_data[i].input,
+                 -float_data[i].input,
+                 output,
+                 output);
+         failed = true;
+      }
+   }
+
+   for (i = 0; i < ARRAY_SIZE(double_data); i++) {
+      double output = _mesa_roundeven(double_data[i].input);
+      if (memcmp(&double_data[i].expected, &output, sizeof(double))) {
+         fprintf(stderr, "%d double: expected %f (%a) from "
+                         "_mesa_roundeven(%f (%a)) but got %f (%a)\n",
+                 i,
+                 double_data[i].expected,
+                 double_data[i].expected,
+                 double_data[i].input,
+                 double_data[i].input,
+                 output,
+                 output);
+         failed = true;
+      }
+   }
+
+   /* Test negated values */
+   for (i = 0; i < ARRAY_SIZE(double_data); i++) {
+      double output = _mesa_roundeven(-double_data[i].input);
+      double negated_expected = -double_data[i].expected;
+      if (memcmp(&negated_expected, &output, sizeof(double))) {
+         fprintf(stderr, "%d double: expected %f (%a) from "
+                         "_mesa_roundeven(%f (%a)) but got %f (%a)\n",
+                 i,
+                 negated_expected,
+                 negated_expected,
+                 -double_data[i].input,
+                 -double_data[i].input,
+                 output,
+                 output);
+         failed = true;
+      }
+   }
+
+   return failed;
+}
diff --git a/mesalib/src/util/rounding.h b/mesalib/src/util/rounding.h
new file mode 100644
index 000000000..0cbe9269f
--- /dev/null
+++ b/mesalib/src/util/rounding.h
@@ -0,0 +1,78 @@
+/*
+ * 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 <math.h>
+
+#ifdef __SSE4_1__
+#include <smmintrin.h>
+#endif
+
+/* The C standard library has functions round()/rint()/nearbyint() that round
+ * their arguments according to the rounding mode set in the floating-point
+ * control register. While there are trunc()/ceil()/floor() functions that do
+ * a specific operation without modifying the rounding mode, there is no
+ * roundeven() in any version of C.
+ *
+ * Technical Specification 18661 (ISO/IEC TS 18661-1:2014) adds roundeven(),
+ * but it's unfortunately not implemented by glibc.
+ *
+ * This implementation differs in that it does not raise the inexact exception.
+ *
+ * We use rint() to implement these functions, with the assumption that the
+ * floating-point rounding mode has not been changed from the default Round
+ * to Nearest.
+ */
+
+/**
+ * \brief Rounds \c x to the nearest integer, with ties to the even integer.
+ */
+static inline float
+_mesa_roundevenf(float x)
+{
+#ifdef __SSE4_1__
+   float ret;
+   __m128 m = _mm_load_ss(&x);
+   m = _mm_round_ss(m, m, _MM_FROUND_CUR_DIRECTION | _MM_FROUND_NO_EXC);
+   _mm_store_ss(&ret, m);
+   return ret;
+#else
+   return rintf(x);
+#endif
+}
+
+/**
+ * \brief Rounds \c x to the nearest integer, with ties to the even integer.
+ */
+static inline double
+_mesa_roundeven(double x)
+{
+#ifdef __SSE4_1__
+   double ret;
+   __m128d m = _mm_load_sd(&x);
+   m = _mm_round_sd(m, m, _MM_FROUND_CUR_DIRECTION | _MM_FROUND_NO_EXC);
+   _mm_store_sd(&ret, m);
+   return ret;
+#else
+   return rint(x);
+#endif
+}
diff --git a/mesalib/src/util/u_atomic.h b/mesalib/src/util/u_atomic.h
index d15398e1e..e38395ac6 100644
--- a/mesalib/src/util/u_atomic.h
+++ b/mesalib/src/util/u_atomic.h
@@ -6,6 +6,8 @@
  *
  */
 
+#include "no_extern_c.h"
+
 #ifndef U_ATOMIC_H
 #define U_ATOMIC_H