Merge remote-tracking branch 'origin/released'

Conflicts:
	fontconfig/src/fcint.h
	fontconfig/src/fcstat.c
	mesalib/src/mapi/glapi/gen/GL3x.xml
	xorg-server/glx/glxext.h

14 files changed, 722 insertions, 609 deletions

diff --git a/mesalib/src/mesa/Makefile b/mesalib/src/mesa/Makefile
index 845b524e6..b0b461fdd 100644
--- a/mesalib/src/mesa/Makefile
+++ b/mesalib/src/mesa/Makefile
@@ -135,7 +135,7 @@ asm_subdirs:
 ######################################################################
 # Dependency generation
 
-depend: $(ALL_FILES)
+depend: $(ALL_FILES) main/git_sha1.h
 	@ echo "running $(MKDEP)"
 	@ touch depend
 	@$(MKDEP) $(MKDEP_OPTIONS) $(MESA_CPPFLAGS) \
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index 8d7e90126..be7141a58 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -319,6 +319,10 @@ struct gl_meta_state
    struct drawtex_state DrawTex;  /**< For _mesa_meta_DrawTex() */
 };
 
+static void meta_glsl_blit_cleanup(struct gl_context *ctx, struct blit_state *blit);
+static void cleanup_temp_texture(struct gl_context *ctx, struct temp_texture *tex);
+static void meta_glsl_clear_cleanup(struct gl_context *ctx, struct clear_state *clear);
+
 static GLuint
 compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB *source)
 {
@@ -335,12 +339,16 @@ compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB
       return shader;
 
    _mesa_GetShaderiv(shader, GL_INFO_LOG_LENGTH, &size);
-   if (size == 0)
+   if (size == 0) {
+      _mesa_DeleteObjectARB(shader);
       return 0;
+   }
 
    info = malloc(size);
-   if (!info)
+   if (!info) {
+      _mesa_DeleteObjectARB(shader);
       return 0;
+   }
 
    _mesa_GetProgramInfoLog(shader, size, NULL, info);
    _mesa_problem(ctx,
@@ -349,6 +357,7 @@ compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB
 		 info, source);
 
    free(info);
+   _mesa_DeleteObjectARB(shader);
 
    return 0;
 }
@@ -401,10 +410,15 @@ _mesa_meta_init(struct gl_context *ctx)
 void
 _mesa_meta_free(struct gl_context *ctx)
 {
-   /* Note: Any textures, VBOs, etc, that we allocate should get
-    * freed by the normal context destruction code.  But this would be
-    * the place to free other meta data someday.
-    */
+   GET_CURRENT_CONTEXT(old_context);
+   _mesa_make_current(ctx, NULL, NULL);
+   meta_glsl_blit_cleanup(ctx, &ctx->Meta->Blit);
+   meta_glsl_clear_cleanup(ctx, &ctx->Meta->Clear);
+   cleanup_temp_texture(ctx, &ctx->Meta->TempTex);
+   if (old_context)
+      _mesa_make_current(old_context, old_context->WinSysDrawBuffer, old_context->WinSysReadBuffer);
+   else
+      _mesa_make_current(NULL, NULL, NULL);
    free(ctx->Meta);
    ctx->Meta = NULL;
 }
@@ -1068,6 +1082,15 @@ init_temp_texture(struct gl_context *ctx, struct temp_texture *tex)
    _mesa_GenTextures(1, &tex->TexObj);
 }
 
+static void
+cleanup_temp_texture(struct gl_context *ctx, struct temp_texture *tex)
+{
+   if (!tex->TexObj)
+     return;
+   _mesa_DeleteTextures(1, &tex->TexObj);
+   tex->TexObj = 0;
+}
+
 
 /**
  * Return pointer to temp_texture info for non-bitmap ops.
@@ -1604,6 +1627,21 @@ _mesa_meta_BlitFramebuffer(struct gl_context *ctx,
    }
 }
 
+static void
+meta_glsl_blit_cleanup(struct gl_context *ctx, struct blit_state *blit)
+{
+   if (blit->ArrayObj) {
+      _mesa_DeleteVertexArraysAPPLE(1, &blit->ArrayObj);
+      blit->ArrayObj = 0;
+      _mesa_DeleteBuffersARB(1, &blit->VBO);
+      blit->VBO = 0;
+   }
+   if (blit->DepthFP) {
+      _mesa_DeletePrograms(1, &blit->DepthFP);
+      blit->DepthFP = 0;
+   }
+}
+
 
 /**
  * Meta implementation of ctx->Driver.Clear() in terms of polygon rendering.
@@ -1786,7 +1824,9 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
 
    clear->ShaderProg = _mesa_CreateProgramObjectARB();
    _mesa_AttachShader(clear->ShaderProg, fs);
+   _mesa_DeleteObjectARB(fs);
    _mesa_AttachShader(clear->ShaderProg, vs);
+   _mesa_DeleteObjectARB(vs);
    _mesa_BindAttribLocationARB(clear->ShaderProg, 0, "position");
    _mesa_LinkProgramARB(clear->ShaderProg);
 
@@ -1799,7 +1839,9 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
 
       clear->IntegerShaderProg = _mesa_CreateProgramObjectARB();
       _mesa_AttachShader(clear->IntegerShaderProg, fs);
+      _mesa_DeleteObjectARB(fs);
       _mesa_AttachShader(clear->IntegerShaderProg, vs);
+      _mesa_DeleteObjectARB(vs);
       _mesa_BindAttribLocationARB(clear->IntegerShaderProg, 0, "position");
 
       /* Note that user-defined out attributes get automatically assigned
@@ -1814,6 +1856,24 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
    }
 }
 
+static void
+meta_glsl_clear_cleanup(struct gl_context *ctx, struct clear_state *clear)
+{
+   if (clear->ArrayObj == 0)
+      return;
+   _mesa_DeleteVertexArraysAPPLE(1, &clear->ArrayObj);
+   clear->ArrayObj = 0;
+   _mesa_DeleteBuffersARB(1, &clear->VBO);
+   clear->VBO = 0;
+   _mesa_DeleteObjectARB(clear->ShaderProg);
+   clear->ShaderProg = 0;
+
+   if (clear->IntegerShaderProg) {
+      _mesa_DeleteObjectARB(clear->IntegerShaderProg);
+      clear->IntegerShaderProg = 0;
+   }
+}
+
 /**
  * Meta implementation of ctx->Driver.Clear() in terms of polygon rendering.
  */
diff --git a/mesalib/src/mesa/drivers/windows/gdi/mesa.def b/mesalib/src/mesa/drivers/windows/gdi/mesa.def
index d5c98801d..fec7bbac0 100644
--- a/mesalib/src/mesa/drivers/windows/gdi/mesa.def
+++ b/mesalib/src/mesa/drivers/windows/gdi/mesa.def
@@ -556,21 +556,6 @@ EXPORTS
 	glFogCoorddvEXT
 	glFogCoordPointerEXT
 	glBlendFuncSeparateEXT
-	glFlushVertexArrayRangeNV
-	glVertexArrayRangeNV
-	glCombinerParameterfvNV
-	glCombinerParameterfNV
-	glCombinerParameterivNV
-	glCombinerParameteriNV
-	glCombinerInputNV
-	glCombinerOutputNV
-	glFinalCombinerInputNV
-	glGetCombinerInputParameterfvNV
-	glGetCombinerInputParameterivNV
-	glGetCombinerOutputParameterfvNV
-	glGetCombinerOutputParameterivNV
-	glGetFinalCombinerInputParameterfvNV
-	glGetFinalCombinerInputParameterivNV
 	glResizeBuffersMESA
 	glWindowPos2dMESA
 	glWindowPos2dvMESA
diff --git a/mesalib/src/mesa/main/shaderobj.c b/mesalib/src/mesa/main/shaderobj.c
index 7eb6f0bda..16800ae5e 100644
--- a/mesalib/src/mesa/main/shaderobj.c
+++ b/mesalib/src/mesa/main/shaderobj.c
@@ -278,7 +278,9 @@ _mesa_clear_shader_program_data(struct gl_context *ctx,
                                 struct gl_shader_program *shProg)
 {
    if (shProg->UniformStorage) {
-      _mesa_uniform_detach_all_driver_storage(shProg->UniformStorage);
+      unsigned i;
+      for (i = 0; i < shProg->NumUserUniformStorage; ++i)
+         _mesa_uniform_detach_all_driver_storage(&shProg->UniformStorage[i]);
       ralloc_free(shProg->UniformStorage);
       shProg->NumUserUniformStorage = 0;
       shProg->UniformStorage = NULL;
diff --git a/mesalib/src/mesa/program/hash_table.c b/mesalib/src/mesa/program/hash_table.c
index dc8563a33..7dabadc50 100644
--- a/mesalib/src/mesa/program/hash_table.c
+++ b/mesalib/src/mesa/program/hash_table.c
@@ -149,7 +149,7 @@ hash_table_insert(struct hash_table *ht, void *data, const void *key)
     insert_at_head(& ht->buckets[bucket], & node->link);
 }
 
-void
+bool
 hash_table_replace(struct hash_table *ht, void *data, const void *key)
 {
     const unsigned hash_value = (*ht->hash)(key);
@@ -162,7 +162,7 @@ hash_table_replace(struct hash_table *ht, void *data, const void *key)
 
        if ((*ht->compare)(hn->key, key) == 0) {
 	  hn->data = data;
-	  return;
+	  return true;
        }
     }
 
@@ -172,6 +172,7 @@ hash_table_replace(struct hash_table *ht, void *data, const void *key)
     hn->key = key;
 
     insert_at_head(& ht->buckets[bucket], & hn->link);
+    return false;
 }
 
 void
diff --git a/mesalib/src/mesa/program/hash_table.h b/mesalib/src/mesa/program/hash_table.h
index e1c234ea4..b814497a6 100644
--- a/mesalib/src/mesa/program/hash_table.h
+++ b/mesalib/src/mesa/program/hash_table.h
@@ -32,6 +32,7 @@
 #define HASH_TABLE_H
 
 #include <string.h>
+#include <stdbool.h>
 #include <stdlib.h>
 #include <stdint.h>
 #include <limits.h>
@@ -115,6 +116,10 @@ extern void hash_table_insert(struct hash_table *ht, void *data,
 /**
  * Add an element to a hash table with replacement
  *
+ * \return
+ * 1 if it did replace the the value (in which case the old key is kept), 0 if
+ * it did not replace the value (in which case the new key is kept).
+ *
  * \warning
  * If \c key is already in the hash table, \c data will \b replace the most
  * recently inserted \c data (see the warning in \c hash_table_insert) for
@@ -122,7 +127,7 @@ extern void hash_table_insert(struct hash_table *ht, void *data,
  *
  * \sa hash_table_insert
  */
-extern void hash_table_replace(struct hash_table *ht, void *data,
+extern bool hash_table_replace(struct hash_table *ht, void *data,
     const void *key);
 
 /**
@@ -220,6 +225,7 @@ public:
     */
    void clear()
    {
+      hash_table_call_foreach(this->ht, delete_key, NULL);
       hash_table_clear(this->ht);
    }
 
@@ -259,9 +265,12 @@ public:
        * because UINT_MAX+1 = 0.
        */
       assert(value != UINT_MAX);
-      hash_table_replace(this->ht,
-			 (void *) (intptr_t) (value + 1),
-			 strdup(key));
+      char *dup_key = strdup(key);
+      bool result = hash_table_replace(this->ht,
+				       (void *) (intptr_t) (value + 1),
+				       dup_key);
+      if (result)
+	 free(dup_key);
    }
 
 private:
diff --git a/mesalib/src/mesa/sources.mak b/mesalib/src/mesa/sources.mak
index 608aa79ab..16b1c39c0 100644
--- a/mesalib/src/mesa/sources.mak
+++ b/mesalib/src/mesa/sources.mak
@@ -198,6 +198,7 @@ VBO_FILES = \
 
 STATETRACKER_FILES = \
 	$(SRCDIR)/state_tracker/st_atom.c \
+	$(SRCDIR)/state_tracker/st_atom_array.c \
 	$(SRCDIR)/state_tracker/st_atom_blend.c \
 	$(SRCDIR)/state_tracker/st_atom_clip.c \
 	$(SRCDIR)/state_tracker/st_atom_constbuf.c \
diff --git a/mesalib/src/mesa/state_tracker/st_atom.c b/mesalib/src/mesa/state_tracker/st_atom.c
index d9cd4aab4..e6fc114f1 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.c
+++ b/mesalib/src/mesa/state_tracker/st_atom.c
@@ -63,7 +63,10 @@ static const struct st_tracked_state *atoms[] =
    &st_update_vs_constants,
    &st_update_gs_constants,
    &st_update_fs_constants,
-   &st_update_pixel_transfer
+   &st_update_pixel_transfer,
+
+   /* this must be done after the vertex program update */
+   &st_update_array
 };
 
 
@@ -122,6 +125,22 @@ static void check_program_state( struct st_context *st )
       st->dirty.st |= ST_NEW_GEOMETRY_PROGRAM;
 }
 
+static void check_attrib_edgeflag(struct st_context *st)
+{
+   const struct gl_client_array **arrays = st->ctx->Array._DrawArrays;
+   GLboolean vertDataEdgeFlags;
+
+   if (!arrays)
+      return;
+
+   vertDataEdgeFlags = arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj &&
+                       arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj->Name;
+   if (vertDataEdgeFlags != st->vertdata_edgeflags) {
+      st->vertdata_edgeflags = vertDataEdgeFlags;
+      st->dirty.st |= ST_NEW_EDGEFLAGS_DATA;
+   }
+}
+
 
 /***********************************************************************
  * Update all derived state:
@@ -132,6 +151,12 @@ void st_validate_state( struct st_context *st )
    struct st_state_flags *state = &st->dirty;
    GLuint i;
 
+   /* Get Mesa driver state. */
+   st->dirty.st |= st->ctx->NewDriverState;
+   st->ctx->NewDriverState = 0;
+
+   check_attrib_edgeflag(st);
+
    /* The bitmap cache is immune to pixel unpack changes.
     * Note that GLUT makes several calls to glPixelStore for each
     * bitmap char it draws so this is an important check.
diff --git a/mesalib/src/mesa/state_tracker/st_atom.h b/mesalib/src/mesa/state_tracker/st_atom.h
index 930a08444..703bc2ab3 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.h
+++ b/mesalib/src/mesa/state_tracker/st_atom.h
@@ -46,6 +46,7 @@ void st_destroy_atoms( struct st_context *st );
 void st_validate_state( struct st_context *st );
 
 
+extern const struct st_tracked_state st_update_array;
 extern const struct st_tracked_state st_update_framebuffer;
 extern const struct st_tracked_state st_update_clip;
 extern const struct st_tracked_state st_update_depth_stencil_alpha;
@@ -70,4 +71,8 @@ extern const struct st_tracked_state st_update_pixel_transfer;
 
 GLuint st_compare_func_to_pipe(GLenum func);
 
+enum pipe_format
+st_pipe_vertex_format(GLenum type, GLuint size, GLenum format,
+                      GLboolean normalized, GLboolean integer);
+
 #endif
diff --git a/mesalib/src/mesa/state_tracker/st_atom_array.c b/mesalib/src/mesa/state_tracker/st_atom_array.c
new file mode 100644
index 000000000..d60b0d7a9
--- /dev/null
+++ b/mesalib/src/mesa/state_tracker/st_atom_array.c
@@ -0,0 +1,588 @@
+/**************************************************************************
+ *
+ * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * Copyright 2012 Marek Olšák <maraeo@gmail.com>
+ * All Rights Reserved.
+ *
+ * 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, sub license, 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 NON-INFRINGEMENT.
+ * IN NO EVENT SHALL AUTHORS AND/OR ITS SUPPLIERS 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.
+ *
+ **************************************************************************/
+
+/*
+ * This converts the VBO's vertex attribute/array information into
+ * Gallium vertex state and binds it.
+ *
+ * Authors:
+ *   Keith Whitwell <keith@tungstengraphics.com>
+ *   Marek Olšák <maraeo@gmail.com>
+ */
+
+#include "st_context.h"
+#include "st_atom.h"
+#include "st_cb_bufferobjects.h"
+#include "st_draw.h"
+#include "st_program.h"
+
+#include "cso_cache/cso_context.h"
+#include "util/u_math.h"
+
+#include "main/bufferobj.h"
+#include "main/image.h"
+
+
+static GLuint double_types[4] = {
+   PIPE_FORMAT_R64_FLOAT,
+   PIPE_FORMAT_R64G64_FLOAT,
+   PIPE_FORMAT_R64G64B64_FLOAT,
+   PIPE_FORMAT_R64G64B64A64_FLOAT
+};
+
+static GLuint float_types[4] = {
+   PIPE_FORMAT_R32_FLOAT,
+   PIPE_FORMAT_R32G32_FLOAT,
+   PIPE_FORMAT_R32G32B32_FLOAT,
+   PIPE_FORMAT_R32G32B32A32_FLOAT
+};
+
+static GLuint half_float_types[4] = {
+   PIPE_FORMAT_R16_FLOAT,
+   PIPE_FORMAT_R16G16_FLOAT,
+   PIPE_FORMAT_R16G16B16_FLOAT,
+   PIPE_FORMAT_R16G16B16A16_FLOAT
+};
+
+static GLuint uint_types_norm[4] = {
+   PIPE_FORMAT_R32_UNORM,
+   PIPE_FORMAT_R32G32_UNORM,
+   PIPE_FORMAT_R32G32B32_UNORM,
+   PIPE_FORMAT_R32G32B32A32_UNORM
+};
+
+static GLuint uint_types_scale[4] = {
+   PIPE_FORMAT_R32_USCALED,
+   PIPE_FORMAT_R32G32_USCALED,
+   PIPE_FORMAT_R32G32B32_USCALED,
+   PIPE_FORMAT_R32G32B32A32_USCALED
+};
+
+static GLuint uint_types_int[4] = {
+   PIPE_FORMAT_R32_UINT,
+   PIPE_FORMAT_R32G32_UINT,
+   PIPE_FORMAT_R32G32B32_UINT,
+   PIPE_FORMAT_R32G32B32A32_UINT
+};
+
+static GLuint int_types_norm[4] = {
+   PIPE_FORMAT_R32_SNORM,
+   PIPE_FORMAT_R32G32_SNORM,
+   PIPE_FORMAT_R32G32B32_SNORM,
+   PIPE_FORMAT_R32G32B32A32_SNORM
+};
+
+static GLuint int_types_scale[4] = {
+   PIPE_FORMAT_R32_SSCALED,
+   PIPE_FORMAT_R32G32_SSCALED,
+   PIPE_FORMAT_R32G32B32_SSCALED,
+   PIPE_FORMAT_R32G32B32A32_SSCALED
+};
+
+static GLuint int_types_int[4] = {
+   PIPE_FORMAT_R32_SINT,
+   PIPE_FORMAT_R32G32_SINT,
+   PIPE_FORMAT_R32G32B32_SINT,
+   PIPE_FORMAT_R32G32B32A32_SINT
+};
+
+static GLuint ushort_types_norm[4] = {
+   PIPE_FORMAT_R16_UNORM,
+   PIPE_FORMAT_R16G16_UNORM,
+   PIPE_FORMAT_R16G16B16_UNORM,
+   PIPE_FORMAT_R16G16B16A16_UNORM
+};
+
+static GLuint ushort_types_scale[4] = {
+   PIPE_FORMAT_R16_USCALED,
+   PIPE_FORMAT_R16G16_USCALED,
+   PIPE_FORMAT_R16G16B16_USCALED,
+   PIPE_FORMAT_R16G16B16A16_USCALED
+};
+
+static GLuint ushort_types_int[4] = {
+   PIPE_FORMAT_R16_UINT,
+   PIPE_FORMAT_R16G16_UINT,
+   PIPE_FORMAT_R16G16B16_UINT,
+   PIPE_FORMAT_R16G16B16A16_UINT
+};
+
+static GLuint short_types_norm[4] = {
+   PIPE_FORMAT_R16_SNORM,
+   PIPE_FORMAT_R16G16_SNORM,
+   PIPE_FORMAT_R16G16B16_SNORM,
+   PIPE_FORMAT_R16G16B16A16_SNORM
+};
+
+static GLuint short_types_scale[4] = {
+   PIPE_FORMAT_R16_SSCALED,
+   PIPE_FORMAT_R16G16_SSCALED,
+   PIPE_FORMAT_R16G16B16_SSCALED,
+   PIPE_FORMAT_R16G16B16A16_SSCALED
+};
+
+static GLuint short_types_int[4] = {
+   PIPE_FORMAT_R16_SINT,
+   PIPE_FORMAT_R16G16_SINT,
+   PIPE_FORMAT_R16G16B16_SINT,
+   PIPE_FORMAT_R16G16B16A16_SINT
+};
+
+static GLuint ubyte_types_norm[4] = {
+   PIPE_FORMAT_R8_UNORM,
+   PIPE_FORMAT_R8G8_UNORM,
+   PIPE_FORMAT_R8G8B8_UNORM,
+   PIPE_FORMAT_R8G8B8A8_UNORM
+};
+
+static GLuint ubyte_types_scale[4] = {
+   PIPE_FORMAT_R8_USCALED,
+   PIPE_FORMAT_R8G8_USCALED,
+   PIPE_FORMAT_R8G8B8_USCALED,
+   PIPE_FORMAT_R8G8B8A8_USCALED
+};
+
+static GLuint ubyte_types_int[4] = {
+   PIPE_FORMAT_R8_UINT,
+   PIPE_FORMAT_R8G8_UINT,
+   PIPE_FORMAT_R8G8B8_UINT,
+   PIPE_FORMAT_R8G8B8A8_UINT
+};
+
+static GLuint byte_types_norm[4] = {
+   PIPE_FORMAT_R8_SNORM,
+   PIPE_FORMAT_R8G8_SNORM,
+   PIPE_FORMAT_R8G8B8_SNORM,
+   PIPE_FORMAT_R8G8B8A8_SNORM
+};
+
+static GLuint byte_types_scale[4] = {
+   PIPE_FORMAT_R8_SSCALED,
+   PIPE_FORMAT_R8G8_SSCALED,
+   PIPE_FORMAT_R8G8B8_SSCALED,
+   PIPE_FORMAT_R8G8B8A8_SSCALED
+};
+
+static GLuint byte_types_int[4] = {
+   PIPE_FORMAT_R8_SINT,
+   PIPE_FORMAT_R8G8_SINT,
+   PIPE_FORMAT_R8G8B8_SINT,
+   PIPE_FORMAT_R8G8B8A8_SINT
+};
+
+static GLuint fixed_types[4] = {
+   PIPE_FORMAT_R32_FIXED,
+   PIPE_FORMAT_R32G32_FIXED,
+   PIPE_FORMAT_R32G32B32_FIXED,
+   PIPE_FORMAT_R32G32B32A32_FIXED
+};
+
+
+/**
+ * Return a PIPE_FORMAT_x for the given GL datatype and size.
+ */
+enum pipe_format
+st_pipe_vertex_format(GLenum type, GLuint size, GLenum format,
+                      GLboolean normalized, GLboolean integer)
+{
+   assert((type >= GL_BYTE && type <= GL_DOUBLE) ||
+          type == GL_FIXED || type == GL_HALF_FLOAT ||
+          type == GL_INT_2_10_10_10_REV ||
+          type == GL_UNSIGNED_INT_2_10_10_10_REV);
+   assert(size >= 1);
+   assert(size <= 4);
+   assert(format == GL_RGBA || format == GL_BGRA);
+
+   if (type == GL_INT_2_10_10_10_REV ||
+       type == GL_UNSIGNED_INT_2_10_10_10_REV) {
+      assert(size == 4);
+      assert(!integer);
+
+      if (format == GL_BGRA) {
+         if (type == GL_INT_2_10_10_10_REV) {
+            if (normalized)
+               return PIPE_FORMAT_B10G10R10A2_SNORM;
+            else
+               return PIPE_FORMAT_B10G10R10A2_SSCALED;
+         } else {
+            if (normalized)
+               return PIPE_FORMAT_B10G10R10A2_UNORM;
+            else
+               return PIPE_FORMAT_B10G10R10A2_USCALED;
+         }
+      } else {
+         if (type == GL_INT_2_10_10_10_REV) {
+            if (normalized)
+               return PIPE_FORMAT_R10G10B10A2_SNORM;
+            else
+               return PIPE_FORMAT_R10G10B10A2_SSCALED;
+         } else {
+            if (normalized)
+               return PIPE_FORMAT_R10G10B10A2_UNORM;
+            else
+               return PIPE_FORMAT_R10G10B10A2_USCALED;
+         }
+      }
+   }
+
+   if (format == GL_BGRA) {
+      /* this is an odd-ball case */
+      assert(type == GL_UNSIGNED_BYTE);
+      assert(normalized);
+      return PIPE_FORMAT_B8G8R8A8_UNORM;
+   }
+
+   if (integer) {
+      switch (type) {
+      case GL_INT: return int_types_int[size-1];
+      case GL_SHORT: return short_types_int[size-1];
+      case GL_BYTE: return byte_types_int[size-1];
+      case GL_UNSIGNED_INT: return uint_types_int[size-1];
+      case GL_UNSIGNED_SHORT: return ushort_types_int[size-1];
+      case GL_UNSIGNED_BYTE: return ubyte_types_int[size-1];
+      default: assert(0); return 0;
+      }
+   }
+   else if (normalized) {
+      switch (type) {
+      case GL_DOUBLE: return double_types[size-1];
+      case GL_FLOAT: return float_types[size-1];
+      case GL_HALF_FLOAT: return half_float_types[size-1];
+      case GL_INT: return int_types_norm[size-1];
+      case GL_SHORT: return short_types_norm[size-1];
+      case GL_BYTE: return byte_types_norm[size-1];
+      case GL_UNSIGNED_INT: return uint_types_norm[size-1];
+      case GL_UNSIGNED_SHORT: return ushort_types_norm[size-1];
+      case GL_UNSIGNED_BYTE: return ubyte_types_norm[size-1];
+      case GL_FIXED: return fixed_types[size-1];
+      default: assert(0); return 0;
+      }
+   }
+   else {
+      switch (type) {
+      case GL_DOUBLE: return double_types[size-1];
+      case GL_FLOAT: return float_types[size-1];
+      case GL_HALF_FLOAT: return half_float_types[size-1];
+      case GL_INT: return int_types_scale[size-1];
+      case GL_SHORT: return short_types_scale[size-1];
+      case GL_BYTE: return byte_types_scale[size-1];
+      case GL_UNSIGNED_INT: return uint_types_scale[size-1];
+      case GL_UNSIGNED_SHORT: return ushort_types_scale[size-1];
+      case GL_UNSIGNED_BYTE: return ubyte_types_scale[size-1];
+      case GL_FIXED: return fixed_types[size-1];
+      default: assert(0); return 0;
+      }
+   }
+   return PIPE_FORMAT_NONE; /* silence compiler warning */
+}
+
+/**
+ * Examine the active arrays to determine if we have interleaved
+ * vertex arrays all living in one VBO, or all living in user space.
+ */
+static GLboolean
+is_interleaved_arrays(const struct st_vertex_program *vp,
+                      const struct st_vp_variant *vpv,
+                      const struct gl_client_array **arrays)
+{
+   GLuint attr;
+   const struct gl_buffer_object *firstBufObj = NULL;
+   GLint firstStride = -1;
+   const GLubyte *firstPtr = NULL;
+   GLboolean userSpaceBuffer = GL_FALSE;
+
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
+      const GLuint mesaAttr = vp->index_to_input[attr];
+      const struct gl_client_array *array = arrays[mesaAttr];
+      const struct gl_buffer_object *bufObj = array->BufferObj;
+      const GLsizei stride = array->StrideB; /* in bytes */
+
+      if (attr == 0) {
+         /* save info about the first array */
+         firstStride = stride;
+         firstPtr = array->Ptr;
+         firstBufObj = bufObj;
+         userSpaceBuffer = !bufObj || !bufObj->Name;
+      }
+      else {
+         /* check if other arrays interleave with the first, in same buffer */
+         if (stride != firstStride)
+            return GL_FALSE; /* strides don't match */
+
+         if (bufObj != firstBufObj)
+            return GL_FALSE; /* arrays in different VBOs */
+
+         if (abs(array->Ptr - firstPtr) > firstStride)
+            return GL_FALSE; /* arrays start too far apart */
+
+         if ((!_mesa_is_bufferobj(bufObj)) != userSpaceBuffer)
+            return GL_FALSE; /* mix of VBO and user-space arrays */
+      }
+   }
+
+   return GL_TRUE;
+}
+
+/**
+ * Set up for drawing interleaved arrays that all live in one VBO
+ * or all live in user space.
+ * \param vbuffer  returns vertex buffer info
+ * \param velements  returns vertex element info
+ */
+static boolean
+setup_interleaved_attribs(const struct st_vertex_program *vp,
+                          const struct st_vp_variant *vpv,
+                          const struct gl_client_array **arrays,
+                          struct pipe_vertex_buffer *vbuffer,
+                          struct pipe_vertex_element velements[])
+{
+   GLuint attr;
+   const GLubyte *low_addr = NULL;
+   GLboolean usingVBO;      /* all arrays in a VBO? */
+   struct gl_buffer_object *bufobj;
+   GLsizei stride;
+
+   /* Find the lowest address of the arrays we're drawing,
+    * Init bufobj and stride.
+    */
+   if (vpv->num_inputs) {
+      const GLuint mesaAttr0 = vp->index_to_input[0];
+      const struct gl_client_array *array = arrays[mesaAttr0];
+
+      /* Since we're doing interleaved arrays, we know there'll be at most
+       * one buffer object and the stride will be the same for all arrays.
+       * Grab them now.
+       */
+      bufobj = array->BufferObj;
+      stride = array->StrideB;
+
+      low_addr = arrays[vp->index_to_input[0]]->Ptr;
+
+      for (attr = 1; attr < vpv->num_inputs; attr++) {
+         const GLubyte *start = arrays[vp->index_to_input[attr]]->Ptr;
+         low_addr = MIN2(low_addr, start);
+      }
+   }
+   else {
+      /* not sure we'll ever have zero inputs, but play it safe */
+      bufobj = NULL;
+      stride = 0;
+      low_addr = 0;
+   }
+
+   /* are the arrays in user space? */
+   usingVBO = _mesa_is_bufferobj(bufobj);
+
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
+      const GLuint mesaAttr = vp->index_to_input[attr];
+      const struct gl_client_array *array = arrays[mesaAttr];
+      unsigned src_offset = (unsigned) (array->Ptr - low_addr);
+      GLuint element_size = array->_ElementSize;
+
+      assert(element_size == array->Size * _mesa_sizeof_type(array->Type));
+
+      velements[attr].src_offset = src_offset;
+      velements[attr].instance_divisor = array->InstanceDivisor;
+      velements[attr].vertex_buffer_index = 0;
+      velements[attr].src_format = st_pipe_vertex_format(array->Type,
+                                                         array->Size,
+                                                         array->Format,
+                                                         array->Normalized,
+                                                         array->Integer);
+      assert(velements[attr].src_format);
+   }
+
+   /*
+    * Return the vbuffer info and setup user-space attrib info, if needed.
+    */
+   if (vpv->num_inputs == 0) {
+      /* just defensive coding here */
+      vbuffer->buffer = NULL;
+      vbuffer->user_buffer = NULL;
+      vbuffer->buffer_offset = 0;
+      vbuffer->stride = 0;
+   }
+   else if (usingVBO) {
+      /* all interleaved arrays in a VBO */
+      struct st_buffer_object *stobj = st_buffer_object(bufobj);
+
+      if (!stobj || !stobj->buffer) {
+         return FALSE; /* out-of-memory error probably */
+      }
+
+      vbuffer->buffer = stobj->buffer;
+      vbuffer->user_buffer = NULL;
+      vbuffer->buffer_offset = pointer_to_offset(low_addr);
+      vbuffer->stride = stride;
+   }
+   else {
+      /* all interleaved arrays in user memory */
+      vbuffer->buffer = NULL;
+      vbuffer->user_buffer = low_addr;
+      vbuffer->buffer_offset = 0;
+      vbuffer->stride = stride;
+   }
+   return TRUE;
+}
+
+/**
+ * Set up a separate pipe_vertex_buffer and pipe_vertex_element for each
+ * vertex attribute.
+ * \param vbuffer  returns vertex buffer info
+ * \param velements  returns vertex element info
+ */
+static boolean
+setup_non_interleaved_attribs(struct st_context *st,
+                              const struct st_vertex_program *vp,
+                              const struct st_vp_variant *vpv,
+                              const struct gl_client_array **arrays,
+                              struct pipe_vertex_buffer vbuffer[],
+                              struct pipe_vertex_element velements[])
+{
+   struct gl_context *ctx = st->ctx;
+   GLuint attr;
+
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
+      const GLuint mesaAttr = vp->index_to_input[attr];
+      const struct gl_client_array *array = arrays[mesaAttr];
+      struct gl_buffer_object *bufobj = array->BufferObj;
+      GLsizei stride = array->StrideB;
+
+      assert(array->_ElementSize == array->Size * _mesa_sizeof_type(array->Type));
+
+      if (_mesa_is_bufferobj(bufobj)) {
+         /* Attribute data is in a VBO.
+          * Recall that for VBOs, the gl_client_array->Ptr field is
+          * really an offset from the start of the VBO, not a pointer.
+          */
+         struct st_buffer_object *stobj = st_buffer_object(bufobj);
+
+         if (!stobj || !stobj->buffer) {
+            return FALSE; /* out-of-memory error probably */
+         }
+
+         vbuffer[attr].buffer = stobj->buffer;
+         vbuffer[attr].user_buffer = NULL;
+         vbuffer[attr].buffer_offset = pointer_to_offset(array->Ptr);
+      }
+      else {
+         /* wrap user data */
+         void *ptr;
+
+         if (array->Ptr) {
+            ptr = (void *) array->Ptr;
+         }
+         else {
+            /* no array, use ctx->Current.Attrib[] value */
+            ptr = (void *) ctx->Current.Attrib[mesaAttr];
+            stride = 0;
+         }
+
+         assert(ptr);
+
+         vbuffer[attr].buffer = NULL;
+         vbuffer[attr].user_buffer = ptr;
+         vbuffer[attr].buffer_offset = 0;
+      }
+
+      /* common-case setup */
+      vbuffer[attr].stride = stride; /* in bytes */
+
+      velements[attr].src_offset = 0;
+      velements[attr].instance_divisor = array->InstanceDivisor;
+      velements[attr].vertex_buffer_index = attr;
+      velements[attr].src_format = st_pipe_vertex_format(array->Type,
+                                                         array->Size,
+                                                         array->Format,
+                                                         array->Normalized,
+                                                         array->Integer);
+      assert(velements[attr].src_format);
+   }
+   return TRUE;
+}
+
+static void update_array(struct st_context *st)
+{
+   struct gl_context *ctx = st->ctx;
+   const struct gl_client_array **arrays = ctx->Array._DrawArrays;
+   const struct st_vertex_program *vp;
+   const struct st_vp_variant *vpv;
+   struct pipe_vertex_buffer vbuffer[PIPE_MAX_SHADER_INPUTS];
+   struct pipe_vertex_element velements[PIPE_MAX_ATTRIBS];
+   unsigned num_vbuffers, num_velements;
+
+   st->vertex_array_out_of_memory = FALSE;
+
+   /* No drawing has been done yet, so do nothing. */
+   if (!arrays)
+      return;
+
+   /* vertex program validation must be done before this */
+   vp = st->vp;
+   vpv = st->vp_variant;
+
+   memset(velements, 0, sizeof(struct pipe_vertex_element) * vpv->num_inputs);
+
+   /*
+    * Setup the vbuffer[] and velements[] arrays.
+    */
+   if (is_interleaved_arrays(vp, vpv, arrays)) {
+      if (!setup_interleaved_attribs(vp, vpv, arrays, vbuffer, velements)) {
+         st->vertex_array_out_of_memory = TRUE;
+         return;
+      }
+
+      num_vbuffers = 1;
+      num_velements = vpv->num_inputs;
+      if (num_velements == 0)
+         num_vbuffers = 0;
+   }
+   else {
+      if (!setup_non_interleaved_attribs(st, vp, vpv, arrays, vbuffer,
+                                         velements)) {
+         st->vertex_array_out_of_memory = TRUE;
+         return;
+      }
+
+      num_vbuffers = vpv->num_inputs;
+      num_velements = vpv->num_inputs;
+   }
+
+   cso_set_vertex_buffers(st->cso_context, num_vbuffers, vbuffer);
+   cso_set_vertex_elements(st->cso_context, num_velements, velements);
+}
+
+
+const struct st_tracked_state st_update_array = {
+   "st_update_array",					/* name */
+   {							/* dirty */
+      (_NEW_PROGRAM | _NEW_BUFFER_OBJECT),		/* mesa */
+      ST_NEW_VERTEX_ARRAYS | ST_NEW_VERTEX_PROGRAM,     /* st */
+   },
+   update_array						/* update */
+};
diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
index 9a3f22465..d3f8d1368 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -217,7 +217,7 @@ st_make_drawpix_z_stencil_program(struct st_context *st,
    if (!p)
       return NULL;
 
-   p->NumInstructions = write_depth ? 2 : 1;
+   p->NumInstructions = write_depth ? 3 : 1;
    p->NumInstructions += write_stencil ? 1 : 0;
 
    p->Instructions = _mesa_alloc_instructions(p->NumInstructions);
@@ -238,6 +238,13 @@ st_make_drawpix_z_stencil_program(struct st_context *st,
       p->Instructions[ic].TexSrcUnit = 0;
       p->Instructions[ic].TexSrcTarget = TEXTURE_2D_INDEX;
       ic++;
+      /* MOV result.color, fragment.color; */
+      p->Instructions[ic].Opcode = OPCODE_MOV;
+      p->Instructions[ic].DstReg.File = PROGRAM_OUTPUT;
+      p->Instructions[ic].DstReg.Index = FRAG_RESULT_COLOR;
+      p->Instructions[ic].SrcReg[0].File = PROGRAM_INPUT;
+      p->Instructions[ic].SrcReg[0].Index = FRAG_ATTRIB_COL0;
+      ic++;
    }
 
    if (write_stencil) {
@@ -260,8 +267,10 @@ st_make_drawpix_z_stencil_program(struct st_context *st,
 
    p->InputsRead = FRAG_BIT_TEX0 | FRAG_BIT_COL0;
    p->OutputsWritten = 0;
-   if (write_depth)
+   if (write_depth) {
       p->OutputsWritten |= BITFIELD64_BIT(FRAG_RESULT_DEPTH);
+      p->OutputsWritten |= BITFIELD64_BIT(FRAG_RESULT_COLOR);
+   }
    if (write_stencil)
       p->OutputsWritten |= BITFIELD64_BIT(FRAG_RESULT_STENCIL);
 
diff --git a/mesalib/src/mesa/state_tracker/st_context.h b/mesalib/src/mesa/state_tracker/st_context.h
index 55ae65b3c..734b4d9c1 100644
--- a/mesalib/src/mesa/state_tracker/st_context.h
+++ b/mesalib/src/mesa/state_tracker/st_context.h
@@ -88,6 +88,8 @@ struct st_context
     */
    boolean invalidate_on_gl_viewport;
 
+   boolean vertex_array_out_of_memory;
+
    /* Some state is contained in constant objects.
     * Other state is just parameter values.
     */
diff --git a/mesalib/src/mesa/state_tracker/st_draw.c b/mesalib/src/mesa/state_tracker/st_draw.c
index 0a06e9995..db8caa566 100644
--- a/mesalib/src/mesa/state_tracker/st_draw.c
+++ b/mesalib/src/mesa/state_tracker/st_draw.c
@@ -31,10 +31,6 @@
  * this function whether the user called glBegin/End, glDrawArrays,
  * glDrawElements, glEvalMesh, or glCalList, etc.
  *
- * We basically convert the VBO's vertex attribute/array information into
- * Gallium vertex state, bind the vertex buffer objects and call
- * pipe->draw_vbo().
- *
  * Authors:
  *   Keith Whitwell <keith@tungstengraphics.com>
  */
@@ -68,261 +64,6 @@
 #include "../glsl/ir_uniform.h"
 
 
-static GLuint double_types[4] = {
-   PIPE_FORMAT_R64_FLOAT,
-   PIPE_FORMAT_R64G64_FLOAT,
-   PIPE_FORMAT_R64G64B64_FLOAT,
-   PIPE_FORMAT_R64G64B64A64_FLOAT
-};
-
-static GLuint float_types[4] = {
-   PIPE_FORMAT_R32_FLOAT,
-   PIPE_FORMAT_R32G32_FLOAT,
-   PIPE_FORMAT_R32G32B32_FLOAT,
-   PIPE_FORMAT_R32G32B32A32_FLOAT
-};
-
-static GLuint half_float_types[4] = {
-   PIPE_FORMAT_R16_FLOAT,
-   PIPE_FORMAT_R16G16_FLOAT,
-   PIPE_FORMAT_R16G16B16_FLOAT,
-   PIPE_FORMAT_R16G16B16A16_FLOAT
-};
-
-static GLuint uint_types_norm[4] = {
-   PIPE_FORMAT_R32_UNORM,
-   PIPE_FORMAT_R32G32_UNORM,
-   PIPE_FORMAT_R32G32B32_UNORM,
-   PIPE_FORMAT_R32G32B32A32_UNORM
-};
-
-static GLuint uint_types_scale[4] = {
-   PIPE_FORMAT_R32_USCALED,
-   PIPE_FORMAT_R32G32_USCALED,
-   PIPE_FORMAT_R32G32B32_USCALED,
-   PIPE_FORMAT_R32G32B32A32_USCALED
-};
-
-static GLuint uint_types_int[4] = {
-   PIPE_FORMAT_R32_UINT,
-   PIPE_FORMAT_R32G32_UINT,
-   PIPE_FORMAT_R32G32B32_UINT,
-   PIPE_FORMAT_R32G32B32A32_UINT
-};
-
-static GLuint int_types_norm[4] = {
-   PIPE_FORMAT_R32_SNORM,
-   PIPE_FORMAT_R32G32_SNORM,
-   PIPE_FORMAT_R32G32B32_SNORM,
-   PIPE_FORMAT_R32G32B32A32_SNORM
-};
-
-static GLuint int_types_scale[4] = {
-   PIPE_FORMAT_R32_SSCALED,
-   PIPE_FORMAT_R32G32_SSCALED,
-   PIPE_FORMAT_R32G32B32_SSCALED,
-   PIPE_FORMAT_R32G32B32A32_SSCALED
-};
-
-static GLuint int_types_int[4] = {
-   PIPE_FORMAT_R32_SINT,
-   PIPE_FORMAT_R32G32_SINT,
-   PIPE_FORMAT_R32G32B32_SINT,
-   PIPE_FORMAT_R32G32B32A32_SINT
-};
-
-static GLuint ushort_types_norm[4] = {
-   PIPE_FORMAT_R16_UNORM,
-   PIPE_FORMAT_R16G16_UNORM,
-   PIPE_FORMAT_R16G16B16_UNORM,
-   PIPE_FORMAT_R16G16B16A16_UNORM
-};
-
-static GLuint ushort_types_scale[4] = {
-   PIPE_FORMAT_R16_USCALED,
-   PIPE_FORMAT_R16G16_USCALED,
-   PIPE_FORMAT_R16G16B16_USCALED,
-   PIPE_FORMAT_R16G16B16A16_USCALED
-};
-
-static GLuint ushort_types_int[4] = {
-   PIPE_FORMAT_R16_UINT,
-   PIPE_FORMAT_R16G16_UINT,
-   PIPE_FORMAT_R16G16B16_UINT,
-   PIPE_FORMAT_R16G16B16A16_UINT
-};
-
-static GLuint short_types_norm[4] = {
-   PIPE_FORMAT_R16_SNORM,
-   PIPE_FORMAT_R16G16_SNORM,
-   PIPE_FORMAT_R16G16B16_SNORM,
-   PIPE_FORMAT_R16G16B16A16_SNORM
-};
-
-static GLuint short_types_scale[4] = {
-   PIPE_FORMAT_R16_SSCALED,
-   PIPE_FORMAT_R16G16_SSCALED,
-   PIPE_FORMAT_R16G16B16_SSCALED,
-   PIPE_FORMAT_R16G16B16A16_SSCALED
-};
-
-static GLuint short_types_int[4] = {
-   PIPE_FORMAT_R16_SINT,
-   PIPE_FORMAT_R16G16_SINT,
-   PIPE_FORMAT_R16G16B16_SINT,
-   PIPE_FORMAT_R16G16B16A16_SINT
-};
-
-static GLuint ubyte_types_norm[4] = {
-   PIPE_FORMAT_R8_UNORM,
-   PIPE_FORMAT_R8G8_UNORM,
-   PIPE_FORMAT_R8G8B8_UNORM,
-   PIPE_FORMAT_R8G8B8A8_UNORM
-};
-
-static GLuint ubyte_types_scale[4] = {
-   PIPE_FORMAT_R8_USCALED,
-   PIPE_FORMAT_R8G8_USCALED,
-   PIPE_FORMAT_R8G8B8_USCALED,
-   PIPE_FORMAT_R8G8B8A8_USCALED
-};
-
-static GLuint ubyte_types_int[4] = {
-   PIPE_FORMAT_R8_UINT,
-   PIPE_FORMAT_R8G8_UINT,
-   PIPE_FORMAT_R8G8B8_UINT,
-   PIPE_FORMAT_R8G8B8A8_UINT
-};
-
-static GLuint byte_types_norm[4] = {
-   PIPE_FORMAT_R8_SNORM,
-   PIPE_FORMAT_R8G8_SNORM,
-   PIPE_FORMAT_R8G8B8_SNORM,
-   PIPE_FORMAT_R8G8B8A8_SNORM
-};
-
-static GLuint byte_types_scale[4] = {
-   PIPE_FORMAT_R8_SSCALED,
-   PIPE_FORMAT_R8G8_SSCALED,
-   PIPE_FORMAT_R8G8B8_SSCALED,
-   PIPE_FORMAT_R8G8B8A8_SSCALED
-};
-
-static GLuint byte_types_int[4] = {
-   PIPE_FORMAT_R8_SINT,
-   PIPE_FORMAT_R8G8_SINT,
-   PIPE_FORMAT_R8G8B8_SINT,
-   PIPE_FORMAT_R8G8B8A8_SINT
-};
-
-static GLuint fixed_types[4] = {
-   PIPE_FORMAT_R32_FIXED,
-   PIPE_FORMAT_R32G32_FIXED,
-   PIPE_FORMAT_R32G32B32_FIXED,
-   PIPE_FORMAT_R32G32B32A32_FIXED
-};
-
-
-
-/**
- * Return a PIPE_FORMAT_x for the given GL datatype and size.
- */
-enum pipe_format
-st_pipe_vertex_format(GLenum type, GLuint size, GLenum format,
-                      GLboolean normalized, GLboolean integer)
-{
-   assert((type >= GL_BYTE && type <= GL_DOUBLE) ||
-          type == GL_FIXED || type == GL_HALF_FLOAT ||
-          type == GL_INT_2_10_10_10_REV ||
-          type == GL_UNSIGNED_INT_2_10_10_10_REV);
-   assert(size >= 1);
-   assert(size <= 4);
-   assert(format == GL_RGBA || format == GL_BGRA);
-
-   if (type == GL_INT_2_10_10_10_REV ||
-       type == GL_UNSIGNED_INT_2_10_10_10_REV) {
-      assert(size == 4);
-      assert(!integer);
-
-      if (format == GL_BGRA) {
-         if (type == GL_INT_2_10_10_10_REV) {
-            if (normalized)
-               return PIPE_FORMAT_B10G10R10A2_SNORM;
-            else
-               return PIPE_FORMAT_B10G10R10A2_SSCALED;
-         } else {
-            if (normalized)
-               return PIPE_FORMAT_B10G10R10A2_UNORM;
-            else
-               return PIPE_FORMAT_B10G10R10A2_USCALED;
-         }
-      } else {
-         if (type == GL_INT_2_10_10_10_REV) {
-            if (normalized)
-               return PIPE_FORMAT_R10G10B10A2_SNORM;
-            else
-               return PIPE_FORMAT_R10G10B10A2_SSCALED;
-         } else {
-            if (normalized)
-               return PIPE_FORMAT_R10G10B10A2_UNORM;
-            else
-               return PIPE_FORMAT_R10G10B10A2_USCALED;
-         }
-      }
-   }
-
-   if (format == GL_BGRA) {
-      /* this is an odd-ball case */
-      assert(type == GL_UNSIGNED_BYTE);
-      assert(normalized);
-      return PIPE_FORMAT_B8G8R8A8_UNORM;
-   }
-
-   if (integer) {
-      switch (type) {
-      case GL_INT: return int_types_int[size-1];
-      case GL_SHORT: return short_types_int[size-1];
-      case GL_BYTE: return byte_types_int[size-1];
-      case GL_UNSIGNED_INT: return uint_types_int[size-1];
-      case GL_UNSIGNED_SHORT: return ushort_types_int[size-1];
-      case GL_UNSIGNED_BYTE: return ubyte_types_int[size-1];
-      default: assert(0); return 0;
-      }
-   }
-   else if (normalized) {
-      switch (type) {
-      case GL_DOUBLE: return double_types[size-1];
-      case GL_FLOAT: return float_types[size-1];
-      case GL_HALF_FLOAT: return half_float_types[size-1];
-      case GL_INT: return int_types_norm[size-1];
-      case GL_SHORT: return short_types_norm[size-1];
-      case GL_BYTE: return byte_types_norm[size-1];
-      case GL_UNSIGNED_INT: return uint_types_norm[size-1];
-      case GL_UNSIGNED_SHORT: return ushort_types_norm[size-1];
-      case GL_UNSIGNED_BYTE: return ubyte_types_norm[size-1];
-      case GL_FIXED: return fixed_types[size-1];
-      default: assert(0); return 0;
-      }
-   }
-   else {
-      switch (type) {
-      case GL_DOUBLE: return double_types[size-1];
-      case GL_FLOAT: return float_types[size-1];
-      case GL_HALF_FLOAT: return half_float_types[size-1];
-      case GL_INT: return int_types_scale[size-1];
-      case GL_SHORT: return short_types_scale[size-1];
-      case GL_BYTE: return byte_types_scale[size-1];
-      case GL_UNSIGNED_INT: return uint_types_scale[size-1];
-      case GL_UNSIGNED_SHORT: return ushort_types_scale[size-1];
-      case GL_UNSIGNED_BYTE: return ubyte_types_scale[size-1];
-      case GL_FIXED: return fixed_types[size-1];
-      default: assert(0); return 0;
-      }
-   }
-   return PIPE_FORMAT_NONE; /* silence compiler warning */
-}
-
-
 /**
  * This is very similar to vbo_all_varyings_in_vbos() but we are
  * only interested in per-vertex data.  See bug 38626.
@@ -342,240 +83,6 @@ all_varyings_in_vbos(const struct gl_client_array *arrays[])
 }
 
 
-/**
- * Examine the active arrays to determine if we have interleaved
- * vertex arrays all living in one VBO, or all living in user space.
- */
-static GLboolean
-is_interleaved_arrays(const struct st_vertex_program *vp,
-                      const struct st_vp_variant *vpv,
-                      const struct gl_client_array **arrays)
-{
-   GLuint attr;
-   const struct gl_buffer_object *firstBufObj = NULL;
-   GLint firstStride = -1;
-   const GLubyte *firstPtr = NULL;
-   GLboolean userSpaceBuffer = GL_FALSE;
-
-   for (attr = 0; attr < vpv->num_inputs; attr++) {
-      const GLuint mesaAttr = vp->index_to_input[attr];
-      const struct gl_client_array *array = arrays[mesaAttr];
-      const struct gl_buffer_object *bufObj = array->BufferObj;
-      const GLsizei stride = array->StrideB; /* in bytes */
-
-      if (attr == 0) {
-         /* save info about the first array */
-         firstStride = stride;
-         firstPtr = array->Ptr;         
-         firstBufObj = bufObj;
-         userSpaceBuffer = !bufObj || !bufObj->Name;
-      }
-      else {
-         /* check if other arrays interleave with the first, in same buffer */
-         if (stride != firstStride)
-            return GL_FALSE; /* strides don't match */
-
-         if (bufObj != firstBufObj)
-            return GL_FALSE; /* arrays in different VBOs */
-
-         if (abs(array->Ptr - firstPtr) > firstStride)
-            return GL_FALSE; /* arrays start too far apart */
-
-         if ((!_mesa_is_bufferobj(bufObj)) != userSpaceBuffer)
-            return GL_FALSE; /* mix of VBO and user-space arrays */
-      }
-   }
-
-   return GL_TRUE;
-}
-
-
-/**
- * Set up for drawing interleaved arrays that all live in one VBO
- * or all live in user space.
- * \param vbuffer  returns vertex buffer info
- * \param velements  returns vertex element info
- * \return GL_TRUE for success, GL_FALSE otherwise (probably out of memory)
- */
-static GLboolean
-setup_interleaved_attribs(struct gl_context *ctx,
-                          const struct st_vertex_program *vp,
-                          const struct st_vp_variant *vpv,
-                          const struct gl_client_array **arrays,
-                          struct pipe_vertex_buffer *vbuffer,
-                          struct pipe_vertex_element velements[])
-{
-   GLuint attr;
-   const GLubyte *low_addr = NULL;
-   GLboolean usingVBO;      /* all arrays in a VBO? */
-   struct gl_buffer_object *bufobj;
-   GLsizei stride;
-
-   /* Find the lowest address of the arrays we're drawing,
-    * Init bufobj and stride.
-    */
-   if (vpv->num_inputs) {
-      const GLuint mesaAttr0 = vp->index_to_input[0];
-      const struct gl_client_array *array = arrays[mesaAttr0];
-
-      /* Since we're doing interleaved arrays, we know there'll be at most
-       * one buffer object and the stride will be the same for all arrays.
-       * Grab them now.
-       */
-      bufobj = array->BufferObj;
-      stride = array->StrideB;
-
-      low_addr = arrays[vp->index_to_input[0]]->Ptr;
-
-      for (attr = 1; attr < vpv->num_inputs; attr++) {
-         const GLubyte *start = arrays[vp->index_to_input[attr]]->Ptr;
-         low_addr = MIN2(low_addr, start);
-      }
-   }
-   else {
-      /* not sure we'll ever have zero inputs, but play it safe */
-      bufobj = NULL;
-      stride = 0;
-      low_addr = 0;
-   }
-
-   /* are the arrays in user space? */
-   usingVBO = _mesa_is_bufferobj(bufobj);
-
-   for (attr = 0; attr < vpv->num_inputs; attr++) {
-      const GLuint mesaAttr = vp->index_to_input[attr];
-      const struct gl_client_array *array = arrays[mesaAttr];
-      unsigned src_offset = (unsigned) (array->Ptr - low_addr);
-      GLuint element_size = array->_ElementSize;
-
-      assert(element_size == array->Size * _mesa_sizeof_type(array->Type));
-
-      velements[attr].src_offset = src_offset;
-      velements[attr].instance_divisor = array->InstanceDivisor;
-      velements[attr].vertex_buffer_index = 0;
-      velements[attr].src_format = st_pipe_vertex_format(array->Type,
-                                                         array->Size,
-                                                         array->Format,
-                                                         array->Normalized,
-                                                         array->Integer);
-      assert(velements[attr].src_format);
-   }
-
-   /*
-    * Return the vbuffer info and setup user-space attrib info, if needed.
-    */
-   if (vpv->num_inputs == 0) {
-      /* just defensive coding here */
-      vbuffer->buffer = NULL;
-      vbuffer->user_buffer = NULL;
-      vbuffer->buffer_offset = 0;
-      vbuffer->stride = 0;
-   }
-   else if (usingVBO) {
-      /* all interleaved arrays in a VBO */
-      struct st_buffer_object *stobj = st_buffer_object(bufobj);
-
-      if (!stobj || !stobj->buffer) {
-         /* probably out of memory (or zero-sized buffer) */
-         return GL_FALSE;
-      }
-
-      vbuffer->buffer = stobj->buffer;
-      vbuffer->user_buffer = NULL;
-      vbuffer->buffer_offset = pointer_to_offset(low_addr);
-      vbuffer->stride = stride;
-   }
-   else {
-      /* all interleaved arrays in user memory */
-      vbuffer->buffer = NULL;
-      vbuffer->user_buffer = low_addr;
-      vbuffer->buffer_offset = 0;
-      vbuffer->stride = stride;
-   }
-
-   return GL_TRUE;
-}
-
-
-/**
- * Set up a separate pipe_vertex_buffer and pipe_vertex_element for each
- * vertex attribute.
- * \param vbuffer  returns vertex buffer info
- * \param velements  returns vertex element info
- * \return GL_TRUE for success, GL_FALSE otherwise (probably out of memory)
- */
-static GLboolean
-setup_non_interleaved_attribs(struct gl_context *ctx,
-                              const struct st_vertex_program *vp,
-                              const struct st_vp_variant *vpv,
-                              const struct gl_client_array **arrays,
-                              struct pipe_vertex_buffer vbuffer[],
-                              struct pipe_vertex_element velements[])
-{
-   GLuint attr;
-
-   for (attr = 0; attr < vpv->num_inputs; attr++) {
-      const GLuint mesaAttr = vp->index_to_input[attr];
-      const struct gl_client_array *array = arrays[mesaAttr];
-      struct gl_buffer_object *bufobj = array->BufferObj;
-      GLsizei stride = array->StrideB;
-
-      assert(array->_ElementSize == array->Size * _mesa_sizeof_type(array->Type));
-
-      if (_mesa_is_bufferobj(bufobj)) {
-         /* Attribute data is in a VBO.
-          * Recall that for VBOs, the gl_client_array->Ptr field is
-          * really an offset from the start of the VBO, not a pointer.
-          */
-         struct st_buffer_object *stobj = st_buffer_object(bufobj);
-
-         if (!stobj || !stobj->buffer) {
-            /* probably out of memory (or zero-sized buffer) */
-            return GL_FALSE;
-         }
-
-         vbuffer[attr].buffer = stobj->buffer;
-         vbuffer[attr].user_buffer = NULL;
-         vbuffer[attr].buffer_offset = pointer_to_offset(array->Ptr);
-      }
-      else {
-         /* wrap user data */
-         void *ptr;
-
-         if (array->Ptr) {
-            ptr = (void *) array->Ptr;
-         }
-         else {
-            /* no array, use ctx->Current.Attrib[] value */
-            ptr = (void *) ctx->Current.Attrib[mesaAttr];
-            stride = 0;
-         }
-
-         assert(ptr);
-
-         vbuffer[attr].buffer = NULL;
-         vbuffer[attr].user_buffer = ptr;
-         vbuffer[attr].buffer_offset = 0;
-      }
-
-      /* common-case setup */
-      vbuffer[attr].stride = stride; /* in bytes */
-
-      velements[attr].src_offset = 0;
-      velements[attr].instance_divisor = array->InstanceDivisor;
-      velements[attr].vertex_buffer_index = attr;
-      velements[attr].src_format = st_pipe_vertex_format(array->Type,
-                                                         array->Size,
-                                                         array->Format,
-                                                         array->Normalized,
-                                                         array->Integer);
-      assert(velements[attr].src_format);
-   }
-
-   return GL_TRUE;
-}
-
-
 static void
 setup_index_buffer(struct st_context *st,
                    const struct _mesa_index_buffer *ib,
@@ -665,58 +172,6 @@ translate_prim(const struct gl_context *ctx, unsigned prim)
 
 
 /**
- * Setup vertex arrays and buffers prior to drawing.
- * \return GL_TRUE for success, GL_FALSE otherwise (probably out of memory)
- */
-static GLboolean
-st_validate_varrays(struct gl_context *ctx,
-                    const struct gl_client_array **arrays)
-{
-   struct st_context *st = st_context(ctx);
-   const struct st_vertex_program *vp;
-   const struct st_vp_variant *vpv;
-   struct pipe_vertex_buffer vbuffer[PIPE_MAX_SHADER_INPUTS];
-   struct pipe_vertex_element velements[PIPE_MAX_ATTRIBS];
-   unsigned num_vbuffers, num_velements;
-
-   /* must get these after state validation! */
-   vp = st->vp;
-   vpv = st->vp_variant;
-
-   memset(velements, 0, sizeof(struct pipe_vertex_element) * vpv->num_inputs);
-
-   /*
-    * Setup the vbuffer[] and velements[] arrays.
-    */
-   if (is_interleaved_arrays(vp, vpv, arrays)) {
-      if (!setup_interleaved_attribs(ctx, vp, vpv, arrays, vbuffer,
-                                     velements)) {
-         return GL_FALSE;
-      }
-
-      num_vbuffers = 1;
-      num_velements = vpv->num_inputs;
-      if (num_velements == 0)
-         num_vbuffers = 0;
-   }
-   else {
-      if (!setup_non_interleaved_attribs(ctx, vp, vpv, arrays,
-                                         vbuffer, velements)) {
-         return GL_FALSE;
-      }
-
-      num_vbuffers = vpv->num_inputs;
-      num_velements = vpv->num_inputs;
-   }
-
-   cso_set_vertex_buffers(st->cso_context, num_vbuffers, vbuffer);
-   cso_set_vertex_elements(st->cso_context, num_velements, velements);
-
-   return GL_TRUE;
-}
-
-
-/**
  * This function gets plugged into the VBO module and is called when
  * we have something to render.
  * Basically, translate the information into the format expected by gallium.
@@ -736,38 +191,16 @@ st_draw_vbo(struct gl_context *ctx,
    struct pipe_draw_info info;
    const struct gl_client_array **arrays = ctx->Array._DrawArrays;
    unsigned i;
-   GLboolean new_array;
 
    /* Mesa core state should have been validated already */
    assert(ctx->NewState == 0x0);
 
-   /* Get Mesa driver state. */
-   st->dirty.st |= ctx->NewDriverState;
-   ctx->NewDriverState = 0;
-
-   new_array =
-      (st->dirty.st & (ST_NEW_VERTEX_ARRAYS | ST_NEW_VERTEX_PROGRAM)) ||
-      (st->dirty.mesa & (_NEW_PROGRAM | _NEW_BUFFER_OBJECT)) != 0;
-
    /* Validate state. */
-   if (st->dirty.st) {
-      GLboolean vertDataEdgeFlags;
-
-      vertDataEdgeFlags = arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj &&
-                          arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj->Name;
-      if (vertDataEdgeFlags != st->vertdata_edgeflags) {
-         st->vertdata_edgeflags = vertDataEdgeFlags;
-         st->dirty.st |= ST_NEW_EDGEFLAGS_DATA;
-      }
-
+   if (st->dirty.st || ctx->NewDriverState) {
       st_validate_state(st);
 
-      if (new_array) {
-         if (!st_validate_varrays(ctx, arrays)) {
-            /* probably out of memory, no-op the draw call */
-            return;
-         }
-      }
+      if (st->vertex_array_out_of_memory)
+         return;
 
 #if 0
       if (MESA_VERBOSE & VERBOSE_GLSL) {
diff --git a/mesalib/src/mesa/state_tracker/st_draw.h b/mesalib/src/mesa/state_tracker/st_draw.h
index c608051eb..3313fc8c7 100644
--- a/mesalib/src/mesa/state_tracker/st_draw.h
+++ b/mesalib/src/mesa/state_tracker/st_draw.h
@@ -67,13 +67,6 @@ st_feedback_draw_vbo(struct gl_context *ctx,
                      GLuint max_index,
                      struct gl_transform_feedback_object *tfb_vertcount);
 
-/* Internal function:
- */
-extern enum pipe_format
-st_pipe_vertex_format(GLenum type, GLuint size, GLenum format,
-                      GLboolean normalized, GLboolean integer);
-
-
 /**
  * When drawing with VBOs, the addresses specified with
  * glVertex/Color/TexCoordPointer() are really offsets into the VBO, not real