Merge remote-tracking branch 'origin/released'

* origin/released:
  libX11 libxcb mesa xserver xcb-proto xkeyboard-config git update 4 Mar 2014

Conflicts:
	mesalib/src/mapi/glapi/glapi.h
	mesalib/src/mapi/glapi/glthread.c
	mesalib/src/mesa/drivers/dri/common/dri_util.c
	mesalib/src/mesa/main/bufferobj.c
	xorg-server/dix/dispatch.c
	xorg-server/hw/xwin/glx/gen_gl_wrappers.py
	xorg-server/hw/xwin/winmultiwindowwm.c

149 files changed, 7099 insertions, 2863 deletions

diff --git a/mesalib/src/gallium/Automake.inc b/mesalib/src/gallium/Automake.inc
index 1e4a34f5d..1acc99ef5 100644
--- a/mesalib/src/gallium/Automake.inc
+++ b/mesalib/src/gallium/Automake.inc
@@ -51,12 +51,14 @@ GALLIUM_VIDEO_CFLAGS = \
 	$(VISIBILITY_CFLAGS)
 
 
-# TODO: add -export-symbols-regex
+DRI_VERSION_SCRIPT ?= $(top_srcdir)/src/gallium/state_trackers/dri/dri.link
+
 GALLIUM_DRI_LINKER_FLAGS = \
 	-module \
 	-avoid-version \
+	-Wl,--version-script=$(DRI_VERSION_SCRIPT) \
 	-shared \
-	-Wl,-Bsymbolic
+	-no-undefined
 
 GALLIUM_VDPAU_LINKER_FLAGS = \
 	-module \
@@ -76,7 +78,11 @@ GALLIUM_VDPAU_LIB_DEPS = \
 	$(top_builddir)/src/gallium/auxiliary/libgallium.la \
 	$(top_builddir)/src/gallium/state_trackers/vdpau/libvdpautracker.la \
 	$(VDPAU_LIBS) \
-	$(LIBDRM_LIBS)
+	$(LIBDRM_LIBS) \
+	-lm \
+	$(CLOCK_LIB) \
+	$(PTHREAD_LIBS) \
+	$(DLOPEN_LIBS)
 
 GALLIUM_XVMC_LIB_DEPS = \
 	$(top_builddir)/src/gallium/auxiliary/libgallium.la \
diff --git a/mesalib/src/gallium/auxiliary/Makefile.sources b/mesalib/src/gallium/auxiliary/Makefile.sources
index c89cbddd0..8919783c2 100644
--- a/mesalib/src/gallium/auxiliary/Makefile.sources
+++ b/mesalib/src/gallium/auxiliary/Makefile.sources
@@ -91,6 +91,7 @@ C_SOURCES := \
 	translate/translate_sse.c \
 	util/u_debug.c \
 	util/u_debug_describe.c \
+	util/u_debug_flush.c \
 	util/u_debug_memory.c \
 	util/u_debug_refcnt.c \
 	util/u_debug_stack.c \
@@ -155,7 +156,8 @@ C_SOURCES := \
         vl/vl_idct.c \
 	vl/vl_mc.c \
         vl/vl_vertex_buffers.c \
-        vl/vl_video_buffer.c
+        vl/vl_video_buffer.c \
+	vl/vl_deint_filter.c
 
 GENERATED_SOURCES := \
 	indices/u_indices_gen.c \
diff --git a/mesalib/src/gallium/auxiliary/util/u_debug_flush.c b/mesalib/src/gallium/auxiliary/util/u_debug_flush.c
new file mode 100644
index 000000000..9cf70db58
--- /dev/null
+++ b/mesalib/src/gallium/auxiliary/util/u_debug_flush.c
@@ -0,0 +1,391 @@
+/**************************************************************************
+ *
+ * Copyright 2012 VMware, Inc.
+ * 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 VMWARE 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.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * u_debug_flush.c Debug flush and map-related issues:
+ * - Flush while synchronously mapped.
+ * - Command stream reference while synchronously mapped.
+ * - Synchronous map while referenced on command stream.
+ * - Recursive maps.
+ * - Unmap while not mapped.
+ *
+ * @author Thomas Hellstrom <thellstrom@vmware.com>
+ */
+
+#ifdef DEBUG
+#include "pipe/p_compiler.h"
+#include "util/u_debug_stack.h"
+#include "util/u_debug.h"
+#include "util/u_memory.h"
+#include "util/u_debug_flush.h"
+#include "util/u_hash_table.h"
+#include "util/u_double_list.h"
+#include "util/u_inlines.h"
+#include "os/os_thread.h"
+#include <stdio.h>
+
+struct debug_flush_buf {
+   /* Atomic */
+   struct pipe_reference reference; /* Must be the first member. */
+   pipe_mutex mutex;
+   /* Immutable */
+   boolean supports_unsync;
+   unsigned bt_depth;
+   /* Protected by mutex */
+   boolean mapped;
+   boolean mapped_sync;
+   struct debug_stack_frame *map_frame;
+};
+
+struct debug_flush_item {
+   struct debug_flush_buf *fbuf;
+   unsigned bt_depth;
+   struct debug_stack_frame *ref_frame;
+};
+
+struct debug_flush_ctx {
+   /* Contexts are used by a single thread at a time */
+   unsigned bt_depth;
+   boolean catch_map_of_referenced;
+   struct util_hash_table *ref_hash;
+   struct list_head head;
+};
+
+pipe_static_mutex(list_mutex);
+static struct list_head ctx_list = {&ctx_list, &ctx_list};
+
+static struct debug_stack_frame *
+debug_flush_capture_frame(int start, int depth)
+{
+   struct debug_stack_frame *frames;
+
+   frames = CALLOC(depth, sizeof(*frames));
+   if (!frames)
+      return NULL;
+
+   debug_backtrace_capture(frames, start, depth);
+   return frames;
+}
+
+static int
+debug_flush_pointer_compare(void *key1, void *key2)
+{
+   return (key1 == key2) ? 0 : 1;
+}
+
+static unsigned
+debug_flush_pointer_hash(void *key)
+{
+   return (unsigned) (unsigned long) key;
+}
+
+struct debug_flush_buf *
+debug_flush_buf_create(boolean supports_unsync, unsigned bt_depth)
+{
+   struct debug_flush_buf *fbuf = CALLOC_STRUCT(debug_flush_buf);
+
+   if (!fbuf)
+      goto out_no_buf;
+
+   fbuf->supports_unsync = supports_unsync;
+   fbuf->bt_depth = bt_depth;
+   pipe_reference_init(&fbuf->reference, 1);
+   pipe_mutex_init(fbuf->mutex);
+
+   return fbuf;
+out_no_buf:
+   debug_printf("Debug flush buffer creation failed.\n");
+   debug_printf("Debug flush checking for this buffer will be incomplete.\n");
+   return NULL;
+}
+
+void
+debug_flush_buf_reference(struct debug_flush_buf **dst,
+                          struct debug_flush_buf *src)
+{
+   struct debug_flush_buf *fbuf = *dst;
+
+   if (pipe_reference(&(*dst)->reference, &src->reference)) {
+      if (fbuf->map_frame)
+         FREE(fbuf->map_frame);
+
+      FREE(fbuf);
+   }
+
+   *dst = src;
+}
+
+static void
+debug_flush_item_destroy(struct debug_flush_item *item)
+{
+   debug_flush_buf_reference(&item->fbuf, NULL);
+
+   if (item->ref_frame)
+      FREE(item->ref_frame);
+
+   FREE(item);
+}
+
+struct debug_flush_ctx *
+debug_flush_ctx_create(boolean catch_reference_of_mapped, unsigned bt_depth)
+{
+   struct debug_flush_ctx *fctx = CALLOC_STRUCT(debug_flush_ctx);
+
+   if (!fctx)
+      goto out_no_ctx;
+
+   fctx->ref_hash = util_hash_table_create(debug_flush_pointer_hash,
+                                           debug_flush_pointer_compare);
+
+   if (!fctx->ref_hash)
+      goto out_no_ref_hash;
+
+   fctx->bt_depth = bt_depth;
+   pipe_mutex_lock(list_mutex);
+   list_addtail(&fctx->head, &ctx_list);
+   pipe_mutex_unlock(list_mutex);
+
+   return fctx;
+
+ out_no_ref_hash:
+   FREE(fctx);
+out_no_ctx:
+   debug_printf("Debug flush context creation failed.\n");
+   debug_printf("Debug flush checking for this context will be incomplete.\n");
+   return NULL;
+}
+
+static void
+debug_flush_alert(const char *s, const char *op,
+                  unsigned start, unsigned depth,
+                  boolean continued,
+                  boolean capture,
+                  const struct debug_stack_frame *frame)
+{
+   if (capture)
+      frame = debug_flush_capture_frame(start, depth);
+
+   if (s)
+      debug_printf("%s ", s);
+   if (frame) {
+      debug_printf("%s backtrace follows:\n", op);
+      debug_backtrace_dump(frame, depth);
+   } else
+      debug_printf("No %s backtrace was captured.\n", op);
+
+   if (continued)
+      debug_printf("**********************************\n");
+   else
+      debug_printf("*********END OF MESSAGE***********\n\n\n");
+
+   if (capture)
+      FREE((void *)frame);
+}
+
+
+void
+debug_flush_map(struct debug_flush_buf *fbuf, unsigned flags)
+{
+   boolean mapped_sync = FALSE;
+
+   if (!fbuf)
+      return;
+
+   pipe_mutex_lock(fbuf->mutex);
+   if (fbuf->mapped) {
+      debug_flush_alert("Recursive map detected.", "Map",
+                        2, fbuf->bt_depth, TRUE, TRUE, NULL);
+      debug_flush_alert(NULL, "Previous map", 0, fbuf->bt_depth, FALSE,
+                        FALSE, fbuf->map_frame);
+   } else if (!(flags & PIPE_TRANSFER_UNSYNCHRONIZED) ||
+              !fbuf->supports_unsync) {
+      fbuf->mapped_sync = mapped_sync = TRUE;
+   }
+   fbuf->map_frame = debug_flush_capture_frame(1, fbuf->bt_depth);
+   fbuf->mapped = TRUE;
+   pipe_mutex_unlock(fbuf->mutex);
+
+   if (mapped_sync) {
+      struct debug_flush_ctx *fctx;
+
+      pipe_mutex_lock(list_mutex);
+      LIST_FOR_EACH_ENTRY(fctx, &ctx_list, head) {
+         struct debug_flush_item *item =
+            util_hash_table_get(fctx->ref_hash, fbuf);
+
+         if (item && fctx->catch_map_of_referenced) {
+            debug_flush_alert("Already referenced map detected.",
+                              "Map", 2, fbuf->bt_depth, TRUE, TRUE, NULL);
+            debug_flush_alert(NULL, "Reference", 0, item->bt_depth,
+                              FALSE, FALSE, item->ref_frame);
+         }
+      }
+      pipe_mutex_unlock(list_mutex);
+   }
+}
+
+void
+debug_flush_unmap(struct debug_flush_buf *fbuf)
+{
+   if (!fbuf)
+      return;
+
+   pipe_mutex_lock(fbuf->mutex);
+   if (!fbuf->mapped)
+      debug_flush_alert("Unmap not previously mapped detected.", "Map",
+                        2, fbuf->bt_depth, FALSE, TRUE, NULL);
+
+   fbuf->mapped_sync = FALSE;
+   fbuf->mapped = FALSE;
+   if (fbuf->map_frame) {
+      FREE(fbuf->map_frame);
+      fbuf->map_frame = NULL;
+   }
+   pipe_mutex_unlock(fbuf->mutex);
+}
+
+void
+debug_flush_cb_reference(struct debug_flush_ctx *fctx,
+                         struct debug_flush_buf *fbuf)
+{
+   struct debug_flush_item *item;
+
+   if (!fctx || !fbuf)
+      return;
+
+   item = util_hash_table_get(fctx->ref_hash, fbuf);
+
+   pipe_mutex_lock(fbuf->mutex);
+   if (fbuf->mapped_sync) {
+      debug_flush_alert("Reference of mapped buffer detected.", "Reference",
+                        2, fctx->bt_depth, TRUE, TRUE, NULL);
+      debug_flush_alert(NULL, "Map", 0, fbuf->bt_depth, FALSE,
+                        FALSE, fbuf->map_frame);
+   }
+   pipe_mutex_unlock(fbuf->mutex);
+
+   if (!item) {
+      item = CALLOC_STRUCT(debug_flush_item);
+      if (item) {
+         debug_flush_buf_reference(&item->fbuf, fbuf);
+         item->bt_depth = fctx->bt_depth;
+         item->ref_frame = debug_flush_capture_frame(2, item->bt_depth);
+         if (util_hash_table_set(fctx->ref_hash, fbuf, item) != PIPE_OK) {
+            debug_flush_item_destroy(item);
+            goto out_no_item;
+         }
+         return;
+      }
+      goto out_no_item;
+   }
+   return;
+
+out_no_item:
+   debug_printf("Debug flush command buffer reference creation failed.\n");
+   debug_printf("Debug flush checking will be incomplete "
+                "for this command batch.\n");
+}
+
+static enum pipe_error
+debug_flush_might_flush_cb(void *key, void *value, void *data)
+{
+   struct debug_flush_item *item =
+      (struct debug_flush_item *) value;
+   struct debug_flush_buf *fbuf = item->fbuf;
+   const char *reason = (const char *) data;
+   char message[80];
+
+   snprintf(message, sizeof(message),
+            "%s referenced mapped buffer detected.", reason);
+
+   pipe_mutex_lock(fbuf->mutex);
+   if (fbuf->mapped_sync) {
+      debug_flush_alert(message, reason, 3, item->bt_depth, TRUE, TRUE, NULL);
+      debug_flush_alert(NULL, "Map", 0, fbuf->bt_depth, TRUE, FALSE,
+                        fbuf->map_frame);
+      debug_flush_alert(NULL, "First reference", 0, item->bt_depth, FALSE,
+                        FALSE, item->ref_frame);
+   }
+   pipe_mutex_unlock(fbuf->mutex);
+
+   return PIPE_OK;
+}
+
+void
+debug_flush_might_flush(struct debug_flush_ctx *fctx)
+{
+   if (!fctx)
+      return;
+
+   util_hash_table_foreach(fctx->ref_hash,
+                           debug_flush_might_flush_cb,
+                           "Might flush");
+}
+
+static enum pipe_error
+debug_flush_flush_cb(void *key, void *value, void *data)
+{
+   struct debug_flush_item *item =
+      (struct debug_flush_item *) value;
+
+   debug_flush_item_destroy(item);
+
+   return PIPE_OK;
+}
+
+
+void
+debug_flush_flush(struct debug_flush_ctx *fctx)
+{
+   if (!fctx)
+      return;
+
+   util_hash_table_foreach(fctx->ref_hash,
+                           debug_flush_might_flush_cb,
+                           "Flush");
+   util_hash_table_foreach(fctx->ref_hash,
+                           debug_flush_flush_cb,
+                           NULL);
+   util_hash_table_clear(fctx->ref_hash);
+}
+
+void
+debug_flush_ctx_destroy(struct debug_flush_ctx *fctx)
+{
+   if (!fctx)
+      return;
+
+   list_del(&fctx->head);
+   util_hash_table_foreach(fctx->ref_hash,
+                           debug_flush_flush_cb,
+                           NULL);
+   util_hash_table_clear(fctx->ref_hash);
+   util_hash_table_destroy(fctx->ref_hash);
+   FREE(fctx);
+}
+#endif
diff --git a/mesalib/src/gallium/auxiliary/util/u_debug_flush.h b/mesalib/src/gallium/auxiliary/util/u_debug_flush.h
new file mode 100644
index 000000000..a604167f0
--- /dev/null
+++ b/mesalib/src/gallium/auxiliary/util/u_debug_flush.h
@@ -0,0 +1,138 @@
+/**************************************************************************
+ *
+ * Copyright 2012 VMware, Inc.
+ * 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 VMWARE 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.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * u_debug_flush.h - Header for debugging flush- and map- related issues.
+ * - Flush while synchronously mapped.
+ * - Command stream reference while synchronously mapped.
+ * - Synchronous map while referenced on command stream.
+ * - Recursive maps.
+ * - Unmap while not mapped.
+ *
+ * @author Thomas Hellstrom <thellstrom@vmware.com>
+ */
+#ifdef DEBUG
+
+#ifndef U_DEBUG_FLUSH_H_
+#define U_DEBUG_FLUSH_H_
+
+struct debug_flush_buf;
+struct debug_flush_ctx;
+
+/**
+ * Create a buffer (AKA allocation) representation.
+ *
+ * @param support_unsync Whether unsynchronous maps are truly supported.
+ * @param bt_depth Depth of backtrace to be captured for this buffer
+ * representation.
+ */
+struct debug_flush_buf *
+debug_flush_buf_create(boolean supports_unsync, unsigned bt_depth);
+
+/**
+ * Reference a buffer representation.
+ *
+ * @param dst Pointer copy destination
+ * @param src Pointer copy source (may be NULL).
+ *
+ * Replace a pointer to a buffer representation with proper refcounting.
+ */
+void
+debug_flush_buf_reference(struct debug_flush_buf **dst,
+                          struct debug_flush_buf *src);
+
+/**
+ * Create a context representation.
+ *
+ * @param catch_map_of_referenced Whether to catch synchronous maps of buffers
+ * already present on the command stream.
+ * @param bt_depth Depth of backtrace to be captured for this context
+ * representation.
+ */
+struct debug_flush_ctx *
+debug_flush_ctx_create(boolean catch_map_of_referenced, unsigned bt_depth);
+
+/**
+ * Destroy a context representation.
+ *
+ * @param fctx The context representation to destroy.
+ */
+void
+debug_flush_ctx_destroy(struct debug_flush_ctx *fctx);
+
+/**
+ * Map annotation
+ *
+ * @param fbuf The buffer representation to map.
+ * @param flags Pipebuffer flags for the map.
+ *
+ * Used to annotate a map of the buffer described by the buffer representation.
+ */
+void debug_flush_map(struct debug_flush_buf *fbuf, unsigned flags);
+
+/**
+ * Unmap annotation
+ *
+ * @param fbuf The buffer representation to map.
+ *
+ * Used to annotate an unmap of the buffer described by the
+ * buffer representation.
+ */
+void debug_flush_unmap(struct debug_flush_buf *fbuf);
+
+/**
+ * Might flush annotation
+ *
+ * @param fctx The context representation that might be flushed.
+ *
+ * Used to annotate a conditional (possible) flush of the given context.
+ */
+void debug_flush_might_flush(struct debug_flush_ctx *fctx);
+
+/**
+ * Flush annotation
+ *
+ * @param fctx The context representation that is flushed.
+ *
+ * Used to annotate a real flush of the given context.
+ */
+void debug_flush_flush(struct debug_flush_ctx *fctx);
+
+
+/**
+ * Flush annotation
+ *
+ * @param fctx The context representation that is flushed.
+ *
+ * Used to annotate a real flush of the given context.
+ */
+void debug_flush_cb_reference(struct debug_flush_ctx *fctx,
+                              struct debug_flush_buf *fbuf);
+
+#endif
+#endif
diff --git a/mesalib/src/gallium/auxiliary/util/u_format.c b/mesalib/src/gallium/auxiliary/util/u_format.c
index 6b602bf32..056f82f72 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format.c
+++ b/mesalib/src/gallium/auxiliary/util/u_format.c
@@ -527,7 +527,7 @@ util_format_fits_8unorm(const struct util_format_description *format_desc)
 }
 
 
-void
+boolean
 util_format_translate(enum pipe_format dst_format,
                       void *dst, unsigned dst_stride,
                       unsigned dst_x, unsigned dst_y,
@@ -555,7 +555,7 @@ util_format_translate(enum pipe_format dst_format,
       util_copy_rect(dst, dst_format, dst_stride,  dst_x, dst_y,
                      width, height, src, (int)src_stride,
                      src_x, src_y);
-      return;
+      return TRUE;
    }
 
    assert(dst_x % dst_format_desc->block.width == 0);
@@ -621,7 +621,7 @@ util_format_translate(enum pipe_format dst_format,
 
       FREE(tmp_z);
 
-      return;
+      return TRUE;
    }
 
    if (util_format_fits_8unorm(src_format_desc) ||
@@ -629,10 +629,15 @@ util_format_translate(enum pipe_format dst_format,
       unsigned tmp_stride;
       uint8_t *tmp_row;
 
+      if (!src_format_desc->unpack_rgba_8unorm ||
+          !dst_format_desc->pack_rgba_8unorm) {
+         return FALSE;
+      }
+
       tmp_stride = MAX2(width, x_step) * 4 * sizeof *tmp_row;
       tmp_row = MALLOC(y_step * tmp_stride);
       if (!tmp_row)
-         return;
+         return FALSE;
 
       while (height >= y_step) {
          src_format_desc->unpack_rgba_8unorm(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
@@ -654,10 +659,15 @@ util_format_translate(enum pipe_format dst_format,
       unsigned tmp_stride;
       float *tmp_row;
 
+      if (!src_format_desc->unpack_rgba_float ||
+          !dst_format_desc->pack_rgba_float) {
+         return FALSE;
+      }
+
       tmp_stride = MAX2(width, x_step) * 4 * sizeof *tmp_row;
       tmp_row = MALLOC(y_step * tmp_stride);
       if (!tmp_row)
-         return;
+         return FALSE;
 
       while (height >= y_step) {
          src_format_desc->unpack_rgba_float(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
@@ -675,6 +685,7 @@ util_format_translate(enum pipe_format dst_format,
 
       FREE(tmp_row);
    }
+   return TRUE;
 }
 
 void util_format_compose_swizzles(const unsigned char swz1[4],
diff --git a/mesalib/src/gallium/auxiliary/util/u_format.h b/mesalib/src/gallium/auxiliary/util/u_format.h
index 5f86e2d2c..e36a9e237 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format.h
+++ b/mesalib/src/gallium/auxiliary/util/u_format.h
@@ -1190,7 +1190,7 @@ util_format_write_4i(enum pipe_format format,
 boolean
 util_format_fits_8unorm(const struct util_format_description *format_desc);
 
-void
+boolean
 util_format_translate(enum pipe_format dst_format,
                       void *dst, unsigned dst_stride,
                       unsigned dst_x, unsigned dst_y,
diff --git a/mesalib/src/gallium/auxiliary/util/u_math.h b/mesalib/src/gallium/auxiliary/util/u_math.h
index b5e06630a..ec03e4e58 100644
--- a/mesalib/src/gallium/auxiliary/util/u_math.h
+++ b/mesalib/src/gallium/auxiliary/util/u_math.h
@@ -112,10 +112,13 @@ static INLINE float logf( float f )
 #define logf(x) ((float)log((double)(x)))
 #endif /* logf */
 
+#if _MSC_VER < 1800
 #define isfinite(x) _finite((double)(x))
 #define isnan(x) _isnan((double)(x))
+#endif /* _MSC_VER < 1800 */
 #endif /* _MSC_VER < 1400 && !defined(__cplusplus) */
 
+#if _MSC_VER < 1800
 static INLINE double log2( double x )
 {
    const double invln2 = 1.442695041;
@@ -133,6 +136,7 @@ roundf(float x)
 {
    return x >= 0.0f ? floorf(x + 0.5f) : ceilf(x - 0.5f);
 }
+#endif
 
 #define INFINITY (DBL_MAX + DBL_MAX)
 #define NAN (INFINITY - INFINITY)
@@ -717,13 +721,18 @@ util_bitcount(unsigned n)
  */
 
 #ifdef PIPE_ARCH_BIG_ENDIAN
+#define util_le64_to_cpu(x) util_bswap64(x)
 #define util_le32_to_cpu(x) util_bswap32(x)
 #define util_le16_to_cpu(x) util_bswap16(x)
 #else
+#define util_le64_to_cpu(x) (x)
 #define util_le32_to_cpu(x) (x)
 #define util_le16_to_cpu(x) (x)
 #endif
 
+#define util_cpu_to_le64(x) util_le64_to_cpu(x)
+#define util_cpu_to_le32(x) util_le32_to_cpu(x)
+#define util_cpu_to_le16(x) util_le16_to_cpu(x)
 
 /**
  * Reverse byte order of a 32 bit word.
@@ -731,7 +740,8 @@ util_bitcount(unsigned n)
 static INLINE uint32_t
 util_bswap32(uint32_t n)
 {
-#if defined(PIPE_CC_GCC) && (PIPE_CC_GCC_VERSION >= 403)
+/* We need the gcc version checks for non-autoconf build system */
+#if defined(HAVE___BUILTIN_BSWAP32) || (defined(PIPE_CC_GCC) && (PIPE_CC_GCC_VERSION >= 403))
    return __builtin_bswap32(n);
 #else
    return (n >> 24) |
@@ -741,6 +751,20 @@ util_bswap32(uint32_t n)
 #endif
 }
 
+/**
+ * Reverse byte order of a 64bit word.
+ */
+static INLINE uint64_t
+util_bswap64(uint64_t n)
+{
+#if defined(HAVE___BUILTIN_BSWAP64)
+   return __builtin_bswap64(n);
+#else
+   return ((uint64_t)util_bswap32(n) << 32) |
+          util_bswap32((n >> 32));
+#endif
+}
+
 
 /**
  * Reverse byte order of a 16 bit word.
diff --git a/mesalib/src/gallium/auxiliary/util/u_range.h b/mesalib/src/gallium/auxiliary/util/u_range.h
index 4b1d0d1be..efe25ef5e 100644
--- a/mesalib/src/gallium/auxiliary/util/u_range.h
+++ b/mesalib/src/gallium/auxiliary/util/u_range.h
@@ -36,6 +36,8 @@
 
 #include "os/os_thread.h"
 
+#include "util/u_math.h"
+
 struct util_range {
    unsigned start; /* inclusive */
    unsigned end; /* exclusive */
diff --git a/mesalib/src/gallium/auxiliary/util/u_upload_mgr.c b/mesalib/src/gallium/auxiliary/util/u_upload_mgr.c
index 7349d0068..744ea2e5e 100644
--- a/mesalib/src/gallium/auxiliary/util/u_upload_mgr.c
+++ b/mesalib/src/gallium/auxiliary/util/u_upload_mgr.c
@@ -44,11 +44,12 @@ struct u_upload_mgr {
    unsigned default_size;  /* Minimum size of the upload buffer, in bytes. */
    unsigned alignment;     /* Alignment of each sub-allocation. */
    unsigned bind;          /* Bitmask of PIPE_BIND_* flags. */
+   unsigned map_flags;     /* Bitmask of PIPE_TRANSFER_* flags. */
+   boolean map_persistent; /* If persistent mappings are supported. */
 
    struct pipe_resource *buffer;   /* Upload buffer. */
    struct pipe_transfer *transfer; /* Transfer object for the upload buffer. */
    uint8_t *map;    /* Pointer to the mapped upload buffer. */
-   unsigned size;   /* Actual size of the upload buffer. */
    unsigned offset; /* Aligned offset to the upload buffer, pointing
                      * at the first unused byte. */
 };
@@ -67,20 +68,39 @@ struct u_upload_mgr *u_upload_create( struct pipe_context *pipe,
    upload->default_size = default_size;
    upload->alignment = alignment;
    upload->bind = bind;
-   upload->buffer = NULL;
+
+   upload->map_persistent =
+      pipe->screen->get_param(pipe->screen,
+                              PIPE_CAP_BUFFER_MAP_PERSISTENT_COHERENT);
+
+   if (upload->map_persistent) {
+      upload->map_flags = PIPE_TRANSFER_WRITE |
+                          PIPE_TRANSFER_PERSISTENT |
+                          PIPE_TRANSFER_COHERENT;
+   }
+   else {
+      upload->map_flags = PIPE_TRANSFER_WRITE |
+                          PIPE_TRANSFER_UNSYNCHRONIZED |
+                          PIPE_TRANSFER_FLUSH_EXPLICIT;
+   }
 
    return upload;
 }
 
-void u_upload_unmap( struct u_upload_mgr *upload )
+
+static void upload_unmap_internal(struct u_upload_mgr *upload, boolean destroying)
 {
+   if (!destroying && upload->map_persistent)
+      return;
+
    if (upload->transfer) {
       struct pipe_box *box = &upload->transfer->box;
-      if ((int) upload->offset > box->x) {
 
+      if (!upload->map_persistent && (int) upload->offset > box->x) {
          pipe_buffer_flush_mapped_range(upload->pipe, upload->transfer,
                                         box->x, upload->offset - box->x);
       }
+
       pipe_transfer_unmap(upload->pipe, upload->transfer);
       upload->transfer = NULL;
       upload->map = NULL;
@@ -88,12 +108,17 @@ void u_upload_unmap( struct u_upload_mgr *upload )
 }
 
 
+void u_upload_unmap( struct u_upload_mgr *upload )
+{
+   upload_unmap_internal(upload, FALSE);
+}
+
+
 static void u_upload_release_buffer(struct u_upload_mgr *upload)
 {
    /* Unmap and unreference the upload buffer. */
-   u_upload_unmap(upload);
+   upload_unmap_internal(upload, TRUE);
    pipe_resource_reference( &upload->buffer, NULL );
-   upload->size = 0;
 }
 
 
@@ -108,6 +133,8 @@ static enum pipe_error
 u_upload_alloc_buffer( struct u_upload_mgr *upload,
                        unsigned min_size )
 {
+   struct pipe_screen *screen = upload->pipe->screen;
+   struct pipe_resource buffer;
    unsigned size;
 
    /* Release the old buffer, if present:
@@ -118,28 +145,36 @@ u_upload_alloc_buffer( struct u_upload_mgr *upload,
     */
    size = align(MAX2(upload->default_size, min_size), 4096);
 
-   upload->buffer = pipe_buffer_create( upload->pipe->screen,
-                                        upload->bind,
-                                        PIPE_USAGE_STREAM,
-                                        size );
+   memset(&buffer, 0, sizeof buffer);
+   buffer.target = PIPE_BUFFER;
+   buffer.format = PIPE_FORMAT_R8_UNORM; /* want TYPELESS or similar */
+   buffer.bind = upload->bind;
+   buffer.usage = PIPE_USAGE_STREAM;
+   buffer.width0 = size;
+   buffer.height0 = 1;
+   buffer.depth0 = 1;
+   buffer.array_size = 1;
+
+   if (upload->map_persistent) {
+      buffer.flags = PIPE_RESOURCE_FLAG_MAP_PERSISTENT |
+                     PIPE_RESOURCE_FLAG_MAP_COHERENT;
+   }
+
+   upload->buffer = screen->resource_create(screen, &buffer);
    if (upload->buffer == NULL) {
       return PIPE_ERROR_OUT_OF_MEMORY;
    }
 
    /* Map the new buffer. */
    upload->map = pipe_buffer_map_range(upload->pipe, upload->buffer,
-                                       0, size,
-                                       PIPE_TRANSFER_WRITE |
-                                       PIPE_TRANSFER_FLUSH_EXPLICIT,
+                                       0, size, upload->map_flags,
                                        &upload->transfer);
    if (upload->map == NULL) {
       upload->transfer = NULL;
-      upload->size = 0;
       pipe_resource_reference(&upload->buffer, NULL);
       return PIPE_ERROR_OUT_OF_MEMORY;
    }
 
-   upload->size = size;
    upload->offset = 0;
    return PIPE_OK;
 }
@@ -164,7 +199,8 @@ enum pipe_error u_upload_alloc( struct u_upload_mgr *upload,
 
    /* Make sure we have enough space in the upload buffer
     * for the sub-allocation. */
-   if (MAX2(upload->offset, alloc_offset) + alloc_size > upload->size) {
+   if (!upload->buffer ||
+       MAX2(upload->offset, alloc_offset) + alloc_size > upload->buffer->width0) {
       enum pipe_error ret = u_upload_alloc_buffer(upload,
                                                   alloc_offset + alloc_size);
       if (ret != PIPE_OK)
@@ -175,10 +211,9 @@ enum pipe_error u_upload_alloc( struct u_upload_mgr *upload,
 
    if (!upload->map) {
       upload->map = pipe_buffer_map_range(upload->pipe, upload->buffer,
-					  offset, upload->size - offset,
-					  PIPE_TRANSFER_WRITE |
-					  PIPE_TRANSFER_FLUSH_EXPLICIT |
-					  PIPE_TRANSFER_UNSYNCHRONIZED,
+                                          offset,
+                                          upload->buffer->width0 - offset,
+                                          upload->map_flags,
 					  &upload->transfer);
       if (!upload->map) {
          upload->transfer = NULL;
diff --git a/mesalib/src/glsl/Makefile.sources b/mesalib/src/glsl/Makefile.sources
index e69c1ac61..5945590a5 100644
--- a/mesalib/src/glsl/Makefile.sources
+++ b/mesalib/src/glsl/Makefile.sources
@@ -67,6 +67,7 @@ LIBGLSL_FILES = \
 	$(GLSL_SRCDIR)/lower_jumps.cpp \
 	$(GLSL_SRCDIR)/lower_mat_op_to_vec.cpp \
 	$(GLSL_SRCDIR)/lower_noise.cpp \
+	$(GLSL_SRCDIR)/lower_offset_array.cpp \
 	$(GLSL_SRCDIR)/lower_packed_varyings.cpp \
 	$(GLSL_SRCDIR)/lower_named_interface_blocks.cpp \
 	$(GLSL_SRCDIR)/lower_packing_builtins.cpp \
diff --git a/mesalib/src/glsl/ast.h b/mesalib/src/glsl/ast.h
index e4c00a408..c7710e5d6 100644
--- a/mesalib/src/glsl/ast.h
+++ b/mesalib/src/glsl/ast.h
@@ -284,14 +284,14 @@ class ast_array_specifier : public ast_node {
 public:
    /** Unsized array specifier ([]) */
    explicit ast_array_specifier(const struct YYLTYPE &locp)
-     : dimension_count(1), is_unsized_array(true)
+     : is_unsized_array(true)
    {
       set_location(locp);
    }
 
    /** Sized array specifier ([dim]) */
    ast_array_specifier(const struct YYLTYPE &locp, ast_expression *dim)
-     : dimension_count(1), is_unsized_array(false)
+     : is_unsized_array(false)
    {
       set_location(locp);
       array_dimensions.push_tail(&dim->link);
@@ -300,14 +300,10 @@ public:
    void add_dimension(ast_expression *dim)
    {
       array_dimensions.push_tail(&dim->link);
-      dimension_count++;
    }
 
    virtual void print(void) const;
 
-   /* Count including sized and unsized dimensions */
-   unsigned dimension_count;
-
    /* If true, this means that the array has an unsized outermost dimension. */
    bool is_unsized_array;
 
@@ -470,17 +466,36 @@ struct ast_type_qualifier {
           * local_size_x, and so on.
           */
          unsigned local_size:3;
+
+	 /** \name Layout and memory qualifiers for ARB_shader_image_load_store. */
+	 /** \{ */
+	 unsigned early_fragment_tests:1;
+	 unsigned explicit_image_format:1;
+	 unsigned coherent:1;
+	 unsigned _volatile:1;
+	 unsigned restrict_flag:1;
+	 unsigned read_only:1; /**< "readonly" qualifier. */
+	 unsigned write_only:1; /**< "writeonly" qualifier. */
+	 /** \} */
+
+         /** \name Layout qualifiers for GL_ARB_gpu_shader5 */
+         /** \{ */
+         unsigned invocations:1;
+         /** \} */
       }
       /** \brief Set of flags, accessed by name. */
       q;
 
       /** \brief Set of flags, accessed as a bitmask. */
-      unsigned i;
+      uint64_t i;
    } flags;
 
    /** Precision of the type (highp/medium/lowp). */
    unsigned precision:2;
 
+   /** Geometry shader invocations for GL_ARB_gpu_shader5. */
+   int invocations;
+
    /**
     * Location specified via GL_ARB_explicit_attrib_location layout
     *
@@ -527,6 +542,25 @@ struct ast_type_qualifier {
    int local_size[3];
 
    /**
+    * Image format specified with an ARB_shader_image_load_store
+    * layout qualifier.
+    *
+    * \note
+    * This field is only valid if \c explicit_image_format is set.
+    */
+   GLenum image_format;
+
+   /**
+    * Base type of the data read from or written to this image.  Only
+    * the following enumerants are allowed: GLSL_TYPE_UINT,
+    * GLSL_TYPE_INT, GLSL_TYPE_FLOAT.
+    *
+    * \note
+    * This field is only valid if \c explicit_image_format is set.
+    */
+   glsl_base_type image_base_type;
+
+   /**
     * Return true if and only if an interpolation qualifier is present.
     */
    bool has_interpolation() const;
@@ -561,6 +595,12 @@ struct ast_type_qualifier {
    bool merge_qualifier(YYLTYPE *loc,
 			_mesa_glsl_parse_state *state,
 			ast_type_qualifier q);
+
+   bool merge_in_qualifier(YYLTYPE *loc,
+                           _mesa_glsl_parse_state *state,
+                           ast_type_qualifier q,
+                           ast_node* &node);
+
 };
 
 class ast_declarator_list;
diff --git a/mesalib/src/glsl/ast_function.cpp b/mesalib/src/glsl/ast_function.cpp
index 4c5b0e4aa..4b8447067 100644
--- a/mesalib/src/glsl/ast_function.cpp
+++ b/mesalib/src/glsl/ast_function.cpp
@@ -93,6 +93,57 @@ prototype_string(const glsl_type *return_type, const char *name,
    return str;
 }
 
+static bool
+verify_image_parameter(YYLTYPE *loc, _mesa_glsl_parse_state *state,
+                       const ir_variable *formal, const ir_variable *actual)
+{
+   /**
+    * From the ARB_shader_image_load_store specification:
+    *
+    * "The values of image variables qualified with coherent,
+    *  volatile, restrict, readonly, or writeonly may not be passed
+    *  to functions whose formal parameters lack such
+    *  qualifiers. [...] It is legal to have additional qualifiers
+    *  on a formal parameter, but not to have fewer."
+    */
+   if (actual->data.image.coherent && !formal->data.image.coherent) {
+      _mesa_glsl_error(loc, state,
+                       "function call parameter `%s' drops "
+                       "`coherent' qualifier", formal->name);
+      return false;
+   }
+
+   if (actual->data.image._volatile && !formal->data.image._volatile) {
+      _mesa_glsl_error(loc, state,
+                       "function call parameter `%s' drops "
+                       "`volatile' qualifier", formal->name);
+      return false;
+   }
+
+   if (actual->data.image.restrict_flag && !formal->data.image.restrict_flag) {
+      _mesa_glsl_error(loc, state,
+                       "function call parameter `%s' drops "
+                       "`restrict' qualifier", formal->name);
+      return false;
+   }
+
+   if (actual->data.image.read_only && !formal->data.image.read_only) {
+      _mesa_glsl_error(loc, state,
+                       "function call parameter `%s' drops "
+                       "`readonly' qualifier", formal->name);
+      return false;
+   }
+
+   if (actual->data.image.write_only && !formal->data.image.write_only) {
+      _mesa_glsl_error(loc, state,
+                       "function call parameter `%s' drops "
+                       "`writeonly' qualifier", formal->name);
+      return false;
+   }
+
+   return true;
+}
+
 /**
  * Verify that 'out' and 'inout' actual parameters are lvalues.  Also, verify
  * that 'const_in' formal parameters (an extension in our IR) correspond to
@@ -180,6 +231,13 @@ verify_parameter_modes(_mesa_glsl_parse_state *state,
 	 }
       }
 
+      if (formal->type->is_image() &&
+          actual->variable_referenced()) {
+         if (!verify_image_parameter(&loc, state, formal,
+                                     actual->variable_referenced()))
+            return false;
+      }
+
       actual_ir_node  = actual_ir_node->next;
       actual_ast_node = actual_ast_node->next;
    }
diff --git a/mesalib/src/glsl/ast_to_hir.cpp b/mesalib/src/glsl/ast_to_hir.cpp
index c89a26bf9..8f6e90174 100644
--- a/mesalib/src/glsl/ast_to_hir.cpp
+++ b/mesalib/src/glsl/ast_to_hir.cpp
@@ -969,6 +969,7 @@ do_comparison(void *mem_ctx, int operation, ir_rvalue *op0, ir_rvalue *op1)
    case GLSL_TYPE_ERROR:
    case GLSL_TYPE_VOID:
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
    case GLSL_TYPE_INTERFACE:
    case GLSL_TYPE_ATOMIC_UINT:
       /* I assume a comparison of a struct containing a sampler just
@@ -1796,7 +1797,7 @@ ast_compound_statement::hir(exec_list *instructions,
  * Evaluate the given exec_node (which should be an ast_node representing
  * a single array dimension) and return its integer value.
  */
-static const unsigned
+static unsigned
 process_array_size(exec_node *node,
                    struct _mesa_glsl_parse_state *state)
 {
@@ -2122,11 +2123,16 @@ validate_explicit_location(const struct ast_type_qualifier *qual,
 {
    bool fail = false;
 
-   /* In the vertex shader only shader inputs can be given explicit
-    * locations.
+   /* Between GL_ARB_explicit_attrib_location an
+    * GL_ARB_separate_shader_objects, the inputs and outputs of any shader
+    * stage can be assigned explicit locations.  The checking here associates
+    * the correct extension with the correct stage's input / output:
     *
-    * In the fragment shader only shader outputs can be given explicit
-    * locations.
+    *                     input            output
+    *                     -----            ------
+    * vertex              explicit_loc     sso
+    * geometry            sso              sso
+    * fragment            sso              explicit_loc
     */
    switch (state->stage) {
    case MESA_SHADER_VERTEX:
@@ -2137,16 +2143,35 @@ validate_explicit_location(const struct ast_type_qualifier *qual,
          break;
       }
 
+      if (var->data.mode == ir_var_shader_out) {
+         if (!state->check_separate_shader_objects_allowed(loc, var))
+            return;
+
+         break;
+      }
+
       fail = true;
       break;
 
    case MESA_SHADER_GEOMETRY:
-      _mesa_glsl_error(loc, state,
-                       "geometry shader variables cannot be given "
-                       "explicit locations");
-      return;
+      if (var->data.mode == ir_var_shader_in || var->data.mode == ir_var_shader_out) {
+         if (!state->check_separate_shader_objects_allowed(loc, var))
+            return;
+
+         break;
+      }
+
+      fail = true;
+      break;
 
    case MESA_SHADER_FRAGMENT:
+      if (var->data.mode == ir_var_shader_in) {
+         if (!state->check_separate_shader_objects_allowed(loc, var))
+            return;
+
+         break;
+      }
+
       if (var->data.mode == ir_var_shader_out) {
          if (!state->check_explicit_attrib_location_allowed(loc, var))
             return;
@@ -2180,9 +2205,26 @@ validate_explicit_location(const struct ast_type_qualifier *qual,
        * ensures that negative values stay negative.
        */
       if (qual->location >= 0) {
-         var->data.location = (state->stage == MESA_SHADER_VERTEX)
-            ? (qual->location + VERT_ATTRIB_GENERIC0)
-            : (qual->location + FRAG_RESULT_DATA0);
+         switch (state->stage) {
+         case MESA_SHADER_VERTEX:
+            var->data.location = (var->data.mode == ir_var_shader_in)
+               ? (qual->location + VERT_ATTRIB_GENERIC0)
+               : (qual->location + VARYING_SLOT_VAR0);
+            break;
+
+         case MESA_SHADER_GEOMETRY:
+            var->data.location = qual->location + VARYING_SLOT_VAR0;
+            break;
+
+         case MESA_SHADER_FRAGMENT:
+            var->data.location = (var->data.mode == ir_var_shader_out)
+               ? (qual->location + FRAG_RESULT_DATA0)
+               : (qual->location + VARYING_SLOT_VAR0);
+            break;
+         case MESA_SHADER_COMPUTE:
+            assert(!"Unexpected shader type");
+            break;
+         }
       } else {
          var->data.location = qual->location;
       }
@@ -2206,8 +2248,54 @@ validate_explicit_location(const struct ast_type_qualifier *qual,
          }
       }
    }
+}
+
+static void
+apply_image_qualifier_to_variable(const struct ast_type_qualifier *qual,
+                                  ir_variable *var,
+                                  struct _mesa_glsl_parse_state *state,
+                                  YYLTYPE *loc)
+{
+   const glsl_type *base_type =
+      (var->type->is_array() ? var->type->element_type() : var->type);
+
+   if (base_type->is_image()) {
+      if (var->data.mode != ir_var_uniform &&
+          var->data.mode != ir_var_function_in) {
+         _mesa_glsl_error(loc, state, "image variables may only be declared as "
+                          "function parameters or uniform-qualified "
+                          "global variables");
+      }
+
+      var->data.image.read_only |= qual->flags.q.read_only;
+      var->data.image.write_only |= qual->flags.q.write_only;
+      var->data.image.coherent |= qual->flags.q.coherent;
+      var->data.image._volatile |= qual->flags.q._volatile;
+      var->data.image.restrict_flag |= qual->flags.q.restrict_flag;
+      var->data.read_only = true;
+
+      if (qual->flags.q.explicit_image_format) {
+         if (var->data.mode == ir_var_function_in) {
+            _mesa_glsl_error(loc, state, "format qualifiers cannot be "
+                             "used on image function parameters");
+         }
+
+         if (qual->image_base_type != base_type->sampler_type) {
+            _mesa_glsl_error(loc, state, "format qualifier doesn't match the "
+                             "base data type of the image");
+         }
+
+         var->data.image.format = qual->image_format;
+      } else {
+         if (var->data.mode == ir_var_uniform && !qual->flags.q.write_only) {
+            _mesa_glsl_error(loc, state, "uniforms not qualified with "
+                             "`writeonly' must have a format layout "
+                             "qualifier");
+         }
 
-   return;
+         var->data.image.format = GL_NONE;
+      }
+   }
 }
 
 static void
@@ -2500,6 +2588,9 @@ apply_type_qualifier_to_variable(const struct ast_type_qualifier *qual,
    if (qual->flags.q.row_major || qual->flags.q.column_major) {
       validate_matrix_layout_for_type(state, loc, var->type, var);
    }
+
+   if (var->type->contains_image())
+      apply_image_qualifier_to_variable(qual, var, state, loc);
 }
 
 /**
@@ -2659,9 +2750,15 @@ process_initializer(ir_variable *var, ast_declaration *decl,
                            "cannot initialize uniforms");
    }
 
-   if (var->type->is_sampler()) {
+   /* From section 4.1.7 of the GLSL 4.40 spec:
+    *
+    *    "Opaque variables [...] are initialized only through the
+    *     OpenGL API; they cannot be declared with an initializer in a
+    *     shader."
+    */
+   if (var->type->contains_opaque()) {
       _mesa_glsl_error(& initializer_loc, state,
-		       "cannot initialize samplers");
+		       "cannot initialize opaque variable");
    }
 
    if ((var->data.mode == ir_var_shader_in) && (state->current_function == NULL)) {
@@ -2784,7 +2881,7 @@ handle_geometry_shader_input_decl(struct _mesa_glsl_parse_state *state,
 {
    unsigned num_vertices = 0;
    if (state->gs_input_prim_type_specified) {
-      num_vertices = vertices_per_prim(state->gs_input_prim_type);
+      num_vertices = vertices_per_prim(state->in_qualifier->prim_type);
    }
 
    /* Geometry shader input variables must be arrays.  Caller should have
@@ -2870,10 +2967,17 @@ validate_identifier(const char *identifier, YYLTYPE loc,
        *     "In addition, all identifiers containing two
        *      consecutive underscores (__) are reserved as
        *      possible future keywords."
+       *
+       * The intention is that names containing __ are reserved for internal
+       * use by the implementation, and names prefixed with GL_ are reserved
+       * for use by Khronos.  Names simply containing __ are dangerous to use,
+       * but should be allowed.
+       *
+       * A future version of the GLSL specification will clarify this.
        */
-      _mesa_glsl_error(&loc, state,
-                       "identifier `%s' uses reserved `__' string",
-                       identifier);
+      _mesa_glsl_warning(&loc, state,
+                         "identifier `%s' uses reserved `__' string",
+                         identifier);
    }
 }
 
@@ -3082,6 +3186,7 @@ ast_declarator_list::hir(exec_list *instructions,
        */
       if (!state->is_version(130, 300)
 	  && !state->has_explicit_attrib_location()
+	  && !state->has_separate_shader_objects()
 	  && !state->ARB_fragment_coord_conventions_enable) {
 	 if (this->type->qualifier.flags.q.out) {
 	    _mesa_glsl_error(& loc, state,
@@ -3421,15 +3526,15 @@ ast_declarator_list::hir(exec_list *instructions,
                           ", integer and sampler types");
       }
 
-      /* From page 17 (page 23 of the PDF) of the GLSL 1.20 spec:
+      /* From section 4.1.7 of the GLSL 4.40 spec:
        *
-       *    "[Sampler types] can only be declared as function
-       *    parameters or uniform variables (see Section 4.3.5
-       *    "Uniform")".
+       *    "[Opaque types] can only be declared as function
+       *     parameters or uniform-qualified variables."
        */
-      if (var_type->contains_sampler() &&
+      if (var_type->contains_opaque() &&
           !this->type->qualifier.flags.q.uniform) {
-         _mesa_glsl_error(&loc, state, "samplers must be declared uniform");
+         _mesa_glsl_error(&loc, state,
+                          "opaque variables must be declared uniform");
       }
 
       /* Process the initializer and add its instructions to a temporary
@@ -3621,15 +3726,16 @@ ast_parameter_declarator::hir(exec_list *instructions,
    apply_type_qualifier_to_variable(& this->type->qualifier, var, state, & loc,
 				    true);
 
-   /* From page 17 (page 23 of the PDF) of the GLSL 1.20 spec:
+   /* From section 4.1.7 of the GLSL 4.40 spec:
     *
-    *    "Samplers cannot be treated as l-values; hence cannot be used
-    *    as out or inout function parameters, nor can they be assigned
-    *    into."
+    *   "Opaque variables cannot be treated as l-values; hence cannot
+    *    be used as out or inout function parameters, nor can they be
+    *    assigned into."
     */
    if ((var->data.mode == ir_var_function_inout || var->data.mode == ir_var_function_out)
-       && type->contains_sampler()) {
-      _mesa_glsl_error(&loc, state, "out and inout parameters cannot contain samplers");
+       && type->contains_opaque()) {
+      _mesa_glsl_error(&loc, state, "out and inout parameters cannot "
+                       "contain opaque variables");
       type = glsl_type::error_type;
    }
 
@@ -3784,15 +3890,15 @@ ast_function::hir(exec_list *instructions,
 		       "sized", name);
    }
 
-   /* From page 17 (page 23 of the PDF) of the GLSL 1.20 spec:
+   /* From section 4.1.7 of the GLSL 4.40 spec:
     *
-    *    "[Sampler types] can only be declared as function parameters
-    *    or uniform variables (see Section 4.3.5 "Uniform")".
+    *    "[Opaque types] can only be declared as function parameters
+    *     or uniform-qualified variables."
     */
-   if (return_type->contains_sampler()) {
+   if (return_type->contains_opaque()) {
       YYLTYPE loc = this->get_location();
       _mesa_glsl_error(&loc, state,
-                       "function `%s' return type can't contain a sampler",
+                       "function `%s' return type can't contain an opaque type",
                        name);
    }
 
@@ -4693,12 +4799,9 @@ ast_process_structure_or_interface_block(exec_list *instructions,
          if (!allow_reserved_names)
             validate_identifier(decl->identifier, loc, state);
 
-         /* From the GL_ARB_uniform_buffer_object spec:
+         /* From section 4.3.9 of the GLSL 4.40 spec:
           *
-          *     "Sampler types are not allowed inside of uniform
-          *      blocks. All other types, arrays, and structures
-          *      allowed for uniforms are allowed within a uniform
-          *      block."
+          *    "[In interface blocks] opaque types are not allowed."
           *
           * It should be impossible for decl_type to be NULL here.  Cases that
           * might naturally lead to decl_type being NULL, especially for the
@@ -4708,10 +4811,11 @@ ast_process_structure_or_interface_block(exec_list *instructions,
          const struct glsl_type *field_type =
             decl_type != NULL ? decl_type : glsl_type::error_type;
 
-         if (is_interface && field_type->contains_sampler()) {
+         if (is_interface && field_type->contains_opaque()) {
             YYLTYPE loc = decl_list->get_location();
             _mesa_glsl_error(&loc, state,
-                             "uniform in non-default uniform block contains sampler");
+                             "uniform in non-default uniform block contains "
+                             "opaque variable");
          }
 
          if (field_type->contains_atomic()) {
@@ -4725,6 +4829,16 @@ ast_process_structure_or_interface_block(exec_list *instructions,
                              "uniform block");
          }
 
+         if (field_type->contains_image()) {
+            /* FINISHME: Same problem as with atomic counters.
+             * FINISHME: Request clarification from Khronos and add
+             * FINISHME: spec quotation here.
+             */
+            YYLTYPE loc = decl_list->get_location();
+            _mesa_glsl_error(&loc, state,
+                             "image in structure or uniform block");
+         }
+
          const struct ast_type_qualifier *const qual =
             & decl_list->type->qualifier;
          if (qual->flags.q.std140 ||
@@ -5258,7 +5372,7 @@ ast_gs_input_layout::hir(exec_list *instructions,
     * was consistent with this one.
     */
    if (state->gs_input_prim_type_specified &&
-       state->gs_input_prim_type != this->prim_type) {
+       state->in_qualifier->prim_type != this->prim_type) {
       _mesa_glsl_error(&loc, state,
                        "geometry shader input layout does not match"
                        " previous declaration");
@@ -5279,7 +5393,6 @@ ast_gs_input_layout::hir(exec_list *instructions,
    }
 
    state->gs_input_prim_type_specified = true;
-   state->gs_input_prim_type = this->prim_type;
 
    /* If any shader inputs occurred before this declaration and did not
     * specify an array size, their size is determined now.
diff --git a/mesalib/src/glsl/ast_type.cpp b/mesalib/src/glsl/ast_type.cpp
index bbc430808..0ee2c495a 100644
--- a/mesalib/src/glsl/ast_type.cpp
+++ b/mesalib/src/glsl/ast_type.cpp
@@ -190,6 +190,117 @@ ast_type_qualifier::merge_qualifier(YYLTYPE *loc,
    if (q.precision != ast_precision_none)
       this->precision = q.precision;
 
+   if (q.flags.q.explicit_image_format) {
+      this->image_format = q.image_format;
+      this->image_base_type = q.image_base_type;
+   }
+
    return true;
 }
 
+bool
+ast_type_qualifier::merge_in_qualifier(YYLTYPE *loc,
+                                       _mesa_glsl_parse_state *state,
+                                       ast_type_qualifier q,
+                                       ast_node* &node)
+{
+   void *mem_ctx = state;
+   bool create_gs_ast = false;
+   bool create_cs_ast = false;
+   ast_type_qualifier valid_in_mask;
+   valid_in_mask.flags.i = 0;
+
+   switch (state->stage) {
+   case MESA_SHADER_GEOMETRY:
+      if (q.flags.q.prim_type) {
+         /* Make sure this is a valid input primitive type. */
+         switch (q.prim_type) {
+         case GL_POINTS:
+         case GL_LINES:
+         case GL_LINES_ADJACENCY:
+         case GL_TRIANGLES:
+         case GL_TRIANGLES_ADJACENCY:
+            break;
+         default:
+            _mesa_glsl_error(loc, state,
+                             "invalid geometry shader input primitive type");
+            break;
+         }
+      }
+
+      create_gs_ast |=
+         q.flags.q.prim_type &&
+         !state->in_qualifier->flags.q.prim_type;
+
+      valid_in_mask.flags.q.prim_type = 1;
+      valid_in_mask.flags.q.invocations = 1;
+      break;
+   case MESA_SHADER_FRAGMENT:
+      if (q.flags.q.early_fragment_tests) {
+         state->early_fragment_tests = true;
+      } else {
+         _mesa_glsl_error(loc, state, "invalid input layout qualifier");
+      }
+      break;
+   case MESA_SHADER_COMPUTE:
+      create_cs_ast |=
+         q.flags.q.local_size != 0 &&
+         state->in_qualifier->flags.q.local_size == 0;
+
+      valid_in_mask.flags.q.local_size = 1;
+      break;
+   default:
+      _mesa_glsl_error(loc, state,
+                       "input layout qualifiers only valid in "
+                       "geometry, fragment and compute shaders");
+      break;
+   }
+
+   /* Generate an error when invalid input layout qualifiers are used. */
+   if ((q.flags.i & ~valid_in_mask.flags.i) != 0) {
+      _mesa_glsl_error(loc, state,
+		       "invalid input layout qualifiers used");
+      return false;
+   }
+
+   /* Input layout qualifiers can be specified multiple
+    * times in separate declarations, as long as they match.
+    */
+   if (this->flags.q.prim_type) {
+      if (q.flags.q.prim_type &&
+          this->prim_type != q.prim_type) {
+         _mesa_glsl_error(loc, state,
+                          "conflicting input primitive types specified");
+      }
+   } else if (q.flags.q.prim_type) {
+      state->in_qualifier->flags.q.prim_type = 1;
+      state->in_qualifier->prim_type = q.prim_type;
+   }
+
+   if (this->flags.q.invocations &&
+       q.flags.q.invocations &&
+       this->invocations != q.invocations) {
+      _mesa_glsl_error(loc, state,
+                       "conflicting invocations counts specified");
+      return false;
+   } else if (q.flags.q.invocations) {
+      this->flags.q.invocations = 1;
+      this->invocations = q.invocations;
+   }
+
+   if (create_gs_ast) {
+      node = new(mem_ctx) ast_gs_input_layout(*loc, q.prim_type);
+   } else if (create_cs_ast) {
+      /* Infer a local_size of 1 for every unspecified dimension */
+      unsigned local_size[3];
+      for (int i = 0; i < 3; i++) {
+         if (q.flags.q.local_size & (1 << i))
+            local_size[i] = q.local_size[i];
+         else
+            local_size[i] = 1;
+      }
+      node = new(mem_ctx) ast_cs_input_layout(*loc, local_size);
+   }
+
+   return true;
+}
diff --git a/mesalib/src/glsl/builtin_functions.cpp b/mesalib/src/glsl/builtin_functions.cpp
index ed58ddddb..c4f39abb7 100755
--- a/mesalib/src/glsl/builtin_functions.cpp
+++ b/mesalib/src/glsl/builtin_functions.cpp
@@ -340,6 +340,13 @@ shader_trinary_minmax(const _mesa_glsl_parse_state *state)
    return state->AMD_shader_trinary_minmax_enable;
 }
 
+static bool
+shader_image_load_store(const _mesa_glsl_parse_state *state)
+{
+   return (state->is_version(420, 0) ||
+           state->ARB_shader_image_load_store_enable);
+}
+
 /** @} */
 
 /******************************************************************************/
@@ -413,6 +420,33 @@ private:
    /** Create a new function and add the given signatures. */
    void add_function(const char *name, ...);
 
+   enum image_function_flags {
+      IMAGE_FUNCTION_EMIT_STUB = (1 << 0),
+      IMAGE_FUNCTION_RETURNS_VOID = (1 << 1),
+      IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE = (1 << 2),
+      IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE = (1 << 3),
+      IMAGE_FUNCTION_READ_ONLY = (1 << 4),
+      IMAGE_FUNCTION_WRITE_ONLY = (1 << 5)
+   };
+
+   /**
+    * Create a new image built-in function for all known image types.
+    * \p flags is a bitfield of \c image_function_flags flags.
+    */
+   void add_image_function(const char *name,
+                           const char *intrinsic_name,
+                           unsigned num_arguments,
+                           unsigned flags);
+
+   /**
+    * Create new functions for all known image built-ins and types.
+    * If \p glsl is \c true, use the GLSL built-in names and emit code
+    * to call into the actual compiler intrinsic.  If \p glsl is
+    * false, emit a function prototype with no body for each image
+    * intrinsic name.
+    */
+   void add_image_functions(bool glsl);
+
    ir_function_signature *new_sig(const glsl_type *return_type,
                                   builtin_available_predicate avail,
                                   int num_params, ...);
@@ -580,6 +614,20 @@ private:
    B1(max3)
    B1(mid3)
 
+   ir_function_signature *_image_prototype(const glsl_type *image_type,
+                                           const char *intrinsic_name,
+                                           unsigned num_arguments,
+                                           unsigned flags);
+   ir_function_signature *_image(const glsl_type *image_type,
+                                 const char *intrinsic_name,
+                                 unsigned num_arguments,
+                                 unsigned flags);
+
+   ir_function_signature *_memory_barrier_intrinsic(
+      builtin_available_predicate avail);
+   ir_function_signature *_memory_barrier(
+      builtin_available_predicate avail);
+
 #undef B0
 #undef B1
 #undef B2
@@ -694,6 +742,12 @@ builtin_builder::create_intrinsics()
    add_function("__intrinsic_atomic_predecrement",
                 _atomic_intrinsic(shader_atomic_counters),
                 NULL);
+
+   add_image_functions(false);
+
+   add_function("__intrinsic_memory_barrier",
+                _memory_barrier_intrinsic(shader_image_load_store),
+                NULL);
 }
 
 /**
@@ -2167,6 +2221,12 @@ builtin_builder::create_builtins()
                 _mid3(glsl_type::uvec4_type),
                 NULL);
 
+   add_image_functions(true);
+
+   add_function("memoryBarrier",
+                _memory_barrier(shader_image_load_store),
+                NULL);
+
 #undef F
 #undef FI
 #undef FIU
@@ -2200,6 +2260,104 @@ builtin_builder::add_function(const char *name, ...)
    shader->symbols->add_function(f);
 }
 
+void
+builtin_builder::add_image_function(const char *name,
+                                    const char *intrinsic_name,
+                                    unsigned num_arguments,
+                                    unsigned flags)
+{
+   static const glsl_type *const types[] = {
+      glsl_type::image1D_type,
+      glsl_type::image2D_type,
+      glsl_type::image3D_type,
+      glsl_type::image2DRect_type,
+      glsl_type::imageCube_type,
+      glsl_type::imageBuffer_type,
+      glsl_type::image1DArray_type,
+      glsl_type::image2DArray_type,
+      glsl_type::imageCubeArray_type,
+      glsl_type::image2DMS_type,
+      glsl_type::image2DMSArray_type,
+      glsl_type::iimage1D_type,
+      glsl_type::iimage2D_type,
+      glsl_type::iimage3D_type,
+      glsl_type::iimage2DRect_type,
+      glsl_type::iimageCube_type,
+      glsl_type::iimageBuffer_type,
+      glsl_type::iimage1DArray_type,
+      glsl_type::iimage2DArray_type,
+      glsl_type::iimageCubeArray_type,
+      glsl_type::iimage2DMS_type,
+      glsl_type::iimage2DMSArray_type,
+      glsl_type::uimage1D_type,
+      glsl_type::uimage2D_type,
+      glsl_type::uimage3D_type,
+      glsl_type::uimage2DRect_type,
+      glsl_type::uimageCube_type,
+      glsl_type::uimageBuffer_type,
+      glsl_type::uimage1DArray_type,
+      glsl_type::uimage2DArray_type,
+      glsl_type::uimageCubeArray_type,
+      glsl_type::uimage2DMS_type,
+      glsl_type::uimage2DMSArray_type
+   };
+   ir_function *f = new(mem_ctx) ir_function(name);
+
+   for (unsigned i = 0; i < Elements(types); ++i) {
+      if (types[i]->sampler_type != GLSL_TYPE_FLOAT ||
+          (flags & IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE))
+         f->add_signature(_image(types[i], intrinsic_name,
+                                 num_arguments, flags));
+   }
+
+   shader->symbols->add_function(f);
+}
+
+void
+builtin_builder::add_image_functions(bool glsl)
+{
+   const unsigned flags = (glsl ? IMAGE_FUNCTION_EMIT_STUB : 0);
+
+   add_image_function(glsl ? "imageLoad" : "__intrinsic_image_load",
+                      "__intrinsic_image_load", 0,
+                      (flags | IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE |
+                       IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE |
+                       IMAGE_FUNCTION_READ_ONLY));
+
+   add_image_function(glsl ? "imageStore" : "__intrinsic_image_store",
+                      "__intrinsic_image_store", 1,
+                      (flags | IMAGE_FUNCTION_RETURNS_VOID |
+                       IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE |
+                       IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE |
+                       IMAGE_FUNCTION_WRITE_ONLY));
+
+   add_image_function(glsl ? "imageAtomicAdd" : "__intrinsic_image_atomic_add",
+                      "__intrinsic_image_atomic_add", 1, flags);
+
+   add_image_function(glsl ? "imageAtomicMin" : "__intrinsic_image_atomic_min",
+                      "__intrinsic_image_atomic_min", 1, flags);
+
+   add_image_function(glsl ? "imageAtomicMax" : "__intrinsic_image_atomic_max",
+                      "__intrinsic_image_atomic_max", 1, flags);
+
+   add_image_function(glsl ? "imageAtomicAnd" : "__intrinsic_image_atomic_and",
+                      "__intrinsic_image_atomic_and", 1, flags);
+
+   add_image_function(glsl ? "imageAtomicOr" : "__intrinsic_image_atomic_or",
+                      "__intrinsic_image_atomic_or", 1, flags);
+
+   add_image_function(glsl ? "imageAtomicXor" : "__intrinsic_image_atomic_xor",
+                      "__intrinsic_image_atomic_xor", 1, flags);
+
+   add_image_function((glsl ? "imageAtomicExchange" :
+                       "__intrinsic_image_atomic_exchange"),
+                      "__intrinsic_image_atomic_exchange", 1, flags);
+
+   add_image_function((glsl ? "imageAtomicCompSwap" :
+                       "__intrinsic_image_atomic_comp_swap"),
+                      "__intrinsic_image_atomic_comp_swap", 2, flags);
+}
+
 ir_variable *
 builtin_builder::in_var(const glsl_type *type, const char *name)
 {
@@ -3549,7 +3707,7 @@ builtin_builder::_texture(ir_texture_opcode opcode,
    ir_texture *tex = new(mem_ctx) ir_texture(opcode);
    tex->set_sampler(var_ref(s), return_type);
 
-   const int coord_size = sampler_type->sampler_coordinate_components();
+   const int coord_size = sampler_type->coordinate_components();
 
    if (coord_size == coord_type->vector_elements) {
       tex->coordinate = var_ref(P);
@@ -4094,12 +4252,106 @@ builtin_builder::_mid3(const glsl_type *type)
    return sig;
 }
 
+ir_function_signature *
+builtin_builder::_image_prototype(const glsl_type *image_type,
+                                  const char *intrinsic_name,
+                                  unsigned num_arguments,
+                                  unsigned flags)
+{
+   const glsl_type *data_type = glsl_type::get_instance(
+      image_type->sampler_type,
+      (flags & IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE ? 4 : 1),
+      1);
+   const glsl_type *ret_type = (flags & IMAGE_FUNCTION_RETURNS_VOID ?
+                                glsl_type::void_type : data_type);
+
+   /* Addressing arguments that are always present. */
+   ir_variable *image = in_var(image_type, "image");
+   ir_variable *coord = in_var(
+      glsl_type::ivec(image_type->coordinate_components()), "coord");
+
+   ir_function_signature *sig = new_sig(
+      ret_type, shader_image_load_store, 2, image, coord);
+
+   /* Sample index for multisample images. */
+   if (image_type->sampler_dimensionality == GLSL_SAMPLER_DIM_MS)
+      sig->parameters.push_tail(in_var(glsl_type::int_type, "sample"));
+
+   /* Data arguments. */
+   for (unsigned i = 0; i < num_arguments; ++i)
+      sig->parameters.push_tail(in_var(data_type,
+                                       ralloc_asprintf(NULL, "arg%d", i)));
+
+   /* Set the maximal set of qualifiers allowed for this image
+    * built-in.  Function calls with arguments having fewer
+    * qualifiers than present in the prototype are allowed by the
+    * spec, but not with more, i.e. this will make the compiler
+    * accept everything that needs to be accepted, and reject cases
+    * like loads from write-only or stores to read-only images.
+    */
+   image->data.image.read_only = flags & IMAGE_FUNCTION_READ_ONLY;
+   image->data.image.write_only = flags & IMAGE_FUNCTION_WRITE_ONLY;
+   image->data.image.coherent = true;
+   image->data.image._volatile = true;
+   image->data.image.restrict_flag = true;
+
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_image(const glsl_type *image_type,
+                        const char *intrinsic_name,
+                        unsigned num_arguments,
+                        unsigned flags)
+{
+   ir_function_signature *sig = _image_prototype(image_type, intrinsic_name,
+                                                 num_arguments, flags);
+
+   if (flags & IMAGE_FUNCTION_EMIT_STUB) {
+      ir_factory body(&sig->body, mem_ctx);
+      ir_function *f = shader->symbols->get_function(intrinsic_name);
+
+      if (flags & IMAGE_FUNCTION_RETURNS_VOID) {
+         body.emit(call(f, NULL, sig->parameters));
+      } else {
+         ir_variable *ret_val =
+            body.make_temp(sig->return_type, "_ret_val");
+         body.emit(call(f, ret_val, sig->parameters));
+         body.emit(ret(ret_val));
+      }
+
+      sig->is_defined = true;
+
+   } else {
+      sig->is_intrinsic = true;
+   }
+
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_memory_barrier_intrinsic(builtin_available_predicate avail)
+{
+   MAKE_INTRINSIC(glsl_type::void_type, avail, 0);
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_memory_barrier(builtin_available_predicate avail)
+{
+   MAKE_SIG(glsl_type::void_type, avail, 0);
+   body.emit(call(shader->symbols->get_function("__intrinsic_memory_barrier"),
+                  NULL, sig->parameters));
+   return sig;
+}
+
 /** @} */
 
 /******************************************************************************/
 
 /* The singleton instance of builtin_builder. */
 static builtin_builder builtins;
+static mtx_t builtins_lock = _MTX_INITIALIZER_NP;
 
 /**
  * External API (exposing the built-in module to the rest of the compiler):
@@ -4108,20 +4360,28 @@ static builtin_builder builtins;
 void
 _mesa_glsl_initialize_builtin_functions()
 {
+   mtx_lock(&builtins_lock);
    builtins.initialize();
+   mtx_unlock(&builtins_lock);
 }
 
 void
 _mesa_glsl_release_builtin_functions()
 {
+   mtx_lock(&builtins_lock);
    builtins.release();
+   mtx_unlock(&builtins_lock);
 }
 
 ir_function_signature *
 _mesa_glsl_find_builtin_function(_mesa_glsl_parse_state *state,
                                  const char *name, exec_list *actual_parameters)
 {
-   return builtins.find(state, name, actual_parameters);
+   ir_function_signature * s;
+   mtx_lock(&builtins_lock);
+   s = builtins.find(state, name, actual_parameters);
+   mtx_unlock(&builtins_lock);
+   return s;
 }
 
 gl_shader *
diff --git a/mesalib/src/glsl/builtin_type_macros.h b/mesalib/src/glsl/builtin_type_macros.h
index 06b4dbd0f..236e1ce8c 100644
--- a/mesalib/src/glsl/builtin_type_macros.h
+++ b/mesalib/src/glsl/builtin_type_macros.h
@@ -64,51 +64,85 @@ DECL_TYPE(mat3x4, GL_FLOAT_MAT3x4, GLSL_TYPE_FLOAT, 4, 3)
 DECL_TYPE(mat4x2, GL_FLOAT_MAT4x2, GLSL_TYPE_FLOAT, 2, 4)
 DECL_TYPE(mat4x3, GL_FLOAT_MAT4x3, GLSL_TYPE_FLOAT, 3, 4)
 
-DECL_TYPE(sampler1D,         GL_SAMPLER_1D,                   GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2D,         GL_SAMPLER_2D,                   GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler3D,         GL_SAMPLER_3D,                   GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(samplerCube,       GL_SAMPLER_CUBE,                 GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler1DArray,    GL_SAMPLER_1D_ARRAY,             GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DArray,    GL_SAMPLER_2D_ARRAY,             GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(samplerCubeArray,  GL_SAMPLER_CUBE_MAP_ARRAY,       GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DRect,     GL_SAMPLER_2D_RECT,              GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(samplerBuffer,     GL_SAMPLER_BUFFER,               GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DMS,       GL_SAMPLER_2D_MULTISAMPLE,       GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DMSArray,  GL_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_FLOAT)
-
-DECL_TYPE(isampler1D,        GL_INT_SAMPLER_1D,                   GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isampler2D,        GL_INT_SAMPLER_2D,                   GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isampler3D,        GL_INT_SAMPLER_3D,                   GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isamplerCube,      GL_INT_SAMPLER_CUBE,                 GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isampler1DArray,   GL_INT_SAMPLER_1D_ARRAY,             GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_INT)
-DECL_TYPE(isampler2DArray,   GL_INT_SAMPLER_2D_ARRAY,             GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_INT)
-DECL_TYPE(isamplerCubeArray, GL_INT_SAMPLER_CUBE_MAP_ARRAY,       GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_INT)
-DECL_TYPE(isampler2DRect,    GL_INT_SAMPLER_2D_RECT,              GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isamplerBuffer,    GL_INT_SAMPLER_BUFFER,               GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isampler2DMS,      GL_INT_SAMPLER_2D_MULTISAMPLE,       GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_INT)
-DECL_TYPE(isampler2DMSArray, GL_INT_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_INT)
-
-DECL_TYPE(usampler1D,        GL_UNSIGNED_INT_SAMPLER_1D,                   GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usampler2D,        GL_UNSIGNED_INT_SAMPLER_2D,                   GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usampler3D,        GL_UNSIGNED_INT_SAMPLER_3D,                   GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usamplerCube,      GL_UNSIGNED_INT_SAMPLER_CUBE,                 GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usampler1DArray,   GL_UNSIGNED_INT_SAMPLER_1D_ARRAY,             GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_UINT)
-DECL_TYPE(usampler2DArray,   GL_UNSIGNED_INT_SAMPLER_2D_ARRAY,             GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_UINT)
-DECL_TYPE(usamplerCubeArray, GL_UNSIGNED_INT_SAMPLER_CUBE_MAP_ARRAY,       GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_UINT)
-DECL_TYPE(usampler2DRect,    GL_UNSIGNED_INT_SAMPLER_2D_RECT,              GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usamplerBuffer,    GL_UNSIGNED_INT_SAMPLER_BUFFER,               GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usampler2DMS,      GL_UNSIGNED_INT_SAMPLER_2D_MULTISAMPLE,       GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_UINT)
-DECL_TYPE(usampler2DMSArray, GL_UNSIGNED_INT_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_UINT)
-
-DECL_TYPE(sampler1DShadow,        GL_SAMPLER_1D_SHADOW,             GLSL_SAMPLER_DIM_1D,       1, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DShadow,        GL_SAMPLER_2D_SHADOW,             GLSL_SAMPLER_DIM_2D,       1, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(samplerCubeShadow,      GL_SAMPLER_CUBE_SHADOW,           GLSL_SAMPLER_DIM_CUBE,     1, 0, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler1DArrayShadow,   GL_SAMPLER_1D_ARRAY_SHADOW,       GLSL_SAMPLER_DIM_1D,       1, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DArrayShadow,   GL_SAMPLER_2D_ARRAY_SHADOW,       GLSL_SAMPLER_DIM_2D,       1, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(samplerCubeArrayShadow, GL_SAMPLER_CUBE_MAP_ARRAY_SHADOW, GLSL_SAMPLER_DIM_CUBE,     1, 1, GLSL_TYPE_FLOAT)
-DECL_TYPE(sampler2DRectShadow,    GL_SAMPLER_2D_RECT_SHADOW,        GLSL_SAMPLER_DIM_RECT,     1, 0, GLSL_TYPE_FLOAT)
-
-DECL_TYPE(samplerExternalOES,     GL_SAMPLER_EXTERNAL_OES,          GLSL_SAMPLER_DIM_EXTERNAL, 0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler1D,         GL_SAMPLER_1D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2D,         GL_SAMPLER_2D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler3D,         GL_SAMPLER_3D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(samplerCube,       GL_SAMPLER_CUBE,                 GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler1DArray,    GL_SAMPLER_1D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DArray,    GL_SAMPLER_2D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(samplerCubeArray,  GL_SAMPLER_CUBE_MAP_ARRAY,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DRect,     GL_SAMPLER_2D_RECT,              GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(samplerBuffer,     GL_SAMPLER_BUFFER,               GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DMS,       GL_SAMPLER_2D_MULTISAMPLE,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DMSArray,  GL_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_FLOAT)
+
+DECL_TYPE(isampler1D,        GL_INT_SAMPLER_1D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isampler2D,        GL_INT_SAMPLER_2D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isampler3D,        GL_INT_SAMPLER_3D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isamplerCube,      GL_INT_SAMPLER_CUBE,                 GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isampler1DArray,   GL_INT_SAMPLER_1D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_INT)
+DECL_TYPE(isampler2DArray,   GL_INT_SAMPLER_2D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_INT)
+DECL_TYPE(isamplerCubeArray, GL_INT_SAMPLER_CUBE_MAP_ARRAY,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_INT)
+DECL_TYPE(isampler2DRect,    GL_INT_SAMPLER_2D_RECT,              GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isamplerBuffer,    GL_INT_SAMPLER_BUFFER,               GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isampler2DMS,      GL_INT_SAMPLER_2D_MULTISAMPLE,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_INT)
+DECL_TYPE(isampler2DMSArray, GL_INT_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_INT)
+
+DECL_TYPE(usampler1D,        GL_UNSIGNED_INT_SAMPLER_1D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usampler2D,        GL_UNSIGNED_INT_SAMPLER_2D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usampler3D,        GL_UNSIGNED_INT_SAMPLER_3D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_3D,   0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usamplerCube,      GL_UNSIGNED_INT_SAMPLER_CUBE,                 GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usampler1DArray,   GL_UNSIGNED_INT_SAMPLER_1D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 1, GLSL_TYPE_UINT)
+DECL_TYPE(usampler2DArray,   GL_UNSIGNED_INT_SAMPLER_2D_ARRAY,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 1, GLSL_TYPE_UINT)
+DECL_TYPE(usamplerCubeArray, GL_UNSIGNED_INT_SAMPLER_CUBE_MAP_ARRAY,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE, 0, 1, GLSL_TYPE_UINT)
+DECL_TYPE(usampler2DRect,    GL_UNSIGNED_INT_SAMPLER_2D_RECT,              GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_RECT, 0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usamplerBuffer,    GL_UNSIGNED_INT_SAMPLER_BUFFER,               GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_BUF,  0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usampler2DMS,      GL_UNSIGNED_INT_SAMPLER_2D_MULTISAMPLE,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 0, GLSL_TYPE_UINT)
+DECL_TYPE(usampler2DMSArray, GL_UNSIGNED_INT_SAMPLER_2D_MULTISAMPLE_ARRAY, GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_MS,   0, 1, GLSL_TYPE_UINT)
+
+DECL_TYPE(sampler1DShadow,        GL_SAMPLER_1D_SHADOW,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,       1, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DShadow,        GL_SAMPLER_2D_SHADOW,             GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,       1, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(samplerCubeShadow,      GL_SAMPLER_CUBE_SHADOW,           GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE,     1, 0, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler1DArrayShadow,   GL_SAMPLER_1D_ARRAY_SHADOW,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,       1, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DArrayShadow,   GL_SAMPLER_2D_ARRAY_SHADOW,       GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,       1, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(samplerCubeArrayShadow, GL_SAMPLER_CUBE_MAP_ARRAY_SHADOW, GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_CUBE,     1, 1, GLSL_TYPE_FLOAT)
+DECL_TYPE(sampler2DRectShadow,    GL_SAMPLER_2D_RECT_SHADOW,        GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_RECT,     1, 0, GLSL_TYPE_FLOAT)
+
+DECL_TYPE(samplerExternalOES,     GL_SAMPLER_EXTERNAL_OES,          GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_EXTERNAL, 0, 0, GLSL_TYPE_FLOAT)
+
+DECL_TYPE(image1D,         GL_IMAGE_1D,                                GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(image2D,         GL_IMAGE_2D,                                GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(image3D,         GL_IMAGE_3D,                                GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_3D,     0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(image2DRect,     GL_IMAGE_2D_RECT,                           GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_RECT,   0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(imageCube,       GL_IMAGE_CUBE,                              GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(imageBuffer,     GL_IMAGE_BUFFER,                            GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_BUF,    0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(image1DArray,    GL_IMAGE_1D_ARRAY,                          GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 1, GLSL_TYPE_FLOAT);
+DECL_TYPE(image2DArray,    GL_IMAGE_2D_ARRAY,                          GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 1, GLSL_TYPE_FLOAT);
+DECL_TYPE(imageCubeArray,  GL_IMAGE_CUBE_MAP_ARRAY,                    GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 1, GLSL_TYPE_FLOAT);
+DECL_TYPE(image2DMS,       GL_IMAGE_2D_MULTISAMPLE,                    GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 0, GLSL_TYPE_FLOAT);
+DECL_TYPE(image2DMSArray,  GL_IMAGE_2D_MULTISAMPLE_ARRAY,              GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 1, GLSL_TYPE_FLOAT);
+DECL_TYPE(iimage1D,        GL_INT_IMAGE_1D,                            GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimage2D,        GL_INT_IMAGE_2D,                            GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimage3D,        GL_INT_IMAGE_3D,                            GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_3D,     0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimage2DRect,    GL_INT_IMAGE_2D_RECT,                       GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_RECT,   0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimageCube,      GL_INT_IMAGE_CUBE,                          GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimageBuffer,    GL_INT_IMAGE_BUFFER,                        GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_BUF,    0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimage1DArray,   GL_INT_IMAGE_1D_ARRAY,                      GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 1, GLSL_TYPE_INT);
+DECL_TYPE(iimage2DArray,   GL_INT_IMAGE_2D_ARRAY,                      GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 1, GLSL_TYPE_INT);
+DECL_TYPE(iimageCubeArray, GL_INT_IMAGE_CUBE_MAP_ARRAY,                GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 1, GLSL_TYPE_INT);
+DECL_TYPE(iimage2DMS,      GL_INT_IMAGE_2D_MULTISAMPLE,                GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 0, GLSL_TYPE_INT);
+DECL_TYPE(iimage2DMSArray, GL_INT_IMAGE_2D_MULTISAMPLE_ARRAY,          GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 1, GLSL_TYPE_INT);
+DECL_TYPE(uimage1D,        GL_UNSIGNED_INT_IMAGE_1D,                   GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimage2D,        GL_UNSIGNED_INT_IMAGE_2D,                   GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimage3D,        GL_UNSIGNED_INT_IMAGE_3D,                   GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_3D,     0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimage2DRect,    GL_UNSIGNED_INT_IMAGE_2D_RECT,              GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_RECT,   0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimageCube,      GL_UNSIGNED_INT_IMAGE_CUBE,                 GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimageBuffer,    GL_UNSIGNED_INT_IMAGE_BUFFER,               GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_BUF,    0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimage1DArray,   GL_UNSIGNED_INT_IMAGE_1D_ARRAY,             GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_1D,     0, 1, GLSL_TYPE_UINT);
+DECL_TYPE(uimage2DArray,   GL_UNSIGNED_INT_IMAGE_2D_ARRAY,             GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_2D,     0, 1, GLSL_TYPE_UINT);
+DECL_TYPE(uimageCubeArray, GL_UNSIGNED_INT_IMAGE_CUBE_MAP_ARRAY,       GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_CUBE,   0, 1, GLSL_TYPE_UINT);
+DECL_TYPE(uimage2DMS,      GL_UNSIGNED_INT_IMAGE_2D_MULTISAMPLE,       GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 0, GLSL_TYPE_UINT);
+DECL_TYPE(uimage2DMSArray, GL_UNSIGNED_INT_IMAGE_2D_MULTISAMPLE_ARRAY, GLSL_TYPE_IMAGE, GLSL_SAMPLER_DIM_MS,     0, 1, GLSL_TYPE_UINT);
 
 DECL_TYPE(atomic_uint, GL_UNSIGNED_INT_ATOMIC_COUNTER, GLSL_TYPE_ATOMIC_UINT, 1, 1)
 
diff --git a/mesalib/src/glsl/builtin_types.cpp b/mesalib/src/glsl/builtin_types.cpp
index 92e386057..dd42ecb05 100644
--- a/mesalib/src/glsl/builtin_types.cpp
+++ b/mesalib/src/glsl/builtin_types.cpp
@@ -204,6 +204,40 @@ const static struct builtin_type_versions {
 
    T(struct_gl_DepthRangeParameters,  110, 100)
 
+   T(image1D,                         420, 999)
+   T(image2D,                         420, 999)
+   T(image3D,                         420, 999)
+   T(image2DRect,                     420, 999)
+   T(imageCube,                       420, 999)
+   T(imageBuffer,                     420, 999)
+   T(image1DArray,                    420, 999)
+   T(image2DArray,                    420, 999)
+   T(imageCubeArray,                  420, 999)
+   T(image2DMS,                       420, 999)
+   T(image2DMSArray,                  420, 999)
+   T(iimage1D,                        420, 999)
+   T(iimage2D,                        420, 999)
+   T(iimage3D,                        420, 999)
+   T(iimage2DRect,                    420, 999)
+   T(iimageCube,                      420, 999)
+   T(iimageBuffer,                    420, 999)
+   T(iimage1DArray,                   420, 999)
+   T(iimage2DArray,                   420, 999)
+   T(iimageCubeArray,                 420, 999)
+   T(iimage2DMS,                      420, 999)
+   T(iimage2DMSArray,                 420, 999)
+   T(uimage1D,                        420, 999)
+   T(uimage2D,                        420, 999)
+   T(uimage3D,                        420, 999)
+   T(uimage2DRect,                    420, 999)
+   T(uimageCube,                      420, 999)
+   T(uimageBuffer,                    420, 999)
+   T(uimage1DArray,                   420, 999)
+   T(uimage2DArray,                   420, 999)
+   T(uimageCubeArray,                 420, 999)
+   T(uimage2DMS,                      420, 999)
+   T(uimage2DMSArray,                 420, 999)
+
    T(atomic_uint,                     420, 999)
 };
 
@@ -287,6 +321,42 @@ _mesa_glsl_initialize_types(struct _mesa_glsl_parse_state *state)
       add_type(symbols, glsl_type::sampler3D_type);
    }
 
+   if (state->ARB_shader_image_load_store_enable) {
+      add_type(symbols, glsl_type::image1D_type);
+      add_type(symbols, glsl_type::image2D_type);
+      add_type(symbols, glsl_type::image3D_type);
+      add_type(symbols, glsl_type::image2DRect_type);
+      add_type(symbols, glsl_type::imageCube_type);
+      add_type(symbols, glsl_type::imageBuffer_type);
+      add_type(symbols, glsl_type::image1DArray_type);
+      add_type(symbols, glsl_type::image2DArray_type);
+      add_type(symbols, glsl_type::imageCubeArray_type);
+      add_type(symbols, glsl_type::image2DMS_type);
+      add_type(symbols, glsl_type::image2DMSArray_type);
+      add_type(symbols, glsl_type::iimage1D_type);
+      add_type(symbols, glsl_type::iimage2D_type);
+      add_type(symbols, glsl_type::iimage3D_type);
+      add_type(symbols, glsl_type::iimage2DRect_type);
+      add_type(symbols, glsl_type::iimageCube_type);
+      add_type(symbols, glsl_type::iimageBuffer_type);
+      add_type(symbols, glsl_type::iimage1DArray_type);
+      add_type(symbols, glsl_type::iimage2DArray_type);
+      add_type(symbols, glsl_type::iimageCubeArray_type);
+      add_type(symbols, glsl_type::iimage2DMS_type);
+      add_type(symbols, glsl_type::iimage2DMSArray_type);
+      add_type(symbols, glsl_type::uimage1D_type);
+      add_type(symbols, glsl_type::uimage2D_type);
+      add_type(symbols, glsl_type::uimage3D_type);
+      add_type(symbols, glsl_type::uimage2DRect_type);
+      add_type(symbols, glsl_type::uimageCube_type);
+      add_type(symbols, glsl_type::uimageBuffer_type);
+      add_type(symbols, glsl_type::uimage1DArray_type);
+      add_type(symbols, glsl_type::uimage2DArray_type);
+      add_type(symbols, glsl_type::uimageCubeArray_type);
+      add_type(symbols, glsl_type::uimage2DMS_type);
+      add_type(symbols, glsl_type::uimage2DMSArray_type);
+   }
+
    if (state->ARB_shader_atomic_counters_enable) {
       add_type(symbols, glsl_type::atomic_uint_type);
    }
diff --git a/mesalib/src/glsl/builtin_variables.cpp b/mesalib/src/glsl/builtin_variables.cpp
index cc423383d..4176ae6e6 100644
--- a/mesalib/src/glsl/builtin_variables.cpp
+++ b/mesalib/src/glsl/builtin_variables.cpp
@@ -711,6 +711,26 @@ builtin_variable_generator::generate_constants()
        * in ast_cs_input_layout::hir().
        */
    }
+
+   if (state->is_version(420, 0) ||
+       state->ARB_shader_image_load_store_enable) {
+      add_const("gl_MaxImageUnits",
+                state->Const.MaxImageUnits);
+      add_const("gl_MaxCombinedImageUnitsAndFragmentOutputs",
+                state->Const.MaxCombinedImageUnitsAndFragmentOutputs);
+      add_const("gl_MaxImageSamples",
+                state->Const.MaxImageSamples);
+      add_const("gl_MaxVertexImageUniforms",
+                state->Const.MaxVertexImageUniforms);
+      add_const("gl_MaxTessControlImageUniforms", 0);
+      add_const("gl_MaxTessEvaluationImageUniforms", 0);
+      add_const("gl_MaxGeometryImageUniforms",
+                state->Const.MaxGeometryImageUniforms);
+      add_const("gl_MaxFragmentImageUniforms",
+                state->Const.MaxFragmentImageUniforms);
+      add_const("gl_MaxCombinedImageUniforms",
+                state->Const.MaxCombinedImageUniforms);
+   }
 }
 
 
@@ -834,6 +854,8 @@ builtin_variable_generator::generate_gs_special_vars()
    add_output(VARYING_SLOT_LAYER, int_t, "gl_Layer");
    if (state->ARB_viewport_array_enable)
       add_output(VARYING_SLOT_VIEWPORT, int_t, "gl_ViewportIndex");
+   if (state->ARB_gpu_shader5_enable)
+      add_system_value(SYSTEM_VALUE_INVOCATION_ID, int_t, "gl_InvocationID");
 
    /* Although gl_PrimitiveID appears in tessellation control and tessellation
     * evaluation shaders, it has a different function there than it has in
diff --git a/mesalib/src/glsl/glcpp/glcpp-parse.y b/mesalib/src/glsl/glcpp/glcpp-parse.y
index eeafa4d39..f28d8531e 100644
--- a/mesalib/src/glsl/glcpp/glcpp-parse.y
+++ b/mesalib/src/glsl/glcpp/glcpp-parse.y
@@ -1770,11 +1770,27 @@ static void
 _check_for_reserved_macro_name (glcpp_parser_t *parser, YYLTYPE *loc,
 				const char *identifier)
 {
-	/* According to the GLSL specification, macro names starting with "__"
-	 * or "GL_" are reserved for future use.  So, don't allow them.
+	/* Section 3.3 (Preprocessor) of the GLSL 1.30 spec (and later) and
+	 * the GLSL ES spec (all versions) say:
+	 *
+	 *     "All macro names containing two consecutive underscores ( __ )
+	 *     are reserved for future use as predefined macro names. All
+	 *     macro names prefixed with "GL_" ("GL" followed by a single
+	 *     underscore) are also reserved."
+	 *
+	 * The intention is that names containing __ are reserved for internal
+	 * use by the implementation, and names prefixed with GL_ are reserved
+	 * for use by Khronos.  Since every extension adds a name prefixed
+	 * with GL_ (i.e., the name of the extension), that should be an
+	 * error.  Names simply containing __ are dangerous to use, but should
+	 * be allowed.
+	 *
+	 * A future version of the GLSL specification will clarify this.
 	 */
 	if (strstr(identifier, "__")) {
-		glcpp_error (loc, parser, "Macro names containing \"__\" are reserved.\n");
+		glcpp_warning(loc, parser,
+			      "Macro names containing \"__\" are reserved "
+			      "for use by the implementation.\n");
 	}
 	if (strncmp(identifier, "GL_", 3) == 0) {
 		glcpp_error (loc, parser, "Macro names starting with \"GL_\" are reserved.\n");
@@ -2118,6 +2134,9 @@ _glcpp_parser_handle_version_declaration(glcpp_parser_t *parser, intmax_t versio
 	      if (extensions->ARB_texture_gather)
 	         add_builtin_define(parser, "GL_ARB_texture_gather", 1);
 
+              if (extensions->ARB_separate_shader_objects)
+                 add_builtin_define(parser, "GL_ARB_separate_shader_objects", 1);
+
 	      if (extensions->ARB_shader_atomic_counters)
 	         add_builtin_define(parser, "GL_ARB_shader_atomic_counters", 1);
 
@@ -2126,6 +2145,9 @@ _glcpp_parser_handle_version_declaration(glcpp_parser_t *parser, intmax_t versio
 
               if (extensions->ARB_compute_shader)
                  add_builtin_define(parser, "GL_ARB_compute_shader", 1);
+
+	      if (extensions->ARB_shader_image_load_store)
+	         add_builtin_define(parser, "GL_ARB_shader_image_load_store", 1);
 	   }
 	}
 
diff --git a/mesalib/src/glsl/glcpp/glcpp.c b/mesalib/src/glsl/glcpp/glcpp.c
index 0144be85b..0144be85b 100755..100644
--- a/mesalib/src/glsl/glcpp/glcpp.c
+++ b/mesalib/src/glsl/glcpp/glcpp.c
diff --git a/mesalib/src/glsl/glsl_lexer.ll b/mesalib/src/glsl/glsl_lexer.ll
index 3208b32da..9fd9b80d3 100644
--- a/mesalib/src/glsl/glsl_lexer.ll
+++ b/mesalib/src/glsl/glsl_lexer.ll
@@ -337,6 +337,51 @@ samplerExternalOES		{
 			     return IDENTIFIER;
 		}
 
+   /* keywords available with ARB_shader_image_load_store */
+image1D         KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE1D);
+image2D         KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE2D);
+image3D         KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE3D);
+image2DRect     KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE2DRECT);
+imageCube       KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGECUBE);
+imageBuffer     KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGEBUFFER);
+image1DArray    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE1DARRAY);
+image2DArray    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE2DARRAY);
+imageCubeArray  KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGECUBEARRAY);
+image2DMS       KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE2DMS);
+image2DMSArray  KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IMAGE2DMSARRAY);
+iimage1D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE1D);
+iimage2D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE2D);
+iimage3D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE3D);
+iimage2DRect    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE2DRECT);
+iimageCube      KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGECUBE);
+iimageBuffer    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGEBUFFER);
+iimage1DArray   KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE1DARRAY);
+iimage2DArray   KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE2DARRAY);
+iimageCubeArray KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGECUBEARRAY);
+iimage2DMS      KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE2DMS);
+iimage2DMSArray KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, IIMAGE2DMSARRAY);
+uimage1D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE1D);
+uimage2D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE2D);
+uimage3D        KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE3D);
+uimage2DRect    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE2DRECT);
+uimageCube      KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGECUBE);
+uimageBuffer    KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGEBUFFER);
+uimage1DArray   KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE1DARRAY);
+uimage2DArray   KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE2DARRAY);
+uimageCubeArray KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGECUBEARRAY);
+uimage2DMS      KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE2DMS);
+uimage2DMSArray KEYWORD_WITH_ALT(130, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, UIMAGE2DMSARRAY);
+image1DShadow           KEYWORD(130, 300, 0, 0, IMAGE1DSHADOW);
+image2DShadow           KEYWORD(130, 300, 0, 0, IMAGE2DSHADOW);
+image1DArrayShadow      KEYWORD(130, 300, 0, 0, IMAGE1DARRAYSHADOW);
+image2DArrayShadow      KEYWORD(130, 300, 0, 0, IMAGE2DARRAYSHADOW);
+
+coherent	KEYWORD_WITH_ALT(420, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, COHERENT);
+volatile	KEYWORD_WITH_ALT(110, 100, 420, 0, yyextra->ARB_shader_image_load_store_enable, VOLATILE);
+restrict	KEYWORD_WITH_ALT(420, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, RESTRICT);
+readonly	KEYWORD_WITH_ALT(420, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, READONLY);
+writeonly	KEYWORD_WITH_ALT(420, 300, 420, 0, yyextra->ARB_shader_image_load_store_enable, WRITEONLY);
+
 atomic_uint     KEYWORD_WITH_ALT(420, 300, 420, 0, yyextra->ARB_shader_atomic_counters_enable, ATOMIC_UINT);
 
 struct		return STRUCT;
@@ -347,6 +392,7 @@ layout		{
 		      || yyextra->AMD_conservative_depth_enable
 		      || yyextra->ARB_conservative_depth_enable
 		      || yyextra->ARB_explicit_attrib_location_enable
+                      || yyextra->has_separate_shader_objects()
 		      || yyextra->ARB_uniform_buffer_object_enable
 		      || yyextra->ARB_fragment_coord_conventions_enable
                       || yyextra->ARB_shading_language_420pack_enable
@@ -436,7 +482,6 @@ switch		KEYWORD(110, 100, 130, 300, SWITCH);
 default		KEYWORD(110, 100, 130, 300, DEFAULT);
 inline		KEYWORD(110, 100, 0, 0, INLINE_TOK);
 noinline	KEYWORD(110, 100, 0, 0, NOINLINE);
-volatile	KEYWORD(110, 100, 0, 0, VOLATILE);
 public		KEYWORD(110, 100, 0, 0, PUBLIC_TOK);
 static		KEYWORD(110, 100, 0, 0, STATIC);
 extern		KEYWORD(110, 100, 0, 0, EXTERN);
@@ -481,31 +526,6 @@ active		KEYWORD(130, 300, 0, 0, ACTIVE);
 superp		KEYWORD(130, 100, 0, 0, SUPERP);
 samplerBuffer	KEYWORD(130, 300, 140, 0, SAMPLERBUFFER);
 filter		KEYWORD(130, 300, 0, 0, FILTER);
-image1D		KEYWORD(130, 300, 0, 0, IMAGE1D);
-image2D		KEYWORD(130, 300, 0, 0, IMAGE2D);
-image3D		KEYWORD(130, 300, 0, 0, IMAGE3D);
-imageCube	KEYWORD(130, 300, 0, 0, IMAGECUBE);
-iimage1D	KEYWORD(130, 300, 0, 0, IIMAGE1D);
-iimage2D	KEYWORD(130, 300, 0, 0, IIMAGE2D);
-iimage3D	KEYWORD(130, 300, 0, 0, IIMAGE3D);
-iimageCube	KEYWORD(130, 300, 0, 0, IIMAGECUBE);
-uimage1D	KEYWORD(130, 300, 0, 0, UIMAGE1D);
-uimage2D	KEYWORD(130, 300, 0, 0, UIMAGE2D);
-uimage3D	KEYWORD(130, 300, 0, 0, UIMAGE3D);
-uimageCube	KEYWORD(130, 300, 0, 0, UIMAGECUBE);
-image1DArray	KEYWORD(130, 300, 0, 0, IMAGE1DARRAY);
-image2DArray	KEYWORD(130, 300, 0, 0, IMAGE2DARRAY);
-iimage1DArray	KEYWORD(130, 300, 0, 0, IIMAGE1DARRAY);
-iimage2DArray	KEYWORD(130, 300, 0, 0, IIMAGE2DARRAY);
-uimage1DArray	KEYWORD(130, 300, 0, 0, UIMAGE1DARRAY);
-uimage2DArray	KEYWORD(130, 300, 0, 0, UIMAGE2DARRAY);
-image1DShadow	KEYWORD(130, 300, 0, 0, IMAGE1DSHADOW);
-image2DShadow	KEYWORD(130, 300, 0, 0, IMAGE2DSHADOW);
-image1DArrayShadow KEYWORD(130, 300, 0, 0, IMAGE1DARRAYSHADOW);
-image2DArrayShadow KEYWORD(130, 300, 0, 0, IMAGE2DARRAYSHADOW);
-imageBuffer	KEYWORD(130, 300, 0, 0, IMAGEBUFFER);
-iimageBuffer	KEYWORD(130, 300, 0, 0, IIMAGEBUFFER);
-uimageBuffer	KEYWORD(130, 300, 0, 0, UIMAGEBUFFER);
 row_major	KEYWORD_WITH_ALT(130, 0, 140, 0, yyextra->ARB_uniform_buffer_object_enable && !yyextra->es_shader, ROW_MAJOR);
 
     /* Additional reserved words in GLSL 1.40 */
@@ -515,10 +535,6 @@ isamplerBuffer	KEYWORD(140, 300, 140, 0, ISAMPLERBUFFER);
 usamplerBuffer	KEYWORD(140, 300, 140, 0, USAMPLERBUFFER);
 
     /* Additional reserved words in GLSL ES 3.00 */
-coherent	KEYWORD(0, 300, 0, 0, COHERENT);
-restrict	KEYWORD(0, 300, 0, 0, RESTRICT);
-readonly	KEYWORD(0, 300, 0, 0, READONLY);
-writeonly	KEYWORD(0, 300, 0, 0, WRITEONLY);
 resource	KEYWORD(0, 300, 0, 0, RESOURCE);
 patch		KEYWORD(0, 300, 0, 0, PATCH);
 sample		KEYWORD_WITH_ALT(400, 300, 400, 0, yyextra->ARB_gpu_shader5_enable, SAMPLE);
diff --git a/mesalib/src/glsl/glsl_parser.yy b/mesalib/src/glsl/glsl_parser.yy
index c4d01a8cd..8751524c2 100644
--- a/mesalib/src/glsl/glsl_parser.yy
+++ b/mesalib/src/glsl/glsl_parser.yy
@@ -146,6 +146,14 @@ static bool match_layout_qualifier(const char *s1, const char *s2,
 %token SAMPLER2DMS ISAMPLER2DMS USAMPLER2DMS
 %token SAMPLER2DMSARRAY ISAMPLER2DMSARRAY USAMPLER2DMSARRAY
 %token SAMPLEREXTERNALOES
+%token IMAGE1D IMAGE2D IMAGE3D IMAGE2DRECT IMAGECUBE IMAGEBUFFER
+%token IMAGE1DARRAY IMAGE2DARRAY IMAGECUBEARRAY IMAGE2DMS IMAGE2DMSARRAY
+%token IIMAGE1D IIMAGE2D IIMAGE3D IIMAGE2DRECT IIMAGECUBE IIMAGEBUFFER
+%token IIMAGE1DARRAY IIMAGE2DARRAY IIMAGECUBEARRAY IIMAGE2DMS IIMAGE2DMSARRAY
+%token UIMAGE1D UIMAGE2D UIMAGE3D UIMAGE2DRECT UIMAGECUBE UIMAGEBUFFER
+%token UIMAGE1DARRAY UIMAGE2DARRAY UIMAGECUBEARRAY UIMAGE2DMS UIMAGE2DMSARRAY
+%token IMAGE1DSHADOW IMAGE2DSHADOW IMAGE1DARRAYSHADOW IMAGE2DARRAYSHADOW
+%token COHERENT VOLATILE RESTRICT READONLY WRITEONLY
 %token ATOMIC_UINT
 %token STRUCT VOID_TOK WHILE
 %token <identifier> IDENTIFIER TYPE_IDENTIFIER NEW_IDENTIFIER
@@ -171,23 +179,17 @@ static bool match_layout_qualifier(const char *s1, const char *s2,
    /* Reserved words that are not actually used in the grammar.
     */
 %token ASM CLASS UNION ENUM TYPEDEF TEMPLATE THIS PACKED_TOK GOTO
-%token INLINE_TOK NOINLINE VOLATILE PUBLIC_TOK STATIC EXTERN EXTERNAL
+%token INLINE_TOK NOINLINE PUBLIC_TOK STATIC EXTERN EXTERNAL
 %token LONG_TOK SHORT_TOK DOUBLE_TOK HALF FIXED_TOK UNSIGNED INPUT_TOK OUPTUT
 %token HVEC2 HVEC3 HVEC4 DVEC2 DVEC3 DVEC4 FVEC2 FVEC3 FVEC4
 %token SAMPLER3DRECT
 %token SIZEOF CAST NAMESPACE USING
-%token COHERENT RESTRICT READONLY WRITEONLY RESOURCE PATCH SAMPLE
+%token RESOURCE PATCH SAMPLE
 %token SUBROUTINE
 
 %token ERROR_TOK
 
-%token COMMON PARTITION ACTIVE FILTER
-%token  IMAGE1D  IMAGE2D  IMAGE3D  IMAGECUBE  IMAGE1DARRAY  IMAGE2DARRAY
-%token IIMAGE1D IIMAGE2D IIMAGE3D IIMAGECUBE IIMAGE1DARRAY IIMAGE2DARRAY
-%token UIMAGE1D UIMAGE2D UIMAGE3D UIMAGECUBE UIMAGE1DARRAY UIMAGE2DARRAY
-%token IMAGE1DSHADOW IMAGE2DSHADOW IMAGEBUFFER IIMAGEBUFFER UIMAGEBUFFER
-%token IMAGE1DARRAYSHADOW IMAGE2DARRAYSHADOW
-%token ROW_MAJOR
+%token COMMON PARTITION ACTIVE FILTER ROW_MAJOR
 
 %type <identifier> variable_identifier
 %type <node> statement
@@ -1230,6 +1232,72 @@ layout_qualifier_id:
          }
       }
 
+      /* Layout qualifiers for ARB_shader_image_load_store. */
+      if (state->ARB_shader_image_load_store_enable ||
+          state->is_version(420, 0)) {
+         if (!$$.flags.i) {
+            static const struct {
+               const char *name;
+               GLenum format;
+               glsl_base_type base_type;
+            } map[] = {
+               { "rgba32f", GL_RGBA32F, GLSL_TYPE_FLOAT },
+               { "rgba16f", GL_RGBA16F, GLSL_TYPE_FLOAT },
+               { "rg32f", GL_RG32F, GLSL_TYPE_FLOAT },
+               { "rg16f", GL_RG16F, GLSL_TYPE_FLOAT },
+               { "r11f_g11f_b10f", GL_R11F_G11F_B10F, GLSL_TYPE_FLOAT },
+               { "r32f", GL_R32F, GLSL_TYPE_FLOAT },
+               { "r16f", GL_R16F, GLSL_TYPE_FLOAT },
+               { "rgba32ui", GL_RGBA32UI, GLSL_TYPE_UINT },
+               { "rgba16ui", GL_RGBA16UI, GLSL_TYPE_UINT },
+               { "rgb10_a2ui", GL_RGB10_A2UI, GLSL_TYPE_UINT },
+               { "rgba8ui", GL_RGBA8UI, GLSL_TYPE_UINT },
+               { "rg32ui", GL_RG32UI, GLSL_TYPE_UINT },
+               { "rg16ui", GL_RG16UI, GLSL_TYPE_UINT },
+               { "rg8ui", GL_RG8UI, GLSL_TYPE_UINT },
+               { "r32ui", GL_R32UI, GLSL_TYPE_UINT },
+               { "r16ui", GL_R16UI, GLSL_TYPE_UINT },
+               { "r8ui", GL_R8UI, GLSL_TYPE_UINT },
+               { "rgba32i", GL_RGBA32I, GLSL_TYPE_INT },
+               { "rgba16i", GL_RGBA16I, GLSL_TYPE_INT },
+               { "rgba8i", GL_RGBA8I, GLSL_TYPE_INT },
+               { "rg32i", GL_RG32I, GLSL_TYPE_INT },
+               { "rg16i", GL_RG16I, GLSL_TYPE_INT },
+               { "rg8i", GL_RG8I, GLSL_TYPE_INT },
+               { "r32i", GL_R32I, GLSL_TYPE_INT },
+               { "r16i", GL_R16I, GLSL_TYPE_INT },
+               { "r8i", GL_R8I, GLSL_TYPE_INT },
+               { "rgba16", GL_RGBA16, GLSL_TYPE_FLOAT },
+               { "rgb10_a2", GL_RGB10_A2, GLSL_TYPE_FLOAT },
+               { "rgba8", GL_RGBA8, GLSL_TYPE_FLOAT },
+               { "rg16", GL_RG16, GLSL_TYPE_FLOAT },
+               { "rg8", GL_RG8, GLSL_TYPE_FLOAT },
+               { "r16", GL_R16, GLSL_TYPE_FLOAT },
+               { "r8", GL_R8, GLSL_TYPE_FLOAT },
+               { "rgba16_snorm", GL_RGBA16_SNORM, GLSL_TYPE_FLOAT },
+               { "rgba8_snorm", GL_RGBA8_SNORM, GLSL_TYPE_FLOAT },
+               { "rg16_snorm", GL_RG16_SNORM, GLSL_TYPE_FLOAT },
+               { "rg8_snorm", GL_RG8_SNORM, GLSL_TYPE_FLOAT },
+               { "r16_snorm", GL_R16_SNORM, GLSL_TYPE_FLOAT },
+               { "r8_snorm", GL_R8_SNORM, GLSL_TYPE_FLOAT }
+            };
+
+            for (unsigned i = 0; i < Elements(map); i++) {
+               if (match_layout_qualifier($1, map[i].name, state) == 0) {
+                  $$.flags.q.explicit_image_format = 1;
+                  $$.image_format = map[i].format;
+                  $$.image_base_type = map[i].base_type;
+                  break;
+               }
+            }
+         }
+
+         if (!$$.flags.i &&
+             match_layout_qualifier($1, "early_fragment_tests", state) == 0) {
+            $$.flags.q.early_fragment_tests = 1;
+         }
+      }
+
       if (!$$.flags.i) {
          _mesa_glsl_error(& @1, state, "unrecognized layout identifier "
                           "`%s'", $1);
@@ -1320,6 +1388,29 @@ layout_qualifier_id:
          }
       }
 
+      if (match_layout_qualifier("invocations", $1, state) == 0) {
+         $$.flags.q.invocations = 1;
+
+         if ($3 <= 0) {
+            _mesa_glsl_error(& @3, state,
+                             "invalid invocations %d specified", $3);
+            YYERROR;
+         } else if ($3 > MAX_GEOMETRY_SHADER_INVOCATIONS) {
+            _mesa_glsl_error(& @3, state,
+                             "invocations (%d) exceeds "
+                             "GL_MAX_GEOMETRY_SHADER_INVOCATIONS", $3);
+            YYERROR;
+         } else {
+            $$.invocations = $3;
+            if (!state->is_version(400, 0) &&
+                !state->ARB_gpu_shader5_enable) {
+               _mesa_glsl_error(& @3, state,
+                                "GL_ARB_gpu_shader5 invocations "
+                                "qualifier specified", $3);
+            }
+         }
+      }
+
       /* If the identifier didn't match any known layout identifiers,
        * emit an error.
        */
@@ -1574,6 +1665,32 @@ storage_qualifier:
       memset(& $$, 0, sizeof($$));
       $$.flags.q.uniform = 1;
    }
+   | COHERENT
+   {
+      memset(& $$, 0, sizeof($$));
+      $$.flags.q.coherent = 1;
+   }
+   | VOLATILE
+   {
+      memset(& $$, 0, sizeof($$));
+      $$.flags.q._volatile = 1;
+   }
+   | RESTRICT
+   {
+      STATIC_ASSERT(sizeof($$.flags.q) <= sizeof($$.flags.i));
+      memset(& $$, 0, sizeof($$));
+      $$.flags.q.restrict_flag = 1;
+   }
+   | READONLY
+   {
+      memset(& $$, 0, sizeof($$));
+      $$.flags.q.read_only = 1;
+   }
+   | WRITEONLY
+   {
+      memset(& $$, 0, sizeof($$));
+      $$.flags.q.write_only = 1;
+   }
    ;
 
 array_specifier:
@@ -1713,6 +1830,39 @@ basic_type_specifier_nonarray:
    | SAMPLER2DMSARRAY       { $$ = "sampler2DMSArray"; }
    | ISAMPLER2DMSARRAY      { $$ = "isampler2DMSArray"; }
    | USAMPLER2DMSARRAY      { $$ = "usampler2DMSArray"; }
+   | IMAGE1D                { $$ = "image1D"; }
+   | IMAGE2D                { $$ = "image2D"; }
+   | IMAGE3D                { $$ = "image3D"; }
+   | IMAGE2DRECT            { $$ = "image2DRect"; }
+   | IMAGECUBE              { $$ = "imageCube"; }
+   | IMAGEBUFFER            { $$ = "imageBuffer"; }
+   | IMAGE1DARRAY           { $$ = "image1DArray"; }
+   | IMAGE2DARRAY           { $$ = "image2DArray"; }
+   | IMAGECUBEARRAY         { $$ = "imageCubeArray"; }
+   | IMAGE2DMS              { $$ = "image2DMS"; }
+   | IMAGE2DMSARRAY         { $$ = "image2DMSArray"; }
+   | IIMAGE1D               { $$ = "iimage1D"; }
+   | IIMAGE2D               { $$ = "iimage2D"; }
+   | IIMAGE3D               { $$ = "iimage3D"; }
+   | IIMAGE2DRECT           { $$ = "iimage2DRect"; }
+   | IIMAGECUBE             { $$ = "iimageCube"; }
+   | IIMAGEBUFFER           { $$ = "iimageBuffer"; }
+   | IIMAGE1DARRAY          { $$ = "iimage1DArray"; }
+   | IIMAGE2DARRAY          { $$ = "iimage2DArray"; }
+   | IIMAGECUBEARRAY        { $$ = "iimageCubeArray"; }
+   | IIMAGE2DMS             { $$ = "iimage2DMS"; }
+   | IIMAGE2DMSARRAY        { $$ = "iimage2DMSArray"; }
+   | UIMAGE1D               { $$ = "uimage1D"; }
+   | UIMAGE2D               { $$ = "uimage2D"; }
+   | UIMAGE3D               { $$ = "uimage3D"; }
+   | UIMAGE2DRECT           { $$ = "uimage2DRect"; }
+   | UIMAGECUBE             { $$ = "uimageCube"; }
+   | UIMAGEBUFFER           { $$ = "uimageBuffer"; }
+   | UIMAGE1DARRAY          { $$ = "uimage1DArray"; }
+   | UIMAGE2DARRAY          { $$ = "uimage2DArray"; }
+   | UIMAGECUBEARRAY        { $$ = "uimageCubeArray"; }
+   | UIMAGE2DMS             { $$ = "uimage2DMS"; }
+   | UIMAGE2DMSARRAY        { $$ = "uimage2DMSArray"; }
    | ATOMIC_UINT            { $$ = "atomic_uint"; }
    ;
 
@@ -2229,7 +2379,7 @@ basic_interface_block:
                                "an instance name are not allowed");
       }
 
-      unsigned interface_type_mask;
+      uint64_t interface_type_mask;
       struct ast_type_qualifier temp_type_qualifier;
 
       /* Get a bitmask containing only the in/out/uniform flags, allowing us
@@ -2245,7 +2395,7 @@ basic_interface_block:
        * production rule guarantees that only one bit will be set (and
        * it will be in/out/uniform).
        */
-       unsigned block_interface_qualifier = $1.flags.i;
+      uint64_t block_interface_qualifier = $1.flags.i;
 
       block->layout.flags.i |= block_interface_qualifier;
 
@@ -2361,55 +2511,9 @@ layout_defaults:
 
    | layout_qualifier IN_TOK ';'
    {
-      void *ctx = state;
       $$ = NULL;
-      switch (state->stage) {
-      case MESA_SHADER_GEOMETRY: {
-         if (!$1.flags.q.prim_type) {
-            _mesa_glsl_error(& @1, state,
-                             "input layout qualifiers must specify a primitive"
-                             " type");
-         } else {
-            /* Make sure this is a valid input primitive type. */
-            switch ($1.prim_type) {
-            case GL_POINTS:
-            case GL_LINES:
-            case GL_LINES_ADJACENCY:
-            case GL_TRIANGLES:
-            case GL_TRIANGLES_ADJACENCY:
-               $$ = new(ctx) ast_gs_input_layout(@1, $1.prim_type);
-               break;
-            default:
-               _mesa_glsl_error(&@1, state,
-                                "invalid geometry shader input primitive type");
-               break;
-            }
-         }
-      }
-         break;
-      case MESA_SHADER_COMPUTE: {
-         if ($1.flags.q.local_size == 0) {
-            _mesa_glsl_error(& @1, state,
-                             "input layout qualifiers must specify a local "
-                             "size");
-         } else {
-            /* Infer a local_size of 1 for every unspecified dimension */
-            unsigned local_size[3];
-            for (int i = 0; i < 3; i++) {
-               if ($1.flags.q.local_size & (1 << i))
-                  local_size[i] = $1.local_size[i];
-               else
-                  local_size[i] = 1;
-            }
-            $$ = new(ctx) ast_cs_input_layout(@1, local_size);
-         }
-      }
-         break;
-      default:
-         _mesa_glsl_error(& @1, state,
-                          "input layout qualifiers only valid in "
-                          "geometry and compute shaders");
-         break;
+      if (!state->in_qualifier->merge_in_qualifier(& @1, state, $1, $$)) {
+         YYERROR;
       }
    }
 
diff --git a/mesalib/src/glsl/glsl_parser_extras.cpp b/mesalib/src/glsl/glsl_parser_extras.cpp
index b822d2292..d7f5202fe 100644
--- a/mesalib/src/glsl/glsl_parser_extras.cpp
+++ b/mesalib/src/glsl/glsl_parser_extras.cpp
@@ -130,6 +130,14 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    for (unsigned i = 0; i < Elements(this->Const.MaxComputeWorkGroupSize); i++)
       this->Const.MaxComputeWorkGroupSize[i] = ctx->Const.MaxComputeWorkGroupSize[i];
 
+   this->Const.MaxImageUnits = ctx->Const.MaxImageUnits;
+   this->Const.MaxCombinedImageUnitsAndFragmentOutputs = ctx->Const.MaxCombinedImageUnitsAndFragmentOutputs;
+   this->Const.MaxImageSamples = ctx->Const.MaxImageSamples;
+   this->Const.MaxVertexImageUniforms = ctx->Const.Program[MESA_SHADER_VERTEX].MaxImageUniforms;
+   this->Const.MaxGeometryImageUniforms = ctx->Const.Program[MESA_SHADER_GEOMETRY].MaxImageUniforms;
+   this->Const.MaxFragmentImageUniforms = ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxImageUniforms;
+   this->Const.MaxCombinedImageUniforms = ctx->Const.MaxCombinedImageUniforms;
+
    this->current_function = NULL;
    this->toplevel_ir = NULL;
    this->found_return = false;
@@ -194,9 +202,10 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    this->default_uniform_qualifier->flags.q.column_major = 1;
 
    this->gs_input_prim_type_specified = false;
-   this->gs_input_prim_type = GL_POINTS;
    this->gs_input_size = 0;
+   this->in_qualifier = new(this) ast_type_qualifier();
    this->out_qualifier = new(this) ast_type_qualifier();
+   this->early_fragment_tests = false;
    memset(this->atomic_counter_offsets, 0,
           sizeof(this->atomic_counter_offsets));
 }
@@ -503,6 +512,7 @@ static const _mesa_glsl_extension _mesa_glsl_supported_extensions[] = {
    EXT(ARB_fragment_coord_conventions, true,  false,     ARB_fragment_coord_conventions),
    EXT(ARB_texture_rectangle,          true,  false,     dummy_true),
    EXT(EXT_texture_array,              true,  false,     EXT_texture_array),
+   EXT(ARB_separate_shader_objects,    true,  false,     ARB_separate_shader_objects),
    EXT(ARB_shader_texture_lod,         true,  false,     ARB_shader_texture_lod),
    EXT(ARB_shader_stencil_export,      true,  false,     ARB_shader_stencil_export),
    EXT(AMD_conservative_depth,         true,  false,     ARB_conservative_depth),
@@ -527,6 +537,7 @@ static const _mesa_glsl_extension _mesa_glsl_supported_extensions[] = {
    EXT(AMD_shader_trinary_minmax,      true,  false,     dummy_true),
    EXT(ARB_viewport_array,             true,  false,     ARB_viewport_array),
    EXT(ARB_compute_shader,             true,  false,     ARB_compute_shader),
+   EXT(ARB_shader_image_load_store,    true,  false,     ARB_shader_image_load_store),
 };
 
 #undef EXT
@@ -1338,7 +1349,7 @@ set_shader_inout_layout(struct gl_shader *shader,
 {
    if (shader->Stage != MESA_SHADER_GEOMETRY) {
       /* Should have been prevented by the parser. */
-      assert(!state->gs_input_prim_type_specified);
+      assert(!state->in_qualifier->flags.i);
       assert(!state->out_qualifier->flags.i);
    }
 
@@ -1354,7 +1365,7 @@ set_shader_inout_layout(struct gl_shader *shader,
          shader->Geom.VerticesOut = state->out_qualifier->max_vertices;
 
       if (state->gs_input_prim_type_specified) {
-         shader->Geom.InputType = state->gs_input_prim_type;
+         shader->Geom.InputType = state->in_qualifier->prim_type;
       } else {
          shader->Geom.InputType = PRIM_UNKNOWN;
       }
@@ -1364,6 +1375,10 @@ set_shader_inout_layout(struct gl_shader *shader,
       } else {
          shader->Geom.OutputType = PRIM_UNKNOWN;
       }
+
+      shader->Geom.Invocations = 0;
+      if (state->in_qualifier->flags.q.invocations)
+         shader->Geom.Invocations = state->in_qualifier->invocations;
       break;
 
    case MESA_SHADER_COMPUTE:
@@ -1419,7 +1434,7 @@ _mesa_glsl_compile_shader(struct gl_context *ctx, struct gl_shader *shader,
 
       /* Print out the unoptimized IR. */
       if (dump_hir) {
-         _mesa_print_ir(shader->ir, state);
+         _mesa_print_ir(stdout, shader->ir, state);
       }
    }
 
diff --git a/mesalib/src/glsl/glsl_parser_extras.h b/mesalib/src/glsl/glsl_parser_extras.h
index 7b0fbbfa1..2e22c8c91 100644
--- a/mesalib/src/glsl/glsl_parser_extras.h
+++ b/mesalib/src/glsl/glsl_parser_extras.h
@@ -139,6 +139,23 @@ struct _mesa_glsl_parse_state {
       return true;
    }
 
+   bool check_separate_shader_objects_allowed(YYLTYPE *locp,
+                                              const ir_variable *var)
+   {
+      if (!this->has_separate_shader_objects()) {
+         const char *const requirement = this->es_shader
+            ? "GL_EXT_separate_shader_objects (not supported by this "
+              "implementation)"
+            : "GL_ARB_separate_shader_objects extension or GLSL 420";
+
+         _mesa_glsl_error(locp, this, "%s explicit location requires %s",
+                          mode_string(var), requirement);
+         return false;
+      }
+
+      return true;
+   }
+
    bool has_explicit_attrib_location() const
    {
       return ARB_explicit_attrib_location_enable || is_version(330, 300);
@@ -149,6 +166,11 @@ struct _mesa_glsl_parse_state {
       return ARB_uniform_buffer_object_enable || is_version(140, 300);
    }
 
+   bool has_separate_shader_objects() const
+   {
+      return ARB_separate_shader_objects_enable || is_version(410, 0);
+   }
+
    void process_version_directive(YYLTYPE *locp, int version,
                                   const char *ident);
 
@@ -194,11 +216,8 @@ struct _mesa_glsl_parse_state {
     */
    bool gs_input_prim_type_specified;
 
-   /**
-    * If gs_input_prim_type_specified is true, the primitive type that was
-    * specified.  Otherwise ignored.
-    */
-   GLenum gs_input_prim_type;
+   /** Input layout qualifiers from GLSL 1.50. (geometry shader controls)*/
+   struct ast_type_qualifier *in_qualifier;
 
    /**
     * True if a compute shader input local size was specified using a layout
@@ -273,6 +292,15 @@ struct _mesa_glsl_parse_state {
       /* ARB_compute_shader */
       unsigned MaxComputeWorkGroupCount[3];
       unsigned MaxComputeWorkGroupSize[3];
+
+      /* ARB_shader_image_load_store */
+      unsigned MaxImageUnits;
+      unsigned MaxCombinedImageUnitsAndFragmentOutputs;
+      unsigned MaxImageSamples;
+      unsigned MaxVertexImageUniforms;
+      unsigned MaxGeometryImageUniforms;
+      unsigned MaxFragmentImageUniforms;
+      unsigned MaxCombinedImageUniforms;
    } Const;
 
    /**
@@ -333,6 +361,8 @@ struct _mesa_glsl_parse_state {
    bool ARB_texture_gather_warn;
    bool EXT_texture_array_enable;
    bool EXT_texture_array_warn;
+   bool ARB_separate_shader_objects_enable;
+   bool ARB_separate_shader_objects_warn;
    bool ARB_shader_texture_lod_enable;
    bool ARB_shader_texture_lod_warn;
    bool ARB_shader_stencil_export_enable;
@@ -381,6 +411,8 @@ struct _mesa_glsl_parse_state {
    bool ARB_viewport_array_warn;
    bool ARB_compute_shader_enable;
    bool ARB_compute_shader_warn;
+   bool ARB_shader_image_load_store_enable;
+   bool ARB_shader_image_load_store_warn;
    /*@}*/
 
    /** Extensions supported by the OpenGL implementation. */
@@ -397,6 +429,8 @@ struct _mesa_glsl_parse_state {
     */
    unsigned gs_input_size;
 
+   bool early_fragment_tests;
+
    /** Atomic counter offsets by binding */
    unsigned atomic_counter_offsets[MAX_COMBINED_ATOMIC_BUFFERS];
 };
diff --git a/mesalib/src/glsl/glsl_types.cpp b/mesalib/src/glsl/glsl_types.cpp
index 1b0b3ef88..849a79af4 100644
--- a/mesalib/src/glsl/glsl_types.cpp
+++ b/mesalib/src/glsl/glsl_types.cpp
@@ -64,20 +64,26 @@ glsl_type::glsl_type(GLenum gl_type,
    memset(& fields, 0, sizeof(fields));
 }
 
-glsl_type::glsl_type(GLenum gl_type,
+glsl_type::glsl_type(GLenum gl_type, glsl_base_type base_type,
 		     enum glsl_sampler_dim dim, bool shadow, bool array,
 		     unsigned type, const char *name) :
    gl_type(gl_type),
-   base_type(GLSL_TYPE_SAMPLER),
+   base_type(base_type),
    sampler_dimensionality(dim), sampler_shadow(shadow),
    sampler_array(array), sampler_type(type), interface_packing(0),
-   vector_elements(0), matrix_columns(0),
    length(0)
 {
    init_ralloc_type_ctx();
    assert(name != NULL);
    this->name = ralloc_strdup(this->mem_ctx, name);
    memset(& fields, 0, sizeof(fields));
+
+   if (base_type == GLSL_TYPE_SAMPLER) {
+      /* Samplers take no storage whatsoever. */
+      matrix_columns = vector_elements = 0;
+   } else {
+      matrix_columns = vector_elements = 1;
+   }
 }
 
 glsl_type::glsl_type(const glsl_struct_field *fields, unsigned num_fields,
@@ -174,6 +180,7 @@ bool
 glsl_type::contains_opaque() const {
    switch (base_type) {
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
    case GLSL_TYPE_ATOMIC_UINT:
       return true;
    case GLSL_TYPE_ARRAY:
@@ -219,6 +226,21 @@ glsl_type::sampler_index() const
    }
 }
 
+bool
+glsl_type::contains_image() const
+{
+   if (this->is_array()) {
+      return this->fields.array->contains_image();
+   } else if (this->is_record()) {
+      for (unsigned int i = 0; i < this->length; i++) {
+	 if (this->fields.structure[i].type->contains_image())
+	    return true;
+      }
+      return false;
+   } else {
+      return this->is_image();
+   }
+}
 
 const glsl_type *glsl_type::get_base_type() const
 {
@@ -642,6 +664,9 @@ glsl_type::component_slots() const
    case GLSL_TYPE_ARRAY:
       return this->length * this->fields.array->component_slots();
 
+   case GLSL_TYPE_IMAGE:
+      return 1;
+
    case GLSL_TYPE_SAMPLER:
    case GLSL_TYPE_ATOMIC_UINT:
    case GLSL_TYPE_VOID:
@@ -932,6 +957,7 @@ glsl_type::count_attribute_slots() const
       return this->length * this->fields.array->count_attribute_slots();
 
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
    case GLSL_TYPE_ATOMIC_UINT:
    case GLSL_TYPE_VOID:
    case GLSL_TYPE_ERROR:
@@ -944,10 +970,8 @@ glsl_type::count_attribute_slots() const
 }
 
 int
-glsl_type::sampler_coordinate_components() const
+glsl_type::coordinate_components() const
 {
-   assert(is_sampler());
-
    int size;
 
    switch (sampler_dimensionality) {
diff --git a/mesalib/src/glsl/glsl_types.h b/mesalib/src/glsl/glsl_types.h
index f88758a9a..ae3829f30 100644
--- a/mesalib/src/glsl/glsl_types.h
+++ b/mesalib/src/glsl/glsl_types.h
@@ -53,6 +53,7 @@ enum glsl_base_type {
    GLSL_TYPE_FLOAT,
    GLSL_TYPE_BOOL,
    GLSL_TYPE_SAMPLER,
+   GLSL_TYPE_IMAGE,
    GLSL_TYPE_ATOMIC_UINT,
    GLSL_TYPE_STRUCT,
    GLSL_TYPE_INTERFACE,
@@ -89,8 +90,9 @@ struct glsl_type {
    unsigned sampler_dimensionality:3; /**< \see glsl_sampler_dim */
    unsigned sampler_shadow:1;
    unsigned sampler_array:1;
-   unsigned sampler_type:2;    /**< Type of data returned using this sampler.
-				* only \c GLSL_TYPE_FLOAT, \c GLSL_TYPE_INT,
+   unsigned sampler_type:2;    /**< Type of data returned using this
+				* sampler or image.  Only \c
+				* GLSL_TYPE_FLOAT, \c GLSL_TYPE_INT,
 				* and \c GLSL_TYPE_UINT are valid.
 				*/
    unsigned interface_packing:2;
@@ -402,6 +404,20 @@ struct glsl_type {
    gl_texture_index sampler_index() const;
 
    /**
+    * Query whether or not type is an image, or for struct and array
+    * types, contains an image.
+    */
+   bool contains_image() const;
+
+   /**
+    * Query whether or not a type is an image
+    */
+   bool is_image() const
+   {
+      return base_type == GLSL_TYPE_IMAGE;
+   }
+
+   /**
     * Query whether or not a type is an array
     */
    bool is_array() const
@@ -531,7 +547,8 @@ struct glsl_type {
    }
 
    /**
-    * Return the number of coordinate components needed for this sampler type.
+    * Return the number of coordinate components needed for this
+    * sampler or image type.
     *
     * This is based purely on the sampler's dimensionality.  For example, this
     * returns 1 for sampler1D, and 3 for sampler2DArray.
@@ -540,7 +557,7 @@ struct glsl_type {
     * a texturing built-in function, since those pack additional values (such
     * as the shadow comparitor or projector) into the coordinate type.
     */
-   int sampler_coordinate_components() const;
+   int coordinate_components() const;
 
    /**
     * Compare a record type against another record type.
@@ -564,8 +581,8 @@ private:
 	     glsl_base_type base_type, unsigned vector_elements,
 	     unsigned matrix_columns, const char *name);
 
-   /** Constructor for sampler types */
-   glsl_type(GLenum gl_type,
+   /** Constructor for sampler or image types */
+   glsl_type(GLenum gl_type, glsl_base_type base_type,
 	     enum glsl_sampler_dim dim, bool shadow, bool array,
 	     unsigned type, const char *name);
 
diff --git a/mesalib/src/glsl/ir.cpp b/mesalib/src/glsl/ir.cpp
index 77638ef26..85fbfa9c4 100644
--- a/mesalib/src/glsl/ir.cpp
+++ b/mesalib/src/glsl/ir.cpp
@@ -1311,13 +1311,13 @@ ir_dereference::is_lvalue() const
    if ((var == NULL) || var->data.read_only)
       return false;
 
-   /* From page 17 (page 23 of the PDF) of the GLSL 1.20 spec:
+   /* From section 4.1.7 of the GLSL 4.40 spec:
     *
-    *    "Samplers cannot be treated as l-values; hence cannot be used
-    *     as out or inout function parameters, nor can they be
-    *     assigned into."
+    *   "Opaque variables cannot be treated as l-values; hence cannot
+    *    be used as out or inout function parameters, nor can they be
+    *    assigned into."
     */
-   if (this->type->contains_sampler())
+   if (this->type->contains_opaque())
       return false;
 
    return true;
@@ -1554,6 +1554,11 @@ ir_variable::ir_variable(const struct glsl_type *type, const char *name,
    this->data.max_array_access = 0;
    this->data.atomic.buffer_index = 0;
    this->data.atomic.offset = 0;
+   this->data.image.read_only = false;
+   this->data.image.write_only = false;
+   this->data.image.coherent = false;
+   this->data.image._volatile = false;
+   this->data.image.restrict_flag = false;
 
    if (type != NULL) {
       if (type->base_type == GLSL_TYPE_SAMPLER)
@@ -1658,7 +1663,12 @@ ir_function_signature::qualifiers_match(exec_list *params)
 	  !modes_match(a->data.mode, b->data.mode) ||
 	  a->data.interpolation != b->data.interpolation ||
 	  a->data.centroid != b->data.centroid ||
-         a->data.sample != b->data.sample) {
+          a->data.sample != b->data.sample ||
+          a->data.image.read_only != b->data.image.read_only ||
+          a->data.image.write_only != b->data.image.write_only ||
+          a->data.image.coherent != b->data.image.coherent ||
+          a->data.image._volatile != b->data.image._volatile ||
+          a->data.image.restrict_flag != b->data.image.restrict_flag) {
 
 	 /* parameter a's qualifiers don't match */
 	 return a->name;
diff --git a/mesalib/src/glsl/ir.h b/mesalib/src/glsl/ir.h
index e266328b2..ed3f086ce 100644
--- a/mesalib/src/glsl/ir.h
+++ b/mesalib/src/glsl/ir.h
@@ -106,6 +106,7 @@ public:
 
    /** ir_print_visitor helper for debugging. */
    void print(void) const;
+   void fprint(FILE *f) const;
 
    virtual void accept(ir_visitor *) = 0;
    virtual ir_visitor_status accept(ir_hierarchical_visitor *) = 0;
@@ -691,6 +692,20 @@ public:
       } atomic;
 
       /**
+       * ARB_shader_image_load_store qualifiers.
+       */
+      struct {
+         bool read_only; /**< "readonly" qualifier. */
+         bool write_only; /**< "writeonly" qualifier. */
+         bool coherent;
+         bool _volatile;
+         bool restrict_flag;
+
+         /** Image internal format if specified explicitly, otherwise GL_NONE. */
+         GLenum format;
+      } image;
+
+      /**
        * Highest element accessed with a constant expression array index
        *
        * Not used for non-array variables.
@@ -1444,6 +1459,18 @@ public:
    }
 
    /**
+    * Return whether the expression operates on vectors horizontally.
+    */
+   bool is_horizontal() const
+   {
+      return operation == ir_binop_all_equal ||
+             operation == ir_binop_any_nequal ||
+             operation == ir_unop_any ||
+             operation == ir_binop_dot ||
+             operation == ir_quadop_vector;
+   }
+
+   /**
     * Return a string representing this expression's operator.
     */
    const char *operator_string();
@@ -2339,7 +2366,7 @@ mode_string(const ir_variable *var);
 extern "C" {
 #endif /* __cplusplus */
 
-extern void _mesa_print_ir(struct exec_list *instructions,
+extern void _mesa_print_ir(FILE *f, struct exec_list *instructions,
                            struct _mesa_glsl_parse_state *state);
 
 #ifdef __cplusplus
diff --git a/mesalib/src/glsl/ir_builder.h b/mesalib/src/glsl/ir_builder.h
index f00e6f3b3..f00e6f3b3 100755..100644
--- a/mesalib/src/glsl/ir_builder.h
+++ b/mesalib/src/glsl/ir_builder.h
diff --git a/mesalib/src/glsl/ir_clone.cpp b/mesalib/src/glsl/ir_clone.cpp
index cb732a51a..311c992be 100644
--- a/mesalib/src/glsl/ir_clone.cpp
+++ b/mesalib/src/glsl/ir_clone.cpp
@@ -373,6 +373,7 @@ ir_constant::clone(void *mem_ctx, struct hash_table *ht) const
    }
 
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
    case GLSL_TYPE_ATOMIC_UINT:
    case GLSL_TYPE_VOID:
    case GLSL_TYPE_ERROR:
diff --git a/mesalib/src/glsl/ir_optimization.h b/mesalib/src/glsl/ir_optimization.h
index 055d65547..5f4a2f405 100644
--- a/mesalib/src/glsl/ir_optimization.h
+++ b/mesalib/src/glsl/ir_optimization.h
@@ -36,9 +36,8 @@
 #define LOG_TO_LOG2        0x10
 #define MOD_TO_FRACT       0x20
 #define INT_DIV_TO_MUL_RCP 0x40
-#define LRP_TO_ARITH       0x80
-#define BITFIELD_INSERT_TO_BFM_BFI 0x100
-#define LDEXP_TO_ARITH     0x200
+#define BITFIELD_INSERT_TO_BFM_BFI 0x80
+#define LDEXP_TO_ARITH     0x100
 
 /**
  * \see class lower_packing_builtins_visitor
@@ -120,6 +119,7 @@ bool lower_vector_insert(exec_list *instructions, bool lower_nonconstant_index);
 void lower_named_interface_blocks(void *mem_ctx, gl_shader *shader);
 bool optimize_redundant_jumps(exec_list *instructions);
 bool optimize_split_arrays(exec_list *instructions, bool linked);
+bool lower_offset_arrays(exec_list *instructions);
 
 ir_rvalue *
 compare_index_block(exec_list *instructions, ir_variable *index,
diff --git a/mesalib/src/glsl/ir_print_visitor.cpp b/mesalib/src/glsl/ir_print_visitor.cpp
index 935782184..c7786ba9f 100644
--- a/mesalib/src/glsl/ir_print_visitor.cpp
+++ b/mesalib/src/glsl/ir_print_visitor.cpp
@@ -27,52 +27,59 @@
 #include "main/macros.h"
 #include "program/hash_table.h"
 
-static void print_type(const glsl_type *t);
+static void print_type(FILE *f, const glsl_type *t);
 
 void
 ir_instruction::print(void) const
 {
+   this->fprint(stdout);
+}
+
+void
+ir_instruction::fprint(FILE *f) const
+{
    ir_instruction *deconsted = const_cast<ir_instruction *>(this);
 
-   ir_print_visitor v;
+   ir_print_visitor v(f);
    deconsted->accept(&v);
 }
 
 extern "C" {
 void
-_mesa_print_ir(exec_list *instructions,
+_mesa_print_ir(FILE *f, exec_list *instructions,
 	       struct _mesa_glsl_parse_state *state)
 {
    if (state) {
       for (unsigned i = 0; i < state->num_user_structures; i++) {
 	 const glsl_type *const s = state->user_structures[i];
 
-	 printf("(structure (%s) (%s@%p) (%u) (\n",
-		s->name, s->name, (void *) s, s->length);
+	 fprintf(f, "(structure (%s) (%s@%p) (%u) (\n",
+                 s->name, s->name, (void *) s, s->length);
 
 	 for (unsigned j = 0; j < s->length; j++) {
-	    printf("\t((");
-	    print_type(s->fields.structure[j].type);
-	    printf(")(%s))\n", s->fields.structure[j].name);
+	    fprintf(f, "\t((");
+	    print_type(f, s->fields.structure[j].type);
+	    fprintf(f, ")(%s))\n", s->fields.structure[j].name);
 	 }
 
-	 printf(")\n");
+	 fprintf(f, ")\n");
       }
    }
 
-   printf("(\n");
+   fprintf(f, "(\n");
    foreach_list(n, instructions) {
       ir_instruction *ir = (ir_instruction *) n;
-      ir->print();
+      ir->fprint(f);
       if (ir->ir_type != ir_type_function)
-	 printf("\n");
+	 fprintf(f, "\n");
    }
-   printf("\n)");
+   fprintf(f, "\n)");
 }
 
 } /* extern "C" */
 
-ir_print_visitor::ir_print_visitor()
+ir_print_visitor::ir_print_visitor(FILE *f)
+   : f(f)
 {
    indentation = 0;
    printable_names =
@@ -91,7 +98,7 @@ ir_print_visitor::~ir_print_visitor()
 void ir_print_visitor::indent(void)
 {
    for (int i = 0; i < indentation; i++)
-      printf("  ");
+      fprintf(f, "  ");
 }
 
 const char *
@@ -125,28 +132,28 @@ ir_print_visitor::unique_name(ir_variable *var)
 }
 
 static void
-print_type(const glsl_type *t)
+print_type(FILE *f, const glsl_type *t)
 {
    if (t->base_type == GLSL_TYPE_ARRAY) {
-      printf("(array ");
-      print_type(t->fields.array);
-      printf(" %u)", t->length);
+      fprintf(f, "(array ");
+      print_type(f, t->fields.array);
+      fprintf(f, " %u)", t->length);
    } else if ((t->base_type == GLSL_TYPE_STRUCT)
 	      && (strncmp("gl_", t->name, 3) != 0)) {
-      printf("%s@%p", t->name, (void *) t);
+      fprintf(f, "%s@%p", t->name, (void *) t);
    } else {
-      printf("%s", t->name);
+      fprintf(f, "%s", t->name);
    }
 }
 
 void ir_print_visitor::visit(ir_rvalue *ir)
 {
-   printf("error");
+   fprintf(f, "error");
 }
 
 void ir_print_visitor::visit(ir_variable *ir)
 {
-   printf("(declare ");
+   fprintf(f, "(declare ");
 
    const char *const cent = (ir->data.centroid) ? "centroid " : "";
    const char *const samp = (ir->data.sample) ? "sample " : "";
@@ -158,25 +165,25 @@ void ir_print_visitor::visit(ir_variable *ir)
    const char *const interp[] = { "", "smooth", "flat", "noperspective" };
    STATIC_ASSERT(ARRAY_SIZE(interp) == INTERP_QUALIFIER_COUNT);
 
-   printf("(%s%s%s%s%s) ",
+   fprintf(f, "(%s%s%s%s%s) ",
 	  cent, samp, inv, mode[ir->data.mode], interp[ir->data.interpolation]);
 
-   print_type(ir->type);
-   printf(" %s)", unique_name(ir));
+   print_type(f, ir->type);
+   fprintf(f, " %s)", unique_name(ir));
 }
 
 
 void ir_print_visitor::visit(ir_function_signature *ir)
 {
    _mesa_symbol_table_push_scope(symbols);
-   printf("(signature ");
+   fprintf(f, "(signature ");
    indentation++;
 
-   print_type(ir->return_type);
-   printf("\n");
+   print_type(f, ir->return_type);
+   fprintf(f, "\n");
    indent();
 
-   printf("(parameters\n");
+   fprintf(f, "(parameters\n");
    indentation++;
 
    foreach_list(n, &ir->parameters) {
@@ -184,16 +191,16 @@ void ir_print_visitor::visit(ir_function_signature *ir)
 
       indent();
       inst->accept(this);
-      printf("\n");
+      fprintf(f, "\n");
    }
    indentation--;
 
    indent();
-   printf(")\n");
+   fprintf(f, ")\n");
 
    indent();
 
-   printf("(\n");
+   fprintf(f, "(\n");
    indentation++;
 
    foreach_list(n, &ir->body) {
@@ -201,11 +208,11 @@ void ir_print_visitor::visit(ir_function_signature *ir)
 
       indent();
       inst->accept(this);
-      printf("\n");
+      fprintf(f, "\n");
    }
    indentation--;
    indent();
-   printf("))\n");
+   fprintf(f, "))\n");
    indentation--;
    _mesa_symbol_table_pop_scope(symbols);
 }
@@ -213,58 +220,58 @@ void ir_print_visitor::visit(ir_function_signature *ir)
 
 void ir_print_visitor::visit(ir_function *ir)
 {
-   printf("(function %s\n", ir->name);
+   fprintf(f, "(function %s\n", ir->name);
    indentation++;
    foreach_list(n, &ir->signatures) {
       ir_function_signature *const sig = (ir_function_signature *) n;
       indent();
       sig->accept(this);
-      printf("\n");
+      fprintf(f, "\n");
    }
    indentation--;
    indent();
-   printf(")\n\n");
+   fprintf(f, ")\n\n");
 }
 
 
 void ir_print_visitor::visit(ir_expression *ir)
 {
-   printf("(expression ");
+   fprintf(f, "(expression ");
 
-   print_type(ir->type);
+   print_type(f, ir->type);
 
-   printf(" %s ", ir->operator_string());
+   fprintf(f, " %s ", ir->operator_string());
 
    for (unsigned i = 0; i < ir->get_num_operands(); i++) {
       ir->operands[i]->accept(this);
    }
 
-   printf(") ");
+   fprintf(f, ") ");
 }
 
 
 void ir_print_visitor::visit(ir_texture *ir)
 {
-   printf("(%s ", ir->opcode_string());
+   fprintf(f, "(%s ", ir->opcode_string());
 
-   print_type(ir->type);
-   printf(" ");
+   print_type(f, ir->type);
+   fprintf(f, " ");
 
    ir->sampler->accept(this);
-   printf(" ");
+   fprintf(f, " ");
 
    if (ir->op != ir_txs && ir->op != ir_query_levels) {
       ir->coordinate->accept(this);
 
-      printf(" ");
+      fprintf(f, " ");
 
       if (ir->offset != NULL) {
 	 ir->offset->accept(this);
       } else {
-	 printf("0");
+	 fprintf(f, "0");
       }
 
-      printf(" ");
+      fprintf(f, " ");
    }
 
    if (ir->op != ir_txf && ir->op != ir_txf_ms &&
@@ -273,17 +280,17 @@ void ir_print_visitor::visit(ir_texture *ir)
       if (ir->projector)
 	 ir->projector->accept(this);
       else
-	 printf("1");
+	 fprintf(f, "1");
 
       if (ir->shadow_comparitor) {
-	 printf(" ");
+	 fprintf(f, " ");
 	 ir->shadow_comparitor->accept(this);
       } else {
-	 printf(" ()");
+	 fprintf(f, " ()");
       }
    }
 
-   printf(" ");
+   fprintf(f, " ");
    switch (ir->op)
    {
    case ir_tex:
@@ -302,17 +309,17 @@ void ir_print_visitor::visit(ir_texture *ir)
       ir->lod_info.sample_index->accept(this);
       break;
    case ir_txd:
-      printf("(");
+      fprintf(f, "(");
       ir->lod_info.grad.dPdx->accept(this);
-      printf(" ");
+      fprintf(f, " ");
       ir->lod_info.grad.dPdy->accept(this);
-      printf(")");
+      fprintf(f, ")");
       break;
    case ir_tg4:
       ir->lod_info.component->accept(this);
       break;
    };
-   printf(")");
+   fprintf(f, ")");
 }
 
 
@@ -325,43 +332,43 @@ void ir_print_visitor::visit(ir_swizzle *ir)
       ir->mask.w,
    };
 
-   printf("(swiz ");
+   fprintf(f, "(swiz ");
    for (unsigned i = 0; i < ir->mask.num_components; i++) {
-      printf("%c", "xyzw"[swiz[i]]);
+      fprintf(f, "%c", "xyzw"[swiz[i]]);
    }
-   printf(" ");
+   fprintf(f, " ");
    ir->val->accept(this);
-   printf(")");
+   fprintf(f, ")");
 }
 
 
 void ir_print_visitor::visit(ir_dereference_variable *ir)
 {
    ir_variable *var = ir->variable_referenced();
-   printf("(var_ref %s) ", unique_name(var));
+   fprintf(f, "(var_ref %s) ", unique_name(var));
 }
 
 
 void ir_print_visitor::visit(ir_dereference_array *ir)
 {
-   printf("(array_ref ");
+   fprintf(f, "(array_ref ");
    ir->array->accept(this);
    ir->array_index->accept(this);
-   printf(") ");
+   fprintf(f, ") ");
 }
 
 
 void ir_print_visitor::visit(ir_dereference_record *ir)
 {
-   printf("(record_ref ");
+   fprintf(f, "(record_ref ");
    ir->record->accept(this);
-   printf(" %s) ", ir->field);
+   fprintf(f, " %s) ", ir->field);
 }
 
 
 void ir_print_visitor::visit(ir_assignment *ir)
 {
-   printf("(assign ");
+   fprintf(f, "(assign ");
 
    if (ir->condition)
       ir->condition->accept(this);
@@ -377,22 +384,22 @@ void ir_print_visitor::visit(ir_assignment *ir)
    }
    mask[j] = '\0';
 
-   printf(" (%s) ", mask);
+   fprintf(f, " (%s) ", mask);
 
    ir->lhs->accept(this);
 
-   printf(" ");
+   fprintf(f, " ");
 
    ir->rhs->accept(this);
-   printf(") ");
+   fprintf(f, ") ");
 }
 
 
 void ir_print_visitor::visit(ir_constant *ir)
 {
-   printf("(constant ");
-   print_type(ir->type);
-   printf(" (");
+   fprintf(f, "(constant ");
+   print_type(f, ir->type);
+   fprintf(f, " (");
 
    if (ir->type->is_array()) {
       for (unsigned i = 0; i < ir->type->length; i++)
@@ -400,91 +407,91 @@ void ir_print_visitor::visit(ir_constant *ir)
    } else if (ir->type->is_record()) {
       ir_constant *value = (ir_constant *) ir->components.get_head();
       for (unsigned i = 0; i < ir->type->length; i++) {
-	 printf("(%s ", ir->type->fields.structure[i].name);
+	 fprintf(f, "(%s ", ir->type->fields.structure[i].name);
 	 value->accept(this);
-	 printf(")");
+	 fprintf(f, ")");
 
 	 value = (ir_constant *) value->next;
       }
    } else {
       for (unsigned i = 0; i < ir->type->components(); i++) {
 	 if (i != 0)
-	    printf(" ");
+	    fprintf(f, " ");
 	 switch (ir->type->base_type) {
-	 case GLSL_TYPE_UINT:  printf("%u", ir->value.u[i]); break;
-	 case GLSL_TYPE_INT:   printf("%d", ir->value.i[i]); break;
+	 case GLSL_TYPE_UINT:  fprintf(f, "%u", ir->value.u[i]); break;
+	 case GLSL_TYPE_INT:   fprintf(f, "%d", ir->value.i[i]); break;
 	 case GLSL_TYPE_FLOAT:
             if (ir->value.f[i] == 0.0f)
                /* 0.0 == -0.0, so print with %f to get the proper sign. */
-               printf("%.1f", ir->value.f[i]);
+               fprintf(f, "%.1f", ir->value.f[i]);
             else if (fabs(ir->value.f[i]) < 0.000001f)
-               printf("%a", ir->value.f[i]);
+               fprintf(f, "%a", ir->value.f[i]);
             else if (fabs(ir->value.f[i]) > 1000000.0f)
-               printf("%e", ir->value.f[i]);
+               fprintf(f, "%e", ir->value.f[i]);
             else
-               printf("%f", ir->value.f[i]);
+               fprintf(f, "%f", ir->value.f[i]);
             break;
-	 case GLSL_TYPE_BOOL:  printf("%d", ir->value.b[i]); break;
+	 case GLSL_TYPE_BOOL:  fprintf(f, "%d", ir->value.b[i]); break;
 	 default: assert(0);
 	 }
       }
    }
-   printf(")) ");
+   fprintf(f, ")) ");
 }
 
 
 void
 ir_print_visitor::visit(ir_call *ir)
 {
-   printf("(call %s ", ir->callee_name());
+   fprintf(f, "(call %s ", ir->callee_name());
    if (ir->return_deref)
       ir->return_deref->accept(this);
-   printf(" (");
+   fprintf(f, " (");
    foreach_list(n, &ir->actual_parameters) {
       ir_rvalue *const param = (ir_rvalue *) n;
 
       param->accept(this);
    }
-   printf("))\n");
+   fprintf(f, "))\n");
 }
 
 
 void
 ir_print_visitor::visit(ir_return *ir)
 {
-   printf("(return");
+   fprintf(f, "(return");
 
    ir_rvalue *const value = ir->get_value();
    if (value) {
-      printf(" ");
+      fprintf(f, " ");
       value->accept(this);
    }
 
-   printf(")");
+   fprintf(f, ")");
 }
 
 
 void
 ir_print_visitor::visit(ir_discard *ir)
 {
-   printf("(discard ");
+   fprintf(f, "(discard ");
 
    if (ir->condition != NULL) {
-      printf(" ");
+      fprintf(f, " ");
       ir->condition->accept(this);
    }
 
-   printf(")");
+   fprintf(f, ")");
 }
 
 
 void
 ir_print_visitor::visit(ir_if *ir)
 {
-   printf("(if ");
+   fprintf(f, "(if ");
    ir->condition->accept(this);
 
-   printf("(\n");
+   fprintf(f, "(\n");
    indentation++;
 
    foreach_list(n, &ir->then_instructions) {
@@ -492,16 +499,16 @@ ir_print_visitor::visit(ir_if *ir)
 
       indent();
       inst->accept(this);
-      printf("\n");
+      fprintf(f, "\n");
    }
 
    indentation--;
    indent();
-   printf(")\n");
+   fprintf(f, ")\n");
 
    indent();
    if (!ir->else_instructions.is_empty()) {
-      printf("(\n");
+      fprintf(f, "(\n");
       indentation++;
 
       foreach_list(n, &ir->else_instructions) {
@@ -509,13 +516,13 @@ ir_print_visitor::visit(ir_if *ir)
 
 	 indent();
 	 inst->accept(this);
-	 printf("\n");
+	 fprintf(f, "\n");
       }
       indentation--;
       indent();
-      printf("))\n");
+      fprintf(f, "))\n");
    } else {
-      printf("())\n");
+      fprintf(f, "())\n");
    }
 }
 
@@ -523,7 +530,7 @@ ir_print_visitor::visit(ir_if *ir)
 void
 ir_print_visitor::visit(ir_loop *ir)
 {
-   printf("(loop (\n");
+   fprintf(f, "(loop (\n");
    indentation++;
 
    foreach_list(n, &ir->body_instructions) {
@@ -531,28 +538,28 @@ ir_print_visitor::visit(ir_loop *ir)
 
       indent();
       inst->accept(this);
-      printf("\n");
+      fprintf(f, "\n");
    }
    indentation--;
    indent();
-   printf("))\n");
+   fprintf(f, "))\n");
 }
 
 
 void
 ir_print_visitor::visit(ir_loop_jump *ir)
 {
-   printf("%s", ir->is_break() ? "break" : "continue");
+   fprintf(f, "%s", ir->is_break() ? "break" : "continue");
 }
 
 void
 ir_print_visitor::visit(ir_emit_vertex *ir)
 {
-   printf("(emit-vertex)");
+   fprintf(f, "(emit-vertex)");
 }
 
 void
 ir_print_visitor::visit(ir_end_primitive *ir)
 {
-   printf("(end-primitive)");
+   fprintf(f, "(end-primitive)");
 }
diff --git a/mesalib/src/glsl/ir_print_visitor.h b/mesalib/src/glsl/ir_print_visitor.h
index 865376fe0..98f041d1a 100644
--- a/mesalib/src/glsl/ir_print_visitor.h
+++ b/mesalib/src/glsl/ir_print_visitor.h
@@ -38,7 +38,7 @@ extern "C" {
  */
 class ir_print_visitor : public ir_visitor {
 public:
-   ir_print_visitor();
+   ir_print_visitor(FILE *f);
    virtual ~ir_print_visitor();
 
    void indent(void);
@@ -87,6 +87,7 @@ private:
    _mesa_symbol_table *symbols;
 
    void *mem_ctx;
+   FILE *f;
 
    int indentation;
 };
diff --git a/mesalib/src/glsl/ir_uniform.h b/mesalib/src/glsl/ir_uniform.h
index f678c2c5c..7508f795d 100644
--- a/mesalib/src/glsl/ir_uniform.h
+++ b/mesalib/src/glsl/ir_uniform.h
@@ -78,6 +78,26 @@ struct gl_uniform_driver_storage {
    void *data;
 };
 
+struct gl_opaque_uniform_index {
+   /**
+    * Base opaque uniform index
+    *
+    * If \c gl_uniform_storage::base_type is an opaque type, this
+    * represents its uniform index.  If \c
+    * gl_uniform_storage::array_elements is not zero, the array will
+    * use opaque uniform indices \c index through \c index + \c
+    * gl_uniform_storage::array_elements - 1, inclusive.
+    *
+    * Note that the index may be different in each shader stage.
+    */
+   uint8_t index;
+
+   /**
+    * Whether this opaque uniform is used in this shader stage.
+    */
+   bool active;
+};
+
 struct gl_uniform_storage {
    char *name;
    /** Type of this uniform data stored.
@@ -99,24 +119,9 @@ struct gl_uniform_storage {
     */
    bool initialized;
 
-   struct {
-      /**
-       * Base sampler index
-       *
-       * If \c ::base_type is \c GLSL_TYPE_SAMPLER, this represents the index
-       * of this sampler.  If \c ::array_elements is not zero, the array will
-       * use sampler indices \c ::sampler through \c ::sampler +
-       * \c ::array_elements - 1, inclusive.
-       *
-       * Note that the index may be different in each shader stage.
-       */
-      uint8_t index;
-
-      /**
-       * Whether this sampler is used in this shader stage.
-       */
-      bool active;
-   } sampler[MESA_SHADER_STAGES];
+   struct gl_opaque_uniform_index sampler[MESA_SHADER_STAGES];
+
+   struct gl_opaque_uniform_index image[MESA_SHADER_STAGES];
 
    /**
     * Storage used by the driver for the uniform
diff --git a/mesalib/src/glsl/link_uniform_initializers.cpp b/mesalib/src/glsl/link_uniform_initializers.cpp
index 7d5c1472d..9d6977d57 100644
--- a/mesalib/src/glsl/link_uniform_initializers.cpp
+++ b/mesalib/src/glsl/link_uniform_initializers.cpp
@@ -69,6 +69,7 @@ copy_constant_to_storage(union gl_constant_value *storage,
 	 break;
       case GLSL_TYPE_ARRAY:
       case GLSL_TYPE_STRUCT:
+      case GLSL_TYPE_IMAGE:
       case GLSL_TYPE_ATOMIC_UINT:
       case GLSL_TYPE_INTERFACE:
       case GLSL_TYPE_VOID:
diff --git a/mesalib/src/glsl/link_uniforms.cpp b/mesalib/src/glsl/link_uniforms.cpp
index 2255f7e6e..8d38324fb 100644
--- a/mesalib/src/glsl/link_uniforms.cpp
+++ b/mesalib/src/glsl/link_uniforms.cpp
@@ -240,7 +240,8 @@ class count_uniform_size : public program_resource_visitor {
 public:
    count_uniform_size(struct string_to_uint_map *map)
       : num_active_uniforms(0), num_values(0), num_shader_samplers(0),
-	num_shader_uniform_components(0), is_ubo_var(false), map(map)
+        num_shader_images(0), num_shader_uniform_components(0),
+        is_ubo_var(false), map(map)
    {
       /* empty */
    }
@@ -248,6 +249,7 @@ public:
    void start_shader()
    {
       this->num_shader_samplers = 0;
+      this->num_shader_images = 0;
       this->num_shader_uniform_components = 0;
    }
 
@@ -277,6 +279,11 @@ public:
    unsigned num_shader_samplers;
 
    /**
+    * Number of images used
+    */
+   unsigned num_shader_images;
+
+   /**
     * Number of uniforms used in the current shader
     */
    unsigned num_shader_uniform_components;
@@ -303,6 +310,15 @@ private:
       if (type->contains_sampler()) {
 	 this->num_shader_samplers +=
 	    type->is_array() ? type->array_size() : 1;
+      } else if (type->contains_image()) {
+         this->num_shader_images += values;
+
+         /* As drivers are likely to represent image uniforms as
+          * scalar indices, count them against the limit of uniform
+          * components in the default block.  The spec allows image
+          * uniforms to use up no more than one scalar slot.
+          */
+         this->num_shader_uniform_components += values;
       } else {
 	 /* Accumulate the total number of uniform slots used by this shader.
 	  * Note that samplers do not count against this limit because they
@@ -364,6 +380,7 @@ public:
       this->shader_samplers_used = 0;
       this->shader_shadow_samplers = 0;
       this->next_sampler = 0;
+      this->next_image = 0;
       memset(this->targets, 0, sizeof(this->targets));
    }
 
@@ -460,6 +477,24 @@ private:
       }
    }
 
+   void handle_images(const glsl_type *base_type,
+                      struct gl_uniform_storage *uniform)
+   {
+      if (base_type->is_image()) {
+         uniform->image[shader_type].index = this->next_image;
+         uniform->image[shader_type].active = true;
+
+         /* Increment the image index by 1 for non-arrays and by the
+          * number of array elements for arrays.
+          */
+         this->next_image += MAX2(1, uniform->array_elements);
+
+      } else {
+         uniform->image[shader_type].index = ~0;
+         uniform->image[shader_type].active = false;
+      }
+   }
+
    virtual void visit_field(const glsl_type *type, const char *name,
                             bool row_major)
    {
@@ -495,8 +530,9 @@ private:
 	 base_type = type;
       }
 
-      /* This assigns sampler uniforms to sampler units. */
+      /* This assigns uniform indices to sampler and image uniforms. */
       handle_samplers(base_type, &this->uniforms[id]);
+      handle_images(base_type, &this->uniforms[id]);
 
       /* If there is already storage associated with this uniform, it means
        * that it was set while processing an earlier shader stage.  For
@@ -554,6 +590,7 @@ private:
 
    struct gl_uniform_storage *uniforms;
    unsigned next_sampler;
+   unsigned next_image;
 
 public:
    union gl_constant_value *values;
@@ -720,6 +757,40 @@ link_assign_uniform_block_offsets(struct gl_shader *shader)
    }
 }
 
+/**
+ * Scan the program for image uniforms and store image unit access
+ * information into the gl_shader data structure.
+ */
+static void
+link_set_image_access_qualifiers(struct gl_shader_program *prog)
+{
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      gl_shader *sh = prog->_LinkedShaders[i];
+
+      if (sh == NULL)
+	 continue;
+
+      foreach_list(node, sh->ir) {
+	 ir_variable *var = ((ir_instruction *) node)->as_variable();
+
+         if (var && var->data.mode == ir_var_uniform &&
+             var->type->contains_image()) {
+            unsigned id;
+            bool found = prog->UniformHash->get(id, var->name);
+            assert(found);
+            const gl_uniform_storage *storage = &prog->UniformStorage[id];
+            const unsigned index = storage->image[i].index;
+            const GLenum access = (var->data.image.read_only ? GL_READ_ONLY :
+                                   var->data.image.write_only ? GL_WRITE_ONLY :
+                                   GL_READ_WRITE);
+
+            for (unsigned j = 0; j < MAX2(1, storage->array_elements); ++j)
+               sh->ImageAccess[index + j] = access;
+         }
+      }
+   }
+}
+
 void
 link_assign_uniform_locations(struct gl_shader_program *prog)
 {
@@ -757,6 +828,7 @@ link_assign_uniform_locations(struct gl_shader_program *prog)
        *     types cannot have initializers."
        */
       memset(sh->SamplerUnits, 0, sizeof(sh->SamplerUnits));
+      memset(sh->ImageUnits, 0, sizeof(sh->ImageUnits));
 
       link_update_uniform_buffer_variables(sh);
 
@@ -782,6 +854,7 @@ link_assign_uniform_locations(struct gl_shader_program *prog)
       }
 
       sh->num_samplers = uniform_size.num_shader_samplers;
+      sh->NumImages = uniform_size.num_shader_images;
       sh->num_uniform_components = uniform_size.num_shader_uniform_components;
 
       sh->num_combined_uniform_components = sh->num_uniform_components;
@@ -861,6 +934,7 @@ link_assign_uniform_locations(struct gl_shader_program *prog)
    prog->NumUserUniformStorage = num_user_uniforms;
    prog->UniformStorage = uniforms;
 
+   link_set_image_access_qualifiers(prog);
    link_set_uniform_initializers(prog);
 
    return;
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index bcd739476..f6b266185 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -1206,6 +1206,7 @@ link_gs_inout_layout_qualifiers(struct gl_shader_program *prog,
 				unsigned num_shaders)
 {
    linked_shader->Geom.VerticesOut = 0;
+   linked_shader->Geom.Invocations = 0;
    linked_shader->Geom.InputType = PRIM_UNKNOWN;
    linked_shader->Geom.OutputType = PRIM_UNKNOWN;
 
@@ -1259,6 +1260,18 @@ link_gs_inout_layout_qualifiers(struct gl_shader_program *prog,
 	 }
 	 linked_shader->Geom.VerticesOut = shader->Geom.VerticesOut;
       }
+
+      if (shader->Geom.Invocations != 0) {
+	 if (linked_shader->Geom.Invocations != 0 &&
+	     linked_shader->Geom.Invocations != shader->Geom.Invocations) {
+	    linker_error(prog, "geometry shader defined with conflicting "
+			 "invocation count (%d and %d)\n",
+			 linked_shader->Geom.Invocations,
+			 shader->Geom.Invocations);
+	    return;
+	 }
+	 linked_shader->Geom.Invocations = shader->Geom.Invocations;
+      }
    }
 
    /* Just do the intrastage -> interstage propagation right now,
@@ -1285,6 +1298,11 @@ link_gs_inout_layout_qualifiers(struct gl_shader_program *prog,
       return;
    }
    prog->Geom.VerticesOut = linked_shader->Geom.VerticesOut;
+
+   if (linked_shader->Geom.Invocations == 0)
+      linked_shader->Geom.Invocations = 1;
+
+   prog->Geom.Invocations = linked_shader->Geom.Invocations;
 }
 
 
@@ -2031,6 +2049,46 @@ check_resources(struct gl_context *ctx, struct gl_shader_program *prog)
    }
 }
 
+/**
+ * Validate shader image resources.
+ */
+static void
+check_image_resources(struct gl_context *ctx, struct gl_shader_program *prog)
+{
+   unsigned total_image_units = 0;
+   unsigned fragment_outputs = 0;
+
+   if (!ctx->Extensions.ARB_shader_image_load_store)
+      return;
+
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      struct gl_shader *sh = prog->_LinkedShaders[i];
+
+      if (sh) {
+         if (sh->NumImages > ctx->Const.Program[i].MaxImageUniforms)
+            linker_error(prog, "Too many %s shader image uniforms",
+                         _mesa_shader_stage_to_string(i));
+
+         total_image_units += sh->NumImages;
+
+         if (i == MESA_SHADER_FRAGMENT) {
+            foreach_list(node, sh->ir) {
+               ir_variable *var = ((ir_instruction *)node)->as_variable();
+               if (var && var->data.mode == ir_var_shader_out)
+                  fragment_outputs += var->type->count_attribute_slots();
+            }
+         }
+      }
+   }
+
+   if (total_image_units > ctx->Const.MaxCombinedImageUniforms)
+      linker_error(prog, "Too many combined image uniforms");
+
+   if (total_image_units + fragment_outputs >
+       ctx->Const.MaxCombinedImageUnitsAndFragmentOutputs)
+      linker_error(prog, "Too many combined image uniforms and fragment outputs");
+}
+
 void
 link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
 {
@@ -2394,17 +2452,19 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
    store_fragdepth_layout(prog);
 
    check_resources(ctx, prog);
+   check_image_resources(ctx, prog);
    link_check_atomic_counter_resources(ctx, prog);
 
    if (!prog->LinkStatus)
       goto done;
 
    /* OpenGL ES requires that a vertex shader and a fragment shader both be
-    * present in a linked program.  By checking prog->IsES, we also
-    * catch the GL_ARB_ES2_compatibility case.
+    * present in a linked program. GL_ARB_ES2_compatibility doesn't say
+    * anything about shader linking when one of the shaders (vertex or
+    * fragment shader) is absent. So, the extension shouldn't change the
+    * behavior specified in GLSL specification.
     */
-   if (!prog->InternalSeparateShader &&
-       (ctx->API == API_OPENGLES2 || prog->IsES)) {
+   if (!prog->InternalSeparateShader && ctx->API == API_OPENGLES2) {
       if (prog->_LinkedShaders[MESA_SHADER_VERTEX] == NULL) {
 	 linker_error(prog, "program lacks a vertex shader\n");
       } else if (prog->_LinkedShaders[MESA_SHADER_FRAGMENT] == NULL) {
diff --git a/mesalib/src/glsl/lower_instructions.cpp b/mesalib/src/glsl/lower_instructions.cpp
index 44a6e8021..01ea0f01f 100644
--- a/mesalib/src/glsl/lower_instructions.cpp
+++ b/mesalib/src/glsl/lower_instructions.cpp
@@ -38,7 +38,6 @@
  * - LOG_TO_LOG2
  * - MOD_TO_FRACT
  * - LDEXP_TO_ARITH
- * - LRP_TO_ARITH
  * - BITFIELD_INSERT_TO_BFM_BFI
  *
  * SUB_TO_ADD_NEG:
@@ -87,10 +86,6 @@
  * -------------
  * Converts ir_binop_ldexp to arithmetic and bit operations.
  *
- * LRP_TO_ARITH:
- * -------------
- * Converts ir_triop_lrp to (op0 * (1.0f - op2)) + (op1 * op2).
- *
  * BITFIELD_INSERT_TO_BFM_BFI:
  * ---------------------------
  * Breaks ir_quadop_bitfield_insert into ir_binop_bfm (bitfield mask) and
@@ -130,7 +125,6 @@ private:
    void exp_to_exp2(ir_expression *);
    void pow_to_exp2(ir_expression *);
    void log_to_log2(ir_expression *);
-   void lrp_to_arith(ir_expression *);
    void bitfield_insert_to_bfm_bfi(ir_expression *);
    void ldexp_to_arith(ir_expression *);
 };
@@ -299,27 +293,6 @@ lower_instructions_visitor::mod_to_fract(ir_expression *ir)
 }
 
 void
-lower_instructions_visitor::lrp_to_arith(ir_expression *ir)
-{
-   /* (lrp x y a) -> x*(1-a) + y*a */
-
-   /* Save op2 */
-   ir_variable *temp = new(ir) ir_variable(ir->operands[2]->type, "lrp_factor",
-					   ir_var_temporary);
-   this->base_ir->insert_before(temp);
-   this->base_ir->insert_before(assign(temp, ir->operands[2]));
-
-   ir_constant *one = new(ir) ir_constant(1.0f);
-
-   ir->operation = ir_binop_add;
-   ir->operands[0] = mul(ir->operands[0], sub(one, temp));
-   ir->operands[1] = mul(ir->operands[1], temp);
-   ir->operands[2] = NULL;
-
-   this->progress = true;
-}
-
-void
 lower_instructions_visitor::bitfield_insert_to_bfm_bfi(ir_expression *ir)
 {
    /* Translates
@@ -499,11 +472,6 @@ lower_instructions_visitor::visit_leave(ir_expression *ir)
 	 pow_to_exp2(ir);
       break;
 
-   case ir_triop_lrp:
-      if (lowering(LRP_TO_ARITH))
-	 lrp_to_arith(ir);
-      break;
-
    case ir_quadop_bitfield_insert:
       if (lowering(BITFIELD_INSERT_TO_BFM_BFI))
          bitfield_insert_to_bfm_bfi(ir);
diff --git a/mesalib/src/glsl/lower_offset_array.cpp b/mesalib/src/glsl/lower_offset_array.cpp
new file mode 100644
index 000000000..0c235eda3
--- /dev/null
+++ b/mesalib/src/glsl/lower_offset_array.cpp
@@ -0,0 +1,90 @@
+/*
+ * Copyright © 2013 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.
+ */
+
+/**
+ * \file brw_lower_offset_array.cpp
+ *
+ * IR lower pass to decompose ir_texture ir_tg4 with an array of offsets
+ * into four ir_tg4s with a single ivec2 offset, select the .w component of each,
+ * and return those four values packed into a gvec4.
+ *
+ * \author Chris Forbes <chrisf@ijw.co.nz>
+ */
+
+#include "glsl_types.h"
+#include "ir.h"
+#include "ir_builder.h"
+#include "ir_optimization.h"
+#include "ir_rvalue_visitor.h"
+
+using namespace ir_builder;
+
+class brw_lower_offset_array_visitor : public ir_rvalue_visitor {
+public:
+   brw_lower_offset_array_visitor()
+   {
+      progress = false;
+   }
+
+   void handle_rvalue(ir_rvalue **rv);
+
+   bool progress;
+};
+
+void
+brw_lower_offset_array_visitor::handle_rvalue(ir_rvalue **rv)
+{
+   if (*rv == NULL || (*rv)->ir_type != ir_type_texture)
+      return;
+
+   ir_texture *ir = (ir_texture *) *rv;
+   if (ir->op != ir_tg4 || !ir->offset || !ir->offset->type->is_array())
+      return;
+
+   void *mem_ctx = ralloc_parent(ir);
+
+   ir_variable *var = new (mem_ctx) ir_variable(ir->type, "result", ir_var_auto);
+   base_ir->insert_before(var);
+
+   for (int i = 0; i < 4; i++) {
+      ir_texture *tex = ir->clone(mem_ctx, NULL);
+      tex->offset = new (mem_ctx) ir_dereference_array(tex->offset,
+            new (mem_ctx) ir_constant(i));
+
+      base_ir->insert_before(assign(var, swizzle_w(tex), 1 << i));
+   }
+
+   *rv = new (mem_ctx) ir_dereference_variable(var);
+
+   progress = true;
+}
+
+bool
+lower_offset_arrays(exec_list *instructions)
+{
+   brw_lower_offset_array_visitor v;
+
+   visit_list_elements(&v, instructions);
+
+   return v.progress;
+}
diff --git a/mesalib/src/glsl/main.cpp b/mesalib/src/glsl/main.cpp
index 48ab1c73a..0d8c01f6a 100644
--- a/mesalib/src/glsl/main.cpp
+++ b/mesalib/src/glsl/main.cpp
@@ -305,7 +305,7 @@ compile_shader(struct gl_context *ctx, struct gl_shader *shader)
 
    /* Print out the resulting IR */
    if (!state->error && dump_lir) {
-      _mesa_print_ir(shader->ir, state);
+      _mesa_print_ir(stdout, shader->ir, state);
    }
 
    return;
diff --git a/mesalib/src/glsl/opt_algebraic.cpp b/mesalib/src/glsl/opt_algebraic.cpp
index 1b4d31936..5c49a785c 100644
--- a/mesalib/src/glsl/opt_algebraic.cpp
+++ b/mesalib/src/glsl/opt_algebraic.cpp
@@ -568,6 +568,12 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
          return ir->operands[1];
       } else if (ir->operands[0]->equals(ir->operands[1])) {
          return ir->operands[0];
+      } else if (is_vec_zero(op_const[0])) {
+         return mul(ir->operands[1], ir->operands[2]);
+      } else if (is_vec_zero(op_const[1])) {
+         unsigned op2_components = ir->operands[2]->type->vector_elements;
+         ir_constant *one = new(mem_ctx) ir_constant(1.0f, op2_components);
+         return mul(ir->operands[0], add(one, neg(ir->operands[2])));
       }
       break;
 
diff --git a/mesalib/src/glsl/opt_array_splitting.cpp b/mesalib/src/glsl/opt_array_splitting.cpp
index f37d09022..97d3a57e9 100644
--- a/mesalib/src/glsl/opt_array_splitting.cpp
+++ b/mesalib/src/glsl/opt_array_splitting.cpp
@@ -400,7 +400,7 @@ optimize_split_arrays(exec_list *instructions, bool linked)
    visit_list_elements(&split, instructions);
 
    if (debug)
-      _mesa_print_ir(instructions, NULL);
+      _mesa_print_ir(stdout, instructions, NULL);
 
    ralloc_free(mem_ctx);
 
diff --git a/mesalib/src/glsl/opt_vectorize.cpp b/mesalib/src/glsl/opt_vectorize.cpp
index 8ee81f1a3..13faac05e 100644
--- a/mesalib/src/glsl/opt_vectorize.cpp
+++ b/mesalib/src/glsl/opt_vectorize.cpp
@@ -82,6 +82,8 @@ public:
 
    virtual ir_visitor_status visit_enter(ir_assignment *);
    virtual ir_visitor_status visit_enter(ir_swizzle *);
+   virtual ir_visitor_status visit_enter(ir_dereference_array *);
+   virtual ir_visitor_status visit_enter(ir_expression *);
    virtual ir_visitor_status visit_enter(ir_if *);
    virtual ir_visitor_status visit_enter(ir_loop *);
 
@@ -289,6 +291,33 @@ ir_vectorize_visitor::visit_enter(ir_swizzle *ir)
    return visit_continue;
 }
 
+/* Upon entering an ir_array_dereference, remove the current assignment from
+ * further consideration. Since the index of an array dereference must scalar,
+ * we are not able to vectorize it.
+ *
+ * FINISHME: If all of scalar indices are identical we could vectorize.
+ */
+ir_visitor_status
+ir_vectorize_visitor::visit_enter(ir_dereference_array *ir)
+{
+   this->current_assignment = NULL;
+   return visit_continue_with_parent;
+}
+
+/**
+ * Upon entering an ir_expression, remove the current assignment from further
+ * consideration if the expression operates horizontally on vectors.
+ */
+ir_visitor_status
+ir_vectorize_visitor::visit_enter(ir_expression *ir)
+{
+   if (ir->is_horizontal()) {
+      this->current_assignment = NULL;
+      return visit_continue_with_parent;
+   }
+   return visit_continue;
+}
+
 /* Since there is no statement to visit between the "then" and "else"
  * instructions try to vectorize before, in between, and after them to avoid
  * combining statements from different basic blocks.
diff --git a/mesalib/src/glsl/test_optpass.cpp b/mesalib/src/glsl/test_optpass.cpp
index 1a15f3c63..f1b9579cd 100644
--- a/mesalib/src/glsl/test_optpass.cpp
+++ b/mesalib/src/glsl/test_optpass.cpp
@@ -235,7 +235,7 @@ int test_optpass(int argc, char **argv)
    /* Print out the initial IR */
    if (!state->error && !quiet) {
       printf("*** pre-optimization IR:\n");
-      _mesa_print_ir(shader->ir, state);
+      _mesa_print_ir(stdout, shader->ir, state);
       printf("\n--\n");
    }
 
@@ -255,7 +255,7 @@ int test_optpass(int argc, char **argv)
       if (!quiet) {
          printf("*** resulting IR:\n");
       }
-      _mesa_print_ir(shader->ir, state);
+      _mesa_print_ir(stdout, shader->ir, state);
       if (!quiet) {
          printf("\n--\n");
       }
diff --git a/mesalib/src/loader/loader.c b/mesalib/src/loader/loader.c
index 811f8a257..811f8a257 100755..100644
--- a/mesalib/src/loader/loader.c
+++ b/mesalib/src/loader/loader.c
diff --git a/mesalib/src/mapi/glapi/Makefile.am b/mesalib/src/mapi/glapi/Makefile.am
index bf653a305..e425bacea 100644
--- a/mesalib/src/mapi/glapi/Makefile.am
+++ b/mesalib/src/mapi/glapi/Makefile.am
@@ -27,6 +27,7 @@ include ../Makefile.sources
 
 AM_CPPFLAGS = \
         $(DEFINES) \
+        $(SELINUX_CFLAGS) \
 	-I$(top_srcdir)/include \
 	-I$(top_srcdir)/src/mapi \
         -I$(top_builddir)/src/mapi \
diff --git a/mesalib/src/mapi/glapi/Makefile.sources b/mesalib/src/mapi/glapi/Makefile.sources
index 58d28c52c..100e6347e 100644
--- a/mesalib/src/mapi/glapi/Makefile.sources
+++ b/mesalib/src/mapi/glapi/Makefile.sources
@@ -6,7 +6,6 @@ GLAPI_SOURCES = \
 	glapi_gentable.c \
 	glapi_getproc.c \
 	glapi_nop.c \
-	glthread.c \
 	glapi.c
 
 X86_API =			\
diff --git a/mesalib/src/mapi/glapi/SConscript b/mesalib/src/mapi/glapi/SConscript
index 152818d22..bc1c43aa8 100644
--- a/mesalib/src/mapi/glapi/SConscript
+++ b/mesalib/src/mapi/glapi/SConscript
@@ -34,7 +34,6 @@ glapi_sources = [
     'glapi_entrypoint.c',
     'glapi_getproc.c',
     'glapi_nop.c',
-    'glthread.c',
     'glapi.c',
 ]
 
diff --git a/mesalib/src/mapi/glapi/gen/ARB_separate_shader_objects.xml b/mesalib/src/mapi/glapi/gen/ARB_separate_shader_objects.xml
new file mode 100644
index 000000000..80234ea5e
--- /dev/null
+++ b/mesalib/src/mapi/glapi/gen/ARB_separate_shader_objects.xml
@@ -0,0 +1,401 @@
+<?xml version="1.0"?>
+<!DOCTYPE OpenGLAPI SYSTEM "gl_API.dtd">
+
+<!-- Note: no GLX protocol info yet. -->
+
+<OpenGLAPI>
+   <category name="ARB_separate_shader_objects" number="97">
+      <enum   name="ACTIVE_PROGRAM"                               value="0x8259"/>
+      <enum   name="PROGRAM_PIPELINE_BINDING"                     value="0x825A"/>
+      <enum   name="VERTEX_SHADER_BIT"                            value="0x00000001"/>
+      <enum   name="FRAGMENT_SHADER_BIT"                          value="0x00000002"/>
+      <enum   name="GEOMETRY_SHADER_BIT"                          value="0x00000004"/>
+      <enum   name="TESS_CONTROL_SHADER_BIT"                      value="0x00000008"/>
+      <enum   name="TESS_EVALUATION_SHADER_BIT"                   value="0x00000010"/>
+      <enum   name="ALL_SHADER_BITS"                              value="0xFFFFFFFF"/>
+      <enum   name="PROGRAM_SEPARABLE"                            value="0x8258"/>
+
+      <function name="UseProgramStages" offset="assign">
+         <param name="pipeline" type="GLuint" />
+         <param name="stages" type="GLbitfield" />
+         <param name="program" type="GLuint" />
+      </function>
+      <function name="ActiveShaderProgram" offset="assign">
+         <param name="pipeline" type="GLuint" />
+         <param name="program" type="GLuint" />
+      </function>
+      <function name="CreateShaderProgramv" offset="assign">
+         <param name="type" type="GLenum" />
+         <param name="count" type="GLsizei" />
+         <param name="strings" type="const GLchar * const *" />
+         <return type="GLuint"/>
+      </function>
+      <function name="BindProgramPipeline" offset="assign">
+         <param name="pipeline" type="GLuint" />
+      </function>
+      <function name="DeleteProgramPipelines" offset="assign">
+         <param name="n" type="GLsizei" />
+         <param name="pipelines" type="const GLuint *" />
+      </function>
+      <function name="GenProgramPipelines" offset="assign">
+         <param name="n" type="GLsizei" />
+         <param name="pipelines" type="GLuint *" />
+      </function>
+      <function name="IsProgramPipeline" offset="assign">
+         <param name="pipeline" type="GLuint" />
+         <return type="GLboolean"/>
+      </function>
+      <!-- Function already included on ARB_get_program_binary.xml. Keep a commented
+      version here for completeness -->
+      <!--
+      <function name="ProgramParameteri" offset="assign" es2="3.0">
+         <param name="program" type="GLuint"/>
+         <param name="pname" type="GLenum"/>
+         <param name="value" type="GLint"/>
+      </function>
+      -->
+      <function name="GetProgramPipelineiv" offset="assign">
+         <param name="pipeline" type="GLuint" />
+         <param name="pname" type="GLenum" />
+         <param name="params" type="GLint *" />
+      </function>
+      <function name="ProgramUniform1i" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLint" />
+      </function>
+      <function name="ProgramUniform2i" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLint" />
+         <param name="y" type="GLint" />
+      </function>
+      <function name="ProgramUniform3i" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLint" />
+         <param name="y" type="GLint" />
+         <param name="z" type="GLint" />
+      </function>
+      <function name="ProgramUniform4i" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLint" />
+         <param name="y" type="GLint" />
+         <param name="z" type="GLint" />
+         <param name="w" type="GLint" />
+      </function>
+      <function name="ProgramUniform1ui" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLuint" />
+      </function>
+      <function name="ProgramUniform2ui" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLuint" />
+         <param name="y" type="GLuint" />
+      </function>
+      <function name="ProgramUniform3ui" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLuint" />
+         <param name="y" type="GLuint" />
+         <param name="z" type="GLuint" />
+      </function>
+      <function name="ProgramUniform4ui" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLuint" />
+         <param name="y" type="GLuint" />
+         <param name="z" type="GLuint" />
+         <param name="w" type="GLuint" />
+      </function>
+      <function name="ProgramUniform1f" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLfloat" />
+      </function>
+      <function name="ProgramUniform2f" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLfloat" />
+         <param name="y" type="GLfloat" />
+      </function>
+      <function name="ProgramUniform3f" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLfloat" />
+         <param name="y" type="GLfloat" />
+         <param name="z" type="GLfloat" />
+      </function>
+      <function name="ProgramUniform4f" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLfloat" />
+         <param name="y" type="GLfloat" />
+         <param name="z" type="GLfloat" />
+         <param name="w" type="GLfloat" />
+      </function>
+      <function name="ProgramUniform1iv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLint *" />
+      </function>
+      <function name="ProgramUniform2iv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLint *" />
+      </function>
+      <function name="ProgramUniform3iv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLint *" />
+      </function>
+      <function name="ProgramUniform4iv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLint *" />
+      </function>
+      <function name="ProgramUniform1uiv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLuint *" />
+      </function>
+      <function name="ProgramUniform2uiv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLuint *" />
+      </function>
+      <function name="ProgramUniform3uiv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLuint *" />
+      </function>
+      <function name="ProgramUniform4uiv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLuint *" />
+      </function>
+      <function name="ProgramUniform1fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniform2fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniform3fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniform4fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix2fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix3fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix4fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix2x3fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix3x2fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix2x4fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix4x2fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix3x4fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ProgramUniformMatrix4x3fv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLfloat *" />
+      </function>
+      <function name="ValidateProgramPipeline" offset="assign">
+         <param name="pipeline" type="GLuint" />
+      </function>
+      <function name="GetProgramPipelineInfoLog" offset="assign">
+         <param name="pipeline" type="GLuint" />
+         <param name="bufSize" type="GLsizei" />
+         <param name="length" type="GLsizei *" />
+         <param name="infoLog" type="GLchar *" />
+      </function>
+
+      <!-- depends on GL_ARB_gpu_shader_fp64
+      <function name="ProgramUniform1d" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLdouble" />
+      </function>
+      <function name="ProgramUniform2d" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLdouble" />
+         <param name="y" type="GLdouble" />
+      </function>
+      <function name="ProgramUniform3d" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLdouble" />
+         <param name="y" type="GLdouble" />
+         <param name="z" type="GLdouble" />
+      </function>
+      <function name="ProgramUniform4d" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="x" type="GLdouble" />
+         <param name="y" type="GLdouble" />
+         <param name="z" type="GLdouble" />
+         <param name="w" type="GLdouble" />
+      </function>
+      <function name="ProgramUniformMatrix2x3dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix3x2dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix2x4dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix4x2dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix3x4dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix4x3dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix2dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix3dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniformMatrix4dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="transpose" type="GLboolean" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniform1dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniform2dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniform3dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      <function name="ProgramUniform4dv" offset="assign">
+         <param name="program" type="GLuint" />
+         <param name="location" type="GLint" />
+         <param name="count" type="GLsizei" />
+         <param name="value" type="const GLdouble *" />
+      </function>
+      -->
+   </category>
+</OpenGLAPI>
diff --git a/mesalib/src/mapi/glapi/gen/Makefile.am b/mesalib/src/mapi/glapi/gen/Makefile.am
index 9e6fe5261..7b3c118f0 100644
--- a/mesalib/src/mapi/glapi/gen/Makefile.am
+++ b/mesalib/src/mapi/glapi/gen/Makefile.am
@@ -113,6 +113,7 @@ API_XML = \
 	ARB_sample_shading.xml \
 	ARB_sampler_objects.xml \
 	ARB_seamless_cube_map.xml \
+	ARB_separate_shader_objects.xml \
 	ARB_shader_atomic_counters.xml \
 	ARB_shader_image_load_store.xml \
 	ARB_sync.xml \
diff --git a/mesalib/src/mapi/glapi/gen/NV_vdpau_interop.xml b/mesalib/src/mapi/glapi/gen/NV_vdpau_interop.xml
index cf5f0eddd..0b19e1a85 100644
--- a/mesalib/src/mapi/glapi/gen/NV_vdpau_interop.xml
+++ b/mesalib/src/mapi/glapi/gen/NV_vdpau_interop.xml
@@ -29,6 +29,7 @@
     </function>
 
     <function name="VDPAUIsSurfaceNV" offset="assign">
+        <return type="GLboolean"/>
 	<param name="surface" type="GLintptr"/>
     </function>
 
diff --git a/mesalib/src/mapi/glapi/gen/glX_proto_recv.py b/mesalib/src/mapi/glapi/gen/glX_proto_recv.py
index 84526581a..086fc97d7 100755
--- a/mesalib/src/mapi/glapi/gen/glX_proto_recv.py
+++ b/mesalib/src/mapi/glapi/gen/glX_proto_recv.py
@@ -97,7 +97,6 @@ class PrintGlxDispatchFunctions(glX_proto_common.glx_print_proto):
         print '#include "singlesize.h"'
         print '#include "glapi.h"'
         print '#include "glapitable.h"'
-        print '#include "glthread.h"'
         print '#include "dispatch.h"'
         print ''
         print '#define __GLX_PAD(x)  (((x) + 3) & ~3)'
diff --git a/mesalib/src/mapi/glapi/gen/glX_proto_send.py b/mesalib/src/mapi/glapi/gen/glX_proto_send.py
index 8b804418b..a98f63e99 100755
--- a/mesalib/src/mapi/glapi/gen/glX_proto_send.py
+++ b/mesalib/src/mapi/glapi/gen/glX_proto_send.py
@@ -167,7 +167,6 @@ class PrintGlxProtoStubs(glX_proto_common.glx_print_proto):
         print '#include "glxclient.h"'
         print '#include "indirect_size.h"'
         print '#include "glapi.h"'
-        print '#include "glthread.h"'
         print '#include <GL/glxproto.h>'
         print '#include <X11/Xlib-xcb.h>'
         print '#include <xcb/xcb.h>'
diff --git a/mesalib/src/mapi/glapi/gen/gl_API.xml b/mesalib/src/mapi/glapi/gen/gl_API.xml
index d5467a823..8f175e6be 100755
--- a/mesalib/src/mapi/glapi/gen/gl_API.xml
+++ b/mesalib/src/mapi/glapi/gen/gl_API.xml
@@ -8275,7 +8275,11 @@
   </function>
 </category>
 
-<!-- ARB extensions #95...#108 -->
+<!-- ARB extensions #95...#96 -->
+
+<xi:include href="ARB_separate_shader_objects.xml" xmlns:xi="http://www.w3.org/2001/XInclude"/>
+
+<!-- ARB extensions #98...#108 -->
 
 <xi:include href="ARB_ES2_compatibility.xml" xmlns:xi="http://www.w3.org/2001/XInclude"/>
 
@@ -8492,6 +8496,25 @@
 
 <xi:include href="ARB_texture_storage_multisample.xml" xmlns:xi="http://www.w3.org/2001/XInclude"/>
 
+<!-- ARB extension 142 - 143. -->
+
+<category name="GL_ARB_buffer_storage" number="144">
+    <enum name="MAP_PERSISTENT_BIT" value="0x40" />
+    <enum name="MAP_COHERENT_BIT" value="0x80" />
+    <enum name="DYNAMIC_STORAGE_BIT" value="0x100" />
+    <enum name="CLIENT_STORAGE_BIT" value="0x200" />
+    <enum name="BUFFER_IMMUTABLE_STORAGE" value="0x821F" />
+    <enum name="BUFFER_STORAGE_FLAGS" value="0x8220" />
+    <enum name="CLIENT_MAPPED_BUFFER_BARRIER_BIT" value="0x4000" />
+
+    <function name="BufferStorage" offset="assign">
+        <param name="target" type="GLenum"/>
+        <param name="size" type="GLsizeiptr"/>
+        <param name="data" type="const GLvoid *"/>
+        <param name="flags" type="GLbitfield"/>
+    </function>
+</category>
+
 <!-- Non-ARB extensions sorted by extension number. -->
 
 <category name="GL_EXT_blend_color" number="2">
diff --git a/mesalib/src/mapi/glapi/gen/gl_functions.py b/mesalib/src/mapi/glapi/gen/gl_functions.py
index 3db72bba8..cd408ade2 100755
--- a/mesalib/src/mapi/glapi/gen/gl_functions.py
+++ b/mesalib/src/mapi/glapi/gen/gl_functions.py
@@ -43,7 +43,7 @@ class PrintFunctionTable(gl_XML.gl_print_base):
         print """
 #include "glapitable.h"
 #include "glapi.h"
-#include "glthread.h"
+#include "u_thread.h"
 #include "dispatch.h"
 """
         return
diff --git a/mesalib/src/mapi/glapi/gen/gl_genexec.py b/mesalib/src/mapi/glapi/gen/gl_genexec.py
index 1765a51ad..460919397 100644
--- a/mesalib/src/mapi/glapi/gen/gl_genexec.py
+++ b/mesalib/src/mapi/glapi/gen/gl_genexec.py
@@ -86,6 +86,7 @@ header = """/**
 #include "main/multisample.h"
 #include "main/objectlabel.h"
 #include "main/performance_monitor.h"
+#include "main/pipelineobj.h"
 #include "main/pixel.h"
 #include "main/pixelstore.h"
 #include "main/points.h"
diff --git a/mesalib/src/mapi/glapi/glapi.h b/mesalib/src/mapi/glapi/glapi.h
index b10f19a63..67bf1f42e 100644
--- a/mesalib/src/mapi/glapi/glapi.h
+++ b/mesalib/src/mapi/glapi/glapi.h
@@ -44,13 +44,14 @@
 #ifndef _GLAPI_H
 #define _GLAPI_H
 
-#include "glthread.h"
+#include "u_thread.h"
 
 
 #ifdef __cplusplus
 extern "C" {
 #endif
 
+
 #ifdef _GLAPI_NO_EXPORTS
 #  define _GLAPI_EXPORT
 #else /* _GLAPI_NO_EXPORTS */
@@ -69,7 +70,6 @@ extern "C" {
 
 #include "GL/gl.h"
 #include "GL/glext.h"
-#include "glthread.h"
 
 
 struct _glapi_table;
@@ -84,8 +84,8 @@ typedef void (*_glapi_warning_func)(void *ctx, const char *str, ...);
 #define _glapi_get_dispatch _mglapi_get_dispatch
 #define _glapi_set_context _mglapi_set_context
 #define _glapi_get_context _mglapi_get_context
-#define _glapi_Context _mglapi_Context
 #define _glapi_Dispatch _mglapi_Dispatch
+#define _glapi_Context _mglapi_Context
 #endif
 
 /*
@@ -100,12 +100,14 @@ typedef void (*_glapi_warning_func)(void *ctx, const char *str, ...);
  **/
 #if defined (GLX_USE_TLS)
 
-_GLAPI_EXPORT extern __thread struct _glapi_table * _glapi_tls_Dispatch;
+_GLAPI_EXPORT extern __thread struct _glapi_table * _glapi_tls_Dispatch
+  ;
 
-_GLAPI_EXPORT extern const void *_glapi_Context;
-_GLAPI_EXPORT extern const struct _glapi_table *_glapi_Dispatch;
+_GLAPI_EXPORT extern __thread void * _glapi_tls_Context
+  ;
 
-_GLAPI_EXPORT extern __thread void * _glapi_tls_Context;
+_GLAPI_EXPORT extern const struct _glapi_table *_glapi_Dispatch;
+_GLAPI_EXPORT extern const void *_glapi_Context;
 
 # define GET_DISPATCH() _glapi_tls_Dispatch
 # define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) _glapi_tls_Context
@@ -118,20 +120,22 @@ _GLAPI_EXPORT extern __thread void * _glapi_tls_Context;
 #define SERVEXTERN _declspec(dllexport)
 #endif
 
-SERVEXTERN void *_glapi_Context;
 SERVEXTERN struct _glapi_table *_glapi_Dispatch;
+SERVEXTERN void *_glapi_Context;
 
 # ifdef THREADS
 
 #  define GET_DISPATCH() \
      (likely(_glapi_Dispatch) ? _glapi_Dispatch : _glapi_get_dispatch())
 
-
 #  define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) \
      (likely(_glapi_Context) ? _glapi_Context : _glapi_get_context())
+
 # else
+
 #  define GET_DISPATCH() _glapi_Dispatch
 #  define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) _glapi_Context
+
 # endif
 
 #endif /* defined (GLX_USE_TLS) */
@@ -183,8 +187,10 @@ _glapi_get_proc_address(const char *funcName);
 extern struct _glapi_table *
 _glapi_create_table_from_handle(void *handle, const char *symbol_prefix);
 
-#endif
+
 
 #ifdef __cplusplus
 }
 #endif
+
+#endif /* _GLAPI_H */
diff --git a/mesalib/src/mapi/glapi/glthread.c b/mesalib/src/mapi/glapi/glthread.c
deleted file mode 100644
index 239090ecf..000000000
--- a/mesalib/src/mapi/glapi/glthread.c
+++ /dev/null
@@ -1,7 +0,0 @@
-#include "glapi/glapi.h"
-
-unsigned long _GLAPI_EXPORT
-_glthread_GetID(void)
-{
-   return u_thread_self();
-}
diff --git a/mesalib/src/mapi/glapi/glthread.h b/mesalib/src/mapi/glapi/glthread.h
deleted file mode 100644
index 2399abb2b..000000000
--- a/mesalib/src/mapi/glapi/glthread.h
+++ /dev/null
@@ -1,28 +0,0 @@
-#ifndef GLTHREAD_H
-#define GLTHREAD_H
-
-#include "u_thread.h"
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#define _glthread_DECLARE_STATIC_MUTEX(name) u_mutex_declare_static(name)
-#define _glthread_INIT_MUTEX(name)           u_mutex_init(name)
-#define _glthread_DESTROY_MUTEX(name)        u_mutex_destroy(name)
-#define _glthread_LOCK_MUTEX(name)           u_mutex_lock(name)
-#define _glthread_UNLOCK_MUTEX(name)         u_mutex_unlock(name)
-
-#define _glthread_InitTSD(tsd)               u_tsd_init(tsd);
-#define _glthread_DestroyTSD(tsd)            u_tsd_destroy(tsd);
-#define _glthread_GetTSD(tsd)                u_tsd_get(tsd);
-#define _glthread_SetTSD(tsd, ptr)           u_tsd_set(tsd, ptr);
-
-typedef struct u_tsd _glthread_TSD;
-typedef u_mutex _glthread_Mutex;
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /* GLTHREAD_H */
diff --git a/mesalib/src/mesa/Makefile.sources b/mesalib/src/mesa/Makefile.sources
index bd02d3ee8..f6f4062a4 100644
--- a/mesalib/src/mesa/Makefile.sources
+++ b/mesalib/src/mesa/Makefile.sources
@@ -69,6 +69,7 @@ MAIN_FILES = \
 	$(SRCDIR)main/pack.c \
 	$(SRCDIR)main/pbo.c \
 	$(SRCDIR)main/performance_monitor.c \
+	$(SRCDIR)main/pipelineobj.c \
 	$(SRCDIR)main/pixel.c \
 	$(SRCDIR)main/pixelstore.c \
 	$(SRCDIR)main/pixeltransfer.c \
@@ -319,6 +320,7 @@ SPARC_FILES =			\
 
 COMMON_DRIVER_FILES =			\
 	$(SRCDIR)drivers/common/driverfuncs.c	\
+	$(SRCDIR)drivers/common/meta_blit.c	\
 	$(SRCDIR)drivers/common/meta.c
 
 
diff --git a/mesalib/src/mesa/SConscript b/mesalib/src/mesa/SConscript
index b52bbdc23..7f4e41983 100644
--- a/mesalib/src/mesa/SConscript
+++ b/mesalib/src/mesa/SConscript
@@ -98,6 +98,7 @@ main_sources = [
     'main/pack.c',
     'main/pbo.c',
     'main/performance_monitor.c',
+    'main/pipelineobj.c',
     'main/pixel.c',
     'main/pixelstore.c',
     'main/pixeltransfer.c',
@@ -325,6 +326,7 @@ program_sources = [
 common_driver_sources =	[
     'drivers/common/driverfuncs.c',
     'drivers/common/meta.c',
+    'drivers/common/meta_blit.c'
 ]
 
 mesa_sources = (
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index f12bcaab1..cde34f99f 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -51,6 +51,8 @@
 #include "main/macros.h"
 #include "main/matrix.h"
 #include "main/mipmap.h"
+#include "main/multisample.h"
+#include "main/objectlabel.h"
 #include "main/pixel.h"
 #include "main/pbo.h"
 #include "main/polygon.h"
@@ -82,283 +84,18 @@
 /** Return offset in bytes of the field within a vertex struct */
 #define OFFSET(FIELD) ((void *) offsetof(struct vertex, FIELD))
 
-/**
- * State which we may save/restore across meta ops.
- * XXX this may be incomplete...
- */
-struct save_state
-{
-   GLbitfield SavedState;  /**< bitmask of MESA_META_* flags */
-
-   /** MESA_META_CLEAR (and others?) */
-   struct gl_query_object *CurrentOcclusionObject;
-
-   /** MESA_META_ALPHA_TEST */
-   GLboolean AlphaEnabled;
-   GLenum AlphaFunc;
-   GLclampf AlphaRef;
-
-   /** MESA_META_BLEND */
-   GLbitfield BlendEnabled;
-   GLboolean ColorLogicOpEnabled;
-
-   /** MESA_META_COLOR_MASK */
-   GLubyte ColorMask[MAX_DRAW_BUFFERS][4];
-
-   /** MESA_META_DEPTH_TEST */
-   struct gl_depthbuffer_attrib Depth;
-
-   /** MESA_META_FOG */
-   GLboolean Fog;
-
-   /** MESA_META_PIXEL_STORE */
-   struct gl_pixelstore_attrib Pack, Unpack;
-
-   /** MESA_META_PIXEL_TRANSFER */
-   GLfloat RedBias, RedScale;
-   GLfloat GreenBias, GreenScale;
-   GLfloat BlueBias, BlueScale;
-   GLfloat AlphaBias, AlphaScale;
-   GLfloat DepthBias, DepthScale;
-   GLboolean MapColorFlag;
-
-   /** MESA_META_RASTERIZATION */
-   GLenum FrontPolygonMode, BackPolygonMode;
-   GLboolean PolygonOffset;
-   GLboolean PolygonSmooth;
-   GLboolean PolygonStipple;
-   GLboolean PolygonCull;
-
-   /** MESA_META_SCISSOR */
-   struct gl_scissor_attrib Scissor;
-
-   /** MESA_META_SHADER */
-   GLboolean VertexProgramEnabled;
-   struct gl_vertex_program *VertexProgram;
-   GLboolean FragmentProgramEnabled;
-   struct gl_fragment_program *FragmentProgram;
-   GLboolean ATIFragmentShaderEnabled;
-   struct gl_shader_program *Shader[MESA_SHADER_STAGES];
-   struct gl_shader_program *ActiveShader;
-
-   /** MESA_META_STENCIL_TEST */
-   struct gl_stencil_attrib Stencil;
-
-   /** MESA_META_TRANSFORM */
-   GLenum MatrixMode;
-   GLfloat ModelviewMatrix[16];
-   GLfloat ProjectionMatrix[16];
-   GLfloat TextureMatrix[16];
-
-   /** MESA_META_CLIP */
-   GLbitfield ClipPlanesEnabled;
-
-   /** MESA_META_TEXTURE */
-   GLuint ActiveUnit;
-   GLuint ClientActiveUnit;
-   /** for unit[0] only */
-   struct gl_texture_object *CurrentTexture[NUM_TEXTURE_TARGETS];
-   /** mask of TEXTURE_2D_BIT, etc */
-   GLbitfield TexEnabled[MAX_TEXTURE_UNITS];
-   GLbitfield TexGenEnabled[MAX_TEXTURE_UNITS];
-   GLuint EnvMode;  /* unit[0] only */
-
-   /** MESA_META_VERTEX */
-   struct gl_vertex_array_object *VAO;
-   struct gl_buffer_object *ArrayBufferObj;
-
-   /** MESA_META_VIEWPORT */
-   GLfloat ViewportX, ViewportY, ViewportW, ViewportH;
-   GLclampd DepthNear, DepthFar;
-
-   /** MESA_META_CLAMP_FRAGMENT_COLOR */
-   GLenum ClampFragmentColor;
-
-   /** MESA_META_CLAMP_VERTEX_COLOR */
-   GLenum ClampVertexColor;
-
-   /** MESA_META_CONDITIONAL_RENDER */
-   struct gl_query_object *CondRenderQuery;
-   GLenum CondRenderMode;
-
-   /** MESA_META_SELECT_FEEDBACK */
-   GLenum RenderMode;
-   struct gl_selection Select;
-   struct gl_feedback Feedback;
-
-   /** MESA_META_MULTISAMPLE */
-   GLboolean MultisampleEnabled;
-
-   /** MESA_META_FRAMEBUFFER_SRGB */
-   GLboolean sRGBEnabled;
-
-   /** Miscellaneous (always disabled) */
-   GLboolean Lighting;
-   GLboolean RasterDiscard;
-   GLboolean TransformFeedbackNeedsResume;
-};
-
-/**
- * Temporary texture used for glBlitFramebuffer, glDrawPixels, etc.
- * This is currently shared by all the meta ops.  But we could create a
- * separate one for each of glDrawPixel, glBlitFramebuffer, glCopyPixels, etc.
- */
-struct temp_texture
-{
-   GLuint TexObj;
-   GLenum Target;         /**< GL_TEXTURE_2D or GL_TEXTURE_RECTANGLE */
-   GLsizei MinSize;       /**< Min texture size to allocate */
-   GLsizei MaxSize;       /**< Max possible texture size */
-   GLboolean NPOT;        /**< Non-power of two size OK? */
-   GLsizei Width, Height; /**< Current texture size */
-   GLenum IntFormat;
-   GLfloat Sright, Ttop;  /**< right, top texcoords */
-};
-
-
-/**
- * State for glBlitFramebufer()
- */
-struct blit_state
-{
-   GLuint VAO;
-   GLuint VBO;
-   GLuint DepthFP;
-   GLuint ShaderProg;
-   GLuint RectShaderProg;
-   struct temp_texture depthTex;
-};
-
-
-/**
- * State for glClear()
- */
-struct clear_state
-{
-   GLuint VAO;
-   GLuint VBO;
-   GLuint ShaderProg;
-   GLint ColorLocation;
-   GLint LayerLocation;
-
-   GLuint IntegerShaderProg;
-   GLint IntegerColorLocation;
-   GLint IntegerLayerLocation;
-};
+static struct blit_shader *
+choose_blit_shader(GLenum target, struct blit_shader_table *table);
 
-
-/**
- * State for glCopyPixels()
- */
-struct copypix_state
-{
-   GLuint VAO;
-   GLuint VBO;
-};
-
-
-/**
- * State for glDrawPixels()
- */
-struct drawpix_state
-{
-   GLuint VAO;
-
-   GLuint StencilFP;  /**< Fragment program for drawing stencil images */
-   GLuint DepthFP;  /**< Fragment program for drawing depth images */
-};
-
-
-/**
- * State for glBitmap()
- */
-struct bitmap_state
-{
-   GLuint VAO;
-   GLuint VBO;
-   struct temp_texture Tex;  /**< separate texture from other meta ops */
-};
-
-/**
- * State for GLSL texture sampler which is used to generate fragment
- * shader in _mesa_meta_generate_mipmap().
- */
-struct glsl_sampler {
-   const char *type;
-   const char *func;
-   const char *texcoords;
-   GLuint shader_prog;
-};
-
-/**
- * State for _mesa_meta_generate_mipmap()
- */
-struct gen_mipmap_state
-{
-   GLuint VAO;
-   GLuint VBO;
-   GLuint FBO;
-   GLuint Sampler;
-   GLuint ShaderProg;
-   struct glsl_sampler sampler_1d;
-   struct glsl_sampler sampler_2d;
-   struct glsl_sampler sampler_3d;
-   struct glsl_sampler sampler_cubemap;
-   struct glsl_sampler sampler_1d_array;
-   struct glsl_sampler sampler_2d_array;
-};
-
-/**
- * State for texture decompression
- */
-struct decompress_state
-{
-   GLuint VAO;
-   GLuint VBO, FBO, RBO, Sampler;
-   GLint Width, Height;
-};
-
-/**
- * State for glDrawTex()
- */
-struct drawtex_state
-{
-   GLuint VAO;
-   GLuint VBO;
-};
-
-#define MAX_META_OPS_DEPTH      8
-/**
- * All per-context meta state.
- */
-struct gl_meta_state
-{
-   /** Stack of state saved during meta-ops */
-   struct save_state Save[MAX_META_OPS_DEPTH];
-   /** Save stack depth */
-   GLuint SaveStackDepth;
-
-   struct temp_texture TempTex;
-
-   struct blit_state Blit;    /**< For _mesa_meta_BlitFramebuffer() */
-   struct clear_state Clear;  /**< For _mesa_meta_Clear() */
-   struct copypix_state CopyPix;  /**< For _mesa_meta_CopyPixels() */
-   struct drawpix_state DrawPix;  /**< For _mesa_meta_DrawPixels() */
-   struct bitmap_state Bitmap;    /**< For _mesa_meta_Bitmap() */
-   struct gen_mipmap_state Mipmap;    /**< For _mesa_meta_GenerateMipmap() */
-   struct decompress_state Decompress;  /**< For texture decompression */
-   struct drawtex_state DrawTex;  /**< For _mesa_meta_DrawTex() */
-};
-
-static void meta_glsl_blit_cleanup(struct blit_state *blit);
 static void cleanup_temp_texture(struct temp_texture *tex);
 static void meta_glsl_clear_cleanup(struct clear_state *clear);
 static void meta_glsl_generate_mipmap_cleanup(struct gen_mipmap_state *mipmap);
 static void meta_decompress_cleanup(struct decompress_state *decompress);
 static void meta_drawpix_cleanup(struct drawpix_state *drawpix);
 
-static GLuint
-compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB *source)
+GLuint
+_mesa_meta_compile_shader_with_debug(struct gl_context *ctx, GLenum target,
+                                     const GLcharARB *source)
 {
    GLuint shader;
    GLint ok, size;
@@ -384,7 +121,7 @@ compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB
       return 0;
    }
 
-   _mesa_GetProgramInfoLog(shader, size, NULL, info);
+   _mesa_GetShaderInfoLog(shader, size, NULL, info);
    _mesa_problem(ctx,
 		 "meta program compile failed:\n%s\n"
 		 "source:\n%s\n",
@@ -396,8 +133,8 @@ compile_shader_with_debug(struct gl_context *ctx, GLenum target, const GLcharARB
    return 0;
 }
 
-static GLuint
-link_program_with_debug(struct gl_context *ctx, GLuint program)
+GLuint
+_mesa_meta_link_program_with_debug(struct gl_context *ctx, GLuint program)
 {
    GLint ok, size;
    GLchar *info;
@@ -425,6 +162,182 @@ link_program_with_debug(struct gl_context *ctx, GLuint program)
 }
 
 /**
+ * Generate a generic shader to blit from a texture to a framebuffer
+ *
+ * \param ctx       Current GL context
+ * \param texTarget Texture target that will be the source of the blit
+ *
+ * \returns a handle to a shader program on success or zero on failure.
+ */
+void
+_mesa_meta_setup_blit_shader(struct gl_context *ctx,
+                             GLenum target,
+                             struct blit_shader_table *table)
+{
+   const char *vs_source;
+   char *fs_source;
+   GLuint vs, fs;
+   void *const mem_ctx = ralloc_context(NULL);
+   struct blit_shader *shader = choose_blit_shader(target, table);
+   char *name;
+
+   assert(shader != NULL);
+
+   if (shader->shader_prog != 0) {
+      _mesa_UseProgram(shader->shader_prog);
+      return;
+   }
+
+   if (ctx->Const.GLSLVersion < 130) {
+      vs_source =
+         "attribute vec2 position;\n"
+         "attribute vec4 textureCoords;\n"
+         "varying vec4 texCoords;\n"
+         "void main()\n"
+         "{\n"
+         "   texCoords = textureCoords;\n"
+         "   gl_Position = vec4(position, 0.0, 1.0);\n"
+         "}\n";
+
+      fs_source = ralloc_asprintf(mem_ctx,
+                                  "#extension GL_EXT_texture_array : enable\n"
+                                  "#extension GL_ARB_texture_cube_map_array: enable\n"
+                                  "uniform %s texSampler;\n"
+                                  "varying vec4 texCoords;\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   gl_FragColor = %s(texSampler, %s);\n"
+                                  "   gl_FragDepth = gl_FragColor.x;\n"
+                                  "}\n",
+                                  shader->type,
+                                  shader->func, shader->texcoords);
+   }
+   else {
+      vs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "in vec2 position;\n"
+                                  "in vec4 textureCoords;\n"
+                                  "out vec4 texCoords;\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   texCoords = textureCoords;\n"
+                                  "   gl_Position = vec4(position, 0.0, 1.0);\n"
+                                  "}\n");
+      fs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "#extension GL_ARB_texture_cube_map_array: enable\n"
+                                  "uniform %s texSampler;\n"
+                                  "in vec4 texCoords;\n"
+                                  "out vec4 out_color;\n"
+                                  "\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   out_color = texture(texSampler, %s);\n"
+                                  "   gl_FragDepth = out_color.x;\n"
+                                  "}\n",
+                                  shader->type,
+                                  shader->texcoords);
+   }
+
+   vs = _mesa_meta_compile_shader_with_debug(ctx, GL_VERTEX_SHADER, vs_source);
+   fs = _mesa_meta_compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER, fs_source);
+
+   shader->shader_prog = _mesa_CreateProgramObjectARB();
+   _mesa_AttachShader(shader->shader_prog, fs);
+   _mesa_DeleteObjectARB(fs);
+   _mesa_AttachShader(shader->shader_prog, vs);
+   _mesa_DeleteObjectARB(vs);
+   _mesa_BindAttribLocation(shader->shader_prog, 0, "position");
+   _mesa_BindAttribLocation(shader->shader_prog, 1, "texcoords");
+   _mesa_meta_link_program_with_debug(ctx, shader->shader_prog);
+   name = ralloc_asprintf(mem_ctx, "%s blit", shader->type);
+   _mesa_ObjectLabel(GL_PROGRAM, shader->shader_prog, -1, name);
+   ralloc_free(mem_ctx);
+
+   _mesa_UseProgram(shader->shader_prog);
+}
+
+/**
+ * Configure vertex buffer and vertex array objects for tests
+ *
+ * Regardless of whether a new VAO and new VBO are created, the objects
+ * referenced by \c VAO and \c VBO will be bound into the GL state vector
+ * when this function terminates.
+ *
+ * \param VAO       Storage for vertex array object handle.  If 0, a new VAO
+ *                  will be created.
+ * \param VBO       Storage for vertex buffer object handle.  If 0, a new VBO
+ *                  will be created.  The new VBO will have storage for 4
+ *                  \c vertex structures.
+ * \param use_generic_attributes  Should generic attributes 0 and 1 be used,
+ *                  or should traditional, fixed-function color and texture
+ *                  coordinate be used?
+ * \param vertex_size  Number of components for attribute 0 / vertex.
+ * \param texcoord_size  Number of components for attribute 1 / texture
+ *                  coordinate.  If this is 0, attribute 1 will not be set or
+ *                  enabled.
+ * \param color_size  Number of components for attribute 1 / primary color.
+ *                  If this is 0, attribute 1 will not be set or enabled.
+ *
+ * \note If \c use_generic_attributes is \c true, \c color_size must be zero.
+ * Use \c texcoord_size instead.
+ */
+void
+_mesa_meta_setup_vertex_objects(GLuint *VAO, GLuint *VBO,
+                                bool use_generic_attributes,
+                                unsigned vertex_size, unsigned texcoord_size,
+                                unsigned color_size)
+{
+   if (*VAO == 0) {
+      assert(*VBO == 0);
+
+      /* create vertex array object */
+      _mesa_GenVertexArrays(1, VAO);
+      _mesa_BindVertexArray(*VAO);
+
+      /* create vertex array buffer */
+      _mesa_GenBuffers(1, VBO);
+      _mesa_BindBuffer(GL_ARRAY_BUFFER, *VBO);
+      _mesa_BufferData(GL_ARRAY_BUFFER, 4 * sizeof(struct vertex), NULL,
+                       GL_DYNAMIC_DRAW);
+
+      /* setup vertex arrays */
+      if (use_generic_attributes) {
+         assert(color_size == 0);
+
+         _mesa_VertexAttribPointer(0, vertex_size, GL_FLOAT, GL_FALSE,
+                                   sizeof(struct vertex), OFFSET(x));
+         _mesa_EnableVertexAttribArray(0);
+
+         if (texcoord_size > 0) {
+            _mesa_VertexAttribPointer(1, texcoord_size, GL_FLOAT, GL_FALSE,
+                                      sizeof(struct vertex), OFFSET(tex));
+            _mesa_EnableVertexAttribArray(1);
+         }
+      } else {
+         _mesa_VertexPointer(vertex_size, GL_FLOAT, sizeof(struct vertex),
+                             OFFSET(x));
+         _mesa_EnableClientState(GL_VERTEX_ARRAY);
+
+         if (texcoord_size > 0) {
+            _mesa_TexCoordPointer(texcoord_size, GL_FLOAT,
+                                  sizeof(struct vertex), OFFSET(tex));
+            _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
+         }
+
+         if (color_size > 0) {
+            _mesa_ColorPointer(color_size, GL_FLOAT,
+                               sizeof(struct vertex), OFFSET(r));
+            _mesa_EnableClientState(GL_COLOR_ARRAY);
+         }
+      }
+   } else {
+      _mesa_BindVertexArray(*VAO);
+      _mesa_BindBuffer(GL_ARRAY_BUFFER, *VBO);
+   }
+}
+
+/**
  * Initialize meta-ops for a context.
  * To be called once during context creation.
  */
@@ -446,7 +359,7 @@ _mesa_meta_free(struct gl_context *ctx)
 {
    GET_CURRENT_CONTEXT(old_context);
    _mesa_make_current(ctx, NULL, NULL);
-   meta_glsl_blit_cleanup(&ctx->Meta->Blit);
+   _mesa_meta_glsl_blit_cleanup(&ctx->Meta->Blit);
    meta_glsl_clear_cleanup(&ctx->Meta->Clear);
    meta_glsl_generate_mipmap_cleanup(&ctx->Meta->Mipmap);
    cleanup_temp_texture(&ctx->Meta->TempTex);
@@ -480,6 +393,13 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
    memset(save, 0, sizeof(*save));
    save->SavedState = state;
 
+   /* We always push into desktop GL mode and pop out at the end.  No sense in
+    * writing our shaders varying based on the user's context choice, when
+    * Mesa can handle either.
+    */
+   save->API = ctx->API;
+   ctx->API = API_OPENGL_COMPAT;
+
    /* Pausing transform feedback needs to be done early, or else we won't be
     * able to change other state.
     */
@@ -539,9 +459,7 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
          _mesa_set_enable(ctx, GL_DEPTH_TEST, GL_FALSE);
    }
 
-   if ((state & MESA_META_FOG)
-       && ctx->API != API_OPENGL_CORE
-       && ctx->API != API_OPENGLES2) {
+   if (state & MESA_META_FOG) {
       save->Fog = ctx->Fog.Enabled;
       if (ctx->Fog.Enabled)
          _mesa_set_enable(ctx, GL_FOG, GL_FALSE);
@@ -586,10 +504,8 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
       save->PolygonCull = ctx->Polygon.CullFlag;
       _mesa_PolygonMode(GL_FRONT_AND_BACK, GL_FILL);
       _mesa_set_enable(ctx, GL_POLYGON_OFFSET_FILL, GL_FALSE);
-      if (ctx->API == API_OPENGL_COMPAT) {
-         _mesa_set_enable(ctx, GL_POLYGON_SMOOTH, GL_FALSE);
-         _mesa_set_enable(ctx, GL_POLYGON_STIPPLE, GL_FALSE);
-      }
+      _mesa_set_enable(ctx, GL_POLYGON_SMOOTH, GL_FALSE);
+      _mesa_set_enable(ctx, GL_POLYGON_STIPPLE, GL_FALSE);
       _mesa_set_enable(ctx, GL_CULL_FACE, GL_FALSE);
    }
 
@@ -601,21 +517,21 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
    if (state & MESA_META_SHADER) {
       int i;
 
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ARB_vertex_program) {
+      if (ctx->Extensions.ARB_vertex_program) {
          save->VertexProgramEnabled = ctx->VertexProgram.Enabled;
          _mesa_reference_vertprog(ctx, &save->VertexProgram,
 				  ctx->VertexProgram.Current);
          _mesa_set_enable(ctx, GL_VERTEX_PROGRAM_ARB, GL_FALSE);
       }
 
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ARB_fragment_program) {
+      if (ctx->Extensions.ARB_fragment_program) {
          save->FragmentProgramEnabled = ctx->FragmentProgram.Enabled;
          _mesa_reference_fragprog(ctx, &save->FragmentProgram,
 				  ctx->FragmentProgram.Current);
          _mesa_set_enable(ctx, GL_FRAGMENT_PROGRAM_ARB, GL_FALSE);
       }
 
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ATI_fragment_shader) {
+      if (ctx->Extensions.ATI_fragment_shader) {
          save->ATIFragmentShaderEnabled = ctx->ATIFragmentShader.Enabled;
          _mesa_set_enable(ctx, GL_FRAGMENT_SHADER_ATI, GL_FALSE);
       }
@@ -645,33 +561,24 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
       save->EnvMode = ctx->Texture.Unit[0].EnvMode;
 
       /* Disable all texture units */
-      if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES) {
-         for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
-            save->TexEnabled[u] = ctx->Texture.Unit[u].Enabled;
-            save->TexGenEnabled[u] = ctx->Texture.Unit[u].TexGenEnabled;
-            if (ctx->Texture.Unit[u].Enabled ||
-                ctx->Texture.Unit[u].TexGenEnabled) {
-               _mesa_ActiveTexture(GL_TEXTURE0 + u);
-               _mesa_set_enable(ctx, GL_TEXTURE_2D, GL_FALSE);
-               if (ctx->Extensions.ARB_texture_cube_map)
-                  _mesa_set_enable(ctx, GL_TEXTURE_CUBE_MAP, GL_FALSE);
-               if (_mesa_is_gles(ctx) &&
-                   ctx->Extensions.OES_EGL_image_external)
-                  _mesa_set_enable(ctx, GL_TEXTURE_EXTERNAL_OES, GL_FALSE);
-
-               if (ctx->API == API_OPENGL_COMPAT) {
-                  _mesa_set_enable(ctx, GL_TEXTURE_1D, GL_FALSE);
-                  _mesa_set_enable(ctx, GL_TEXTURE_3D, GL_FALSE);
-                  if (ctx->Extensions.NV_texture_rectangle)
-                     _mesa_set_enable(ctx, GL_TEXTURE_RECTANGLE, GL_FALSE);
-                  _mesa_set_enable(ctx, GL_TEXTURE_GEN_S, GL_FALSE);
-                  _mesa_set_enable(ctx, GL_TEXTURE_GEN_T, GL_FALSE);
-                  _mesa_set_enable(ctx, GL_TEXTURE_GEN_R, GL_FALSE);
-                  _mesa_set_enable(ctx, GL_TEXTURE_GEN_Q, GL_FALSE);
-               } else {
-                  _mesa_set_enable(ctx, GL_TEXTURE_GEN_STR_OES, GL_FALSE);
-               }
-            }
+      for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
+         save->TexEnabled[u] = ctx->Texture.Unit[u].Enabled;
+         save->TexGenEnabled[u] = ctx->Texture.Unit[u].TexGenEnabled;
+         if (ctx->Texture.Unit[u].Enabled ||
+             ctx->Texture.Unit[u].TexGenEnabled) {
+            _mesa_ActiveTexture(GL_TEXTURE0 + u);
+            _mesa_set_enable(ctx, GL_TEXTURE_2D, GL_FALSE);
+            if (ctx->Extensions.ARB_texture_cube_map)
+               _mesa_set_enable(ctx, GL_TEXTURE_CUBE_MAP, GL_FALSE);
+
+            _mesa_set_enable(ctx, GL_TEXTURE_1D, GL_FALSE);
+            _mesa_set_enable(ctx, GL_TEXTURE_3D, GL_FALSE);
+            if (ctx->Extensions.NV_texture_rectangle)
+               _mesa_set_enable(ctx, GL_TEXTURE_RECTANGLE, GL_FALSE);
+            _mesa_set_enable(ctx, GL_TEXTURE_GEN_S, GL_FALSE);
+            _mesa_set_enable(ctx, GL_TEXTURE_GEN_T, GL_FALSE);
+            _mesa_set_enable(ctx, GL_TEXTURE_GEN_R, GL_FALSE);
+            _mesa_set_enable(ctx, GL_TEXTURE_GEN_Q, GL_FALSE);
          }
       }
 
@@ -684,9 +591,7 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
       /* set defaults for unit[0] */
       _mesa_ActiveTexture(GL_TEXTURE0);
       _mesa_ClientActiveTexture(GL_TEXTURE0);
-      if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES) {
-         _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
-      }
+      _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
    }
 
    if (state & MESA_META_TRANSFORM) {
@@ -799,9 +704,20 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
    }
 
    if (state & MESA_META_MULTISAMPLE) {
-      save->MultisampleEnabled = ctx->Multisample.Enabled;
+      save->Multisample = ctx->Multisample; /* struct copy */
+
       if (ctx->Multisample.Enabled)
          _mesa_set_multisample(ctx, GL_FALSE);
+      if (ctx->Multisample.SampleCoverage)
+         _mesa_set_enable(ctx, GL_SAMPLE_COVERAGE, GL_FALSE);
+      if (ctx->Multisample.SampleAlphaToCoverage)
+         _mesa_set_enable(ctx, GL_SAMPLE_ALPHA_TO_COVERAGE, GL_FALSE);
+      if (ctx->Multisample.SampleAlphaToOne)
+         _mesa_set_enable(ctx, GL_SAMPLE_ALPHA_TO_ONE, GL_FALSE);
+      if (ctx->Multisample.SampleShading)
+         _mesa_set_enable(ctx, GL_SAMPLE_SHADING, GL_FALSE);
+      if (ctx->Multisample.SampleMask)
+         _mesa_set_enable(ctx, GL_SAMPLE_MASK, GL_FALSE);
    }
 
    if (state & MESA_META_FRAMEBUFFER_SRGB) {
@@ -892,9 +808,7 @@ _mesa_meta_end(struct gl_context *ctx)
       _mesa_DepthMask(save->Depth.Mask);
    }
 
-   if ((state & MESA_META_FOG)
-       && ctx->API != API_OPENGL_CORE
-       && ctx->API != API_OPENGLES2) {
+   if (state & MESA_META_FOG) {
       _mesa_set_enable(ctx, GL_FOG, save->Fog);
    }
 
@@ -918,18 +832,10 @@ _mesa_meta_end(struct gl_context *ctx)
    }
 
    if (state & MESA_META_RASTERIZATION) {
-      /* Core context requires that front and back mode be the same.
-       */
-      if (ctx->API == API_OPENGL_CORE) {
-         _mesa_PolygonMode(GL_FRONT_AND_BACK, save->FrontPolygonMode);
-      } else {
-         _mesa_PolygonMode(GL_FRONT, save->FrontPolygonMode);
-         _mesa_PolygonMode(GL_BACK, save->BackPolygonMode);
-      }
-      if (ctx->API == API_OPENGL_COMPAT) {
-         _mesa_set_enable(ctx, GL_POLYGON_STIPPLE, save->PolygonStipple);
-         _mesa_set_enable(ctx, GL_POLYGON_SMOOTH, save->PolygonSmooth);
-      }
+      _mesa_PolygonMode(GL_FRONT, save->FrontPolygonMode);
+      _mesa_PolygonMode(GL_BACK, save->BackPolygonMode);
+      _mesa_set_enable(ctx, GL_POLYGON_STIPPLE, save->PolygonStipple);
+      _mesa_set_enable(ctx, GL_POLYGON_SMOOTH, save->PolygonSmooth);
       _mesa_set_enable(ctx, GL_POLYGON_OFFSET_FILL, save->PolygonOffset);
       _mesa_set_enable(ctx, GL_CULL_FACE, save->PolygonCull);
    }
@@ -949,7 +855,7 @@ _mesa_meta_end(struct gl_context *ctx)
    }
 
    if (state & MESA_META_SHADER) {
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ARB_vertex_program) {
+      if (ctx->Extensions.ARB_vertex_program) {
          _mesa_set_enable(ctx, GL_VERTEX_PROGRAM_ARB,
                           save->VertexProgramEnabled);
          _mesa_reference_vertprog(ctx, &ctx->VertexProgram.Current, 
@@ -957,7 +863,7 @@ _mesa_meta_end(struct gl_context *ctx)
 	 _mesa_reference_vertprog(ctx, &save->VertexProgram, NULL);
       }
 
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ARB_fragment_program) {
+      if (ctx->Extensions.ARB_fragment_program) {
          _mesa_set_enable(ctx, GL_FRAGMENT_PROGRAM_ARB,
                           save->FragmentProgramEnabled);
          _mesa_reference_fragprog(ctx, &ctx->FragmentProgram.Current,
@@ -965,7 +871,7 @@ _mesa_meta_end(struct gl_context *ctx)
 	 _mesa_reference_fragprog(ctx, &save->FragmentProgram, NULL);
       }
 
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.ATI_fragment_shader) {
+      if (ctx->Extensions.ATI_fragment_shader) {
          _mesa_set_enable(ctx, GL_FRAGMENT_SHADER_ATI,
                           save->ATIFragmentShaderEnabled);
       }
@@ -996,7 +902,7 @@ _mesa_meta_end(struct gl_context *ctx)
 
       _mesa_set_enable(ctx, GL_STENCIL_TEST, stencil->Enabled);
       _mesa_ClearStencil(stencil->Clear);
-      if (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.EXT_stencil_two_side) {
+      if (ctx->Extensions.EXT_stencil_two_side) {
          _mesa_set_enable(ctx, GL_STENCIL_TEST_TWO_SIDE_EXT,
                           stencil->TestTwoSide);
          _mesa_ActiveStencilFaceEXT(stencil->ActiveFace
@@ -1028,9 +934,7 @@ _mesa_meta_end(struct gl_context *ctx)
       ASSERT(ctx->Texture.CurrentUnit == 0);
 
       /* restore texenv for unit[0] */
-      if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES) {
-         _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, save->EnvMode);
-      }
+      _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, save->EnvMode);
 
       /* restore texture objects for unit[0] only */
       for (tgt = 0; tgt < NUM_TEXTURE_TARGETS; tgt++) {
@@ -1043,17 +947,15 @@ _mesa_meta_end(struct gl_context *ctx)
       }
 
       /* Restore fixed function texture enables, texgen */
-      if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES) {
-         for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
-            if (ctx->Texture.Unit[u].Enabled != save->TexEnabled[u]) {
-               FLUSH_VERTICES(ctx, _NEW_TEXTURE);
-               ctx->Texture.Unit[u].Enabled = save->TexEnabled[u];
-            }
+      for (u = 0; u < ctx->Const.MaxTextureUnits; u++) {
+         if (ctx->Texture.Unit[u].Enabled != save->TexEnabled[u]) {
+            FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+            ctx->Texture.Unit[u].Enabled = save->TexEnabled[u];
+         }
 
-            if (ctx->Texture.Unit[u].TexGenEnabled != save->TexGenEnabled[u]) {
-               FLUSH_VERTICES(ctx, _NEW_TEXTURE);
-               ctx->Texture.Unit[u].TexGenEnabled = save->TexGenEnabled[u];
-            }
+         if (ctx->Texture.Unit[u].TexGenEnabled != save->TexGenEnabled[u]) {
+            FLUSH_VERTICES(ctx, _NEW_TEXTURE);
+            ctx->Texture.Unit[u].TexGenEnabled = save->TexGenEnabled[u];
          }
       }
 
@@ -1137,8 +1039,30 @@ _mesa_meta_end(struct gl_context *ctx)
    }
 
    if (state & MESA_META_MULTISAMPLE) {
-      if (ctx->Multisample.Enabled != save->MultisampleEnabled)
-         _mesa_set_multisample(ctx, save->MultisampleEnabled);
+      struct gl_multisample_attrib *ctx_ms = &ctx->Multisample;
+      struct gl_multisample_attrib *save_ms = &save->Multisample;
+
+      if (ctx_ms->Enabled != save_ms->Enabled)
+         _mesa_set_multisample(ctx, save_ms->Enabled);
+      if (ctx_ms->SampleCoverage != save_ms->SampleCoverage)
+         _mesa_set_enable(ctx, GL_SAMPLE_COVERAGE, save_ms->SampleCoverage);
+      if (ctx_ms->SampleAlphaToCoverage != save_ms->SampleAlphaToCoverage)
+         _mesa_set_enable(ctx, GL_SAMPLE_ALPHA_TO_COVERAGE, save_ms->SampleAlphaToCoverage);
+      if (ctx_ms->SampleAlphaToOne != save_ms->SampleAlphaToOne)
+         _mesa_set_enable(ctx, GL_SAMPLE_ALPHA_TO_ONE, save_ms->SampleAlphaToOne);
+      if (ctx_ms->SampleCoverageValue != save_ms->SampleCoverageValue ||
+          ctx_ms->SampleCoverageInvert != save_ms->SampleCoverageInvert) {
+         _mesa_SampleCoverage(save_ms->SampleCoverageValue,
+                              save_ms->SampleCoverageInvert);
+      }
+      if (ctx_ms->SampleShading != save_ms->SampleShading)
+         _mesa_set_enable(ctx, GL_SAMPLE_SHADING, save_ms->SampleShading);
+      if (ctx_ms->SampleMask != save_ms->SampleMask)
+         _mesa_set_enable(ctx, GL_SAMPLE_MASK, save_ms->SampleMask);
+      if (ctx_ms->SampleMaskValue != save_ms->SampleMaskValue)
+         _mesa_SampleMaski(0, save_ms->SampleMaskValue);
+      if (ctx_ms->MinSampleShadingValue != save_ms->MinSampleShadingValue)
+         _mesa_MinSampleShading(save_ms->MinSampleShadingValue);
    }
 
    if (state & MESA_META_FRAMEBUFFER_SRGB) {
@@ -1157,6 +1081,8 @@ _mesa_meta_end(struct gl_context *ctx)
       _mesa_ResumeTransformFeedback();
 
    ctx->Meta->SaveStackDepth--;
+
+   ctx->API = save->API;
 }
 
 
@@ -1224,8 +1150,8 @@ cleanup_temp_texture(struct temp_texture *tex)
  * Return pointer to temp_texture info for non-bitmap ops.
  * This does some one-time init if needed.
  */
-static struct temp_texture *
-get_temp_texture(struct gl_context *ctx)
+struct temp_texture *
+_mesa_meta_get_temp_texture(struct gl_context *ctx)
 {
    struct temp_texture *tex = &ctx->Meta->TempTex;
 
@@ -1258,8 +1184,8 @@ get_bitmap_temp_texture(struct gl_context *ctx)
  * Return pointer to depth temp_texture.
  * This does some one-time init if needed.
  */
-static struct temp_texture *
-get_temp_depth_texture(struct gl_context *ctx)
+struct temp_texture *
+_mesa_meta_get_temp_depth_texture(struct gl_context *ctx)
 {
    struct temp_texture *tex = &ctx->Meta->Blit.depthTex;
 
@@ -1279,9 +1205,9 @@ get_temp_depth_texture(struct gl_context *ctx)
  *
  * \return GL_TRUE if new texture is needed, GL_FALSE otherwise
  */
-static GLboolean
-alloc_texture(struct temp_texture *tex,
-              GLsizei width, GLsizei height, GLenum intFormat)
+GLboolean
+_mesa_meta_alloc_texture(struct temp_texture *tex,
+                         GLsizei width, GLsizei height, GLenum intFormat)
 {
    GLboolean newTex = GL_FALSE;
 
@@ -1332,19 +1258,22 @@ alloc_texture(struct temp_texture *tex,
 /**
  * Setup/load texture for glCopyPixels or glBlitFramebuffer.
  */
-static void
-setup_copypix_texture(struct gl_context *ctx,
-                      struct temp_texture *tex,
-                      GLboolean newTex,
-                      GLint srcX, GLint srcY,
-                      GLsizei width, GLsizei height, GLenum intFormat,
-                      GLenum filter)
+void
+_mesa_meta_setup_copypix_texture(struct gl_context *ctx,
+                                 struct temp_texture *tex,
+                                 GLint srcX, GLint srcY,
+                                 GLsizei width, GLsizei height,
+                                 GLenum intFormat,
+                                 GLenum filter)
 {
+   bool newTex;
+
    _mesa_BindTexture(tex->Target, tex->TexObj);
    _mesa_TexParameteri(tex->Target, GL_TEXTURE_MIN_FILTER, filter);
    _mesa_TexParameteri(tex->Target, GL_TEXTURE_MAG_FILTER, filter);
-   if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES)
-      _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
+   _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
+
+   newTex = _mesa_meta_alloc_texture(tex, width, height, intFormat);
 
    /* copy framebuffer image to texture */
    if (newTex) {
@@ -1375,20 +1304,18 @@ setup_copypix_texture(struct gl_context *ctx,
 /**
  * Setup/load texture for glDrawPixels.
  */
-static void
-setup_drawpix_texture(struct gl_context *ctx,
-		      struct temp_texture *tex,
-                      GLboolean newTex,
-                      GLenum texIntFormat,
-                      GLsizei width, GLsizei height,
-                      GLenum format, GLenum type,
-                      const GLvoid *pixels)
+void
+_mesa_meta_setup_drawpix_texture(struct gl_context *ctx,
+                                 struct temp_texture *tex,
+                                 GLboolean newTex,
+                                 GLsizei width, GLsizei height,
+                                 GLenum format, GLenum type,
+                                 const GLvoid *pixels)
 {
    _mesa_BindTexture(tex->Target, tex->TexObj);
    _mesa_TexParameteri(tex->Target, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
    _mesa_TexParameteri(tex->Target, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
-   if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES)
-      _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
+   _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
 
    /* copy pixel data to texture */
    if (newTex) {
@@ -1422,617 +1349,24 @@ setup_drawpix_texture(struct gl_context *ctx,
    }
 }
 
-
-
-/**
- * One-time init for drawing depth pixels.
- */
-static void
-init_blit_depth_pixels(struct gl_context *ctx)
-{
-   static const char *program =
-      "!!ARBfp1.0\n"
-      "TEX result.depth, fragment.texcoord[0], texture[0], %s; \n"
-      "END \n";
-   char program2[200];
-   struct blit_state *blit = &ctx->Meta->Blit;
-   struct temp_texture *tex = get_temp_texture(ctx);
-   const char *texTarget;
-
-   assert(blit->DepthFP == 0);
-
-   /* replace %s with "RECT" or "2D" */
-   assert(strlen(program) + 4 < sizeof(program2));
-   if (tex->Target == GL_TEXTURE_RECTANGLE)
-      texTarget = "RECT";
-   else
-      texTarget = "2D";
-   _mesa_snprintf(program2, sizeof(program2), program, texTarget);
-
-   _mesa_GenProgramsARB(1, &blit->DepthFP);
-   _mesa_BindProgramARB(GL_FRAGMENT_PROGRAM_ARB, blit->DepthFP);
-   _mesa_ProgramStringARB(GL_FRAGMENT_PROGRAM_ARB, GL_PROGRAM_FORMAT_ASCII_ARB,
-                          strlen(program2), (const GLubyte *) program2);
-}
-
-static void
-setup_ff_blit_framebuffer(struct blit_state *blit)
+void
+_mesa_meta_setup_ff_tnl_for_blit(GLuint *VAO, GLuint *VBO,
+                                 unsigned texcoord_size)
 {
-   struct vertex {
-      GLfloat x, y, s, t;
-   };
-   struct vertex verts[4];
-
-   if (blit->VAO == 0) {
-      /* one-time setup */
-
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &blit->VAO);
-      _mesa_BindVertexArray(blit->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &blit->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, blit->VBO);
-      _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
-                          NULL, GL_DYNAMIC_DRAW_ARB);
-
-      /* setup vertex arrays */
-      _mesa_VertexPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_TexCoordPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(s));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-   }
+   _mesa_meta_setup_vertex_objects(VAO, VBO, false, 2, texcoord_size, 0);
 
    /* setup projection matrix */
    _mesa_MatrixMode(GL_PROJECTION);
    _mesa_LoadIdentity();
-   _mesa_Ortho(-1.0, 1.0, -1.0, 1.0, -1.0, 1.0);
-
-}
-
-static void
-setup_glsl_blit_framebuffer(struct gl_context *ctx,
-                            struct blit_state *blit,
-                            GLenum target)
-{
-   struct vertex {
-      GLfloat x, y, s, t;
-   };
-   struct vertex verts[4];
-   const char *vs_source;
-   char *fs_source;
-   GLuint vs, fs;
-   void *mem_ctx;
-   GLuint ShaderProg;
-   GLboolean texture_2d = (target == GL_TEXTURE_2D);
-
-   /* target = GL_TEXTURE_RECTANGLE is not supported in GLES 3.0 */
-   assert(_mesa_is_desktop_gl(ctx) || texture_2d);
-
-   /* Check if already initialized */
-   if (blit->VAO == 0) {
-
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &blit->VAO);
-      _mesa_BindVertexArray(blit->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &blit->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, blit->VBO);
-      _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
-                          NULL, GL_DYNAMIC_DRAW_ARB);
-
-      /* setup vertex arrays */
-      _mesa_VertexAttribPointer(0, 2, GL_FLOAT, GL_FALSE,
-                                   sizeof(struct vertex), OFFSET(x));
-      _mesa_VertexAttribPointer(1, 2, GL_FLOAT, GL_FALSE,
-                                   sizeof(struct vertex), OFFSET(s));
-
-      _mesa_EnableVertexAttribArray(0);
-      _mesa_EnableVertexAttribArray(1);
-   }
-
-   /* Generate a relevant fragment shader program for the texture target */
-   if ((target == GL_TEXTURE_2D && blit->ShaderProg != 0) ||
-       (target == GL_TEXTURE_RECTANGLE && blit->RectShaderProg != 0)) {
-      return;
-   }
-
-   mem_ctx = ralloc_context(NULL);
-
-   if (ctx->Const.GLSLVersion < 130) {
-      vs_source =
-         "attribute vec2 position;\n"
-         "attribute vec2 textureCoords;\n"
-         "varying vec2 texCoords;\n"
-         "void main()\n"
-         "{\n"
-         "   texCoords = textureCoords;\n"
-         "   gl_Position = vec4(position, 0.0, 1.0);\n"
-         "}\n";
-
-      fs_source = ralloc_asprintf(mem_ctx,
-                                  "#ifdef GL_ES\n"
-                                  "precision highp float;\n"
-                                  "#endif\n"
-                                  "uniform %s texSampler;\n"
-                                  "varying vec2 texCoords;\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   gl_FragColor = %s(texSampler, texCoords);\n"
-                                  "   gl_FragDepth = gl_FragColor.r;\n"
-                                  "}\n",
-                                  texture_2d ? "sampler2D" : "sampler2DRect",
-                                  texture_2d ? "texture2D" : "texture2DRect");
-   }
-   else {
-      vs_source = ralloc_asprintf(mem_ctx,
-                                  "#version %s\n"
-                                  "in vec2 position;\n"
-                                  "in vec2 textureCoords;\n"
-                                  "out vec2 texCoords;\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   texCoords = textureCoords;\n"
-                                  "   gl_Position = vec4(position, 0.0, 1.0);\n"
-                                  "}\n",
-                                  _mesa_is_desktop_gl(ctx) ? "130" : "300 es");
-      fs_source = ralloc_asprintf(mem_ctx,
-                                  "#version %s\n"
-                                  "#ifdef GL_ES\n"
-                                  "precision highp float;\n"
-                                  "#endif\n"
-                                  "uniform %s texSampler;\n"
-                                  "in vec2 texCoords;\n"
-                                  "out vec4 out_color;\n"
-                                  "\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   out_color = %s(texSampler, texCoords);\n"
-                                  "   gl_FragDepth = out_color.r;\n"
-                                  "}\n",
-                                  _mesa_is_desktop_gl(ctx) ? "130" : "300 es",
-                                  texture_2d ? "sampler2D" : "sampler2DRect",
-                                  texture_2d ? "texture" : "texture2DRect");
-   }
-
-   vs = compile_shader_with_debug(ctx, GL_VERTEX_SHADER, vs_source);
-   fs = compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER, fs_source);
-
-   ShaderProg = _mesa_CreateProgramObjectARB();
-   _mesa_AttachShader(ShaderProg, fs);
-   _mesa_DeleteObjectARB(fs);
-   _mesa_AttachShader(ShaderProg, vs);
-   _mesa_DeleteObjectARB(vs);
-   _mesa_BindAttribLocation(ShaderProg, 0, "position");
-   _mesa_BindAttribLocation(ShaderProg, 1, "texcoords");
-   link_program_with_debug(ctx, ShaderProg);
-   ralloc_free(mem_ctx);
-   if (texture_2d)
-      blit->ShaderProg = ShaderProg;
-   else
-      blit->RectShaderProg = ShaderProg;
 }
 
 /**
- * Try to do a glBlitFramebuffer using no-copy texturing.
- * We can do this when the src renderbuffer is actually a texture.
- * But if the src buffer == dst buffer we cannot do this.
- *
- * \return new buffer mask indicating the buffers left to blit using the
- *         normal path.
- */
-static GLbitfield
-blitframebuffer_texture(struct gl_context *ctx,
-                        GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
-                        GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,
-                        GLbitfield mask, GLenum filter, GLint flipX,
-                        GLint flipY, GLboolean glsl_version)
-{
-   if (mask & GL_COLOR_BUFFER_BIT) {
-      const struct gl_framebuffer *drawFb = ctx->DrawBuffer;
-      const struct gl_framebuffer *readFb = ctx->ReadBuffer;
-      const struct gl_renderbuffer_attachment *drawAtt;
-      const struct gl_renderbuffer_attachment *readAtt =
-         &readFb->Attachment[readFb->_ColorReadBufferIndex];
-
-      if (readAtt && readAtt->Texture) {
-         struct blit_state *blit = &ctx->Meta->Blit;
-         const GLint dstX = MIN2(dstX0, dstX1);
-         const GLint dstY = MIN2(dstY0, dstY1);
-         const GLint dstW = abs(dstX1 - dstX0);
-         const GLint dstH = abs(dstY1 - dstY0);
-         const struct gl_texture_object *texObj = readAtt->Texture;
-         const GLuint srcLevel = readAtt->TextureLevel;
-         const GLint baseLevelSave = texObj->BaseLevel;
-         const GLint maxLevelSave = texObj->MaxLevel;
-         const GLenum target = texObj->Target;
-         GLuint sampler, samplerSave =
-            ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler ?
-            ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler->Name : 0;
-         int i;
-
-         /* Iterate through all draw buffers */
-         for (i = 0; i < ctx->DrawBuffer->_NumColorDrawBuffers; i++) {
-            int idx = ctx->DrawBuffer->_ColorDrawBufferIndexes[i];
-            if (idx == -1)
-               continue;
-            drawAtt = &drawFb->Attachment[idx];
-
-            if (drawAtt->Texture == readAtt->Texture) {
-               /* Can't use same texture as both the source and dest.  We need
-                * to handle overlapping blits and besides, some hw may not
-                * support this.
-                */
-               return mask;
-            }
-         }
-
-         if (target != GL_TEXTURE_2D && target != GL_TEXTURE_RECTANGLE_ARB) {
-            /* Can't handle other texture types at this time */
-            return mask;
-         }
-
-         /* Choose between glsl version and fixed function version of
-          * BlitFramebuffer function.
-          */
-         if (glsl_version) {
-            setup_glsl_blit_framebuffer(ctx, blit, target);
-            if (target == GL_TEXTURE_2D)
-               _mesa_UseProgram(blit->ShaderProg);
-            else
-               _mesa_UseProgram(blit->RectShaderProg);
-         }
-         else {
-            setup_ff_blit_framebuffer(&ctx->Meta->Blit);
-         }
-
-         _mesa_BindVertexArray(blit->VAO);
-         _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, blit->VBO);
-
-         _mesa_GenSamplers(1, &sampler);
-         _mesa_BindSampler(ctx->Texture.CurrentUnit, sampler);
-
-         /*
-         printf("Blit from texture!\n");
-         printf("  srcAtt %p  dstAtt %p\n", readAtt, drawAtt);
-         printf("  srcTex %p  dstText %p\n", texObj, drawAtt->Texture);
-         */
-
-         /* Prepare src texture state */
-         _mesa_BindTexture(target, texObj->Name);
-         _mesa_SamplerParameteri(sampler, GL_TEXTURE_MIN_FILTER, filter);
-         _mesa_SamplerParameteri(sampler, GL_TEXTURE_MAG_FILTER, filter);
-         if (target != GL_TEXTURE_RECTANGLE_ARB) {
-            _mesa_TexParameteri(target, GL_TEXTURE_BASE_LEVEL, srcLevel);
-            _mesa_TexParameteri(target, GL_TEXTURE_MAX_LEVEL, srcLevel);
-         }
-         _mesa_SamplerParameteri(sampler, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-         _mesa_SamplerParameteri(sampler, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-
-	 /* Always do our blits with no sRGB decode or encode.  Note that
-          * GL_FRAMEBUFFER_SRGB has already been disabled by
-          * _mesa_meta_begin().
-          */
-	 if (ctx->Extensions.EXT_texture_sRGB_decode) {
-	    _mesa_SamplerParameteri(sampler, GL_TEXTURE_SRGB_DECODE_EXT,
-				GL_SKIP_DECODE_EXT);
-	 }
-
-         if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES) {
-            _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
-            _mesa_set_enable(ctx, target, GL_TRUE);
-	 }
-
-         /* Prepare vertex data (the VBO was previously created and bound) */
-         {
-            struct vertex {
-               GLfloat x, y, s, t;
-            };
-            struct vertex verts[4];
-            GLfloat s0, t0, s1, t1;
-
-            if (target == GL_TEXTURE_2D) {
-               const struct gl_texture_image *texImage
-                   = _mesa_select_tex_image(ctx, texObj, target, srcLevel);
-               s0 = srcX0 / (float) texImage->Width;
-               s1 = srcX1 / (float) texImage->Width;
-               t0 = srcY0 / (float) texImage->Height;
-               t1 = srcY1 / (float) texImage->Height;
-            }
-            else {
-               assert(target == GL_TEXTURE_RECTANGLE_ARB);
-               s0 = (float) srcX0;
-               s1 = (float) srcX1;
-               t0 = (float) srcY0;
-               t1 = (float) srcY1;
-            }
-
-            /* setup vertex positions */
-            verts[0].x = -1.0F * flipX;
-            verts[0].y = -1.0F * flipY;
-            verts[1].x =  1.0F * flipX;
-            verts[1].y = -1.0F * flipY;
-            verts[2].x =  1.0F * flipX;
-            verts[2].y =  1.0F * flipY;
-            verts[3].x = -1.0F * flipX;
-            verts[3].y =  1.0F * flipY;
-
-            verts[0].s = s0;
-            verts[0].t = t0;
-            verts[1].s = s1;
-            verts[1].t = t0;
-            verts[2].s = s1;
-            verts[2].t = t1;
-            verts[3].s = s0;
-            verts[3].t = t1;
-
-            _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
-         }
-
-         /* setup viewport */
-         _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
-         _mesa_ColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
-         _mesa_DepthMask(GL_FALSE);
-         _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
-
-         /* Restore texture object state, the texture binding will
-          * be restored by _mesa_meta_end().
-          */
-         if (target != GL_TEXTURE_RECTANGLE_ARB) {
-            _mesa_TexParameteri(target, GL_TEXTURE_BASE_LEVEL, baseLevelSave);
-            _mesa_TexParameteri(target, GL_TEXTURE_MAX_LEVEL, maxLevelSave);
-         }
-
-         _mesa_BindSampler(ctx->Texture.CurrentUnit, samplerSave);
-         _mesa_DeleteSamplers(1, &sampler);
-
-         /* Done with color buffer */
-         mask &= ~GL_COLOR_BUFFER_BIT;
-      }
-   }
-
-   return mask;
-}
-
-
-/**
- * Meta implementation of ctx->Driver.BlitFramebuffer() in terms
- * of texture mapping and polygon rendering.
- */
-void
-_mesa_meta_BlitFramebuffer(struct gl_context *ctx,
-                           GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
-                           GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,
-                           GLbitfield mask, GLenum filter)
-{
-   struct blit_state *blit = &ctx->Meta->Blit;
-   struct temp_texture *tex = get_temp_texture(ctx);
-   struct temp_texture *depthTex = get_temp_depth_texture(ctx);
-   const GLsizei maxTexSize = tex->MaxSize;
-   const GLint srcX = MIN2(srcX0, srcX1);
-   const GLint srcY = MIN2(srcY0, srcY1);
-   const GLint srcW = abs(srcX1 - srcX0);
-   const GLint srcH = abs(srcY1 - srcY0);
-   const GLint dstX = MIN2(dstX0, dstX1);
-   const GLint dstY = MIN2(dstY0, dstY1);
-   const GLint dstW = abs(dstX1 - dstX0);
-   const GLint dstH = abs(dstY1 - dstY0);
-   const GLint srcFlipX = (srcX1 - srcX0) / srcW;
-   const GLint srcFlipY = (srcY1 - srcY0) / srcH;
-   const GLint dstFlipX = (dstX1 - dstX0) / dstW;
-   const GLint dstFlipY = (dstY1 - dstY0) / dstH;
-   const GLint flipX = srcFlipX * dstFlipX;
-   const GLint flipY = srcFlipY * dstFlipY;
-
-   struct vertex {
-      GLfloat x, y, s, t;
-   };
-   struct vertex verts[4];
-   GLboolean newTex;
-   const GLboolean use_glsl_version = ctx->Extensions.ARB_vertex_shader &&
-                                      ctx->Extensions.ARB_fragment_shader &&
-                                      (ctx->API != API_OPENGLES);
-
-   /* In addition to falling back if the blit size is larger than the maximum
-    * texture size, fallback if the source is multisampled.  This fallback can
-    * be removed once Mesa gets support ARB_texture_multisample.
-    */
-   if (srcW > maxTexSize || srcH > maxTexSize
-       || ctx->ReadBuffer->Visual.samples > 0) {
-      /* XXX avoid this fallback */
-      _swrast_BlitFramebuffer(ctx, srcX0, srcY0, srcX1, srcY1,
-                              dstX0, dstY0, dstX1, dstY1, mask, filter);
-      return;
-   }
-
-   /* only scissor effects blit so save/clear all other relevant state */
-   _mesa_meta_begin(ctx, ~MESA_META_SCISSOR);
-
-   /* Try faster, direct texture approach first */
-   mask = blitframebuffer_texture(ctx, srcX0, srcY0, srcX1, srcY1,
-                                  dstX0, dstY0, dstX1, dstY1, mask, filter,
-                                  dstFlipX, dstFlipY, use_glsl_version);
-   if (mask == 0x0) {
-      _mesa_meta_end(ctx);
-      return;
-   }
-
-   /* Choose between glsl version and fixed function version of
-    * BlitFramebuffer function.
-    */
-   if (use_glsl_version) {
-      setup_glsl_blit_framebuffer(ctx, blit, tex->Target);
-      if (tex->Target == GL_TEXTURE_2D)
-         _mesa_UseProgram(blit->ShaderProg);
-      else
-         _mesa_UseProgram(blit->RectShaderProg);
-   }
-   else {
-      setup_ff_blit_framebuffer(blit);
-   }
-
-   _mesa_BindVertexArray(blit->VAO);
-   _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, blit->VBO);
-
-   /* Continue with "normal" approach which involves copying the src rect
-    * into a temporary texture and is "blitted" by drawing a textured quad.
-    */
-   {
-      /* setup vertex positions */
-      verts[0].x = -1.0F * flipX;
-      verts[0].y = -1.0F * flipY;
-      verts[1].x =  1.0F * flipX;
-      verts[1].y = -1.0F * flipY;
-      verts[2].x =  1.0F * flipX;
-      verts[2].y =  1.0F * flipY;
-      verts[3].x = -1.0F * flipX;
-      verts[3].y =  1.0F * flipY;
-
-   }
-
-   /* glEnable() in gles2 and gles3 doesn't allow GL_TEXTURE_{1D, 2D, etc.}
-    * tokens.
-    */
-   if (_mesa_is_desktop_gl(ctx) || ctx->API == API_OPENGLES)
-      _mesa_set_enable(ctx, tex->Target, GL_TRUE);
-
-   if (mask & GL_COLOR_BUFFER_BIT) {
-      const struct gl_framebuffer *readFb = ctx->ReadBuffer;
-      const struct gl_renderbuffer *colorReadRb = readFb->_ColorReadBuffer;
-      const GLenum rb_base_format =
-         _mesa_base_tex_format(ctx, colorReadRb->InternalFormat);
-
-      /* Using  the exact source rectangle to create the texture does incorrect
-       * linear filtering along the edges. So, allocate the texture extended along
-       * edges by one pixel in x, y directions.
-       */
-      newTex = alloc_texture(tex, srcW + 2, srcH + 2, rb_base_format);
-      setup_copypix_texture(ctx, tex, newTex,
-                            srcX - 1, srcY - 1, srcW + 2, srcH + 2,
-                            rb_base_format, filter);
-      /* texcoords (after texture allocation!) */
-      {
-         verts[0].s = 1.0F;
-         verts[0].t = 1.0F;
-         verts[1].s = tex->Sright - 1.0F;
-         verts[1].t = 1.0F;
-         verts[2].s = tex->Sright - 1.0F;
-         verts[2].t = tex->Ttop - 1.0F;
-         verts[3].s = 1.0F;
-         verts[3].t = tex->Ttop - 1.0F;
-
-         /* upload new vertex data */
-         _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
-      }
-
-      _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
-      _mesa_ColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
-      _mesa_set_enable(ctx, GL_DEPTH_TEST, GL_FALSE);
-      _mesa_DepthMask(GL_FALSE);
-      _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
-      mask &= ~GL_COLOR_BUFFER_BIT;
-   }
-
-   if ((mask & GL_DEPTH_BUFFER_BIT) &&
-       _mesa_is_desktop_gl(ctx) &&
-       ctx->Extensions.ARB_depth_texture &&
-       ctx->Extensions.ARB_fragment_program) {
-
-      GLuint *tmp = malloc(srcW * srcH * sizeof(GLuint));
-
-      if (tmp) {
-
-         newTex = alloc_texture(depthTex, srcW, srcH, GL_DEPTH_COMPONENT);
-         _mesa_ReadPixels(srcX, srcY, srcW, srcH, GL_DEPTH_COMPONENT,
-                          GL_UNSIGNED_INT, tmp);
-         setup_drawpix_texture(ctx, depthTex, newTex, GL_DEPTH_COMPONENT,
-                               srcW, srcH, GL_DEPTH_COMPONENT,
-                               GL_UNSIGNED_INT, tmp);
-
-         /* texcoords (after texture allocation!) */
-         {
-            verts[0].s = 0.0F;
-            verts[0].t = 0.0F;
-            verts[1].s = depthTex->Sright;
-            verts[1].t = 0.0F;
-            verts[2].s = depthTex->Sright;
-            verts[2].t = depthTex->Ttop;
-            verts[3].s = 0.0F;
-            verts[3].t = depthTex->Ttop;
-
-            /* upload new vertex data */
-            _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
-         }
-
-         if (!blit->DepthFP)
-            init_blit_depth_pixels(ctx);
-
-         _mesa_BindProgramARB(GL_FRAGMENT_PROGRAM_ARB, blit->DepthFP);
-         _mesa_set_enable(ctx, GL_FRAGMENT_PROGRAM_ARB, GL_TRUE);
-         _mesa_ColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
-         _mesa_set_enable(ctx, GL_DEPTH_TEST, GL_TRUE);
-         _mesa_DepthFunc(GL_ALWAYS);
-         _mesa_DepthMask(GL_TRUE);
-
-         _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
-         _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
-         _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
-         mask &= ~GL_DEPTH_BUFFER_BIT;
-
-         free(tmp);
-      }
-   }
-
-   if (mask & GL_STENCIL_BUFFER_BIT) {
-      /* XXX can't easily do stencil */
-   }
-
-   if (_mesa_is_desktop_gl(ctx) || ctx->API == API_OPENGLES)
-      _mesa_set_enable(ctx, tex->Target, GL_FALSE);
-
-   _mesa_meta_end(ctx);
-
-   if (mask) {
-      _swrast_BlitFramebuffer(ctx, srcX0, srcY0, srcX1, srcY1,
-                              dstX0, dstY0, dstX1, dstY1, mask, filter);
-   }
-}
-
-static void
-meta_glsl_blit_cleanup(struct blit_state *blit)
-{
-   if (blit->VAO) {
-      _mesa_DeleteVertexArrays(1, &blit->VAO);
-      blit->VAO = 0;
-      _mesa_DeleteBuffers(1, &blit->VBO);
-      blit->VBO = 0;
-   }
-   if (blit->DepthFP) {
-      _mesa_DeleteProgramsARB(1, &blit->DepthFP);
-      blit->DepthFP = 0;
-   }
-
-   _mesa_DeleteObjectARB(blit->ShaderProg);
-   blit->ShaderProg = 0;
-   _mesa_DeleteObjectARB(blit->RectShaderProg);
-   blit->RectShaderProg = 0;
-
-   _mesa_DeleteTextures(1, &blit->depthTex.TexObj);
-   blit->depthTex.TexObj = 0;
-}
-
-
-/**
  * Meta implementation of ctx->Driver.Clear() in terms of polygon rendering.
  */
 void
 _mesa_meta_Clear(struct gl_context *ctx, GLbitfield buffers)
 {
    struct clear_state *clear = &ctx->Meta->Clear;
-   struct vertex {
-      GLfloat x, y, z, r, g, b, a;
-   };
    struct vertex verts[4];
    /* save all state but scissor, pixel pack/unpack */
    GLbitfield metaSave = (MESA_META_ALL -
@@ -2049,27 +1383,7 @@ _mesa_meta_Clear(struct gl_context *ctx, GLbitfield buffers)
 
    _mesa_meta_begin(ctx, metaSave);
 
-   if (clear->VAO == 0) {
-      /* one-time setup */
-
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &clear->VAO);
-      _mesa_BindVertexArray(clear->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &clear->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, clear->VBO);
-
-      /* setup vertex arrays */
-      _mesa_VertexPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_ColorPointer(4, GL_FLOAT, sizeof(struct vertex), OFFSET(r));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_COLOR_ARRAY);
-   }
-   else {
-      _mesa_BindVertexArray(clear->VAO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, clear->VBO);
-   }
+   _mesa_meta_setup_vertex_objects(&clear->VAO, &clear->VBO, false, 3, 0, 4);
 
    /* GL_COLOR_BUFFER_BIT */
    if (buffers & BUFFER_BITS_COLOR) {
@@ -2171,9 +1485,6 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
       "  }\n"
       "}\n";
    const char *fs_source =
-      "#ifdef GL_ES\n"
-      "precision highp float;\n"
-      "#endif\n"
       "uniform vec4 color;\n"
       "void main()\n"
       "{\n"
@@ -2182,20 +1493,10 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
    GLuint vs, gs = 0, fs;
    bool has_integer_textures;
 
-   if (clear->VAO != 0)
-      return;
-
-   /* create vertex array object */
-   _mesa_GenVertexArrays(1, &clear->VAO);
-   _mesa_BindVertexArray(clear->VAO);
-
-   /* create vertex array buffer */
-   _mesa_GenBuffers(1, &clear->VBO);
-   _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, clear->VBO);
+   _mesa_meta_setup_vertex_objects(&clear->VAO, &clear->VBO, true, 3, 0, 0);
 
-   /* setup vertex arrays */
-   _mesa_VertexAttribPointer(0, 3, GL_FLOAT, GL_FALSE, 0, (void *)0);
-   _mesa_EnableVertexAttribArray(0);
+   if (clear->ShaderProg != 0)
+      return;
 
    vs = _mesa_CreateShaderObjectARB(GL_VERTEX_SHADER);
    _mesa_ShaderSource(vs, 1, &vs_source, NULL);
@@ -2235,30 +1536,27 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
       void *shader_source_mem_ctx = ralloc_context(NULL);
       const char *vs_int_source =
          ralloc_asprintf(shader_source_mem_ctx,
-                         "#version %s\n"
+                         "#version 130\n"
                          "in vec4 position;\n"
                          "void main()\n"
                          "{\n"
                          "   gl_Position = position;\n"
-                         "}\n",
-                         _mesa_is_desktop_gl(ctx) ? "130" : "300 es");
+                         "}\n");
       const char *fs_int_source =
          ralloc_asprintf(shader_source_mem_ctx,
-                         "#version %s\n"
-                         "#ifdef GL_ES\n"
-                         "precision highp float;\n"
-                         "#endif\n"
+                         "#version 130\n"
                          "uniform ivec4 color;\n"
                          "out ivec4 out_color;\n"
                          "\n"
                          "void main()\n"
                          "{\n"
                          "   out_color = color;\n"
-                         "}\n",
-                         _mesa_is_desktop_gl(ctx) ? "130" : "300 es");
+                         "}\n");
 
-      vs = compile_shader_with_debug(ctx, GL_VERTEX_SHADER, vs_int_source);
-      fs = compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER, fs_int_source);
+      vs = _mesa_meta_compile_shader_with_debug(ctx, GL_VERTEX_SHADER,
+                                                vs_int_source);
+      fs = _mesa_meta_compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER,
+                                                fs_int_source);
       ralloc_free(shader_source_mem_ctx);
 
       clear->IntegerShaderProg = _mesa_CreateProgramObjectARB();
@@ -2275,7 +1573,9 @@ meta_glsl_clear_init(struct gl_context *ctx, struct clear_state *clear)
        * BindFragDataLocation to 0.
        */
 
-      link_program_with_debug(ctx, clear->IntegerShaderProg);
+      _mesa_ObjectLabel(GL_PROGRAM, clear->IntegerShaderProg, -1,
+                        "integer clear");
+      _mesa_meta_link_program_with_debug(ctx, clear->IntegerShaderProg);
 
       clear->IntegerColorLocation =
 	 _mesa_GetUniformLocation(clear->IntegerShaderProg, "color");
@@ -2321,9 +1621,7 @@ _mesa_meta_glsl_Clear(struct gl_context *ctx, GLbitfield buffers)
    const float x1 = ((float)fb->_Xmax / fb->Width)  * 2.0f - 1.0f;
    const float y1 = ((float)fb->_Ymax / fb->Height) * 2.0f - 1.0f;
    const float z = -invert_z(ctx->Depth.Clear);
-   struct vertex {
-      GLfloat x, y, z;
-   } verts[4];
+   struct vertex verts[4];
 
    metaSave = (MESA_META_ALPHA_TEST |
 	       MESA_META_BLEND |
@@ -2359,9 +1657,6 @@ _mesa_meta_glsl_Clear(struct gl_context *ctx, GLbitfield buffers)
 			  ctx->Color.ClearColor.f);
    }
 
-   _mesa_BindVertexArray(clear->VAO);
-   _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, clear->VBO);
-
    /* GL_COLOR_BUFFER_BIT */
    if (buffers & BUFFER_BITS_COLOR) {
       /* leave colormask, glDrawBuffer state as-is */
@@ -2443,13 +1738,8 @@ _mesa_meta_CopyPixels(struct gl_context *ctx, GLint srcX, GLint srcY,
                       GLint dstX, GLint dstY, GLenum type)
 {
    struct copypix_state *copypix = &ctx->Meta->CopyPix;
-   struct temp_texture *tex = get_temp_texture(ctx);
-   struct vertex {
-      GLfloat x, y, z, s, t;
-   };
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
    struct vertex verts[4];
-   GLboolean newTex;
-   GLenum intFormat = GL_RGBA;
 
    if (type != GL_COLOR ||
        ctx->_ImageTransferState ||
@@ -2472,31 +1762,15 @@ _mesa_meta_CopyPixels(struct gl_context *ctx, GLint srcX, GLint srcY,
                           MESA_META_VERTEX |
                           MESA_META_VIEWPORT));
 
-   if (copypix->VAO == 0) {
-      /* one-time setup */
+   _mesa_meta_setup_vertex_objects(&copypix->VAO, &copypix->VBO, false,
+                                   3, 2, 0);
 
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &copypix->VAO);
-      _mesa_BindVertexArray(copypix->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &copypix->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, copypix->VBO);
-      _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
-                          NULL, GL_DYNAMIC_DRAW_ARB);
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
 
-      /* setup vertex arrays */
-      _mesa_VertexPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_TexCoordPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(s));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-   }
-   else {
-      _mesa_BindVertexArray(copypix->VAO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, copypix->VBO);
-   }
-
-   newTex = alloc_texture(tex, width, height, intFormat);
+   /* Alloc/setup texture */
+   _mesa_meta_setup_copypix_texture(ctx, tex, srcX, srcY, width, height,
+                                    GL_RGBA, GL_NEAREST);
 
    /* vertex positions, texcoords (after texture allocation!) */
    {
@@ -2509,32 +1783,28 @@ _mesa_meta_CopyPixels(struct gl_context *ctx, GLint srcX, GLint srcY,
       verts[0].x = dstX0;
       verts[0].y = dstY0;
       verts[0].z = z;
-      verts[0].s = 0.0F;
-      verts[0].t = 0.0F;
+      verts[0].tex[0] = 0.0F;
+      verts[0].tex[1] = 0.0F;
       verts[1].x = dstX1;
       verts[1].y = dstY0;
       verts[1].z = z;
-      verts[1].s = tex->Sright;
-      verts[1].t = 0.0F;
+      verts[1].tex[0] = tex->Sright;
+      verts[1].tex[1] = 0.0F;
       verts[2].x = dstX1;
       verts[2].y = dstY1;
       verts[2].z = z;
-      verts[2].s = tex->Sright;
-      verts[2].t = tex->Ttop;
+      verts[2].tex[0] = tex->Sright;
+      verts[2].tex[1] = tex->Ttop;
       verts[3].x = dstX0;
       verts[3].y = dstY1;
       verts[3].z = z;
-      verts[3].s = 0.0F;
-      verts[3].t = tex->Ttop;
+      verts[3].tex[0] = 0.0F;
+      verts[3].tex[1] = tex->Ttop;
 
       /* upload new vertex data */
       _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
    }
 
-   /* Alloc/setup texture */
-   setup_copypix_texture(ctx, tex, newTex, srcX, srcY, width, height,
-                         GL_RGBA, GL_NEAREST);
-
    _mesa_set_enable(ctx, tex->Target, GL_TRUE);
 
    /* draw textured quad */
@@ -2551,6 +1821,9 @@ meta_drawpix_cleanup(struct drawpix_state *drawpix)
    if (drawpix->VAO != 0) {
       _mesa_DeleteVertexArrays(1, &drawpix->VAO);
       drawpix->VAO = 0;
+
+      _mesa_DeleteBuffers(1, &drawpix->VBO);
+      drawpix->VBO = 0;
    }
 
    if (drawpix->StencilFP != 0) {
@@ -2646,7 +1919,7 @@ init_draw_stencil_pixels(struct gl_context *ctx)
       "END \n";
    char program2[1000];
    struct drawpix_state *drawpix = &ctx->Meta->DrawPix;
-   struct temp_texture *tex = get_temp_texture(ctx);
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
    const char *texTarget;
 
    assert(drawpix->StencilFP == 0);
@@ -2680,7 +1953,7 @@ init_draw_depth_pixels(struct gl_context *ctx)
       "END \n";
    char program2[200];
    struct drawpix_state *drawpix = &ctx->Meta->DrawPix;
-   struct temp_texture *tex = get_temp_texture(ctx);
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
    const char *texTarget;
 
    assert(drawpix->DepthFP == 0);
@@ -2712,17 +1985,13 @@ _mesa_meta_DrawPixels(struct gl_context *ctx,
                       const GLvoid *pixels)
 {
    struct drawpix_state *drawpix = &ctx->Meta->DrawPix;
-   struct temp_texture *tex = get_temp_texture(ctx);
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
    const struct gl_pixelstore_attrib unpackSave = ctx->Unpack;
    const GLuint origStencilMask = ctx->Stencil.WriteMask[0];
-   struct vertex {
-      GLfloat x, y, z, s, t;
-   };
    struct vertex verts[4];
    GLenum texIntFormat;
    GLboolean fallback, newTex;
    GLbitfield metaExtraSave = 0x0;
-   GLuint vbo;
 
    /*
     * Determine if we can do the glDrawPixels with texture mapping.
@@ -2812,7 +2081,13 @@ _mesa_meta_DrawPixels(struct gl_context *ctx,
                           MESA_META_VIEWPORT |
                           metaExtraSave));
 
-   newTex = alloc_texture(tex, width, height, texIntFormat);
+   newTex = _mesa_meta_alloc_texture(tex, width, height, texIntFormat);
+
+   _mesa_meta_setup_vertex_objects(&drawpix->VAO, &drawpix->VBO, false,
+                                   3, 2, 0);
+
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
 
    /* vertex positions, texcoords (after texture allocation!) */
    {
@@ -2825,43 +2100,29 @@ _mesa_meta_DrawPixels(struct gl_context *ctx,
       verts[0].x = x0;
       verts[0].y = y0;
       verts[0].z = z;
-      verts[0].s = 0.0F;
-      verts[0].t = 0.0F;
+      verts[0].tex[0] = 0.0F;
+      verts[0].tex[1] = 0.0F;
       verts[1].x = x1;
       verts[1].y = y0;
       verts[1].z = z;
-      verts[1].s = tex->Sright;
-      verts[1].t = 0.0F;
+      verts[1].tex[0] = tex->Sright;
+      verts[1].tex[1] = 0.0F;
       verts[2].x = x1;
       verts[2].y = y1;
       verts[2].z = z;
-      verts[2].s = tex->Sright;
-      verts[2].t = tex->Ttop;
+      verts[2].tex[0] = tex->Sright;
+      verts[2].tex[1] = tex->Ttop;
       verts[3].x = x0;
       verts[3].y = y1;
       verts[3].z = z;
-      verts[3].s = 0.0F;
-      verts[3].t = tex->Ttop;
-   }
-
-   if (drawpix->VAO == 0) {
-      /* one-time setup: create vertex array object */
-      _mesa_GenVertexArrays(1, &drawpix->VAO);
+      verts[3].tex[0] = 0.0F;
+      verts[3].tex[1] = tex->Ttop;
    }
-   _mesa_BindVertexArray(drawpix->VAO);
 
-   /* create vertex array buffer */
-   _mesa_GenBuffers(1, &vbo);
-   _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, vbo);
+   /* upload new vertex data */
    _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
                        verts, GL_DYNAMIC_DRAW_ARB);
 
-   /* setup vertex arrays */
-   _mesa_VertexPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-   _mesa_TexCoordPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(s));
-   _mesa_EnableClientState(GL_VERTEX_ARRAY);
-   _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-
    /* set given unpack params */
    ctx->Unpack = *unpack;
 
@@ -2874,8 +2135,8 @@ _mesa_meta_DrawPixels(struct gl_context *ctx,
       if (!drawpix->StencilFP)
          init_draw_stencil_pixels(ctx);
 
-      setup_drawpix_texture(ctx, tex, newTex, texIntFormat, width, height,
-                            GL_ALPHA, type, pixels);
+      _mesa_meta_setup_drawpix_texture(ctx, tex, newTex, width, height,
+                                       GL_ALPHA, type, pixels);
 
       _mesa_ColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
 
@@ -2917,22 +2178,20 @@ _mesa_meta_DrawPixels(struct gl_context *ctx,
       _mesa_ProgramLocalParameter4fvARB(GL_FRAGMENT_PROGRAM_ARB, 0,
                                         ctx->Current.RasterColor);
 
-      setup_drawpix_texture(ctx, tex, newTex, texIntFormat, width, height,
-                            format, type, pixels);
+      _mesa_meta_setup_drawpix_texture(ctx, tex, newTex, width, height,
+                                       format, type, pixels);
 
       _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
    }
    else {
       /* Drawing RGBA */
-      setup_drawpix_texture(ctx, tex, newTex, texIntFormat, width, height,
-                            format, type, pixels);
+      _mesa_meta_setup_drawpix_texture(ctx, tex, newTex, width, height,
+                                       format, type, pixels);
       _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
    }
 
    _mesa_set_enable(ctx, tex->Target, GL_FALSE);
 
-   _mesa_DeleteBuffers(1, &vbo);
-
    /* restore unpack params */
    ctx->Unpack = unpackSave;
 
@@ -2984,9 +2243,6 @@ _mesa_meta_Bitmap(struct gl_context *ctx,
    const GLenum texIntFormat = GL_ALPHA;
    const struct gl_pixelstore_attrib unpackSave = *unpack;
    GLubyte fg, bg;
-   struct vertex {
-      GLfloat x, y, z, s, t, r, g, b, a;
-   };
    struct vertex verts[4];
    GLboolean newTex;
    GLubyte *bitmap8;
@@ -3020,33 +2276,12 @@ _mesa_meta_Bitmap(struct gl_context *ctx,
                           MESA_META_VERTEX |
                           MESA_META_VIEWPORT));
 
-   if (bitmap->VAO == 0) {
-      /* one-time setup */
+   _mesa_meta_setup_vertex_objects(&bitmap->VAO, &bitmap->VBO, false, 3, 2, 4);
 
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &bitmap->VAO);
-      _mesa_BindVertexArray(bitmap->VAO);
+   newTex = _mesa_meta_alloc_texture(tex, width, height, texIntFormat);
 
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &bitmap->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, bitmap->VBO);
-      _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
-                          NULL, GL_DYNAMIC_DRAW_ARB);
-
-      /* setup vertex arrays */
-      _mesa_VertexPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_TexCoordPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(s));
-      _mesa_ColorPointer(4, GL_FLOAT, sizeof(struct vertex), OFFSET(r));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-      _mesa_EnableClientState(GL_COLOR_ARRAY);
-   }
-   else {
-      _mesa_BindVertexArray(bitmap->VAO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, bitmap->VBO);
-   }
-
-   newTex = alloc_texture(tex, width, height, texIntFormat);
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
 
    /* vertex positions, texcoords, colors (after texture allocation!) */
    {
@@ -3060,23 +2295,23 @@ _mesa_meta_Bitmap(struct gl_context *ctx,
       verts[0].x = x0;
       verts[0].y = y0;
       verts[0].z = z;
-      verts[0].s = 0.0F;
-      verts[0].t = 0.0F;
+      verts[0].tex[0] = 0.0F;
+      verts[0].tex[1] = 0.0F;
       verts[1].x = x1;
       verts[1].y = y0;
       verts[1].z = z;
-      verts[1].s = tex->Sright;
-      verts[1].t = 0.0F;
+      verts[1].tex[0] = tex->Sright;
+      verts[1].tex[1] = 0.0F;
       verts[2].x = x1;
       verts[2].y = y1;
       verts[2].z = z;
-      verts[2].s = tex->Sright;
-      verts[2].t = tex->Ttop;
+      verts[2].tex[0] = tex->Sright;
+      verts[2].tex[1] = tex->Ttop;
       verts[3].x = x0;
       verts[3].y = y1;
       verts[3].z = z;
-      verts[3].s = 0.0F;
-      verts[3].t = tex->Ttop;
+      verts[3].tex[0] = 0.0F;
+      verts[3].tex[1] = tex->Ttop;
 
       for (i = 0; i < 4; i++) {
          verts[i].r = ctx->Current.RasterColor[0];
@@ -3110,8 +2345,8 @@ _mesa_meta_Bitmap(struct gl_context *ctx,
       _mesa_set_enable(ctx, GL_ALPHA_TEST, GL_TRUE);
       _mesa_AlphaFunc(GL_NOTEQUAL, UBYTE_TO_FLOAT(bg));
 
-      setup_drawpix_texture(ctx, tex, newTex, texIntFormat, width, height,
-                            GL_ALPHA, GL_UNSIGNED_BYTE, bitmap8);
+      _mesa_meta_setup_drawpix_texture(ctx, tex, newTex, width, height,
+                                       GL_ALPHA, GL_UNSIGNED_BYTE, bitmap8);
 
       _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
 
@@ -3236,10 +2471,10 @@ setup_texture_coords(GLenum faceTarget,
                      GLint width,
                      GLint height,
                      GLint depth,
-                     GLfloat coords0[3],
-                     GLfloat coords1[3],
-                     GLfloat coords2[3],
-                     GLfloat coords3[3])
+                     GLfloat coords0[4],
+                     GLfloat coords1[4],
+                     GLfloat coords2[4],
+                     GLfloat coords3[4])
 {
    static const GLfloat st[4][2] = {
       {0.0f, 0.0f}, {1.0f, 0.0f}, {1.0f, 1.0f}, {0.0f, 1.0f}
@@ -3247,6 +2482,13 @@ setup_texture_coords(GLenum faceTarget,
    GLuint i;
    GLfloat r;
 
+   /* Currently all texture targets want the W component to be 1.0.
+    */
+   coords0[3] = 1.0F;
+   coords1[3] = 1.0F;
+   coords2[3] = 1.0F;
+   coords3[3] = 1.0F;
+
    switch (faceTarget) {
    case GL_TEXTURE_1D:
    case GL_TEXTURE_2D:
@@ -3337,6 +2579,8 @@ setup_texture_coords(GLenum faceTarget,
             assert(0);
          }
 
+         coord[3] = (float) (slice / 6);
+
          switch (faceTarget) {
          case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
             coord[0] = 1.0f;
@@ -3378,74 +2622,53 @@ setup_texture_coords(GLenum faceTarget,
    }
 }
 
-
-static void
-setup_ff_generate_mipmap(struct gen_mipmap_state *mipmap)
-{
-   struct vertex {
-      GLfloat x, y, tex[3];
-   };
-
-   if (mipmap->VAO == 0) {
-      /* one-time setup */
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &mipmap->VAO);
-      _mesa_BindVertexArray(mipmap->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &mipmap->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, mipmap->VBO);
-      /* setup vertex arrays */
-      _mesa_VertexPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_TexCoordPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(tex));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-   }
-
-   /* setup projection matrix */
-   _mesa_MatrixMode(GL_PROJECTION);
-   _mesa_LoadIdentity();
-   _mesa_Ortho(-1.0, 1.0, -1.0, 1.0, -1.0, 1.0);
-}
-
-
-static struct glsl_sampler *
-setup_texture_sampler(GLenum target, struct gen_mipmap_state *mipmap)
+static struct blit_shader *
+choose_blit_shader(GLenum target, struct blit_shader_table *table)
 {
    switch(target) {
    case GL_TEXTURE_1D:
-      mipmap->sampler_1d.type = "sampler1D";
-      mipmap->sampler_1d.func = "texture1D";
-      mipmap->sampler_1d.texcoords = "texCoords.x";
-      return &mipmap->sampler_1d;
+      table->sampler_1d.type = "sampler1D";
+      table->sampler_1d.func = "texture1D";
+      table->sampler_1d.texcoords = "texCoords.x";
+      return &table->sampler_1d;
    case GL_TEXTURE_2D:
-      mipmap->sampler_2d.type = "sampler2D";
-      mipmap->sampler_2d.func = "texture2D";
-      mipmap->sampler_2d.texcoords = "texCoords.xy";
-      return &mipmap->sampler_2d;
+      table->sampler_2d.type = "sampler2D";
+      table->sampler_2d.func = "texture2D";
+      table->sampler_2d.texcoords = "texCoords.xy";
+      return &table->sampler_2d;
+   case GL_TEXTURE_RECTANGLE:
+      table->sampler_rect.type = "sampler2DRect";
+      table->sampler_rect.func = "texture2DRect";
+      table->sampler_rect.texcoords = "texCoords.xy";
+      return &table->sampler_rect;
    case GL_TEXTURE_3D:
       /* Code for mipmap generation with 3D textures is not used yet.
        * It's a sw fallback.
        */
-      mipmap->sampler_3d.type = "sampler3D";
-      mipmap->sampler_3d.func = "texture3D";
-      mipmap->sampler_3d.texcoords = "texCoords";
-      return &mipmap->sampler_3d;
+      table->sampler_3d.type = "sampler3D";
+      table->sampler_3d.func = "texture3D";
+      table->sampler_3d.texcoords = "texCoords.xyz";
+      return &table->sampler_3d;
    case GL_TEXTURE_CUBE_MAP:
-      mipmap->sampler_cubemap.type = "samplerCube";
-      mipmap->sampler_cubemap.func = "textureCube";
-      mipmap->sampler_cubemap.texcoords = "texCoords";
-      return &mipmap->sampler_cubemap;
+      table->sampler_cubemap.type = "samplerCube";
+      table->sampler_cubemap.func = "textureCube";
+      table->sampler_cubemap.texcoords = "texCoords.xyz";
+      return &table->sampler_cubemap;
    case GL_TEXTURE_1D_ARRAY:
-      mipmap->sampler_1d_array.type = "sampler1DArray";
-      mipmap->sampler_1d_array.func = "texture1DArray";
-      mipmap->sampler_1d_array.texcoords = "texCoords.xy";
-      return &mipmap->sampler_1d_array;
+      table->sampler_1d_array.type = "sampler1DArray";
+      table->sampler_1d_array.func = "texture1DArray";
+      table->sampler_1d_array.texcoords = "texCoords.xy";
+      return &table->sampler_1d_array;
    case GL_TEXTURE_2D_ARRAY:
-      mipmap->sampler_2d_array.type = "sampler2DArray";
-      mipmap->sampler_2d_array.func = "texture2DArray";
-      mipmap->sampler_2d_array.texcoords = "texCoords";
-      return &mipmap->sampler_2d_array;
+      table->sampler_2d_array.type = "sampler2DArray";
+      table->sampler_2d_array.func = "texture2DArray";
+      table->sampler_2d_array.texcoords = "texCoords.xyz";
+      return &table->sampler_2d_array;
+   case GL_TEXTURE_CUBE_MAP_ARRAY:
+      table->sampler_cubemap_array.type = "samplerCubeArray";
+      table->sampler_cubemap_array.func = "textureCubeArray";
+      table->sampler_cubemap_array.texcoords = "texCoords.xyzw";
+      return &table->sampler_cubemap_array;
    default:
       _mesa_problem(NULL, "Unexpected texture target 0x%x in"
                     " setup_texture_sampler()\n", target);
@@ -3453,122 +2676,28 @@ setup_texture_sampler(GLenum target, struct gen_mipmap_state *mipmap)
    }
 }
 
-
-static void
-setup_glsl_generate_mipmap(struct gl_context *ctx,
-                           struct gen_mipmap_state *mipmap,
-                           GLenum target)
+void
+_mesa_meta_blit_shader_table_cleanup(struct blit_shader_table *table)
 {
-   struct vertex {
-      GLfloat x, y, tex[3];
-   };
-   struct glsl_sampler *sampler;
-   const char *vs_source;
-   char *fs_source;
-   GLuint vs, fs;
-   void *mem_ctx;
-
-   /* Check if already initialized */
-   if (mipmap->VAO == 0) {
-
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &mipmap->VAO);
-      _mesa_BindVertexArray(mipmap->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &mipmap->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, mipmap->VBO);
-
-      /* setup vertex arrays */
-      _mesa_VertexAttribPointer(0, 2, GL_FLOAT, GL_FALSE,
-                                   sizeof(struct vertex), OFFSET(x));
-      _mesa_VertexAttribPointer(1, 3, GL_FLOAT, GL_FALSE,
-                                   sizeof(struct vertex), OFFSET(tex));
-      _mesa_EnableVertexAttribArray(0);
-      _mesa_EnableVertexAttribArray(1);
-   }
-
-   /* Generate a fragment shader program appropriate for the texture target */
-   sampler = setup_texture_sampler(target, mipmap);
-   assert(sampler != NULL);
-   if (sampler->shader_prog != 0) {
-      mipmap->ShaderProg = sampler->shader_prog;
-      return;
-   }
-
-   mem_ctx = ralloc_context(NULL);
-
-   if (ctx->API == API_OPENGLES2 || ctx->Const.GLSLVersion < 130) {
-      vs_source =
-         "attribute vec2 position;\n"
-         "attribute vec3 textureCoords;\n"
-         "varying vec3 texCoords;\n"
-         "void main()\n"
-         "{\n"
-         "   texCoords = textureCoords;\n"
-         "   gl_Position = vec4(position, 0.0, 1.0);\n"
-         "}\n";
-
-      fs_source = ralloc_asprintf(mem_ctx,
-                                  "#extension GL_EXT_texture_array : enable\n"
-                                  "#ifdef GL_ES\n"
-                                  "precision highp float;\n"
-                                  "#endif\n"
-                                  "uniform %s texSampler;\n"
-                                  "varying vec3 texCoords;\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   gl_FragColor = %s(texSampler, %s);\n"
-                                  "}\n",
-                                  sampler->type,
-                                  sampler->func, sampler->texcoords);
-   }
-   else {
-      vs_source = ralloc_asprintf(mem_ctx,
-                                  "#version %s\n"
-                                  "in vec2 position;\n"
-                                  "in vec3 textureCoords;\n"
-                                  "out vec3 texCoords;\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   texCoords = textureCoords;\n"
-                                  "   gl_Position = vec4(position, 0.0, 1.0);\n"
-                                  "}\n",
-                                  _mesa_is_desktop_gl(ctx) ? "130" : "300 es");
-      fs_source = ralloc_asprintf(mem_ctx,
-                                  "#version %s\n"
-                                  "#ifdef GL_ES\n"
-                                  "precision highp float;\n"
-                                  "#endif\n"
-                                  "uniform %s texSampler;\n"
-                                  "in vec3 texCoords;\n"
-                                  "out vec4 out_color;\n"
-                                  "\n"
-                                  "void main()\n"
-                                  "{\n"
-                                  "   out_color = texture(texSampler, %s);\n"
-                                  "}\n",
-                                  _mesa_is_desktop_gl(ctx) ? "130" : "300 es",
-                                  sampler->type,
-                                  sampler->texcoords);
-   }
-
-   vs = compile_shader_with_debug(ctx, GL_VERTEX_SHADER, vs_source);
-   fs = compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER, fs_source);
-
-   mipmap->ShaderProg = _mesa_CreateProgramObjectARB();
-   _mesa_AttachShader(mipmap->ShaderProg, fs);
-   _mesa_DeleteObjectARB(fs);
-   _mesa_AttachShader(mipmap->ShaderProg, vs);
-   _mesa_DeleteObjectARB(vs);
-   _mesa_BindAttribLocation(mipmap->ShaderProg, 0, "position");
-   _mesa_BindAttribLocation(mipmap->ShaderProg, 1, "texcoords");
-   link_program_with_debug(ctx, mipmap->ShaderProg);
-   sampler->shader_prog = mipmap->ShaderProg;
-   ralloc_free(mem_ctx);
+   _mesa_DeleteObjectARB(table->sampler_1d.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_2d.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_3d.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_rect.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_cubemap.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_1d_array.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_2d_array.shader_prog);
+   _mesa_DeleteObjectARB(table->sampler_cubemap_array.shader_prog);
+
+   table->sampler_1d.shader_prog = 0;
+   table->sampler_2d.shader_prog = 0;
+   table->sampler_3d.shader_prog = 0;
+   table->sampler_rect.shader_prog = 0;
+   table->sampler_cubemap.shader_prog = 0;
+   table->sampler_1d_array.shader_prog = 0;
+   table->sampler_2d_array.shader_prog = 0;
+   table->sampler_cubemap_array.shader_prog = 0;
 }
 
-
 static void
 meta_glsl_generate_mipmap_cleanup(struct gen_mipmap_state *mipmap)
 {
@@ -3579,19 +2708,7 @@ meta_glsl_generate_mipmap_cleanup(struct gen_mipmap_state *mipmap)
    _mesa_DeleteBuffers(1, &mipmap->VBO);
    mipmap->VBO = 0;
 
-   _mesa_DeleteObjectARB(mipmap->sampler_1d.shader_prog);
-   _mesa_DeleteObjectARB(mipmap->sampler_2d.shader_prog);
-   _mesa_DeleteObjectARB(mipmap->sampler_3d.shader_prog);
-   _mesa_DeleteObjectARB(mipmap->sampler_cubemap.shader_prog);
-   _mesa_DeleteObjectARB(mipmap->sampler_1d_array.shader_prog);
-   _mesa_DeleteObjectARB(mipmap->sampler_2d_array.shader_prog);
-
-   mipmap->sampler_1d.shader_prog = 0;
-   mipmap->sampler_2d.shader_prog = 0;
-   mipmap->sampler_3d.shader_prog = 0;
-   mipmap->sampler_cubemap.shader_prog = 0;
-   mipmap->sampler_1d_array.shader_prog = 0;
-   mipmap->sampler_2d_array.shader_prog = 0;
+   _mesa_meta_blit_shader_table_cleanup(&mipmap->shaders);
 }
 
 
@@ -3605,9 +2722,6 @@ _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target,
                           struct gl_texture_object *texObj)
 {
    struct gen_mipmap_state *mipmap = &ctx->Meta->Mipmap;
-   struct vertex {
-      GLfloat x, y, tex[3];
-   };
    struct vertex verts[4];
    const GLuint baseLevel = texObj->BaseLevel;
    const GLuint maxLevel = texObj->MaxLevel;
@@ -3616,8 +2730,7 @@ _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target,
    const GLuint fboSave = ctx->DrawBuffer->Name;
    const GLuint currentTexUnitSave = ctx->Texture.CurrentUnit;
    const GLboolean use_glsl_version = ctx->Extensions.ARB_vertex_shader &&
-                                      ctx->Extensions.ARB_fragment_shader &&
-				      (ctx->API != API_OPENGLES);
+                                      ctx->Extensions.ARB_fragment_shader;
    GLenum faceTarget;
    GLuint dstLevel;
    const GLint slice = 0;
@@ -3643,17 +2756,15 @@ _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target,
     * GenerateMipmap function.
     */
    if (use_glsl_version) {
-      setup_glsl_generate_mipmap(ctx, mipmap, target);
-      _mesa_UseProgram(mipmap->ShaderProg);
+      _mesa_meta_setup_vertex_objects(&mipmap->VAO, &mipmap->VBO, true,
+                                      2, 3, 0);
+      _mesa_meta_setup_blit_shader(ctx, target, &mipmap->shaders);
    }
    else {
-      setup_ff_generate_mipmap(mipmap);
+      _mesa_meta_setup_ff_tnl_for_blit(&mipmap->VAO, &mipmap->VBO, 3);
       _mesa_set_enable(ctx, target, GL_TRUE);
    }
 
-   _mesa_BindVertexArray(mipmap->VAO);
-   _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, mipmap->VBO);
-
    samplerSave = ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler ?
       ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler->Name : 0;
 
@@ -3691,10 +2802,10 @@ _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target,
 
    _mesa_BindFramebuffer(GL_FRAMEBUFFER_EXT, mipmap->FBO);
 
-   if (ctx->API == API_OPENGL_COMPAT || ctx->API == API_OPENGLES)
-      _mesa_TexParameteri(target, GL_GENERATE_MIPMAP, GL_FALSE);
-   else
-      assert(!genMipmapSave);
+   _mesa_TexParameteri(target, GL_GENERATE_MIPMAP, GL_FALSE);
+
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
 
    /* Setup texture coordinates */
    setup_texture_coords(faceTarget,
@@ -3895,6 +3006,12 @@ _mesa_meta_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
    GLint bpp;
    void *buf;
 
+   /* The gl_renderbuffer is part of the interface for
+    * dd_function_table::CopyTexSubImage, but this implementation does not use
+    * it.
+    */
+   (void) rb;
+
    /* Choose format/type for temporary image buffer */
    format = _mesa_get_format_base_format(texImage->TexFormat);
    if (format == GL_LUMINANCE ||
@@ -4004,17 +3121,17 @@ decompress_texture_image(struct gl_context *ctx,
    const GLint depth = texImage->Height;
    const GLenum target = texObj->Target;
    GLenum faceTarget;
-   struct vertex {
-      GLfloat x, y, tex[3];
-   };
    struct vertex verts[4];
    GLuint fboDrawSave, fboReadSave;
    GLuint rbSave;
    GLuint samplerSave;
+   const bool use_glsl_version = ctx->Extensions.ARB_vertex_shader &&
+                                      ctx->Extensions.ARB_fragment_shader;
 
    if (slice > 0) {
       assert(target == GL_TEXTURE_3D ||
-             target == GL_TEXTURE_2D_ARRAY);
+             target == GL_TEXTURE_2D_ARRAY ||
+             target == GL_TEXTURE_CUBE_MAP_ARRAY);
    }
 
    switch (target) {
@@ -4027,10 +3144,9 @@ decompress_texture_image(struct gl_context *ctx,
       assert(!"No compressed 3D textures.");
       return;
 
-   case GL_TEXTURE_2D_ARRAY:
    case GL_TEXTURE_CUBE_MAP_ARRAY:
-      /* These targets are just broken currently. */
-      return;
+      faceTarget = GL_TEXTURE_CUBE_MAP_POSITIVE_X + (slice % 6);
+      break;
 
    case GL_TEXTURE_CUBE_MAP:
       faceTarget = GL_TEXTURE_CUBE_MAP_POSITIVE_X + texImage->Face;
@@ -4075,27 +3191,13 @@ decompress_texture_image(struct gl_context *ctx,
       decompress->Height = height;
    }
 
-   /* setup VBO data */
-   if (decompress->VAO == 0) {
-      /* create vertex array object */
-      _mesa_GenVertexArrays(1, &decompress->VAO);
-      _mesa_BindVertexArray(decompress->VAO);
-
-      /* create vertex array buffer */
-      _mesa_GenBuffers(1, &decompress->VBO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, decompress->VBO);
-      _mesa_BufferData(GL_ARRAY_BUFFER_ARB, sizeof(verts),
-                          NULL, GL_DYNAMIC_DRAW_ARB);
+   if (use_glsl_version) {
+      _mesa_meta_setup_vertex_objects(&decompress->VAO, &decompress->VBO, true,
+                                      2, 4, 0);
 
-      /* setup vertex arrays */
-      _mesa_VertexPointer(2, GL_FLOAT, sizeof(struct vertex), OFFSET(x));
-      _mesa_TexCoordPointer(3, GL_FLOAT, sizeof(struct vertex), OFFSET(tex));
-      _mesa_EnableClientState(GL_VERTEX_ARRAY);
-      _mesa_EnableClientState(GL_TEXTURE_COORD_ARRAY);
-   }
-   else {
-      _mesa_BindVertexArray(decompress->VAO);
-      _mesa_BindBuffer(GL_ARRAY_BUFFER_ARB, decompress->VBO);
+      _mesa_meta_setup_blit_shader(ctx, target, &decompress->shaders);
+   } else {
+      _mesa_meta_setup_ff_tnl_for_blit(&decompress->VAO, &decompress->VBO, 3);
    }
 
    if (!decompress->Sampler) {
@@ -4114,6 +3216,9 @@ decompress_texture_image(struct gl_context *ctx,
       _mesa_BindSampler(ctx->Texture.CurrentUnit, decompress->Sampler);
    }
 
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
+
    setup_texture_coords(faceTarget, slice, width, height, depth,
                         verts[0].tex,
                         verts[1].tex,
@@ -4130,8 +3235,6 @@ decompress_texture_image(struct gl_context *ctx,
    verts[3].x = -1.0F;
    verts[3].y =  1.0F;
 
-   _mesa_MatrixMode(GL_PROJECTION);
-   _mesa_LoadIdentity();
    _mesa_set_viewport(ctx, 0, 0, 0, width, height);
 
    /* upload new vertex data */
@@ -4139,7 +3242,9 @@ decompress_texture_image(struct gl_context *ctx,
 
    /* setup texture state */
    _mesa_BindTexture(target, texObj->Name);
-   _mesa_set_enable(ctx, target, GL_TRUE);
+
+   if (!use_glsl_version)
+      _mesa_set_enable(ctx, target, GL_TRUE);
 
    {
       /* save texture object state */
@@ -4199,7 +3304,8 @@ decompress_texture_image(struct gl_context *ctx,
    }
 
    /* disable texture unit */
-   _mesa_set_enable(ctx, target, GL_FALSE);
+   if (!use_glsl_version)
+      _mesa_set_enable(ctx, target, GL_FALSE);
 
    _mesa_BindSampler(ctx->Texture.CurrentUnit, samplerSave);
 
@@ -4231,8 +3337,7 @@ _mesa_meta_GetTexImage(struct gl_context *ctx,
     * unsigned, normalized values.  We could handle signed and unnormalized 
     * with floating point renderbuffers...
     */
-   if (texImage->TexObject->Target != GL_TEXTURE_CUBE_MAP_ARRAY
-       && _mesa_is_format_compressed(texImage->TexFormat) &&
+   if (_mesa_is_format_compressed(texImage->TexFormat) &&
        _mesa_get_format_datatype(texImage->TexFormat)
        == GL_UNSIGNED_NORMALIZED) {
       struct gl_texture_object *texObj = texImage->TexObject;
@@ -4241,7 +3346,8 @@ _mesa_meta_GetTexImage(struct gl_context *ctx,
       _mesa_unlock_texture(ctx, texObj);
       for (slice = 0; slice < texImage->Depth; slice++) {
          void *dst;
-         if (texImage->TexObject->Target == GL_TEXTURE_2D_ARRAY) {
+         if (texImage->TexObject->Target == GL_TEXTURE_2D_ARRAY
+             || texImage->TexObject->Target == GL_TEXTURE_CUBE_MAP_ARRAY) {
             /* Setup pixel packing.  SkipPixels and SkipRows will be applied
              * in the decompress_texture_image() function's call to
              * glReadPixels but we need to compute the dest slice's address
diff --git a/mesalib/src/mesa/drivers/common/meta.h b/mesalib/src/mesa/drivers/common/meta.h
index 53d7b8ab5..fcf45c403 100644
--- a/mesalib/src/mesa/drivers/common/meta.h
+++ b/mesalib/src/mesa/drivers/common/meta.h
@@ -60,6 +60,311 @@
 #define MESA_META_OCCLUSION_QUERY      0x400000
 /**\}*/
 
+/**
+ * State which we may save/restore across meta ops.
+ * XXX this may be incomplete...
+ */
+struct save_state
+{
+   GLbitfield SavedState;  /**< bitmask of MESA_META_* flags */
+
+   /* Always saved/restored with meta. */
+   gl_api API;
+
+   /** MESA_META_CLEAR (and others?) */
+   struct gl_query_object *CurrentOcclusionObject;
+
+   /** MESA_META_ALPHA_TEST */
+   GLboolean AlphaEnabled;
+   GLenum AlphaFunc;
+   GLclampf AlphaRef;
+
+   /** MESA_META_BLEND */
+   GLbitfield BlendEnabled;
+   GLboolean ColorLogicOpEnabled;
+
+   /** MESA_META_COLOR_MASK */
+   GLubyte ColorMask[MAX_DRAW_BUFFERS][4];
+
+   /** MESA_META_DEPTH_TEST */
+   struct gl_depthbuffer_attrib Depth;
+
+   /** MESA_META_FOG */
+   GLboolean Fog;
+
+   /** MESA_META_PIXEL_STORE */
+   struct gl_pixelstore_attrib Pack, Unpack;
+
+   /** MESA_META_PIXEL_TRANSFER */
+   GLfloat RedBias, RedScale;
+   GLfloat GreenBias, GreenScale;
+   GLfloat BlueBias, BlueScale;
+   GLfloat AlphaBias, AlphaScale;
+   GLfloat DepthBias, DepthScale;
+   GLboolean MapColorFlag;
+
+   /** MESA_META_RASTERIZATION */
+   GLenum FrontPolygonMode, BackPolygonMode;
+   GLboolean PolygonOffset;
+   GLboolean PolygonSmooth;
+   GLboolean PolygonStipple;
+   GLboolean PolygonCull;
+
+   /** MESA_META_SCISSOR */
+   struct gl_scissor_attrib Scissor;
+
+   /** MESA_META_SHADER */
+   GLboolean VertexProgramEnabled;
+   struct gl_vertex_program *VertexProgram;
+   GLboolean FragmentProgramEnabled;
+   struct gl_fragment_program *FragmentProgram;
+   GLboolean ATIFragmentShaderEnabled;
+   struct gl_shader_program *Shader[MESA_SHADER_STAGES];
+   struct gl_shader_program *ActiveShader;
+
+   /** MESA_META_STENCIL_TEST */
+   struct gl_stencil_attrib Stencil;
+
+   /** MESA_META_TRANSFORM */
+   GLenum MatrixMode;
+   GLfloat ModelviewMatrix[16];
+   GLfloat ProjectionMatrix[16];
+   GLfloat TextureMatrix[16];
+
+   /** MESA_META_CLIP */
+   GLbitfield ClipPlanesEnabled;
+
+   /** MESA_META_TEXTURE */
+   GLuint ActiveUnit;
+   GLuint ClientActiveUnit;
+   /** for unit[0] only */
+   struct gl_texture_object *CurrentTexture[NUM_TEXTURE_TARGETS];
+   /** mask of TEXTURE_2D_BIT, etc */
+   GLbitfield TexEnabled[MAX_TEXTURE_UNITS];
+   GLbitfield TexGenEnabled[MAX_TEXTURE_UNITS];
+   GLuint EnvMode;  /* unit[0] only */
+
+   /** MESA_META_VERTEX */
+   struct gl_vertex_array_object *VAO;
+   struct gl_buffer_object *ArrayBufferObj;
+
+   /** MESA_META_VIEWPORT */
+   GLfloat ViewportX, ViewportY, ViewportW, ViewportH;
+   GLclampd DepthNear, DepthFar;
+
+   /** MESA_META_CLAMP_FRAGMENT_COLOR */
+   GLenum ClampFragmentColor;
+
+   /** MESA_META_CLAMP_VERTEX_COLOR */
+   GLenum ClampVertexColor;
+
+   /** MESA_META_CONDITIONAL_RENDER */
+   struct gl_query_object *CondRenderQuery;
+   GLenum CondRenderMode;
+
+   /** MESA_META_SELECT_FEEDBACK */
+   GLenum RenderMode;
+   struct gl_selection Select;
+   struct gl_feedback Feedback;
+
+   /** MESA_META_MULTISAMPLE */
+   struct gl_multisample_attrib Multisample;
+
+   /** MESA_META_FRAMEBUFFER_SRGB */
+   GLboolean sRGBEnabled;
+
+   /** Miscellaneous (always disabled) */
+   GLboolean Lighting;
+   GLboolean RasterDiscard;
+   GLboolean TransformFeedbackNeedsResume;
+};
+
+/**
+ * Temporary texture used for glBlitFramebuffer, glDrawPixels, etc.
+ * This is currently shared by all the meta ops.  But we could create a
+ * separate one for each of glDrawPixel, glBlitFramebuffer, glCopyPixels, etc.
+ */
+struct temp_texture
+{
+   GLuint TexObj;
+   GLenum Target;         /**< GL_TEXTURE_2D or GL_TEXTURE_RECTANGLE */
+   GLsizei MinSize;       /**< Min texture size to allocate */
+   GLsizei MaxSize;       /**< Max possible texture size */
+   GLboolean NPOT;        /**< Non-power of two size OK? */
+   GLsizei Width, Height; /**< Current texture size */
+   GLenum IntFormat;
+   GLfloat Sright, Ttop;  /**< right, top texcoords */
+};
+
+/**
+ * State for GLSL texture sampler which is used to generate fragment
+ * shader in _mesa_meta_generate_mipmap().
+ */
+struct blit_shader {
+   const char *type;
+   const char *func;
+   const char *texcoords;
+   GLuint shader_prog;
+};
+
+/**
+ * Table of all sampler types and shaders for accessing them.
+ */
+struct blit_shader_table {
+   struct blit_shader sampler_1d;
+   struct blit_shader sampler_2d;
+   struct blit_shader sampler_3d;
+   struct blit_shader sampler_rect;
+   struct blit_shader sampler_cubemap;
+   struct blit_shader sampler_1d_array;
+   struct blit_shader sampler_2d_array;
+   struct blit_shader sampler_cubemap_array;
+};
+
+/**
+ * Indices in the blit_state->msaa_shaders[] array
+ *
+ * Note that setup_glsl_msaa_blit_shader() assumes that the _INT enums are one
+ * more than the non-_INT version and _UINT is one beyond that.
+ */
+enum blit_msaa_shader {
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE_INT,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE_UINT,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_COPY,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_COPY_INT,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_COPY_UINT,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_RESOLVE,
+   BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_COPY,
+   BLIT_MSAA_SHADER_COUNT,
+};
+
+/**
+ * State for glBlitFramebufer()
+ */
+struct blit_state
+{
+   GLuint VAO;
+   GLuint VBO;
+   GLuint DepthFP;
+   struct blit_shader_table shaders;
+   GLuint msaa_shaders[BLIT_MSAA_SHADER_COUNT];
+   struct temp_texture depthTex;
+};
+
+
+/**
+ * State for glClear()
+ */
+struct clear_state
+{
+   GLuint VAO;
+   GLuint VBO;
+   GLuint ShaderProg;
+   GLint ColorLocation;
+   GLint LayerLocation;
+
+   GLuint IntegerShaderProg;
+   GLint IntegerColorLocation;
+   GLint IntegerLayerLocation;
+};
+
+
+/**
+ * State for glCopyPixels()
+ */
+struct copypix_state
+{
+   GLuint VAO;
+   GLuint VBO;
+};
+
+
+/**
+ * State for glDrawPixels()
+ */
+struct drawpix_state
+{
+   GLuint VAO;
+   GLuint VBO;
+
+   GLuint StencilFP;  /**< Fragment program for drawing stencil images */
+   GLuint DepthFP;  /**< Fragment program for drawing depth images */
+};
+
+
+/**
+ * State for glBitmap()
+ */
+struct bitmap_state
+{
+   GLuint VAO;
+   GLuint VBO;
+   struct temp_texture Tex;  /**< separate texture from other meta ops */
+};
+
+/**
+ * State for _mesa_meta_generate_mipmap()
+ */
+struct gen_mipmap_state
+{
+   GLuint VAO;
+   GLuint VBO;
+   GLuint FBO;
+   GLuint Sampler;
+
+   struct blit_shader_table shaders;
+};
+
+/**
+ * State for texture decompression
+ */
+struct decompress_state
+{
+   GLuint VAO;
+   GLuint VBO, FBO, RBO, Sampler;
+   GLint Width, Height;
+
+   struct blit_shader_table shaders;
+};
+
+/**
+ * State for glDrawTex()
+ */
+struct drawtex_state
+{
+   GLuint VAO;
+   GLuint VBO;
+};
+
+#define MAX_META_OPS_DEPTH      8
+/**
+ * All per-context meta state.
+ */
+struct gl_meta_state
+{
+   /** Stack of state saved during meta-ops */
+   struct save_state Save[MAX_META_OPS_DEPTH];
+   /** Save stack depth */
+   GLuint SaveStackDepth;
+
+   struct temp_texture TempTex;
+
+   struct blit_state Blit;    /**< For _mesa_meta_BlitFramebuffer() */
+   struct clear_state Clear;  /**< For _mesa_meta_Clear() */
+   struct copypix_state CopyPix;  /**< For _mesa_meta_CopyPixels() */
+   struct drawpix_state DrawPix;  /**< For _mesa_meta_DrawPixels() */
+   struct bitmap_state Bitmap;    /**< For _mesa_meta_Bitmap() */
+   struct gen_mipmap_state Mipmap;    /**< For _mesa_meta_GenerateMipmap() */
+   struct decompress_state Decompress;  /**< For texture decompression */
+   struct drawtex_state DrawTex;  /**< For _mesa_meta_DrawTex() */
+};
+
+struct vertex {
+   GLfloat x, y, z, tex[4];
+   GLfloat r, g, b, a;
+};
+
 extern void
 _mesa_meta_init(struct gl_context *ctx);
 
@@ -130,4 +435,60 @@ extern void
 _mesa_meta_DrawTex(struct gl_context *ctx, GLfloat x, GLfloat y, GLfloat z,
                    GLfloat width, GLfloat height);
 
+/* meta-internal functions */
+GLuint
+_mesa_meta_compile_shader_with_debug(struct gl_context *ctx, GLenum target,
+                                     const GLcharARB *source);
+
+
+GLuint
+_mesa_meta_link_program_with_debug(struct gl_context *ctx, GLuint program);
+
+GLboolean
+_mesa_meta_alloc_texture(struct temp_texture *tex,
+                         GLsizei width, GLsizei height, GLenum intFormat);
+
+struct temp_texture *
+_mesa_meta_get_temp_texture(struct gl_context *ctx);
+
+struct temp_texture *
+_mesa_meta_get_temp_depth_texture(struct gl_context *ctx);
+
+void
+_mesa_meta_setup_vertex_objects(GLuint *VAO, GLuint *VBO,
+                                bool use_generic_attributes,
+                                unsigned vertex_size, unsigned texcoord_size,
+                                unsigned color_size);
+
+void
+_mesa_meta_setup_ff_tnl_for_blit(GLuint *VAO, GLuint *VBO,
+                                 unsigned texcoord_size);
+
+void
+_mesa_meta_setup_drawpix_texture(struct gl_context *ctx,
+                                 struct temp_texture *tex,
+                                 GLboolean newTex,
+                                 GLsizei width, GLsizei height,
+                                 GLenum format, GLenum type,
+                                 const GLvoid *pixels);
+
+void
+_mesa_meta_setup_copypix_texture(struct gl_context *ctx,
+                                 struct temp_texture *tex,
+                                 GLint srcX, GLint srcY,
+                                 GLsizei width, GLsizei height,
+                                 GLenum intFormat,
+                                 GLenum filter);
+
+void
+_mesa_meta_setup_blit_shader(struct gl_context *ctx,
+                             GLenum target,
+                             struct blit_shader_table *table);
+
+void
+_mesa_meta_glsl_blit_cleanup(struct blit_state *blit);
+
+void
+_mesa_meta_blit_shader_table_cleanup(struct blit_shader_table *table);
+
 #endif /* META_H */
diff --git a/mesalib/src/mesa/drivers/common/meta_blit.c b/mesalib/src/mesa/drivers/common/meta_blit.c
new file mode 100644
index 000000000..907c2cdf3
--- /dev/null
+++ b/mesalib/src/mesa/drivers/common/meta_blit.c
@@ -0,0 +1,825 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright (C) 2009  VMware, Inc.  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, 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 shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "main/glheader.h"
+#include "main/mtypes.h"
+#include "main/imports.h"
+#include "main/arbprogram.h"
+#include "main/arrayobj.h"
+#include "main/blend.h"
+#include "main/condrender.h"
+#include "main/depth.h"
+#include "main/enable.h"
+#include "main/enums.h"
+#include "main/fbobject.h"
+#include "main/macros.h"
+#include "main/matrix.h"
+#include "main/multisample.h"
+#include "main/objectlabel.h"
+#include "main/readpix.h"
+#include "main/shaderapi.h"
+#include "main/texobj.h"
+#include "main/texenv.h"
+#include "main/teximage.h"
+#include "main/texparam.h"
+#include "main/varray.h"
+#include "main/viewport.h"
+#include "swrast/swrast.h"
+#include "drivers/common/meta.h"
+#include "../glsl/ralloc.h"
+
+/** Return offset in bytes of the field within a vertex struct */
+#define OFFSET(FIELD) ((void *) offsetof(struct vertex, FIELD))
+
+/**
+ * One-time init for drawing depth pixels.
+ */
+static void
+init_blit_depth_pixels(struct gl_context *ctx)
+{
+   static const char *program =
+      "!!ARBfp1.0\n"
+      "TEX result.depth, fragment.texcoord[0], texture[0], %s; \n"
+      "END \n";
+   char program2[200];
+   struct blit_state *blit = &ctx->Meta->Blit;
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
+   const char *texTarget;
+
+   assert(blit->DepthFP == 0);
+
+   /* replace %s with "RECT" or "2D" */
+   assert(strlen(program) + 4 < sizeof(program2));
+   if (tex->Target == GL_TEXTURE_RECTANGLE)
+      texTarget = "RECT";
+   else
+      texTarget = "2D";
+   _mesa_snprintf(program2, sizeof(program2), program, texTarget);
+
+   _mesa_GenProgramsARB(1, &blit->DepthFP);
+   _mesa_BindProgramARB(GL_FRAGMENT_PROGRAM_ARB, blit->DepthFP);
+   _mesa_ProgramStringARB(GL_FRAGMENT_PROGRAM_ARB, GL_PROGRAM_FORMAT_ASCII_ARB,
+                          strlen(program2), (const GLubyte *) program2);
+}
+
+static void
+setup_glsl_msaa_blit_shader(struct gl_context *ctx,
+                            struct blit_state *blit,
+                            struct gl_renderbuffer *src_rb,
+                            GLenum target)
+{
+   const char *vs_source;
+   char *fs_source;
+   GLuint vs, fs;
+   void *mem_ctx;
+   enum blit_msaa_shader shader_index;
+   bool dst_is_msaa = false;
+   GLenum src_datatype;
+   const char *vec4_prefix;
+   char *name;
+
+   if (src_rb) {
+      src_datatype = _mesa_get_format_datatype(src_rb->Format);
+   } else {
+      /* depth-or-color glCopyTexImage fallback path that passes a NULL rb and
+       * doesn't handle integer.
+       */
+      src_datatype = GL_UNSIGNED_NORMALIZED;
+   }
+
+   if (ctx->DrawBuffer->Visual.samples > 1) {
+      /* If you're calling meta_BlitFramebuffer with the destination
+       * multisampled, this is the only path that will work -- swrast and
+       * CopyTexImage won't work on it either.
+       */
+      assert(ctx->Extensions.ARB_sample_shading);
+
+      dst_is_msaa = true;
+
+      /* We need shader invocation per sample, not per pixel */
+      _mesa_set_enable(ctx, GL_MULTISAMPLE, GL_TRUE);
+      _mesa_set_enable(ctx, GL_SAMPLE_SHADING, GL_TRUE);
+      _mesa_MinSampleShading(1.0);
+   }
+
+   switch (target) {
+   case GL_TEXTURE_2D_MULTISAMPLE:
+      if (src_rb->_BaseFormat == GL_DEPTH_COMPONENT ||
+          src_rb->_BaseFormat == GL_DEPTH_STENCIL) {
+         if (dst_is_msaa)
+            shader_index = BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_COPY;
+         else
+            shader_index = BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_RESOLVE;
+      } else {
+         if (dst_is_msaa)
+            shader_index = BLIT_MSAA_SHADER_2D_MULTISAMPLE_COPY;
+         else
+            shader_index = BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE;
+      }
+      break;
+   default:
+      _mesa_problem(ctx, "Unkown texture target %s\n",
+                    _mesa_lookup_enum_by_nr(target));
+      shader_index = BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE;
+   }
+
+   /* We rely on the enum being sorted this way. */
+   STATIC_ASSERT(BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE_INT ==
+                 BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE + 1);
+   STATIC_ASSERT(BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE_UINT ==
+                 BLIT_MSAA_SHADER_2D_MULTISAMPLE_RESOLVE + 2);
+   if (src_datatype == GL_INT) {
+      shader_index++;
+      vec4_prefix = "i";
+   } else if (src_datatype == GL_UNSIGNED_INT) {
+      shader_index += 2;
+      vec4_prefix = "u";
+   } else {
+      vec4_prefix = "";
+   }
+
+   if (blit->msaa_shaders[shader_index]) {
+      _mesa_UseProgram(blit->msaa_shaders[shader_index]);
+      return;
+   }
+
+   mem_ctx = ralloc_context(NULL);
+
+   if (shader_index == BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_RESOLVE ||
+       shader_index == BLIT_MSAA_SHADER_2D_MULTISAMPLE_DEPTH_COPY) {
+      char *sample_index;
+      const char *arb_sample_shading_extension_string;
+
+      if (dst_is_msaa) {
+         arb_sample_shading_extension_string = "#extension GL_ARB_sample_shading : enable";
+         sample_index = "gl_SampleID";
+         name = "depth MSAA copy";
+      } else {
+         /* Don't need that extension, since we're drawing to a single-sampled
+          * destination.
+          */
+         arb_sample_shading_extension_string = "";
+         /* From the GL 4.3 spec:
+          *
+          *     "If there is a multisample buffer (the value of SAMPLE_BUFFERS
+          *      is one), then values are obtained from the depth samples in
+          *      this buffer. It is recommended that the depth value of the
+          *      centermost sample be used, though implementations may choose
+          *      any function of the depth sample values at each pixel.
+          *
+          * We're slacking and instead of choosing centermost, we've got 0.
+          */
+         sample_index = "0";
+         name = "depth MSAA resolve";
+      }
+
+      vs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "in vec2 position;\n"
+                                  "in vec2 textureCoords;\n"
+                                  "out vec2 texCoords;\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   texCoords = textureCoords;\n"
+                                  "   gl_Position = vec4(position, 0.0, 1.0);\n"
+                                  "}\n");
+      fs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "#extension GL_ARB_texture_multisample : enable\n"
+                                  "%s\n"
+                                  "uniform sampler2DMS texSampler;\n"
+                                  "in vec2 texCoords;\n"
+                                  "out vec4 out_color;\n"
+                                  "\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   gl_FragDepth = texelFetch(texSampler, ivec2(texCoords), %s).r;\n"
+                                  "}\n",
+                                  arb_sample_shading_extension_string,
+                                  sample_index);
+   } else {
+      /* You can create 2D_MULTISAMPLE textures with 0 sample count (meaning 1
+       * sample).  Yes, this is ridiculous.
+       */
+      int samples;
+      char *sample_resolve;
+      const char *arb_sample_shading_extension_string;
+      const char *merge_function;
+      name = ralloc_asprintf(mem_ctx, "%svec4 MSAA %s",
+                             vec4_prefix,
+                             dst_is_msaa ? "copy" : "resolve");
+
+      samples = MAX2(src_rb->NumSamples, 1);
+
+      if (dst_is_msaa) {
+         arb_sample_shading_extension_string = "#extension GL_ARB_sample_shading : enable";
+         sample_resolve = ralloc_asprintf(mem_ctx, "   out_color = texelFetch(texSampler, ivec2(texCoords), gl_SampleID);");
+         merge_function = "";
+      } else {
+         int i;
+         int step;
+
+         if (src_datatype == GL_INT || src_datatype == GL_UNSIGNED_INT) {
+            merge_function =
+               "gvec4 merge(gvec4 a, gvec4 b) { return (a >> gvec4(1)) + (b >> gvec4(1)) + (a & b & gvec4(1)); }\n";
+         } else {
+            /* The divide will happen at the end for floats. */
+            merge_function =
+               "vec4 merge(vec4 a, vec4 b) { return (a + b); }\n";
+         }
+
+         arb_sample_shading_extension_string = "";
+
+         /* We're assuming power of two samples for this resolution procedure.
+          *
+          * To avoid losing any floating point precision if the samples all
+          * happen to have the same value, we merge pairs of values at a time
+          * (so the floating point exponent just gets increased), rather than
+          * doing a naive sum and dividing.
+          */
+         assert((samples & (samples - 1)) == 0);
+         /* Fetch each individual sample. */
+         sample_resolve = rzalloc_size(mem_ctx, 1);
+         for (i = 0; i < samples; i++) {
+            ralloc_asprintf_append(&sample_resolve,
+                                   "   gvec4 sample_1_%d = texelFetch(texSampler, ivec2(texCoords), %d);\n",
+                                   i, i);
+         }
+         /* Now, merge each pair of samples, then merge each pair of those,
+          * etc.
+          */
+         for (step = 2; step <= samples; step *= 2) {
+            for (i = 0; i < samples; i += step) {
+               ralloc_asprintf_append(&sample_resolve,
+                                      "   gvec4 sample_%d_%d = merge(sample_%d_%d, sample_%d_%d);\n",
+                                      step, i,
+                                      step / 2, i,
+                                      step / 2, i + step / 2);
+            }
+         }
+
+         /* Scale the final result. */
+         if (src_datatype == GL_UNSIGNED_INT || src_datatype == GL_INT) {
+            ralloc_asprintf_append(&sample_resolve,
+                                   "   out_color = sample_%d_0;\n",
+                                   samples);
+         } else {
+            ralloc_asprintf_append(&sample_resolve,
+                                   "   out_color = sample_%d_0 / %f;\n",
+                                   samples, (float)samples);
+         }
+      }
+
+      vs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "in vec2 position;\n"
+                                  "in vec2 textureCoords;\n"
+                                  "out vec2 texCoords;\n"
+                                  "void main()\n"
+                                  "{\n"
+                                  "   texCoords = textureCoords;\n"
+                                  "   gl_Position = vec4(position, 0.0, 1.0);\n"
+                                  "}\n");
+      fs_source = ralloc_asprintf(mem_ctx,
+                                  "#version 130\n"
+                                  "#extension GL_ARB_texture_multisample : enable\n"
+                                  "%s\n"
+                                  "#define gvec4 %svec4\n"
+                                  "uniform %ssampler2DMS texSampler;\n"
+                                  "in vec2 texCoords;\n"
+                                  "out gvec4 out_color;\n"
+                                  "\n"
+                                  "%s" /* merge_function */
+                                  "void main()\n"
+                                  "{\n"
+                                  "%s\n" /* sample_resolve */
+                                  "}\n",
+                                  arb_sample_shading_extension_string,
+                                  vec4_prefix,
+                                  vec4_prefix,
+                                  merge_function,
+                                  sample_resolve);
+   }
+
+   vs = _mesa_meta_compile_shader_with_debug(ctx, GL_VERTEX_SHADER, vs_source);
+   fs = _mesa_meta_compile_shader_with_debug(ctx, GL_FRAGMENT_SHADER, fs_source);
+
+   blit->msaa_shaders[shader_index] = _mesa_CreateProgramObjectARB();
+   _mesa_AttachShader(blit->msaa_shaders[shader_index], fs);
+   _mesa_DeleteObjectARB(fs);
+   _mesa_AttachShader(blit->msaa_shaders[shader_index], vs);
+   _mesa_DeleteObjectARB(vs);
+   _mesa_BindAttribLocation(blit->msaa_shaders[shader_index], 0, "position");
+   _mesa_BindAttribLocation(blit->msaa_shaders[shader_index], 1, "texcoords");
+   _mesa_meta_link_program_with_debug(ctx, blit->msaa_shaders[shader_index]);
+   _mesa_ObjectLabel(GL_PROGRAM, blit->msaa_shaders[shader_index], -1, name);
+   ralloc_free(mem_ctx);
+
+   _mesa_UseProgram(blit->msaa_shaders[shader_index]);
+}
+
+static void
+setup_glsl_blit_framebuffer(struct gl_context *ctx,
+                            struct blit_state *blit,
+                            struct gl_renderbuffer *src_rb,
+                            GLenum target)
+{
+   /* target = GL_TEXTURE_RECTANGLE is not supported in GLES 3.0 */
+   assert(_mesa_is_desktop_gl(ctx) || target == GL_TEXTURE_2D);
+
+   _mesa_meta_setup_vertex_objects(&blit->VAO, &blit->VBO, true, 2, 2, 0);
+
+   if (target == GL_TEXTURE_2D_MULTISAMPLE) {
+      setup_glsl_msaa_blit_shader(ctx, blit, src_rb, target);
+   } else {
+      _mesa_meta_setup_blit_shader(ctx, target, &blit->shaders);
+   }
+}
+
+/**
+ * Try to do a color or depth glBlitFramebuffer using texturing.
+ *
+ * We can do this when the src renderbuffer is actually a texture, or when the
+ * driver exposes BindRenderbufferTexImage().
+ */
+static bool
+blitframebuffer_texture(struct gl_context *ctx,
+                        GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
+                        GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,
+                        GLenum filter, GLint flipX, GLint flipY,
+                        GLboolean glsl_version, GLboolean do_depth)
+{
+   const struct gl_framebuffer *readFb = ctx->ReadBuffer;
+   int att_index = do_depth ? BUFFER_DEPTH : readFb->_ColorReadBufferIndex;
+   const struct gl_renderbuffer_attachment *readAtt =
+      &readFb->Attachment[att_index];
+   struct blit_state *blit = &ctx->Meta->Blit;
+   const GLint dstX = MIN2(dstX0, dstX1);
+   const GLint dstY = MIN2(dstY0, dstY1);
+   const GLint dstW = abs(dstX1 - dstX0);
+   const GLint dstH = abs(dstY1 - dstY0);
+   struct gl_texture_object *texObj;
+   GLuint srcLevel;
+   GLint baseLevelSave;
+   GLint maxLevelSave;
+   GLenum target;
+   GLuint sampler, samplerSave =
+      ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler ?
+      ctx->Texture.Unit[ctx->Texture.CurrentUnit].Sampler->Name : 0;
+   GLuint tempTex = 0;
+   struct gl_renderbuffer *rb = readAtt->Renderbuffer;
+
+   if (rb->NumSamples && !ctx->Extensions.ARB_texture_multisample)
+      return false;
+
+   if (filter == GL_SCALED_RESOLVE_FASTEST_EXT ||
+       filter == GL_SCALED_RESOLVE_NICEST_EXT) {
+      filter = GL_LINEAR;
+   }
+
+   if (readAtt->Texture) {
+      /* If there's a texture attached of a type we can handle, then just use
+       * it directly.
+       */
+      srcLevel = readAtt->TextureLevel;
+      texObj = readAtt->Texture;
+      target = texObj->Target;
+
+      switch (target) {
+      case GL_TEXTURE_2D:
+      case GL_TEXTURE_RECTANGLE:
+      case GL_TEXTURE_2D_MULTISAMPLE:
+         break;
+      default:
+         return false;
+      }
+   } else if (ctx->Driver.BindRenderbufferTexImage) {
+      /* Otherwise, we need the driver to be able to bind a renderbuffer as
+       * a texture image.
+       */
+      struct gl_texture_image *texImage;
+
+      if (rb->NumSamples > 1)
+         target = GL_TEXTURE_2D_MULTISAMPLE;
+      else
+         target = GL_TEXTURE_2D;
+
+      _mesa_GenTextures(1, &tempTex);
+      _mesa_BindTexture(target, tempTex);
+      srcLevel = 0;
+      texObj = _mesa_lookup_texture(ctx, tempTex);
+      texImage = _mesa_get_tex_image(ctx, texObj, target, srcLevel);
+
+      if (!ctx->Driver.BindRenderbufferTexImage(ctx, rb, texImage)) {
+         _mesa_DeleteTextures(1, &tempTex);
+         return false;
+      } else {
+         if (ctx->Driver.FinishRenderTexture &&
+             !rb->NeedsFinishRenderTexture) {
+            rb->NeedsFinishRenderTexture = true;
+            ctx->Driver.FinishRenderTexture(ctx, rb);
+         }
+
+         if (_mesa_is_winsys_fbo(readFb)) {
+            GLint temp = srcY0;
+            srcY0 = rb->Height - srcY1;
+            srcY1 = rb->Height - temp;
+            flipY = -flipY;
+         }
+      }
+   } else {
+      return false;
+   }
+
+   baseLevelSave = texObj->BaseLevel;
+   maxLevelSave = texObj->MaxLevel;
+
+   if (glsl_version) {
+      setup_glsl_blit_framebuffer(ctx, blit, rb, target);
+   }
+   else {
+      _mesa_meta_setup_ff_tnl_for_blit(&ctx->Meta->Blit.VAO,
+                                       &ctx->Meta->Blit.VBO,
+                                       2);
+   }
+
+   _mesa_GenSamplers(1, &sampler);
+   _mesa_BindSampler(ctx->Texture.CurrentUnit, sampler);
+
+   /*
+     printf("Blit from texture!\n");
+     printf("  srcAtt %p  dstAtt %p\n", readAtt, drawAtt);
+     printf("  srcTex %p  dstText %p\n", texObj, drawAtt->Texture);
+   */
+
+   /* Prepare src texture state */
+   _mesa_BindTexture(target, texObj->Name);
+   _mesa_SamplerParameteri(sampler, GL_TEXTURE_MIN_FILTER, filter);
+   _mesa_SamplerParameteri(sampler, GL_TEXTURE_MAG_FILTER, filter);
+   if (target != GL_TEXTURE_RECTANGLE_ARB) {
+      _mesa_TexParameteri(target, GL_TEXTURE_BASE_LEVEL, srcLevel);
+      _mesa_TexParameteri(target, GL_TEXTURE_MAX_LEVEL, srcLevel);
+   }
+   _mesa_SamplerParameteri(sampler, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
+   _mesa_SamplerParameteri(sampler, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+
+   /* Always do our blits with no net sRGB decode or encode.
+    *
+    * However, if both the src and dst can be srgb decode/encoded, enable them
+    * so that we do any blending (from scaling or from MSAA resolves) in the
+    * right colorspace.
+    *
+    * Our choice of not doing any net encode/decode is from the GL 3.0
+    * specification:
+    *
+    *     "Blit operations bypass the fragment pipeline. The only fragment
+    *      operations which affect a blit are the pixel ownership test and the
+    *      scissor test."
+    *
+    * The GL 4.4 specification disagrees and says that the sRGB part of the
+    * fragment pipeline applies, but this was found to break applications.
+    */
+   if (ctx->Extensions.EXT_texture_sRGB_decode) {
+      if (_mesa_get_format_color_encoding(rb->Format) == GL_SRGB &&
+          ctx->DrawBuffer->Visual.sRGBCapable) {
+         _mesa_SamplerParameteri(sampler, GL_TEXTURE_SRGB_DECODE_EXT,
+                                 GL_DECODE_EXT);
+         _mesa_set_framebuffer_srgb(ctx, GL_TRUE);
+      } else {
+         _mesa_SamplerParameteri(sampler, GL_TEXTURE_SRGB_DECODE_EXT,
+                                 GL_SKIP_DECODE_EXT);
+         /* set_framebuffer_srgb was set by _mesa_meta_begin(). */
+      }
+   }
+
+   if (!glsl_version) {
+      _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_REPLACE);
+      _mesa_set_enable(ctx, target, GL_TRUE);
+   }
+
+   /* Prepare vertex data (the VBO was previously created and bound) */
+   {
+      struct vertex verts[4];
+      GLfloat s0, t0, s1, t1;
+
+      if (target == GL_TEXTURE_2D) {
+         const struct gl_texture_image *texImage
+            = _mesa_select_tex_image(ctx, texObj, target, srcLevel);
+         s0 = srcX0 / (float) texImage->Width;
+         s1 = srcX1 / (float) texImage->Width;
+         t0 = srcY0 / (float) texImage->Height;
+         t1 = srcY1 / (float) texImage->Height;
+      }
+      else {
+         assert(target == GL_TEXTURE_RECTANGLE_ARB ||
+                target == GL_TEXTURE_2D_MULTISAMPLE);
+         s0 = (float) srcX0;
+         s1 = (float) srcX1;
+         t0 = (float) srcY0;
+         t1 = (float) srcY1;
+      }
+
+      /* Silence valgrind warnings about reading uninitialized stack. */
+      memset(verts, 0, sizeof(verts));
+
+      /* setup vertex positions */
+      verts[0].x = -1.0F * flipX;
+      verts[0].y = -1.0F * flipY;
+      verts[1].x =  1.0F * flipX;
+      verts[1].y = -1.0F * flipY;
+      verts[2].x =  1.0F * flipX;
+      verts[2].y =  1.0F * flipY;
+      verts[3].x = -1.0F * flipX;
+      verts[3].y =  1.0F * flipY;
+
+      verts[0].tex[0] = s0;
+      verts[0].tex[1] = t0;
+      verts[1].tex[0] = s1;
+      verts[1].tex[1] = t0;
+      verts[2].tex[0] = s1;
+      verts[2].tex[1] = t1;
+      verts[3].tex[0] = s0;
+      verts[3].tex[1] = t1;
+
+      _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
+   }
+
+   /* setup viewport */
+   _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
+   _mesa_ColorMask(!do_depth, !do_depth, !do_depth, !do_depth);
+   _mesa_set_enable(ctx, GL_DEPTH_TEST, do_depth);
+   _mesa_DepthMask(do_depth);
+   _mesa_DepthFunc(GL_ALWAYS);
+
+   _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
+
+   /* Restore texture object state, the texture binding will
+    * be restored by _mesa_meta_end().
+    */
+   if (target != GL_TEXTURE_RECTANGLE_ARB) {
+      _mesa_TexParameteri(target, GL_TEXTURE_BASE_LEVEL, baseLevelSave);
+      _mesa_TexParameteri(target, GL_TEXTURE_MAX_LEVEL, maxLevelSave);
+   }
+
+   _mesa_BindSampler(ctx->Texture.CurrentUnit, samplerSave);
+   _mesa_DeleteSamplers(1, &sampler);
+   if (tempTex)
+      _mesa_DeleteTextures(1, &tempTex);
+
+   return true;
+}
+
+/**
+ * Meta implementation of ctx->Driver.BlitFramebuffer() in terms
+ * of texture mapping and polygon rendering.
+ */
+void
+_mesa_meta_BlitFramebuffer(struct gl_context *ctx,
+                           GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
+                           GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,
+                           GLbitfield mask, GLenum filter)
+{
+   struct blit_state *blit = &ctx->Meta->Blit;
+   struct temp_texture *tex = _mesa_meta_get_temp_texture(ctx);
+   struct temp_texture *depthTex = _mesa_meta_get_temp_depth_texture(ctx);
+   const GLsizei maxTexSize = tex->MaxSize;
+   const GLint srcX = MIN2(srcX0, srcX1);
+   const GLint srcY = MIN2(srcY0, srcY1);
+   const GLint srcW = abs(srcX1 - srcX0);
+   const GLint srcH = abs(srcY1 - srcY0);
+   const GLint dstX = MIN2(dstX0, dstX1);
+   const GLint dstY = MIN2(dstY0, dstY1);
+   const GLint dstW = abs(dstX1 - dstX0);
+   const GLint dstH = abs(dstY1 - dstY0);
+   const GLint srcFlipX = (srcX1 - srcX0) / srcW;
+   const GLint srcFlipY = (srcY1 - srcY0) / srcH;
+   const GLint dstFlipX = (dstX1 - dstX0) / dstW;
+   const GLint dstFlipY = (dstY1 - dstY0) / dstH;
+   const GLint flipX = srcFlipX * dstFlipX;
+   const GLint flipY = srcFlipY * dstFlipY;
+
+   struct vertex verts[4];
+   GLboolean newTex;
+   const GLboolean use_glsl_version = ctx->Extensions.ARB_vertex_shader &&
+                                      ctx->Extensions.ARB_fragment_shader;
+
+   /* In addition to falling back if the blit size is larger than the maximum
+    * texture size, fallback if the source is multisampled.  This fallback can
+    * be removed once Mesa gets support ARB_texture_multisample.
+    */
+   if (srcW > maxTexSize || srcH > maxTexSize) {
+      /* XXX avoid this fallback */
+      goto fallback;
+   }
+
+   /* Multisample texture blit support requires texture multisample. */
+   if (ctx->ReadBuffer->Visual.samples > 0 &&
+       !ctx->Extensions.ARB_texture_multisample) {
+      goto fallback;
+   }
+
+   /* only scissor effects blit so save/clear all other relevant state */
+   _mesa_meta_begin(ctx, ~MESA_META_SCISSOR);
+
+   /* Try faster, direct texture approach first */
+   if (mask & GL_COLOR_BUFFER_BIT) {
+      if (blitframebuffer_texture(ctx, srcX0, srcY0, srcX1, srcY1,
+                                  dstX0, dstY0, dstX1, dstY1,
+                                  filter, dstFlipX, dstFlipY,
+                                  use_glsl_version, false)) {
+         mask &= ~GL_COLOR_BUFFER_BIT;
+         if (mask == 0x0) {
+            _mesa_meta_end(ctx);
+            return;
+         }
+      }
+   }
+
+   if (mask & GL_DEPTH_BUFFER_BIT && use_glsl_version) {
+      if (blitframebuffer_texture(ctx, srcX0, srcY0, srcX1, srcY1,
+                                  dstX0, dstY0, dstX1, dstY1,
+                                  filter, dstFlipX, dstFlipY,
+                                  use_glsl_version, true)) {
+         mask &= ~GL_DEPTH_BUFFER_BIT;
+         if (mask == 0x0) {
+            _mesa_meta_end(ctx);
+            return;
+         }
+      }
+   }
+
+   /* Choose between glsl version and fixed function version of
+    * BlitFramebuffer function.
+    */
+   if (use_glsl_version) {
+      setup_glsl_blit_framebuffer(ctx, blit, NULL, tex->Target);
+   }
+   else {
+      _mesa_meta_setup_ff_tnl_for_blit(&blit->VAO, &blit->VBO, 2);
+   }
+
+   /* Silence valgrind warnings about reading uninitialized stack. */
+   memset(verts, 0, sizeof(verts));
+
+   /* Continue with "normal" approach which involves copying the src rect
+    * into a temporary texture and is "blitted" by drawing a textured quad.
+    */
+   {
+      /* setup vertex positions */
+      verts[0].x = -1.0F * flipX;
+      verts[0].y = -1.0F * flipY;
+      verts[1].x =  1.0F * flipX;
+      verts[1].y = -1.0F * flipY;
+      verts[2].x =  1.0F * flipX;
+      verts[2].y =  1.0F * flipY;
+      verts[3].x = -1.0F * flipX;
+      verts[3].y =  1.0F * flipY;
+
+   }
+
+   if (!use_glsl_version)
+      _mesa_set_enable(ctx, tex->Target, GL_TRUE);
+
+   if (mask & GL_COLOR_BUFFER_BIT) {
+      const struct gl_framebuffer *readFb = ctx->ReadBuffer;
+      const struct gl_renderbuffer *colorReadRb = readFb->_ColorReadBuffer;
+      const GLenum rb_base_format =
+         _mesa_base_tex_format(ctx, colorReadRb->InternalFormat);
+
+      /* Using  the exact source rectangle to create the texture does incorrect
+       * linear filtering along the edges. So, allocate the texture extended along
+       * edges by one pixel in x, y directions.
+       */
+      _mesa_meta_setup_copypix_texture(ctx, tex,
+                                       srcX - 1, srcY - 1, srcW + 2, srcH + 2,
+                                       rb_base_format, filter);
+      /* texcoords (after texture allocation!) */
+      {
+         verts[0].tex[0] = 1.0F;
+         verts[0].tex[1] = 1.0F;
+         verts[1].tex[0] = tex->Sright - 1.0F;
+         verts[1].tex[1] = 1.0F;
+         verts[2].tex[0] = tex->Sright - 1.0F;
+         verts[2].tex[1] = tex->Ttop - 1.0F;
+         verts[3].tex[0] = 1.0F;
+         verts[3].tex[1] = tex->Ttop - 1.0F;
+
+         /* upload new vertex data */
+         _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
+      }
+
+      _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
+      _mesa_ColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
+      _mesa_set_enable(ctx, GL_DEPTH_TEST, GL_FALSE);
+      _mesa_DepthMask(GL_FALSE);
+      _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
+      mask &= ~GL_COLOR_BUFFER_BIT;
+   }
+
+   if ((mask & GL_DEPTH_BUFFER_BIT) &&
+       _mesa_is_desktop_gl(ctx) &&
+       ctx->Extensions.ARB_depth_texture &&
+       ctx->Extensions.ARB_fragment_program) {
+
+      GLuint *tmp = malloc(srcW * srcH * sizeof(GLuint));
+
+      if (tmp) {
+
+         newTex = _mesa_meta_alloc_texture(depthTex, srcW, srcH,
+                                           GL_DEPTH_COMPONENT);
+         _mesa_ReadPixels(srcX, srcY, srcW, srcH, GL_DEPTH_COMPONENT,
+                          GL_UNSIGNED_INT, tmp);
+         _mesa_meta_setup_drawpix_texture(ctx, depthTex, newTex,
+                                          srcW, srcH, GL_DEPTH_COMPONENT,
+                                          GL_UNSIGNED_INT, tmp);
+
+         /* texcoords (after texture allocation!) */
+         {
+            verts[0].tex[0] = 0.0F;
+            verts[0].tex[1] = 0.0F;
+            verts[1].tex[0] = depthTex->Sright;
+            verts[1].tex[1] = 0.0F;
+            verts[2].tex[0] = depthTex->Sright;
+            verts[2].tex[1] = depthTex->Ttop;
+            verts[3].tex[0] = 0.0F;
+            verts[3].tex[1] = depthTex->Ttop;
+
+            /* upload new vertex data */
+            _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
+         }
+
+         if (!blit->DepthFP)
+            init_blit_depth_pixels(ctx);
+
+         _mesa_BindProgramARB(GL_FRAGMENT_PROGRAM_ARB, blit->DepthFP);
+         _mesa_set_enable(ctx, GL_FRAGMENT_PROGRAM_ARB, GL_TRUE);
+         _mesa_ColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
+         _mesa_set_enable(ctx, GL_DEPTH_TEST, GL_TRUE);
+         _mesa_DepthFunc(GL_ALWAYS);
+         _mesa_DepthMask(GL_TRUE);
+
+         _mesa_set_viewport(ctx, 0, dstX, dstY, dstW, dstH);
+         _mesa_BufferSubData(GL_ARRAY_BUFFER_ARB, 0, sizeof(verts), verts);
+         _mesa_DrawArrays(GL_TRIANGLE_FAN, 0, 4);
+         mask &= ~GL_DEPTH_BUFFER_BIT;
+
+         free(tmp);
+      }
+   }
+
+   if (mask & GL_STENCIL_BUFFER_BIT) {
+      /* XXX can't easily do stencil */
+   }
+
+   if (!use_glsl_version)
+      _mesa_set_enable(ctx, tex->Target, GL_FALSE);
+
+   _mesa_meta_end(ctx);
+
+fallback:
+   if (mask) {
+      _swrast_BlitFramebuffer(ctx, srcX0, srcY0, srcX1, srcY1,
+                              dstX0, dstY0, dstX1, dstY1, mask, filter);
+   }
+}
+
+void
+_mesa_meta_glsl_blit_cleanup(struct blit_state *blit)
+{
+   if (blit->VAO) {
+      _mesa_DeleteVertexArrays(1, &blit->VAO);
+      blit->VAO = 0;
+      _mesa_DeleteBuffers(1, &blit->VBO);
+      blit->VBO = 0;
+   }
+   if (blit->DepthFP) {
+      _mesa_DeleteProgramsARB(1, &blit->DepthFP);
+      blit->DepthFP = 0;
+   }
+
+   _mesa_meta_blit_shader_table_cleanup(&blit->shaders);
+
+   _mesa_DeleteTextures(1, &blit->depthTex.TexObj);
+   blit->depthTex.TexObj = 0;
+}
diff --git a/mesalib/src/mesa/drivers/dri/common/dri_util.c b/mesalib/src/mesa/drivers/dri/common/dri_util.c
index 22c0848c8..78cfcf536 100644
--- a/mesalib/src/mesa/drivers/dri/common/dri_util.c
+++ b/mesalib/src/mesa/drivers/dri/common/dri_util.c
@@ -49,6 +49,7 @@
 #include "../glsl/glsl_parser_extras.h"
 #include "main/mtypes.h"
 #include "main/version.h"
+#include "main/errors.h"
 #include "main/macros.h"
 
 PUBLIC const char __dri2ConfigOptions[] =
@@ -452,8 +453,11 @@ driContextSetFlags(struct gl_context *ctx, uint32_t flags)
     if ((flags & __DRI_CTX_FLAG_FORWARD_COMPATIBLE) != 0)
         ctx->Const.ContextFlags |= GL_CONTEXT_FLAG_FORWARD_COMPATIBLE_BIT;
     if ((flags & __DRI_CTX_FLAG_DEBUG) != 0) {
+        struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+        if (debug) {
+            debug->DebugOutput = GL_TRUE;
+        }
         ctx->Const.ContextFlags |= GL_CONTEXT_FLAG_DEBUG_BIT;
-        ctx->Debug.DebugOutput = GL_TRUE;
     }
 }
 
@@ -768,12 +772,13 @@ const __DRIdri2Extension driDRI2Extension = {
 };
 
 const __DRIswrastExtension driSWRastExtension = {
-    { __DRI_SWRAST, 4 },
-    driSWRastCreateNewScreen,
-    driCreateNewDrawable,
-    driCreateNewContextForAPI,
-    driCreateContextAttribs,
-    driSWRastCreateNewScreen2,
+    .base = { __DRI_SWRAST, 4 },
+
+    .createNewScreen            = driSWRastCreateNewScreen,
+    .createNewDrawable          = driCreateNewDrawable,
+    .createNewContextForAPI     = driCreateNewContextForAPI,
+    .createContextAttribs       = driCreateContextAttribs,
+    .createNewScreen2           = driSWRastCreateNewScreen2,
 };
 
 const __DRI2configQueryExtension dri2ConfigQueryExtension = {
@@ -873,7 +878,7 @@ driImageFormatToGLFormat(uint32_t image_format)
 
 /** Image driver interface */
 const __DRIimageDriverExtension driImageDriverExtension = {
-    /*.base = */{ __DRI_IMAGE_DRIVER, __DRI_IMAGE_DRIVER_VERSION },
+    /*.base =*/ { __DRI_IMAGE_DRIVER, 1 },
 
     /*.createNewScreen2           =*/ driCreateNewScreen2,
     /*.createNewDrawable          =*/ driCreateNewDrawable,
@@ -892,6 +897,7 @@ static void driCopySubBuffer(__DRIdrawable *pdp, int x, int y,
 
 /* for swrast only */
 const __DRIcopySubBufferExtension driCopySubBufferExtension = {
-   { __DRI_COPY_SUB_BUFFER, 1 },
-   /*.copySubBuffer =*/ driCopySubBuffer,
+   /*.base =*/ { __DRI_COPY_SUB_BUFFER, 1 },
+
+   /*.copySubBuffer               =*/ driCopySubBuffer,
 };
diff --git a/mesalib/src/mesa/main/api_arrayelt.c b/mesalib/src/mesa/main/api_arrayelt.c
index 29a57c8e5..05cbc0f1f 100644
--- a/mesalib/src/mesa/main/api_arrayelt.c
+++ b/mesalib/src/mesa/main/api_arrayelt.c
@@ -1458,7 +1458,8 @@ _ae_destroy_context(struct gl_context *ctx)
 static void
 check_vbo(AEcontext *actx, struct gl_buffer_object *vbo)
 {
-   if (_mesa_is_bufferobj(vbo) && !_mesa_bufferobj_mapped(vbo)) {
+   if (_mesa_is_bufferobj(vbo) &&
+       !_mesa_bufferobj_mapped(vbo, MAP_INTERNAL)) {
       GLuint i;
       for (i = 0; i < actx->nr_vbos; i++)
          if (actx->vbo[i] == vbo)
@@ -1633,7 +1634,8 @@ _ae_map_vbos(struct gl_context *ctx)
       ctx->Driver.MapBufferRange(ctx, 0,
 				 actx->vbo[i]->Size,
 				 GL_MAP_READ_BIT,
-				 actx->vbo[i]);
+				 actx->vbo[i],
+                                 MAP_INTERNAL);
 
    if (actx->nr_vbos)
       actx->mapped_vbos = GL_TRUE;
@@ -1655,7 +1657,7 @@ _ae_unmap_vbos(struct gl_context *ctx)
    assert (!actx->NewState);
 
    for (i = 0; i < actx->nr_vbos; i++)
-      ctx->Driver.UnmapBuffer(ctx, actx->vbo[i]);
+      ctx->Driver.UnmapBuffer(ctx, actx->vbo[i], MAP_INTERNAL);
 
    actx->mapped_vbos = GL_FALSE;
 }
@@ -1701,7 +1703,8 @@ _ae_ArrayElement(GLint elt)
    /* emit generic attribute elements */
    for (at = actx->attribs; at->func; at++) {
       const GLubyte *src
-         = ADD_POINTERS(at->array->BufferObj->Pointer, at->array->Ptr)
+         = ADD_POINTERS(at->array->BufferObj->Mappings[MAP_INTERNAL].Pointer,
+                        at->array->Ptr)
          + elt * at->array->StrideB;
       at->func(at->index, src);
    }
@@ -1709,7 +1712,8 @@ _ae_ArrayElement(GLint elt)
    /* emit conventional arrays elements */
    for (aa = actx->arrays; aa->offset != -1 ; aa++) {
       const GLubyte *src
-         = ADD_POINTERS(aa->array->BufferObj->Pointer, aa->array->Ptr)
+         = ADD_POINTERS(aa->array->BufferObj->Mappings[MAP_INTERNAL].Pointer,
+                        aa->array->Ptr)
          + elt * aa->array->StrideB;
       CALL_by_offset(disp, (array_func), aa->offset, ((const void *) src));
    }
diff --git a/mesalib/src/mesa/main/api_loopback.c b/mesalib/src/mesa/main/api_loopback.c
index 8a04174c7..d10ae15ea 100644
--- a/mesalib/src/mesa/main/api_loopback.c
+++ b/mesalib/src/mesa/main/api_loopback.c
@@ -34,7 +34,6 @@
 #include "api_loopback.h"
 #include "mtypes.h"
 #include "glapi/glapi.h"
-#include "glapi/glthread.h"
 #include "main/dispatch.h"
 #include "main/context.h"
 
diff --git a/mesalib/src/mesa/main/api_loopback.h b/mesalib/src/mesa/main/api_loopback.h
index 0b014ad5a..2195e01e3 100644
--- a/mesalib/src/mesa/main/api_loopback.h
+++ b/mesalib/src/mesa/main/api_loopback.h
@@ -32,7 +32,6 @@
 #include "main/macros.h" // ?
 #include "main/mtypes.h" // ?
 #include "glapi/glapi.h" // ?
-#include "glapi/glthread.h" // ?
 #include "main/dispatch.h" // ?
 #include "main/context.h" // ?
 
diff --git a/mesalib/src/mesa/main/api_validate.c b/mesalib/src/mesa/main/api_validate.c
index af469e046..f3fd1a475 100644
--- a/mesalib/src/mesa/main/api_validate.c
+++ b/mesalib/src/mesa/main/api_validate.c
@@ -69,7 +69,8 @@ _mesa_max_buffer_index(struct gl_context *ctx, GLuint count, GLenum type,
    if (_mesa_is_bufferobj(elementBuf)) {
       /* elements are in a user-defined buffer object.  need to map it */
       map = ctx->Driver.MapBufferRange(ctx, 0, elementBuf->Size,
-				       GL_MAP_READ_BIT, elementBuf);
+				       GL_MAP_READ_BIT, elementBuf,
+                                       MAP_INTERNAL);
       /* Actual address is the sum of pointers */
       indices = (const GLvoid *) ADD_POINTERS(map, (const GLubyte *) indices);
    }
@@ -92,7 +93,7 @@ _mesa_max_buffer_index(struct gl_context *ctx, GLuint count, GLenum type,
    }
 
    if (map) {
-      ctx->Driver.UnmapBuffer(ctx, elementBuf);
+      ctx->Driver.UnmapBuffer(ctx, elementBuf, MAP_INTERNAL);
    }
 
    return max;
@@ -865,7 +866,7 @@ valid_draw_indirect(struct gl_context *ctx,
       return GL_FALSE;
    }
 
-   if (_mesa_bufferobj_mapped(ctx->DrawIndirectBuffer)) {
+   if (_mesa_check_disallowed_mapping(ctx->DrawIndirectBuffer)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "%s(DRAW_INDIRECT_BUFFER is mapped)", name);
       return GL_FALSE;
diff --git a/mesalib/src/mesa/main/arrayobj.c b/mesalib/src/mesa/main/arrayobj.c
index b33ba8016..efb993012 100644
--- a/mesalib/src/mesa/main/arrayobj.c
+++ b/mesalib/src/mesa/main/arrayobj.c
@@ -119,7 +119,7 @@ _mesa_delete_vao(struct gl_context *ctx, struct gl_vertex_array_object *obj)
 {
    unbind_array_object_vbos(ctx, obj);
    _mesa_reference_buffer_object(ctx, &obj->IndexBufferObj, NULL);
-   _glthread_DESTROY_MUTEX(obj->Mutex);
+   mtx_destroy(&obj->Mutex);
    free(obj->Label);
    free(obj);
 }
@@ -142,7 +142,7 @@ _mesa_reference_vao_(struct gl_context *ctx,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_vertex_array_object *oldObj = *ptr;
 
-      _glthread_LOCK_MUTEX(oldObj->Mutex);
+      mtx_lock(&oldObj->Mutex);
       ASSERT(oldObj->RefCount > 0);
       oldObj->RefCount--;
 #if 0
@@ -150,7 +150,7 @@ _mesa_reference_vao_(struct gl_context *ctx,
              (void *) oldObj, oldObj->Name, oldObj->RefCount);
 #endif
       deleteFlag = (oldObj->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(oldObj->Mutex);
+      mtx_unlock(&oldObj->Mutex);
 
       if (deleteFlag) {
 	 ASSERT(ctx->Driver.DeleteArrayObject);
@@ -163,7 +163,7 @@ _mesa_reference_vao_(struct gl_context *ctx,
 
    if (vao) {
       /* reference new array object */
-      _glthread_LOCK_MUTEX(vao->Mutex);
+      mtx_lock(&vao->Mutex);
       if (vao->RefCount == 0) {
          /* this array's being deleted (look just above) */
          /* Not sure this can every really happen.  Warn if it does. */
@@ -178,7 +178,7 @@ _mesa_reference_vao_(struct gl_context *ctx,
 #endif
          *ptr = vao;
       }
-      _glthread_UNLOCK_MUTEX(vao->Mutex);
+      mtx_unlock(&vao->Mutex);
    }
 }
 
@@ -226,7 +226,7 @@ _mesa_initialize_vao(struct gl_context *ctx,
 
    obj->Name = name;
 
-   _glthread_INIT_MUTEX(obj->Mutex);
+   mtx_init(&obj->Mutex, mtx_plain);
    obj->RefCount = 1;
 
    /* Init the individual arrays */
diff --git a/mesalib/src/mesa/main/attrib.c b/mesalib/src/mesa/main/attrib.c
index 004528044..5a626f2f4 100644
--- a/mesalib/src/mesa/main/attrib.c
+++ b/mesalib/src/mesa/main/attrib.c
@@ -1457,6 +1457,7 @@ copy_array_object(struct gl_context *ctx,
 
    /* _Enabled must be the same than on push */
    dest->_Enabled = src->_Enabled;
+   dest->NewArrays = src->NewArrays;
    dest->_MaxElement = src->_MaxElement;
 }
 
diff --git a/mesalib/src/mesa/main/bufferobj.c b/mesalib/src/mesa/main/bufferobj.c
index 966e61c4c..c4417c922 100644
--- a/mesalib/src/mesa/main/bufferobj.c
+++ b/mesalib/src/mesa/main/bufferobj.c
@@ -207,11 +207,12 @@ static bool
 bufferobj_range_mapped(const struct gl_buffer_object *obj,
                        GLintptr offset, GLsizeiptr size)
 {
-   if (_mesa_bufferobj_mapped(obj)) {
+   if (_mesa_bufferobj_mapped(obj, MAP_USER)) {
       const GLintptr end = offset + size;
-      const GLintptr mapEnd = obj->Offset + obj->Length;
+      const GLintptr mapEnd = obj->Mappings[MAP_USER].Offset +
+                              obj->Mappings[MAP_USER].Length;
 
-      if (!(end <= obj->Offset || offset >= mapEnd)) {
+      if (!(end <= obj->Mappings[MAP_USER].Offset || offset >= mapEnd)) {
          return true;
       }
    }
@@ -269,6 +270,9 @@ buffer_object_subdata_range_good(struct gl_context * ctx, GLenum target,
       return NULL;
    }
 
+   if (bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_PERSISTENT_BIT)
+      return bufObj;
+
    if (mappedRange) {
       if (bufferobj_range_mapped(bufObj, offset, size)) {
          _mesa_error(ctx, GL_INVALID_OPERATION, "%s", caller);
@@ -276,7 +280,7 @@ buffer_object_subdata_range_good(struct gl_context * ctx, GLenum target,
       }
    }
    else {
-      if (_mesa_bufferobj_mapped(bufObj)) {
+      if (_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
          _mesa_error(ctx, GL_INVALID_OPERATION, "%s", caller);
          return NULL;
       }
@@ -413,7 +417,7 @@ _mesa_delete_buffer_object(struct gl_context *ctx,
    bufObj->RefCount = -1000;
    bufObj->Name = ~0;
 
-   _glthread_DESTROY_MUTEX(bufObj->Mutex);
+   mtx_destroy(&bufObj->Mutex);
    free(bufObj->Label);
    free(bufObj);
 }
@@ -435,7 +439,7 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_buffer_object *oldObj = *ptr;
 
-      _glthread_LOCK_MUTEX(oldObj->Mutex);
+      mtx_lock(&oldObj->Mutex);
       ASSERT(oldObj->RefCount > 0);
       oldObj->RefCount--;
 #if 0
@@ -443,7 +447,7 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
              (void *) oldObj, oldObj->Name, oldObj->RefCount);
 #endif
       deleteFlag = (oldObj->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(oldObj->Mutex);
+      mtx_unlock(&oldObj->Mutex);
 
       if (deleteFlag) {
 
@@ -465,7 +469,7 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
 
    if (bufObj) {
       /* reference new buffer */
-      _glthread_LOCK_MUTEX(bufObj->Mutex);
+      mtx_lock(&bufObj->Mutex);
       if (bufObj->RefCount == 0) {
          /* this buffer's being deleted (look just above) */
          /* Not sure this can every really happen.  Warn if it does. */
@@ -480,7 +484,7 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
 #endif
          *ptr = bufObj;
       }
-      _glthread_UNLOCK_MUTEX(bufObj->Mutex);
+      mtx_unlock(&bufObj->Mutex);
    }
 }
 
@@ -496,11 +500,10 @@ _mesa_initialize_buffer_object( struct gl_context *ctx,
    (void) target;
 
    memset(obj, 0, sizeof(struct gl_buffer_object));
-   _glthread_INIT_MUTEX(obj->Mutex);
+   mtx_init(&obj->Mutex, mtx_plain);
    obj->RefCount = 1;
    obj->Name = name;
    obj->Usage = GL_STATIC_DRAW_ARB;
-   obj->AccessFlags = 0;
 }
 
 
@@ -555,21 +558,21 @@ _mesa_total_buffer_object_memory(struct gl_context *ctx)
  */
 GLboolean
 _mesa_buffer_data( struct gl_context *ctx, GLenum target, GLsizeiptrARB size,
-		   const GLvoid * data, GLenum usage,
+		   const GLvoid * data, GLenum usage, GLenum storageFlags,
 		   struct gl_buffer_object * bufObj )
 {
    void * new_data;
 
    (void) target;
 
-   if (bufObj->Data)
-      _mesa_align_free( bufObj->Data );
+   _mesa_align_free( bufObj->Data );
 
    new_data = _mesa_align_malloc( size, ctx->Const.MinMapBufferAlignment );
    if (new_data) {
       bufObj->Data = (GLubyte *) new_data;
       bufObj->Size = size;
       bufObj->Usage = usage;
+      bufObj->StorageFlags = storageFlags;
 
       if (data) {
 	 memcpy( bufObj->Data, data, size );
@@ -672,33 +675,11 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
    GLsizeiptr i;
    GLubyte *dest;
 
-   if (_mesa_bufferobj_mapped(bufObj)) {
-      GLubyte *data = malloc(size);
-      GLubyte *dataStart = data;
-      if (data == NULL) {
-         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glClearBuffer[Sub]Data");
-         return;
-      }
-
-      if (clearValue == NULL) {
-         /* Clear with zeros, per the spec */
-         memset(data, 0, size);
-      }
-      else {
-         for (i = 0; i < size/clearValueSize; ++i) {
-            memcpy(data, clearValue, clearValueSize);
-            data += clearValueSize;
-         }
-      }
-      ctx->Driver.BufferSubData(ctx, offset, size, dataStart, bufObj);
-      return;
-   }
-
    ASSERT(ctx->Driver.MapBufferRange);
    dest = ctx->Driver.MapBufferRange(ctx, offset, size,
                                      GL_MAP_WRITE_BIT |
                                      GL_MAP_INVALIDATE_RANGE_BIT,
-                                     bufObj);
+                                     bufObj, MAP_INTERNAL);
 
    if (!dest) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "glClearBuffer[Sub]Data");
@@ -708,7 +689,7 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
    if (clearValue == NULL) {
       /* Clear with zeros, per the spec */
       memset(dest, 0, size);
-      ctx->Driver.UnmapBuffer(ctx, bufObj);
+      ctx->Driver.UnmapBuffer(ctx, bufObj, MAP_INTERNAL);
       return;
    }
 
@@ -717,7 +698,7 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
       dest += clearValueSize;
    }
 
-   ctx->Driver.UnmapBuffer(ctx, bufObj);
+   ctx->Driver.UnmapBuffer(ctx, bufObj, MAP_INTERNAL);
 }
 
 
@@ -728,16 +709,17 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
 static void *
 _mesa_buffer_map_range( struct gl_context *ctx, GLintptr offset,
                         GLsizeiptr length, GLbitfield access,
-                        struct gl_buffer_object *bufObj )
+                        struct gl_buffer_object *bufObj,
+                        gl_map_buffer_index index)
 {
    (void) ctx;
-   assert(!_mesa_bufferobj_mapped(bufObj));
+   assert(!_mesa_bufferobj_mapped(bufObj, index));
    /* Just return a direct pointer to the data */
-   bufObj->Pointer = bufObj->Data + offset;
-   bufObj->Length = length;
-   bufObj->Offset = offset;
-   bufObj->AccessFlags = access;
-   return bufObj->Pointer;
+   bufObj->Mappings[index].Pointer = bufObj->Data + offset;
+   bufObj->Mappings[index].Length = length;
+   bufObj->Mappings[index].Offset = offset;
+   bufObj->Mappings[index].AccessFlags = access;
+   return bufObj->Mappings[index].Pointer;
 }
 
 
@@ -748,7 +730,8 @@ _mesa_buffer_map_range( struct gl_context *ctx, GLintptr offset,
 static void
 _mesa_buffer_flush_mapped_range( struct gl_context *ctx,
                                  GLintptr offset, GLsizeiptr length,
-                                 struct gl_buffer_object *obj )
+                                 struct gl_buffer_object *obj,
+                                 gl_map_buffer_index index)
 {
    (void) ctx;
    (void) offset;
@@ -766,14 +749,15 @@ _mesa_buffer_flush_mapped_range( struct gl_context *ctx,
  * \sa glUnmapBufferARB, dd_function_table::UnmapBuffer
  */
 GLboolean
-_mesa_buffer_unmap( struct gl_context *ctx, struct gl_buffer_object *bufObj )
+_mesa_buffer_unmap(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 */
-   bufObj->Pointer = NULL;
-   bufObj->Length = 0;
-   bufObj->Offset = 0;
-   bufObj->AccessFlags = 0x0;
+   bufObj->Mappings[index].Pointer = NULL;
+   bufObj->Mappings[index].Length = 0;
+   bufObj->Mappings[index].Offset = 0;
+   bufObj->Mappings[index].AccessFlags = 0x0;
    return GL_TRUE;
 }
 
@@ -791,14 +775,11 @@ _mesa_copy_buffer_subdata(struct gl_context *ctx,
 {
    GLubyte *srcPtr, *dstPtr;
 
-   /* the buffers should not be mapped */
-   assert(!_mesa_bufferobj_mapped(src));
-   assert(!_mesa_bufferobj_mapped(dst));
-
    if (src == dst) {
       srcPtr = dstPtr = ctx->Driver.MapBufferRange(ctx, 0, src->Size,
 						   GL_MAP_READ_BIT |
-						   GL_MAP_WRITE_BIT, src);
+						   GL_MAP_WRITE_BIT, src,
+                                                   MAP_INTERNAL);
 
       if (!srcPtr)
 	 return;
@@ -807,10 +788,12 @@ _mesa_copy_buffer_subdata(struct gl_context *ctx,
       dstPtr += writeOffset;
    } else {
       srcPtr = ctx->Driver.MapBufferRange(ctx, readOffset, size,
-					  GL_MAP_READ_BIT, src);
+					  GL_MAP_READ_BIT, src,
+                                          MAP_INTERNAL);
       dstPtr = ctx->Driver.MapBufferRange(ctx, writeOffset, size,
 					  (GL_MAP_WRITE_BIT |
-					   GL_MAP_INVALIDATE_RANGE_BIT), dst);
+					   GL_MAP_INVALIDATE_RANGE_BIT), dst,
+                                          MAP_INTERNAL);
    }
 
    /* Note: the src and dst regions will never overlap.  Trying to do so
@@ -819,9 +802,9 @@ _mesa_copy_buffer_subdata(struct gl_context *ctx,
    if (srcPtr && dstPtr)
       memcpy(dstPtr, srcPtr, size);
 
-   ctx->Driver.UnmapBuffer(ctx, src);
+   ctx->Driver.UnmapBuffer(ctx, src, MAP_INTERNAL);
    if (dst != src)
-      ctx->Driver.UnmapBuffer(ctx, dst);
+      ctx->Driver.UnmapBuffer(ctx, dst, MAP_INTERNAL);
 }
 
 
@@ -835,7 +818,7 @@ _mesa_init_buffer_objects( struct gl_context *ctx )
    GLuint i;
 
    memset(&DummyBufferObject, 0, sizeof(DummyBufferObject));
-   _glthread_INIT_MUTEX(DummyBufferObject.Mutex);
+   mtx_init(&DummyBufferObject.Mutex, mtx_plain);
    DummyBufferObject.RefCount = 1000*1000*1000; /* never delete */
 
    _mesa_reference_buffer_object(ctx, &ctx->Array.ArrayBufferObj,
@@ -955,10 +938,6 @@ bind_buffer_object(struct gl_context *ctx, GLenum target, GLuint buffer)
    
    /* bind new buffer */
    _mesa_reference_buffer_object(ctx, bindTarget, newBufObj);
-
-   /* Pass BindBuffer call to device driver */
-   if (ctx->Driver.BindBuffer)
-      ctx->Driver.BindBuffer( ctx, target, newBufObj );
 }
 
 
@@ -1023,7 +1002,6 @@ _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->BindBuffer = NULL;
    driver->BufferData = _mesa_buffer_data;
    driver->BufferSubData = _mesa_buffer_subdata;
    driver->GetBufferSubData = _mesa_buffer_get_subdata;
@@ -1041,6 +1019,21 @@ _mesa_init_buffer_object_functions(struct dd_function_table *driver)
 }
 
 
+void
+_mesa_buffer_unmap_all_mappings(struct gl_context *ctx,
+                                struct gl_buffer_object *bufObj)
+{
+   int i;
+
+   for (i = 0; i < MAP_COUNT; i++) {
+      if (_mesa_bufferobj_mapped(bufObj, i)) {
+         ctx->Driver.UnmapBuffer(ctx, bufObj, i);
+         ASSERT(bufObj->Mappings[i].Pointer == NULL);
+         bufObj->Mappings[i].AccessFlags = 0;
+      }
+   }
+}
+
 
 /**********************************************************************/
 /* API Functions                                                      */
@@ -1077,7 +1070,7 @@ _mesa_DeleteBuffers(GLsizei n, const GLuint *ids)
       return;
    }
 
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
 
    for (i = 0; i < n; i++) {
       struct gl_buffer_object *bufObj = _mesa_lookup_bufferobj(ctx, ids[i]);
@@ -1087,12 +1080,7 @@ _mesa_DeleteBuffers(GLsizei n, const GLuint *ids)
 
          ASSERT(bufObj->Name == ids[i] || bufObj == &DummyBufferObject);
 
-         if (_mesa_bufferobj_mapped(bufObj)) {
-            /* if mapped, unmap it now */
-            ctx->Driver.UnmapBuffer(ctx, bufObj);
-            bufObj->AccessFlags = 0;
-            bufObj->Pointer = NULL;
-         }
+         _mesa_buffer_unmap_all_mappings(ctx, bufObj);
 
          /* unbind any vertex pointers bound to this buffer */
          for (j = 0; j < Elements(vao->VertexBinding); j++) {
@@ -1169,7 +1157,7 @@ _mesa_DeleteBuffers(GLsizei n, const GLuint *ids)
       }
    }
 
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 }
 
 
@@ -1201,7 +1189,7 @@ _mesa_GenBuffers(GLsizei n, GLuint *buffer)
    /*
     * This must be atomic (generation and allocation of buffer object IDs)
     */
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
 
    first = _mesa_HashFindFreeKeyBlock(ctx->Shared->BufferObjects, n);
 
@@ -1212,7 +1200,7 @@ _mesa_GenBuffers(GLsizei n, GLuint *buffer)
       buffer[i] = first + i;
    }
 
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 }
 
 
@@ -1230,15 +1218,73 @@ _mesa_IsBuffer(GLuint id)
    GET_CURRENT_CONTEXT(ctx);
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
 
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
    bufObj = _mesa_lookup_bufferobj(ctx, id);
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 
    return bufObj && bufObj != &DummyBufferObject;
 }
 
 
 void GLAPIENTRY
+_mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
+                    GLbitfield flags)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_buffer_object *bufObj;
+
+   if (size <= 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(size <= 0)");
+      return;
+   }
+
+   if (flags & ~(GL_MAP_READ_BIT |
+                 GL_MAP_WRITE_BIT |
+                 GL_MAP_PERSISTENT_BIT |
+                 GL_MAP_COHERENT_BIT |
+                 GL_DYNAMIC_STORAGE_BIT |
+                 GL_CLIENT_STORAGE_BIT)) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(flags)");
+      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)");
+      return;
+   }
+
+   if (flags & GL_MAP_COHERENT_BIT && !(flags & GL_MAP_PERSISTENT_BIT)) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glBufferStorage(flags!=PERSISTENT)");
+      return;
+   }
+
+   bufObj = get_buffer(ctx, "glBufferStorage", target, GL_INVALID_OPERATION);
+   if (!bufObj)
+      return;
+
+   if (bufObj->Immutable) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferStorage(immutable)");
+      return;
+   }
+
+   /* Unmap the existing buffer.  We'll replace it now.  Not an error. */
+   _mesa_buffer_unmap_all_mappings(ctx, bufObj);
+
+   FLUSH_VERTICES(ctx, _NEW_BUFFER_OBJECT);
+
+   bufObj->Written = GL_TRUE;
+   bufObj->Immutable = GL_TRUE;
+
+   ASSERT(ctx->Driver.BufferData);
+   if (!ctx->Driver.BufferData(ctx, target, size, data, GL_DYNAMIC_DRAW,
+                               flags, bufObj)) {
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBufferStorage()");
+   }
+}
+
+
+void GLAPIENTRY
 _mesa_BufferData(GLenum target, GLsizeiptrARB size,
                     const GLvoid * data, GLenum usage)
 {
@@ -1290,12 +1336,13 @@ _mesa_BufferData(GLenum target, GLsizeiptrARB size,
    if (!bufObj)
       return;
 
-   if (_mesa_bufferobj_mapped(bufObj)) {
-      /* Unmap the existing buffer.  We'll replace it now.  Not an error. */
-      ctx->Driver.UnmapBuffer(ctx, bufObj);
-      bufObj->AccessFlags = 0;
-      ASSERT(bufObj->Pointer == NULL);
-   }  
+   if (bufObj->Immutable) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferData(immutable)");
+      return;
+   }
+
+   /* Unmap the existing buffer.  We'll replace it now.  Not an error. */
+   _mesa_buffer_unmap_all_mappings(ctx, bufObj);
 
    FLUSH_VERTICES(ctx, _NEW_BUFFER_OBJECT);
 
@@ -1311,7 +1358,11 @@ _mesa_BufferData(GLenum target, GLsizeiptrARB size,
 #endif
 
    ASSERT(ctx->Driver.BufferData);
-   if (!ctx->Driver.BufferData( ctx, target, size, data, usage, bufObj )) {
+   if (!ctx->Driver.BufferData(ctx, target, size, data, usage,
+                               GL_MAP_READ_BIT |
+                               GL_MAP_WRITE_BIT |
+                               GL_DYNAMIC_STORAGE_BIT,
+                               bufObj)) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBufferDataARB()");
    }
 }
@@ -1332,6 +1383,12 @@ _mesa_BufferSubData(GLenum target, GLintptrARB offset,
       return;
    }
 
+   if (bufObj->Immutable &&
+       !(bufObj->StorageFlags & GL_DYNAMIC_STORAGE_BIT)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glBufferSubData");
+      return;
+   }
+
    if (size == 0)
       return;
 
@@ -1377,7 +1434,7 @@ _mesa_ClearBufferData(GLenum target, GLenum internalformat, GLenum format,
       return;
    }
 
-   if (_mesa_bufferobj_mapped(bufObj)) {
+   if (_mesa_check_disallowed_mapping(bufObj)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glClearBufferData(buffer currently mapped)");
       return;
@@ -1451,8 +1508,10 @@ _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
 
    if (data == NULL) {
       /* clear to zeros, per the spec */
-      ctx->Driver.ClearBufferSubData(ctx, offset, size,
-                                     NULL, 0, bufObj);
+      if (size > 0) {
+         ctx->Driver.ClearBufferSubData(ctx, offset, size,
+                                        NULL, 0, bufObj);
+      }
       return;
    }
 
@@ -1462,8 +1521,10 @@ _mesa_ClearBufferSubData(GLenum target, GLenum internalformat,
       return;
    }
 
-   ctx->Driver.ClearBufferSubData(ctx, offset, size,
-                                  clearValue, clearValueSize, bufObj);
+   if (size > 0) {
+      ctx->Driver.ClearBufferSubData(ctx, offset, size,
+                                     clearValue, clearValueSize, bufObj);
+   }
 }
 
 
@@ -1505,7 +1566,21 @@ _mesa_MapBuffer(GLenum target, GLenum access)
    if (!bufObj)
       return NULL;
 
-   if (_mesa_bufferobj_mapped(bufObj)) {
+   if (accessFlags & GL_MAP_READ_BIT &&
+       !(bufObj->StorageFlags & GL_MAP_READ_BIT)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapBuffer(invalid read flag)");
+      return NULL;
+   }
+
+   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;
    }
@@ -1517,7 +1592,8 @@ _mesa_MapBuffer(GLenum target, GLenum access)
    }
 
    ASSERT(ctx->Driver.MapBufferRange);
-   map = ctx->Driver.MapBufferRange(ctx, 0, bufObj->Size, accessFlags, bufObj);
+   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;
@@ -1527,10 +1603,10 @@ _mesa_MapBuffer(GLenum target, GLenum access)
        * This is important because other modules (like VBO) might call
        * the driver function directly.
        */
-      ASSERT(bufObj->Pointer == map);
-      ASSERT(bufObj->Length == bufObj->Size);
-      ASSERT(bufObj->Offset == 0);
-      bufObj->AccessFlags = accessFlags;
+      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;
    }
 
    if (access == GL_WRITE_ONLY_ARB || access == GL_READ_WRITE_ARB)
@@ -1558,7 +1634,7 @@ _mesa_MapBuffer(GLenum target, GLenum access)
    }
 #endif
 
-   return bufObj->Pointer;
+   return bufObj->Mappings[MAP_USER].Pointer;
 }
 
 
@@ -1574,7 +1650,7 @@ _mesa_UnmapBuffer(GLenum target)
    if (!bufObj)
       return GL_FALSE;
 
-   if (!_mesa_bufferobj_mapped(bufObj)) {
+   if (!_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "glUnmapBufferARB");
       return GL_FALSE;
    }
@@ -1615,11 +1691,11 @@ _mesa_UnmapBuffer(GLenum target)
    }
 #endif
 
-   status = ctx->Driver.UnmapBuffer( ctx, bufObj );
-   bufObj->AccessFlags = 0;
-   ASSERT(bufObj->Pointer == NULL);
-   ASSERT(bufObj->Offset == 0);
-   ASSERT(bufObj->Length == 0);
+   status = ctx->Driver.UnmapBuffer(ctx, bufObj, MAP_USER);
+   bufObj->Mappings[MAP_USER].AccessFlags = 0;
+   ASSERT(bufObj->Mappings[MAP_USER].Pointer == NULL);
+   ASSERT(bufObj->Mappings[MAP_USER].Offset == 0);
+   ASSERT(bufObj->Mappings[MAP_USER].Length == 0);
 
    return status;
 }
@@ -1644,25 +1720,36 @@ _mesa_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params)
       *params = bufObj->Usage;
       return;
    case GL_BUFFER_ACCESS_ARB:
-      *params = simplified_access_mode(ctx, bufObj->AccessFlags);
+      *params = simplified_access_mode(ctx,
+                            bufObj->Mappings[MAP_USER].AccessFlags);
       return;
    case GL_BUFFER_MAPPED_ARB:
-      *params = _mesa_bufferobj_mapped(bufObj);
+      *params = _mesa_bufferobj_mapped(bufObj, MAP_USER);
       return;
    case GL_BUFFER_ACCESS_FLAGS:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = bufObj->AccessFlags;
+      *params = bufObj->Mappings[MAP_USER].AccessFlags;
       return;
    case GL_BUFFER_MAP_OFFSET:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = (GLint) bufObj->Offset;
+      *params = (GLint) bufObj->Mappings[MAP_USER].Offset;
       return;
    case GL_BUFFER_MAP_LENGTH:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = (GLint) bufObj->Length;
+      *params = (GLint) bufObj->Mappings[MAP_USER].Length;
+      return;
+   case GL_BUFFER_IMMUTABLE_STORAGE:
+      if (!ctx->Extensions.ARB_buffer_storage)
+         goto invalid_pname;
+      *params = bufObj->Immutable;
+      return;
+   case GL_BUFFER_STORAGE_FLAGS:
+      if (!ctx->Extensions.ARB_buffer_storage)
+         goto invalid_pname;
+      *params = bufObj->StorageFlags;
       return;
    default:
       ; /* fall-through */
@@ -1698,25 +1785,36 @@ _mesa_GetBufferParameteri64v(GLenum target, GLenum pname, GLint64 *params)
       *params = bufObj->Usage;
       return;
    case GL_BUFFER_ACCESS_ARB:
-      *params = simplified_access_mode(ctx, bufObj->AccessFlags);
+      *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->AccessFlags;
+      *params = bufObj->Mappings[MAP_USER].AccessFlags;
       return;
    case GL_BUFFER_MAPPED_ARB:
-      *params = _mesa_bufferobj_mapped(bufObj);
+      *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->Offset;
+      *params = bufObj->Mappings[MAP_USER].Offset;
       return;
    case GL_BUFFER_MAP_LENGTH:
       if (!ctx->Extensions.ARB_map_buffer_range)
          goto invalid_pname;
-      *params = bufObj->Length;
+      *params = bufObj->Mappings[MAP_USER].Length;
+      return;
+   case GL_BUFFER_IMMUTABLE_STORAGE:
+      if (!ctx->Extensions.ARB_buffer_storage)
+         goto invalid_pname;
+      *params = bufObj->Immutable;
+      return;
+   case GL_BUFFER_STORAGE_FLAGS:
+      if (!ctx->Extensions.ARB_buffer_storage)
+         goto invalid_pname;
+      *params = bufObj->StorageFlags;
       return;
    default:
       ; /* fall-through */
@@ -1744,7 +1842,7 @@ _mesa_GetBufferPointerv(GLenum target, GLenum pname, GLvoid **params)
    if (!bufObj)
       return;
 
-   *params = bufObj->Pointer;
+   *params = bufObj->Mappings[MAP_USER].Pointer;
 }
 
 
@@ -1766,13 +1864,13 @@ _mesa_CopyBufferSubData(GLenum readTarget, GLenum writeTarget,
    if (!dst)
       return;
 
-   if (_mesa_bufferobj_mapped(src)) {
+   if (_mesa_check_disallowed_mapping(src)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glCopyBufferSubData(readBuffer is mapped)");
       return;
    }
 
-   if (_mesa_bufferobj_mapped(dst)) {
+   if (_mesa_check_disallowed_mapping(dst)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glCopyBufferSubData(writeBuffer is mapped)");
       return;
@@ -1839,6 +1937,7 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
    GET_CURRENT_CONTEXT(ctx);
    struct gl_buffer_object *bufObj;
    void *map;
+   GLbitfield allowed_access;
 
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, NULL);
 
@@ -1873,13 +1972,20 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
       return NULL;
    }
 
-   if (access & ~(GL_MAP_READ_BIT |
-                  GL_MAP_WRITE_BIT |
-                  GL_MAP_INVALIDATE_RANGE_BIT |
-                  GL_MAP_INVALIDATE_BUFFER_BIT |
-                  GL_MAP_FLUSH_EXPLICIT_BIT |
-                  GL_MAP_UNSYNCHRONIZED_BIT)) {
-      /* generate an error if any undefind bit is set */
+   allowed_access = GL_MAP_READ_BIT |
+                    GL_MAP_WRITE_BIT |
+                    GL_MAP_INVALIDATE_RANGE_BIT |
+                    GL_MAP_INVALIDATE_BUFFER_BIT |
+                    GL_MAP_FLUSH_EXPLICIT_BIT |
+                    GL_MAP_UNSYNCHRONIZED_BIT;
+
+   if (ctx->Extensions.ARB_buffer_storage) {
+         allowed_access |= GL_MAP_PERSISTENT_BIT |
+                           GL_MAP_COHERENT_BIT;
+   }
+
+   if (access & ~allowed_access) {
+      /* generate an error if any other than allowed bit is set */
       _mesa_error(ctx, GL_INVALID_VALUE, "glMapBufferRange(access)");
       return NULL;
    }
@@ -1910,13 +2016,41 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
    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)");
+      return NULL;
+   }
+
+   if (access & GL_MAP_WRITE_BIT &&
+       !(bufObj->StorageFlags & GL_MAP_WRITE_BIT)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapBufferRange(invalid write flag)");
+      return NULL;
+   }
+
+   if (access & GL_MAP_COHERENT_BIT &&
+       !(bufObj->StorageFlags & GL_MAP_COHERENT_BIT)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapBufferRange(invalid coherent flag)");
+      return NULL;
+   }
+
+   if (access & GL_MAP_PERSISTENT_BIT &&
+       !(bufObj->StorageFlags & GL_MAP_PERSISTENT_BIT)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glMapBufferRange(invalid persistent flag)");
+      return NULL;
+   }
+
    if (offset + length > bufObj->Size) {
       _mesa_error(ctx, GL_INVALID_VALUE,
                   "glMapBufferRange(offset + length > size)");
       return NULL;
    }
 
-   if (_mesa_bufferobj_mapped(bufObj)) {
+   if (_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glMapBufferRange(buffer already mapped)");
       return NULL;
@@ -1931,15 +2065,16 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
    /* Mapping zero bytes should return a non-null pointer. */
    if (!length) {
       static long dummy = 0;
-      bufObj->Pointer = &dummy;
-      bufObj->Length = length;
-      bufObj->Offset = offset;
-      bufObj->AccessFlags = access;
-      return bufObj->Pointer;
+      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 = ctx->Driver.MapBufferRange(ctx, offset, length, access, bufObj,
+                                    MAP_USER);
    if (!map) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "glMapBufferARB(map failed)");
    }
@@ -1948,10 +2083,10 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
        * This is important because other modules (like VBO) might call
        * the driver function directly.
        */
-      ASSERT(bufObj->Pointer == map);
-      ASSERT(bufObj->Length == length);
-      ASSERT(bufObj->Offset == offset);
-      ASSERT(bufObj->AccessFlags == access);
+      ASSERT(bufObj->Mappings[MAP_USER].Pointer == map);
+      ASSERT(bufObj->Mappings[MAP_USER].Length == length);
+      ASSERT(bufObj->Mappings[MAP_USER].Offset == offset);
+      ASSERT(bufObj->Mappings[MAP_USER].AccessFlags == access);
    }
 
    return map;
@@ -1990,30 +2125,33 @@ _mesa_FlushMappedBufferRange(GLenum target, GLintptr offset, GLsizeiptr length)
    if (!bufObj)
       return;
 
-   if (!_mesa_bufferobj_mapped(bufObj)) {
+   if (!_mesa_bufferobj_mapped(bufObj, MAP_USER)) {
       /* buffer is not mapped */
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glFlushMappedBufferRange(buffer is not mapped)");
       return;
    }
 
-   if ((bufObj->AccessFlags & GL_MAP_FLUSH_EXPLICIT_BIT) == 0) {
+   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)");
       return;
    }
 
-   if (offset + length > bufObj->Length) {
+   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->Length);
+		  (long)offset, (long)length,
+                  (long)bufObj->Mappings[MAP_USER].Length);
       return;
    }
 
-   ASSERT(bufObj->AccessFlags & GL_MAP_WRITE_BIT);
+   ASSERT(bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_WRITE_BIT);
 
    if (ctx->Driver.FlushMappedBufferRange)
-      ctx->Driver.FlushMappedBufferRange(ctx, offset, length, bufObj);
+      ctx->Driver.FlushMappedBufferRange(ctx, offset, length, bufObj,
+                                         MAP_USER);
 }
 
 
@@ -2660,13 +2798,15 @@ _mesa_InvalidateBufferSubData(GLuint buffer, GLintptr offset,
       return;
    }
 
-   /* The GL_ARB_invalidate_subdata spec says:
+   /* The OpenGL 4.4 (Core Profile) spec says:
     *
-    *     "An INVALID_OPERATION error is generated if the buffer is currently
-    *     mapped by MapBuffer, or if the invalidate range intersects the range
-    *     currently mapped by MapBufferRange."
+    *     "An INVALID_OPERATION error is generated if buffer is currently
+    *     mapped by MapBuffer or if the invalidate range intersects the range
+    *     currently mapped by MapBufferRange, unless it was mapped
+    *     with MAP_PERSISTENT_BIT set in the MapBufferRange access flags."
     */
-   if (bufferobj_range_mapped(bufObj, offset, length)) {
+   if (!(bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_PERSISTENT_BIT) &&
+       bufferobj_range_mapped(bufObj, offset, length)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glInvalidateBufferSubData(intersection with mapped "
                   "range)");
@@ -2693,13 +2833,14 @@ _mesa_InvalidateBufferData(GLuint buffer)
       return;
    }
 
-   /* The GL_ARB_invalidate_subdata spec says:
+   /* The OpenGL 4.4 (Core Profile) spec says:
     *
-    *     "An INVALID_OPERATION error is generated if the buffer is currently
-    *     mapped by MapBuffer, or if the invalidate range intersects the range
-    *     currently mapped by MapBufferRange."
+    *     "An INVALID_OPERATION error is generated if buffer is currently
+    *     mapped by MapBuffer or if the invalidate range intersects the range
+    *     currently mapped by MapBufferRange, unless it was mapped
+    *     with MAP_PERSISTENT_BIT set in the MapBufferRange access flags."
     */
-   if (_mesa_bufferobj_mapped(bufObj)) {
+   if (_mesa_check_disallowed_mapping(bufObj)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glInvalidateBufferData(intersection with mapped "
                   "range)");
diff --git a/mesalib/src/mesa/main/bufferobj.h b/mesalib/src/mesa/main/bufferobj.h
index 71988b0d9..9814552eb 100644
--- a/mesalib/src/mesa/main/bufferobj.h
+++ b/mesalib/src/mesa/main/bufferobj.h
@@ -37,11 +37,21 @@
  */
 
 
-/** Is the given buffer object currently mapped? */
+/** Is the given buffer object currently mapped by the GL user? */
 static inline GLboolean
-_mesa_bufferobj_mapped(const struct gl_buffer_object *obj)
+_mesa_bufferobj_mapped(const struct gl_buffer_object *obj,
+                       gl_map_buffer_index index)
 {
-   return obj->Pointer != NULL;
+   return obj->Mappings[index].Pointer != NULL;
+}
+
+/** Can we not use this buffer while mapped? */
+static inline GLboolean
+_mesa_check_disallowed_mapping(const struct gl_buffer_object *obj)
+{
+   return _mesa_bufferobj_mapped(obj, MAP_USER) &&
+          !(obj->Mappings[MAP_USER].AccessFlags &
+            GL_MAP_PERSISTENT_BIT);
 }
 
 /**
@@ -101,6 +111,9 @@ _mesa_total_buffer_object_memory(struct gl_context *ctx);
 extern void
 _mesa_init_buffer_object_functions(struct dd_function_table *driver);
 
+extern void
+_mesa_buffer_unmap_all_mappings(struct gl_context *ctx,
+                                struct gl_buffer_object *bufObj);
 
 /*
  * API functions
@@ -118,6 +131,10 @@ GLboolean GLAPIENTRY
 _mesa_IsBuffer(GLuint buffer);
 
 void GLAPIENTRY
+_mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
+                    GLbitfield flags);
+
+void GLAPIENTRY
 _mesa_BufferData(GLenum target, GLsizeiptrARB size,
                  const GLvoid * data, GLenum usage);
 
diff --git a/mesalib/src/mesa/main/context.c b/mesalib/src/mesa/main/context.c
index 9372b3bea..a19180870 100644
--- a/mesalib/src/mesa/main/context.c
+++ b/mesalib/src/mesa/main/context.c
@@ -106,6 +106,7 @@
 #include "matrix.h"
 #include "multisample.h"
 #include "performance_monitor.h"
+#include "pipelineobj.h"
 #include "pixel.h"
 #include "pixelstore.h"
 #include "points.h"
@@ -363,7 +364,7 @@ dummy_enum_func(void)
  *
  * \sa Used by one_time_init().
  */
-_glthread_DECLARE_STATIC_MUTEX(OneTimeLock);
+mtx_t OneTimeLock = _MTX_INITIALIZER_NP;
 
 
 
@@ -381,7 +382,7 @@ one_time_init( struct gl_context *ctx )
 {
    static GLbitfield api_init_mask = 0x0;
 
-   _glthread_LOCK_MUTEX(OneTimeLock);
+   mtx_lock(&OneTimeLock);
 
    /* truly one-time init */
    if (!api_init_mask) {
@@ -422,7 +423,7 @@ one_time_init( struct gl_context *ctx )
 
    api_init_mask |= 1 << ctx->API;
 
-   _glthread_UNLOCK_MUTEX(OneTimeLock);
+   mtx_unlock(&OneTimeLock);
 
    /* Hopefully atexit() is widely available.  If not, we may need some
     * #ifdef tests here.
@@ -603,9 +604,6 @@ _mesa_init_constants(struct gl_context *ctx)
    ctx->Const.ViewportBounds.Min = 0;
    ctx->Const.ViewportBounds.Max = 0;
 
-   /* Driver must override if it supports ARB_viewport_array */
-   ctx->Const.MaxViewports = 1;
-
    /** GL_ARB_uniform_buffer_object */
    ctx->Const.MaxCombinedUniformBlocks = 36;
    ctx->Const.MaxUniformBufferBindings = 36;
@@ -814,6 +812,7 @@ init_attrib_groups(struct gl_context *ctx)
    _mesa_init_matrix( ctx );
    _mesa_init_multisample( ctx );
    _mesa_init_performance_monitors( ctx );
+   _mesa_init_pipeline( ctx );
    _mesa_init_pixel( ctx );
    _mesa_init_pixelstore( ctx );
    _mesa_init_point( ctx );
@@ -1219,6 +1218,7 @@ _mesa_free_context_data( struct gl_context *ctx )
    _mesa_free_texture_data( ctx );
    _mesa_free_matrix_data( ctx );
    _mesa_free_viewport_data( ctx );
+   _mesa_free_pipeline_data(ctx);
    _mesa_free_program_data(ctx);
    _mesa_free_shader_state(ctx);
    _mesa_free_queryobj_data(ctx);
@@ -1769,93 +1769,59 @@ _mesa_check_blend_func_error(struct gl_context *ctx)
    return GL_TRUE;
 }
 
-/**
- * Prior to drawing anything with glBegin, glDrawArrays, etc. this function
- * is called to see if it's valid to render.  This involves checking that
- * the current shader is valid and the framebuffer is complete.
- * If an error is detected it'll be recorded here.
- * \return GL_TRUE if OK to render, GL_FALSE if not
- */
-GLboolean
-_mesa_valid_to_render(struct gl_context *ctx, const char *where)
+static bool
+shader_linked_or_absent(struct gl_context *ctx,
+                        const struct gl_shader_program *shProg,
+                        bool *shader_present, const char *where)
 {
-   bool vert_from_glsl_shader = false;
-   bool geom_from_glsl_shader = false;
-   bool frag_from_glsl_shader = false;
-
-   /* This depends on having up to date derived state (shaders) */
-   if (ctx->NewState)
-      _mesa_update_state(ctx);
+   if (shProg) {
+      *shader_present = true;
 
-   if (ctx->Shader.CurrentProgram[MESA_SHADER_VERTEX]) {
-      vert_from_glsl_shader = true;
-
-      if (!ctx->Shader.CurrentProgram[MESA_SHADER_VERTEX]->LinkStatus) {
-         _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s(shader not linked)", where);
-         return GL_FALSE;
+      if (!shProg->LinkStatus) {
+         _mesa_error(ctx, GL_INVALID_OPERATION, "%s(shader not linked)", where);
+         return false;
       }
 #if 0 /* not normally enabled */
       {
          char errMsg[100];
-         if (!_mesa_validate_shader_program(ctx,
-					    ctx->Shader.CurrentProgram[MESA_SHADER_VERTEX],
-                                            errMsg)) {
+         if (!_mesa_validate_shader_program(ctx, shProg, errMsg)) {
             _mesa_warning(ctx, "Shader program %u is invalid: %s",
-                          ctx->Shader.CurrentProgram[MESA_SHADER_VERTEX]->Name, errMsg);
+                          shProg->Name, errMsg);
          }
       }
 #endif
    }
 
-   if (ctx->Shader.CurrentProgram[MESA_SHADER_GEOMETRY]) {
-      geom_from_glsl_shader = true;
+   return true;
+}
 
-      if (!ctx->Shader.CurrentProgram[MESA_SHADER_GEOMETRY]->LinkStatus) {
-         _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s(shader not linked)", where);
-         return GL_FALSE;
-      }
-#if 0 /* not normally enabled */
-      {
-         char errMsg[100];
-         if (!_mesa_validate_shader_program(ctx,
-					    ctx->Shader.CurrentProgram[MESA_SHADER_GEOMETRY],
-                                            errMsg)) {
-            _mesa_warning(ctx, "Shader program %u is invalid: %s",
-                          ctx->Shader.CurrentProgram[MESA_SHADER_GEOMETRY]->Name,
-                          errMsg);
-         }
-      }
-#endif
-   }
+/**
+ * Prior to drawing anything with glBegin, glDrawArrays, etc. this function
+ * is called to see if it's valid to render.  This involves checking that
+ * the current shader is valid and the framebuffer is complete.
+ * If an error is detected it'll be recorded here.
+ * \return GL_TRUE if OK to render, GL_FALSE if not
+ */
+GLboolean
+_mesa_valid_to_render(struct gl_context *ctx, const char *where)
+{
+   bool from_glsl_shader[MESA_SHADER_COMPUTE] = { false };
+   unsigned i;
 
-   if (ctx->Shader.CurrentProgram[MESA_SHADER_FRAGMENT]) {
-      frag_from_glsl_shader = true;
+   /* This depends on having up to date derived state (shaders) */
+   if (ctx->NewState)
+      _mesa_update_state(ctx);
 
-      if (!ctx->Shader.CurrentProgram[MESA_SHADER_FRAGMENT]->LinkStatus) {
-         _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "%s(shader not linked)", where);
+   for (i = 0; i < MESA_SHADER_COMPUTE; i++) {
+      if (!shader_linked_or_absent(ctx, ctx->Shader.CurrentProgram[i],
+                                   &from_glsl_shader[i], where))
          return GL_FALSE;
-      }
-#if 0 /* not normally enabled */
-      {
-         char errMsg[100];
-         if (!_mesa_validate_shader_program(ctx,
-					    ctx->Shader.CurrentProgram[MESA_SHADER_FRAGMENT],
-                                            errMsg)) {
-            _mesa_warning(ctx, "Shader program %u is invalid: %s",
-                          ctx->Shader.CurrentProgram[MESA_SHADER_FRAGMENT]->Name,
-                          errMsg);
-         }
-      }
-#endif
    }
 
    /* Any shader stages that are not supplied by the GLSL shader and have
     * assembly shaders enabled must now be validated.
     */
-   if (!vert_from_glsl_shader
+   if (!from_glsl_shader[MESA_SHADER_VERTEX]
        && ctx->VertexProgram.Enabled && !ctx->VertexProgram._Enabled) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
 		  "%s(vertex program not valid)", where);
@@ -1865,9 +1831,9 @@ _mesa_valid_to_render(struct gl_context *ctx, const char *where)
    /* FINISHME: If GL_NV_geometry_program4 is ever supported, the current
     * FINISHME: geometry program should validated here.
     */
-   (void) geom_from_glsl_shader;
+   (void) from_glsl_shader[MESA_SHADER_GEOMETRY];
 
-   if (!frag_from_glsl_shader) {
+   if (!from_glsl_shader[MESA_SHADER_FRAGMENT]) {
       if (ctx->FragmentProgram.Enabled && !ctx->FragmentProgram._Enabled) {
 	 _mesa_error(ctx, GL_INVALID_OPERATION,
 		     "%s(fragment program not valid)", where);
diff --git a/mesalib/src/mesa/main/dd.h b/mesalib/src/mesa/main/dd.h
index ac317e399..971524135 100644
--- a/mesalib/src/mesa/main/dd.h
+++ b/mesalib/src/mesa/main/dd.h
@@ -399,6 +399,13 @@ struct dd_function_table {
    void (*UnmapRenderbuffer)(struct gl_context *ctx,
 			     struct gl_renderbuffer *rb);
 
+   /**
+    * Optional driver entrypoint that binds a non-texture renderbuffer's
+    * contents to a texture image.
+    */
+   GLboolean (*BindRenderbufferTexImage)(struct gl_context *ctx,
+                                         struct gl_renderbuffer *rb,
+                                         struct gl_texture_image *texImage);
    /*@}*/
 
 
@@ -564,9 +571,6 @@ struct dd_function_table {
     * \name Vertex/pixel buffer object functions
     */
    /*@{*/
-   void (*BindBuffer)( struct gl_context *ctx, GLenum target,
-		       struct gl_buffer_object *obj );
-
    struct gl_buffer_object * (*NewBufferObject)(struct gl_context *ctx,
                                                 GLuint buffer, GLenum target);
    
@@ -574,7 +578,7 @@ struct dd_function_table {
 
    GLboolean (*BufferData)(struct gl_context *ctx, GLenum target,
                            GLsizeiptrARB size, const GLvoid *data, GLenum usage,
-                           struct gl_buffer_object *obj);
+                           GLenum storageFlags, struct gl_buffer_object *obj);
 
    void (*BufferSubData)( struct gl_context *ctx, GLintptrARB offset,
 			  GLsizeiptrARB size, const GLvoid *data,
@@ -601,14 +605,17 @@ struct dd_function_table {
     */
    void * (*MapBufferRange)( struct gl_context *ctx, GLintptr offset,
                              GLsizeiptr length, GLbitfield access,
-                             struct gl_buffer_object *obj);
+                             struct gl_buffer_object *obj,
+                             gl_map_buffer_index index);
 
    void (*FlushMappedBufferRange)(struct gl_context *ctx,
                                   GLintptr offset, GLsizeiptr length,
-                                  struct gl_buffer_object *obj);
+                                  struct gl_buffer_object *obj,
+                                  gl_map_buffer_index index);
 
    GLboolean (*UnmapBuffer)( struct gl_context *ctx,
-			     struct gl_buffer_object *obj );
+			     struct gl_buffer_object *obj,
+                             gl_map_buffer_index index);
    /*@}*/
 
    /**
diff --git a/mesalib/src/mesa/main/dlist.c b/mesalib/src/mesa/main/dlist.c
index 08943c9f9..d431fd221 100644
--- a/mesalib/src/mesa/main/dlist.c
+++ b/mesalib/src/mesa/main/dlist.c
@@ -914,7 +914,8 @@ unpack_image(struct gl_context *ctx, GLuint dimensions,
 
       map = (GLubyte *)
          ctx->Driver.MapBufferRange(ctx, 0, unpack->BufferObj->Size,
-				    GL_MAP_READ_BIT, unpack->BufferObj);
+				    GL_MAP_READ_BIT, unpack->BufferObj,
+                                    MAP_INTERNAL);
       if (!map) {
          /* unable to map src buffer! */
          _mesa_error(ctx, GL_INVALID_OPERATION, "unable to map PBO");
@@ -928,7 +929,7 @@ unpack_image(struct gl_context *ctx, GLuint dimensions,
          image = _mesa_unpack_image(dimensions, width, height, depth,
                                     format, type, src, unpack);
 
-      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj, MAP_INTERNAL);
 
       if (!image) {
          _mesa_error(ctx, GL_OUT_OF_MEMORY, "display list construction");
@@ -8176,7 +8177,7 @@ _mesa_GenLists(GLsizei range)
    /*
     * Make this an atomic operation
     */
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
 
    base = _mesa_HashFindFreeKeyBlock(ctx->Shared->DisplayList, range);
    if (base) {
@@ -8188,7 +8189,7 @@ _mesa_GenLists(GLsizei range)
       }
    }
 
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 
    return base;
 }
diff --git a/mesalib/src/mesa/main/drawpix.c b/mesalib/src/mesa/main/drawpix.c
index 096615c05..63e5870e6 100644
--- a/mesalib/src/mesa/main/drawpix.c
+++ b/mesalib/src/mesa/main/drawpix.c
@@ -151,7 +151,7 @@ _mesa_DrawPixels( GLsizei width, GLsizei height,
                            "glDrawPixels(invalid PBO access)");
                goto end;
             }
-            if (_mesa_bufferobj_mapped(ctx->Unpack.BufferObj)) {
+            if (_mesa_check_disallowed_mapping(ctx->Unpack.BufferObj)) {
                /* buffer is mapped - that's an error */
                _mesa_error(ctx, GL_INVALID_OPERATION,
                            "glDrawPixels(PBO is mapped)");
@@ -335,7 +335,7 @@ _mesa_Bitmap( GLsizei width, GLsizei height,
                            "glBitmap(invalid PBO access)");
                return;
             }
-            if (_mesa_bufferobj_mapped(ctx->Unpack.BufferObj)) {
+            if (_mesa_check_disallowed_mapping(ctx->Unpack.BufferObj)) {
                /* buffer is mapped - that's an error */
                _mesa_error(ctx, GL_INVALID_OPERATION,
                            "glBitmap(PBO is mapped)");
diff --git a/mesalib/src/mesa/main/enable.c b/mesalib/src/mesa/main/enable.c
index 40508a456..edd4751e1 100644
--- a/mesalib/src/mesa/main/enable.c
+++ b/mesalib/src/mesa/main/enable.c
@@ -32,6 +32,7 @@
 #include "clip.h"
 #include "context.h"
 #include "enable.h"
+#include "errors.h"
 #include "light.h"
 #include "simple_list.h"
 #include "mtypes.h"
@@ -367,14 +368,26 @@ _mesa_set_enable(struct gl_context *ctx, GLenum cap, GLboolean state)
          ctx->Depth.Test = state;
          break;
       case GL_DEBUG_OUTPUT:
-         if (!_mesa_is_desktop_gl(ctx))
+         if (!_mesa_is_desktop_gl(ctx)) {
             goto invalid_enum_error;
-         ctx->Debug.DebugOutput = state;
+         }
+         else {
+            struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+            if (debug) {
+               debug->DebugOutput = state;
+            }
+         }
          break;
       case GL_DEBUG_OUTPUT_SYNCHRONOUS_ARB:
-         if (!_mesa_is_desktop_gl(ctx))
+         if (!_mesa_is_desktop_gl(ctx)) {
             goto invalid_enum_error;
-         ctx->Debug.SyncOutput = state;
+         }
+         else {
+            struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+            if (debug) {
+               debug->SyncOutput = state;
+            }
+         }
          break;
       case GL_DITHER:
          if (ctx->Color.DitherFlag == state)
@@ -1228,11 +1241,19 @@ _mesa_IsEnabled( GLenum cap )
       case GL_DEBUG_OUTPUT:
          if (!_mesa_is_desktop_gl(ctx))
             goto invalid_enum_error;
-         return ctx->Debug.DebugOutput;
+         if (ctx->Debug) {
+            return ctx->Debug->DebugOutput;
+         } else {
+            return GL_FALSE;
+         }
       case GL_DEBUG_OUTPUT_SYNCHRONOUS_ARB:
          if (!_mesa_is_desktop_gl(ctx))
             goto invalid_enum_error;
-         return ctx->Debug.SyncOutput;
+         if (ctx->Debug) {
+            return ctx->Debug->SyncOutput;
+         } else {
+            return GL_FALSE;
+         }
       case GL_DEPTH_TEST:
          return ctx->Depth.Test;
       case GL_DITHER:
diff --git a/mesalib/src/mesa/main/errors.c b/mesalib/src/mesa/main/errors.c
index 28357e0e8..ca73a6686 100644
--- a/mesalib/src/mesa/main/errors.c
+++ b/mesalib/src/mesa/main/errors.c
@@ -37,12 +37,11 @@
 #include "mtypes.h"
 #include "version.h"
 #include "hash_table.h"
-#include "glapi/glthread.h"
 
 #define MESSAGE_LOG 1
 #define MESSAGE_LOG_ARB 2
 
-_glthread_DECLARE_STATIC_MUTEX(DynamicIDMutex);
+static mtx_t DynamicIDMutex = _MTX_INITIALIZER_NP;
 static GLuint NextDynamicID = 1;
 
 struct gl_debug_severity
@@ -81,6 +80,7 @@ static const GLenum debug_severity_enums[] = {
    GL_DEBUG_SEVERITY_NOTIFICATION,
 };
 
+
 static enum mesa_debug_source
 gl_enum_to_debug_source(GLenum e)
 {
@@ -117,6 +117,7 @@ gl_enum_to_debug_severity(GLenum e)
    return i;
 }
 
+
 /**
  * Handles generating a GL_ARB_debug_output message ID generated by the GL or
  * GLSL compiler.
@@ -134,13 +135,14 @@ static void
 debug_get_id(GLuint *id)
 {
    if (!(*id)) {
-      _glthread_LOCK_MUTEX(DynamicIDMutex);
+      mtx_lock(&DynamicIDMutex);
       if (!(*id))
          *id = NextDynamicID++;
-      _glthread_UNLOCK_MUTEX(DynamicIDMutex);
+      mtx_unlock(&DynamicIDMutex);
    }
 }
 
+
 /*
  * We store a bitfield in the hash table, with five possible values total.
  *
@@ -182,6 +184,50 @@ enum {
    ENABLED = ENABLED_BIT | FOUND_BIT
 };
 
+
+/**
+ * Return debug state for the context.  The debug state will be allocated
+ * and initialized upon the first call.
+ */
+struct gl_debug_state *
+_mesa_get_debug_state(struct gl_context *ctx)
+{
+   if (!ctx->Debug) {
+      ctx->Debug = CALLOC_STRUCT(gl_debug_state);
+      if (!ctx->Debug) {
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "allocating debug state");
+      }
+      else {
+         struct gl_debug_state *debug = ctx->Debug;
+         int s, t, sev;
+
+         /* Enable all the messages with severity HIGH or MEDIUM by default. */
+         memset(debug->Defaults[0][MESA_DEBUG_SEVERITY_HIGH], GL_TRUE,
+                sizeof debug->Defaults[0][MESA_DEBUG_SEVERITY_HIGH]);
+         memset(debug->Defaults[0][MESA_DEBUG_SEVERITY_MEDIUM], GL_TRUE,
+                sizeof debug->Defaults[0][MESA_DEBUG_SEVERITY_MEDIUM]);
+         memset(debug->Defaults[0][MESA_DEBUG_SEVERITY_LOW], GL_FALSE,
+                sizeof debug->Defaults[0][MESA_DEBUG_SEVERITY_LOW]);
+
+         /* Initialize state for filtering known debug messages. */
+         for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++) {
+            for (t = 0; t < MESA_DEBUG_TYPE_COUNT; t++) {
+               debug->Namespaces[0][s][t].IDs = _mesa_NewHashTable();
+               assert(debug->Namespaces[0][s][t].IDs);
+
+               for (sev = 0; sev < MESA_DEBUG_SEVERITY_COUNT; sev++) {
+                  make_empty_list(&debug->Namespaces[0][s][t].Severity[sev]);
+               }
+            }
+         }
+      }
+   }
+
+   return ctx->Debug;
+}
+
+
+
 /**
  * Returns the state of the given message source/type/ID tuple.
  */
@@ -192,52 +238,64 @@ should_log(struct gl_context *ctx,
            GLuint id,
            enum mesa_debug_severity severity)
 {
-   GLint gstack = ctx->Debug.GroupStackDepth;
-   struct gl_debug_namespace *nspace =
-         &ctx->Debug.Namespaces[gstack][source][type];
-   uintptr_t state;
+   struct gl_debug_state *debug;
+   uintptr_t state = 0;
 
-   if (!ctx->Debug.DebugOutput)
+   if (!ctx->Debug) {
+      /* no debug state set so far */
       return GL_FALSE;
+   }
 
-   /* In addition to not being able to store zero as a value, HashTable also
-      can't use zero as a key. */
-   if (id)
-      state = (uintptr_t)_mesa_HashLookup(nspace->IDs, id);
-   else
-      state = nspace->ZeroID;
-
-   /* Only do this once for each ID. This makes sure the ID exists in,
-      at most, one list, and does not pointlessly appear multiple times. */
-   if (!(state & KNOWN_SEVERITY)) {
-      struct gl_debug_severity *entry;
-
-      if (state == NOT_FOUND) {
-         if (ctx->Debug.Defaults[gstack][severity][source][type])
-            state = ENABLED;
-         else
-            state = DISABLED;
-      }
-
-      entry = malloc(sizeof *entry);
-      if (!entry)
-         goto out;
+   debug = _mesa_get_debug_state(ctx);
+   if (debug) {
+      const GLint gstack = debug->GroupStackDepth;
+      struct gl_debug_namespace *nspace =
+         &debug->Namespaces[gstack][source][type];
 
-      state |= KNOWN_SEVERITY;
+      if (!debug->DebugOutput)
+         return GL_FALSE;
 
+      /* In addition to not being able to store zero as a value, HashTable also
+       * can't use zero as a key.
+       */
       if (id)
-         _mesa_HashInsert(nspace->IDs, id, (void*)state);
+         state = (uintptr_t)_mesa_HashLookup(nspace->IDs, id);
       else
-         nspace->ZeroID = state;
+         state = nspace->ZeroID;
 
-      entry->ID = id;
-      insert_at_tail(&nspace->Severity[severity], &entry->link);
-   }
+      /* Only do this once for each ID. This makes sure the ID exists in,
+       * at most, one list, and does not pointlessly appear multiple times.
+       */
+      if (!(state & KNOWN_SEVERITY)) {
+         struct gl_debug_severity *entry;
+
+         if (state == NOT_FOUND) {
+            if (debug->Defaults[gstack][severity][source][type])
+               state = ENABLED;
+            else
+               state = DISABLED;
+         }
+
+         entry = malloc(sizeof *entry);
+         if (!entry)
+            goto out;
+
+         state |= KNOWN_SEVERITY;
 
+         if (id)
+            _mesa_HashInsert(nspace->IDs, id, (void*)state);
+         else
+            nspace->ZeroID = state;
+
+         entry->ID = id;
+         insert_at_tail(&nspace->Severity[severity], &entry->link);
+      }
+   }
 out:
    return !!(state & ENABLED_BIT);
 }
 
+
 /**
  * Sets the state of the given message source/type/ID tuple.
  */
@@ -247,33 +305,39 @@ set_message_state(struct gl_context *ctx,
                   enum mesa_debug_type type,
                   GLuint id, GLboolean enabled)
 {
-   GLint gstack = ctx->Debug.GroupStackDepth;
-   struct gl_debug_namespace *nspace =
-         &ctx->Debug.Namespaces[gstack][source][type];
-   uintptr_t state;
-
-   /* In addition to not being able to store zero as a value, HashTable also
-      can't use zero as a key. */
-   if (id)
-      state = (uintptr_t)_mesa_HashLookup(nspace->IDs, id);
-   else
-      state = nspace->ZeroID;
-
-   if (state == NOT_FOUND)
-      state = enabled ? ENABLED : DISABLED;
-   else {
-      if (enabled)
-         state |= ENABLED_BIT;
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+
+   if (debug) {
+      GLint gstack = debug->GroupStackDepth;
+      struct gl_debug_namespace *nspace =
+         &debug->Namespaces[gstack][source][type];
+      uintptr_t state;
+
+      /* In addition to not being able to store zero as a value, HashTable also
+       * can't use zero as a key.
+       */
+      if (id)
+         state = (uintptr_t)_mesa_HashLookup(nspace->IDs, id);
       else
-         state &= ~ENABLED_BIT;
-   }
+         state = nspace->ZeroID;
 
-   if (id)
-      _mesa_HashInsert(nspace->IDs, id, (void*)state);
-   else
-      nspace->ZeroID = state;
+      if (state == NOT_FOUND)
+         state = enabled ? ENABLED : DISABLED;
+      else {
+         if (enabled)
+            state |= ENABLED_BIT;
+         else
+            state &= ~ENABLED_BIT;
+      }
+
+      if (id)
+         _mesa_HashInsert(nspace->IDs, id, (void*)state);
+      else
+         nspace->ZeroID = state;
+   }
 }
 
+
 static void
 store_message_details(struct gl_debug_msg *emptySlot,
                       enum mesa_debug_source source,
@@ -307,7 +371,8 @@ store_message_details(struct gl_debug_msg *emptySlot,
    }
 }
 
- /**
+
+/**
  * Remap any type exclusive to KHR_debug to something suitable
  * for ARB_debug_output
  */
@@ -326,6 +391,7 @@ remap_type(GLenum type) {
   return type;
 }
 
+
 /**
  * Remap severity exclusive to KHR_debug to something suitable
  * for ARB_debug_output
@@ -339,6 +405,7 @@ remap_severity(GLenum severity) {
    return severity;
 }
 
+
 /**
  * 'buf' is not necessarily a null-terminated string. When logging, copy
  * 'len' characters from it, store them in a new, null-terminated string,
@@ -346,49 +413,51 @@ remap_severity(GLenum severity) {
  * the null terminator this time.
  */
 static void
-_mesa_log_msg(struct gl_context *ctx, enum mesa_debug_source source,
-              enum mesa_debug_type type, GLuint id,
-              enum mesa_debug_severity severity, GLint len, const char *buf)
+log_msg(struct gl_context *ctx, enum mesa_debug_source source,
+        enum mesa_debug_type type, GLuint id,
+        enum mesa_debug_severity severity, GLint len, const char *buf)
 {
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
    GLint nextEmpty;
    struct gl_debug_msg *emptySlot;
 
+   if (!debug)
+      return;
+
    assert(len >= 0 && len < MAX_DEBUG_MESSAGE_LENGTH);
 
    if (!should_log(ctx, source, type, id, severity))
       return;
 
-   if (ctx->Debug.Callback) {
+   if (debug->Callback) {
        GLenum gl_type = debug_type_enums[type];
        GLenum gl_severity = debug_severity_enums[severity];
 
-       if (ctx->Debug.ARBCallback) {
+       if (debug->ARBCallback) {
           gl_severity = remap_severity(gl_severity);
           gl_type = remap_type(gl_type);
       }
-      ctx->Debug.Callback(debug_source_enums[source],
-                          gl_type,
-                          id,
-                          gl_severity,
-                          len, buf, ctx->Debug.CallbackData);
+      debug->Callback(debug_source_enums[source], gl_type, id, gl_severity,
+                      len, buf, debug->CallbackData);
       return;
    }
 
-   if (ctx->Debug.NumMessages == MAX_DEBUG_LOGGED_MESSAGES)
+   if (debug->NumMessages == MAX_DEBUG_LOGGED_MESSAGES)
       return;
 
-   nextEmpty = (ctx->Debug.NextMsg + ctx->Debug.NumMessages)
+   nextEmpty = (debug->NextMsg + debug->NumMessages)
                           % MAX_DEBUG_LOGGED_MESSAGES;
-   emptySlot = &ctx->Debug.Log[nextEmpty];
+   emptySlot = &debug->Log[nextEmpty];
 
    store_message_details(emptySlot, source, type, id, severity, len, buf);
 
-   if (ctx->Debug.NumMessages == 0)
-      ctx->Debug.NextMsgLength = ctx->Debug.Log[ctx->Debug.NextMsg].length;
+   if (debug->NumMessages == 0)
+      debug->NextMsgLength = debug->Log[debug->NextMsg].length;
 
-   ctx->Debug.NumMessages++;
+   debug->NumMessages++;
 }
 
+
 /**
  * Pop the oldest debug message out of the log.
  * Writes the message string, including the null terminator, into 'buf',
@@ -400,20 +469,21 @@ _mesa_log_msg(struct gl_context *ctx, enum mesa_debug_source source,
  * indicates failure.
  */
 static GLsizei
-_mesa_get_msg(struct gl_context *ctx, GLenum *source, GLenum *type,
-              GLuint *id, GLenum *severity, GLsizei bufSize, char *buf,
-              unsigned caller)
+get_msg(struct gl_context *ctx, GLenum *source, GLenum *type,
+        GLuint *id, GLenum *severity, GLsizei bufSize, char *buf,
+        unsigned caller)
 {
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
    struct gl_debug_msg *msg;
    GLsizei length;
 
-   if (ctx->Debug.NumMessages == 0)
+   if (!debug || debug->NumMessages == 0)
       return 0;
 
-   msg = &ctx->Debug.Log[ctx->Debug.NextMsg];
+   msg = &debug->Log[debug->NextMsg];
    length = msg->length;
 
-   assert(length > 0 && length == ctx->Debug.NextMsgLength);
+   assert(length > 0 && length == debug->NextMsgLength);
 
    if (bufSize < length && buf != NULL)
       return 0;
@@ -423,15 +493,20 @@ _mesa_get_msg(struct gl_context *ctx, GLenum *source, GLenum *type,
       if (caller == MESSAGE_LOG_ARB)
          *severity = remap_severity(*severity);
    }
-   if (source)
+
+   if (source) {
       *source = debug_source_enums[msg->source];
+   }
+
    if (type) {
       *type = debug_type_enums[msg->type];
       if (caller == MESSAGE_LOG_ARB)
          *type = remap_type(*type);
    }
-   if (id)
+
+   if (id) {
       *id = msg->id;
+   }
 
    if (buf) {
       assert(msg->message[length-1] == '\0');
@@ -443,14 +518,15 @@ _mesa_get_msg(struct gl_context *ctx, GLenum *source, GLenum *type,
    msg->message = NULL;
    msg->length = 0;
 
-   ctx->Debug.NumMessages--;
-   ctx->Debug.NextMsg++;
-   ctx->Debug.NextMsg %= MAX_DEBUG_LOGGED_MESSAGES;
-   ctx->Debug.NextMsgLength = ctx->Debug.Log[ctx->Debug.NextMsg].length;
+   debug->NumMessages--;
+   debug->NextMsg++;
+   debug->NextMsg %= MAX_DEBUG_LOGGED_MESSAGES;
+   debug->NextMsgLength = debug->Log[debug->NextMsg].length;
 
    return length;
 }
 
+
 /**
  * Verify that source, type, and severity are valid enums.
  * glDebugMessageInsertARB only accepts two values for 'source',
@@ -523,14 +599,14 @@ validate_params(struct gl_context *ctx, unsigned caller,
    return GL_TRUE;
 
 error:
-   {
-      _mesa_error(ctx, GL_INVALID_ENUM, "bad values passed to %s"
-                  "(source=0x%x, type=0x%x, severity=0x%x)", callerstr,
-                  source, type, severity);
-   }
+   _mesa_error(ctx, GL_INVALID_ENUM, "bad values passed to %s"
+               "(source=0x%x, type=0x%x, severity=0x%x)", callerstr,
+               source, type, severity);
+
    return GL_FALSE;
 }
 
+
 /**
  * Set the state of all message IDs found in the given intersection of
  * 'source', 'type', and 'severity'.  The _COUNT enum can be used for
@@ -549,8 +625,12 @@ control_messages(struct gl_context *ctx,
                  enum mesa_debug_severity severity,
                  GLboolean enabled)
 {
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
    int s, t, sev, smax, tmax, sevmax;
-   GLint gstack = ctx->Debug.GroupStackDepth;
+   const GLint gstack = debug ? debug->GroupStackDepth : 0;
+
+   if (!debug)
+      return;
 
    if (source == MESA_DEBUG_SOURCE_COUNT) {
       source = 0;
@@ -573,23 +653,26 @@ control_messages(struct gl_context *ctx,
       sevmax = severity+1;
    }
 
-   for (sev = severity; sev < sevmax; sev++)
-      for (s = source; s < smax; s++)
+   for (sev = severity; sev < sevmax; sev++) {
+      for (s = source; s < smax; s++) {
          for (t = type; t < tmax; t++) {
             struct simple_node *node;
             struct gl_debug_severity *entry;
 
             /* change the default for IDs we've never seen before. */
-            ctx->Debug.Defaults[gstack][sev][s][t] = enabled;
+            debug->Defaults[gstack][sev][s][t] = enabled;
 
             /* Now change the state of IDs we *have* seen... */
-            foreach(node, &ctx->Debug.Namespaces[gstack][s][t].Severity[sev]) {
+            foreach(node, &debug->Namespaces[gstack][s][t].Severity[sev]) {
                entry = (struct gl_debug_severity *)node;
                set_message_state(ctx, s, t, entry->ID, enabled);
             }
          }
+      }
+   }
 }
 
+
 /**
  * Debugging-message namespaces with the source APPLICATION or THIRD_PARTY
  * require special handling, since the IDs in them are controlled by clients,
@@ -621,6 +704,7 @@ control_app_messages(struct gl_context *ctx, GLenum esource, GLenum etype,
    control_messages(ctx, source, type, severity, enabled);
 }
 
+
 /**
  * This is a generic message control function for use by both
  * glDebugMessageControlARB and glDebugMessageControl.
@@ -658,6 +742,7 @@ message_control(GLenum gl_source, GLenum gl_type,
                         count, ids, enabled);
 }
 
+
 /**
  * This is a generic message insert function.
  * Validation of source, type and severity parameters should be done
@@ -665,7 +750,7 @@ message_control(GLenum gl_source, GLenum gl_type,
  */
 static void
 message_insert(GLenum source, GLenum type, GLuint id,
-               GLenum severity, GLint length, const GLchar* buf,
+               GLenum severity, GLint length, const GLchar *buf,
                const char *callerstr)
 {
    GET_CURRENT_CONTEXT(ctx);
@@ -681,20 +766,21 @@ message_insert(GLenum source, GLenum type, GLuint id,
       return;
    }
 
-   _mesa_log_msg(ctx,
-                 gl_enum_to_debug_source(source),
-                 gl_enum_to_debug_type(type), id,
-                 gl_enum_to_debug_severity(severity), length, buf);
+   log_msg(ctx,
+           gl_enum_to_debug_source(source),
+           gl_enum_to_debug_type(type), id,
+           gl_enum_to_debug_severity(severity), length, buf);
 }
 
+
 /**
  * This is a generic message insert function for use by both
  * glGetDebugMessageLogARB and glGetDebugMessageLog.
  */
 static GLuint
-get_message_log(GLuint count, GLsizei logSize, GLenum* sources,
-                GLenum* types, GLenum* ids, GLenum* severities,
-                GLsizei* lengths, GLchar* messageLog,
+get_message_log(GLuint count, GLsizei logSize, GLenum *sources,
+                GLenum *types, GLenum *ids, GLenum *severities,
+                GLsizei *lengths, GLchar *messageLog,
                 unsigned caller, const char *callerstr)
 {
    GET_CURRENT_CONTEXT(ctx);
@@ -711,8 +797,8 @@ get_message_log(GLuint count, GLsizei logSize, GLenum* sources,
    }
 
    for (ret = 0; ret < count; ret++) {
-      GLsizei written = _mesa_get_msg(ctx, sources, types, ids, severities,
-                                      logSize, messageLog, caller);
+      GLsizei written = get_msg(ctx, sources, types, ids, severities,
+                                logSize, messageLog, caller);
       if (!written)
          break;
 
@@ -738,41 +824,52 @@ get_message_log(GLuint count, GLsizei logSize, GLenum* sources,
    return ret;
 }
 
+
 static void
 do_nothing(GLuint key, void *data, void *userData)
 {
 }
 
+
+/**
+ * Free context state pertaining to error/debug state for the given stack
+ * depth.
+ */
 static void
 free_errors_data(struct gl_context *ctx, GLint gstack)
 {
+   struct gl_debug_state *debug = ctx->Debug;
    enum mesa_debug_type t;
    enum mesa_debug_source s;
    enum mesa_debug_severity sev;
 
+   assert(debug);
+
    /* Tear down state for filtering debug messages. */
-   for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++)
+   for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++) {
       for (t = 0; t < MESA_DEBUG_TYPE_COUNT; t++) {
-         _mesa_HashDeleteAll(ctx->Debug.Namespaces[gstack][s][t].IDs,
+         _mesa_HashDeleteAll(debug->Namespaces[gstack][s][t].IDs,
                              do_nothing, NULL);
-         _mesa_DeleteHashTable(ctx->Debug.Namespaces[gstack][s][t].IDs);
+         _mesa_DeleteHashTable(debug->Namespaces[gstack][s][t].IDs);
          for (sev = 0; sev < MESA_DEBUG_SEVERITY_COUNT; sev++) {
             struct simple_node *node, *tmp;
             struct gl_debug_severity *entry;
 
             foreach_s(node, tmp,
-                      &ctx->Debug.Namespaces[gstack][s][t].Severity[sev]) {
+                      &debug->Namespaces[gstack][s][t].Severity[sev]) {
                entry = (struct gl_debug_severity *)node;
                free(entry);
             }
          }
       }
+   }
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageInsert(GLenum source, GLenum type, GLuint id,
                          GLenum severity, GLint length,
-                         const GLchar* buf)
+                         const GLchar *buf)
 {
    const char *callerstr = "glDebugMessageInsert";
 
@@ -781,14 +878,14 @@ _mesa_DebugMessageInsert(GLenum source, GLenum type, GLuint id,
    if (!validate_params(ctx, INSERT, callerstr, source, type, severity))
       return; /* GL_INVALID_ENUM */
 
-   message_insert(source, type, id, severity, length, buf,
-                  callerstr);
+   message_insert(source, type, id, severity, length, buf, callerstr);
 }
 
+
 GLuint GLAPIENTRY
-_mesa_GetDebugMessageLog(GLuint count, GLsizei logSize, GLenum* sources,
-                         GLenum* types, GLenum* ids, GLenum* severities,
-                         GLsizei* lengths, GLchar* messageLog)
+_mesa_GetDebugMessageLog(GLuint count, GLsizei logSize, GLenum *sources,
+                         GLenum *types, GLenum *ids, GLenum *severities,
+                         GLsizei *lengths, GLchar *messageLog)
 {
    const char *callerstr = "glGetDebugMessageLog";
 
@@ -796,6 +893,7 @@ _mesa_GetDebugMessageLog(GLuint count, GLsizei logSize, GLenum* sources,
                           lengths, messageLog, MESSAGE_LOG, callerstr);
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageControl(GLenum source, GLenum type, GLenum severity,
                           GLsizei count, const GLuint *ids,
@@ -807,28 +905,36 @@ _mesa_DebugMessageControl(GLenum source, GLenum type, GLenum severity,
                    enabled, CONTROL, callerstr);
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageCallback(GLDEBUGPROC callback, const void *userParam)
 {
    GET_CURRENT_CONTEXT(ctx);
-   ctx->Debug.Callback = callback;
-   ctx->Debug.CallbackData = userParam;
-   ctx->Debug.ARBCallback = GL_FALSE;
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+   if (debug) {
+      debug->Callback = callback;
+      debug->CallbackData = userParam;
+      debug->ARBCallback = GL_FALSE;
+   }
 }
 
+
 void GLAPIENTRY
 _mesa_PushDebugGroup(GLenum source, GLuint id, GLsizei length,
                      const GLchar *message)
 {
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
    const char *callerstr = "glPushDebugGroup";
    int s, t, sev;
    GLint prevStackDepth;
    GLint currStackDepth;
    struct gl_debug_msg *emptySlot;
 
-   GET_CURRENT_CONTEXT(ctx);
+   if (!debug)
+      return;
 
-   if (ctx->Debug.GroupStackDepth >= MAX_DEBUG_GROUP_STACK_DEPTH-1) {
+   if (debug->GroupStackDepth >= MAX_DEBUG_GROUP_STACK_DEPTH-1) {
       _mesa_error(ctx, GL_STACK_OVERFLOW, "%s", callerstr);
       return;
    }
@@ -847,14 +953,14 @@ _mesa_PushDebugGroup(GLenum source, GLuint id, GLsizei length,
                   GL_DEBUG_SEVERITY_NOTIFICATION, length,
                   message, callerstr);
 
-   prevStackDepth = ctx->Debug.GroupStackDepth;
-   ctx->Debug.GroupStackDepth++;
-   currStackDepth = ctx->Debug.GroupStackDepth;
+   prevStackDepth = debug->GroupStackDepth;
+   debug->GroupStackDepth++;
+   currStackDepth = debug->GroupStackDepth;
 
    /* pop reuses the message details from push so we store this */
    if (length < 0)
       length = strlen(message);
-   emptySlot = &ctx->Debug.DebugGroupMsgs[ctx->Debug.GroupStackDepth];
+   emptySlot = &debug->DebugGroupMsgs[debug->GroupStackDepth];
    store_message_details(emptySlot, gl_enum_to_debug_source(source),
                          gl_enum_to_debug_type(GL_DEBUG_TYPE_PUSH_GROUP),
                          id,
@@ -864,60 +970,66 @@ _mesa_PushDebugGroup(GLenum source, GLuint id, GLsizei length,
    /* inherit the control volume of the debug group previously residing on
     * the top of the debug group stack
     */
-   for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++)
+   for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++) {
       for (t = 0; t < MESA_DEBUG_TYPE_COUNT; t++) {
          /* copy id settings */
-         ctx->Debug.Namespaces[currStackDepth][s][t].IDs =
-            _mesa_HashClone(ctx->Debug.Namespaces[prevStackDepth][s][t].IDs);
+         debug->Namespaces[currStackDepth][s][t].IDs =
+            _mesa_HashClone(debug->Namespaces[prevStackDepth][s][t].IDs);
 
          for (sev = 0; sev < MESA_DEBUG_SEVERITY_COUNT; sev++) {
             struct gl_debug_severity *entry, *prevEntry;
             struct simple_node *node;
 
             /* copy default settings for unknown ids */
-            ctx->Debug.Defaults[currStackDepth][sev][s][t] = ctx->Debug.Defaults[prevStackDepth][sev][s][t];
+            debug->Defaults[currStackDepth][sev][s][t] =
+               debug->Defaults[prevStackDepth][sev][s][t];
 
             /* copy known id severity settings */
-            make_empty_list(&ctx->Debug.Namespaces[currStackDepth][s][t].Severity[sev]);
-            foreach(node, &ctx->Debug.Namespaces[prevStackDepth][s][t].Severity[sev]) {
+            make_empty_list(&debug->Namespaces[currStackDepth][s][t].Severity[sev]);
+            foreach(node, &debug->Namespaces[prevStackDepth][s][t].Severity[sev]) {
                prevEntry = (struct gl_debug_severity *)node;
                entry = malloc(sizeof *entry);
                if (!entry)
                   return;
 
                entry->ID = prevEntry->ID;
-               insert_at_tail(&ctx->Debug.Namespaces[currStackDepth][s][t].Severity[sev], &entry->link);
+               insert_at_tail(&debug->Namespaces[currStackDepth][s][t].Severity[sev], &entry->link);
             }
          }
       }
+   }
 }
 
+
 void GLAPIENTRY
-_mesa_PopDebugGroup()
+_mesa_PopDebugGroup(void)
 {
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
    const char *callerstr = "glPopDebugGroup";
    struct gl_debug_msg *gdmessage;
    GLint prevStackDepth;
 
-   GET_CURRENT_CONTEXT(ctx);
+   if (!debug)
+      return;
 
-   if (ctx->Debug.GroupStackDepth <= 0) {
+   if (debug->GroupStackDepth <= 0) {
       _mesa_error(ctx, GL_STACK_UNDERFLOW, "%s", callerstr);
       return;
    }
 
-   prevStackDepth = ctx->Debug.GroupStackDepth;
-   ctx->Debug.GroupStackDepth--;
+   prevStackDepth = debug->GroupStackDepth;
+   debug->GroupStackDepth--;
 
-   gdmessage = &ctx->Debug.DebugGroupMsgs[prevStackDepth];
-   /* using _mesa_log_msg() directly here as verification of parameters
+   gdmessage = &debug->DebugGroupMsgs[prevStackDepth];
+   /* using log_msg() directly here as verification of parameters
     * already done in push
     */
-   _mesa_log_msg(ctx, gdmessage->source,
-                 gl_enum_to_debug_type(GL_DEBUG_TYPE_POP_GROUP),
-                 gdmessage->id,
-                 gl_enum_to_debug_severity(GL_DEBUG_SEVERITY_NOTIFICATION),
-                 gdmessage->length, gdmessage->message);
+   log_msg(ctx, gdmessage->source,
+           gl_enum_to_debug_type(GL_DEBUG_TYPE_POP_GROUP),
+           gdmessage->id,
+           gl_enum_to_debug_severity(GL_DEBUG_SEVERITY_NOTIFICATION),
+           gdmessage->length, gdmessage->message);
 
    if (gdmessage->message != (char*)out_of_memory)
       free(gdmessage->message);
@@ -928,10 +1040,11 @@ _mesa_PopDebugGroup()
    free_errors_data(ctx, prevStackDepth);
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageInsertARB(GLenum source, GLenum type, GLuint id,
                             GLenum severity, GLint length,
-                            const GLcharARB* buf)
+                            const GLcharARB *buf)
 {
    const char *callerstr = "glDebugMessageInsertARB";
 
@@ -940,14 +1053,14 @@ _mesa_DebugMessageInsertARB(GLenum source, GLenum type, GLuint id,
    if (!validate_params(ctx, INSERT_ARB, callerstr, source, type, severity))
       return; /* GL_INVALID_ENUM */
 
-   message_insert(source, type, id, severity, length, buf,
-                  callerstr);
+   message_insert(source, type, id, severity, length, buf, callerstr);
 }
 
+
 GLuint GLAPIENTRY
-_mesa_GetDebugMessageLogARB(GLuint count, GLsizei logSize, GLenum* sources,
-                            GLenum* types, GLenum* ids, GLenum* severities,
-                            GLsizei* lengths, GLcharARB* messageLog)
+_mesa_GetDebugMessageLogARB(GLuint count, GLsizei logSize, GLenum *sources,
+                            GLenum *types, GLenum *ids, GLenum *severities,
+                            GLsizei *lengths, GLcharARB *messageLog)
 {
    const char *callerstr = "glGetDebugMessageLogARB";
 
@@ -955,6 +1068,7 @@ _mesa_GetDebugMessageLogARB(GLuint count, GLsizei logSize, GLenum* sources,
                           lengths, messageLog, MESSAGE_LOG_ARB, callerstr);
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageControlARB(GLenum gl_source, GLenum gl_type,
                              GLenum gl_severity,
@@ -967,47 +1081,27 @@ _mesa_DebugMessageControlARB(GLenum gl_source, GLenum gl_type,
                    enabled, CONTROL_ARB, callerstr);
 }
 
+
 void GLAPIENTRY
 _mesa_DebugMessageCallbackARB(GLDEBUGPROCARB callback, const void *userParam)
 {
    GET_CURRENT_CONTEXT(ctx);
-   ctx->Debug.Callback = callback;
-   ctx->Debug.CallbackData = userParam;
-   ctx->Debug.ARBCallback = GL_TRUE;
+   struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+   if (debug) {
+      debug->Callback = callback;
+      debug->CallbackData = userParam;
+      debug->ARBCallback = GL_TRUE;
+   }
 }
 
+
 void
 _mesa_init_errors(struct gl_context *ctx)
 {
-   int s, t, sev;
-
-   ctx->Debug.Callback = NULL;
-   ctx->Debug.SyncOutput = GL_FALSE;
-   ctx->Debug.Log[0].length = 0;
-   ctx->Debug.NumMessages = 0;
-   ctx->Debug.NextMsg = 0;
-   ctx->Debug.NextMsgLength = 0;
-   ctx->Debug.GroupStackDepth = 0;
-
-   /* Enable all the messages with severity HIGH or MEDIUM by default. */
-   memset(ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_HIGH], GL_TRUE,
-          sizeof ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_HIGH]);
-   memset(ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_MEDIUM], GL_TRUE,
-          sizeof ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_MEDIUM]);
-   memset(ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_LOW], GL_FALSE,
-          sizeof ctx->Debug.Defaults[0][MESA_DEBUG_SEVERITY_LOW]);
-
-   /* Initialize state for filtering known debug messages. */
-   for (s = 0; s < MESA_DEBUG_SOURCE_COUNT; s++)
-      for (t = 0; t < MESA_DEBUG_TYPE_COUNT; t++) {
-         ctx->Debug.Namespaces[0][s][t].IDs = _mesa_NewHashTable();
-         assert(ctx->Debug.Namespaces[0][s][t].IDs);
-
-         for (sev = 0; sev < MESA_DEBUG_SEVERITY_COUNT; sev++)
-            make_empty_list(&ctx->Debug.Namespaces[0][s][t].Severity[sev]);
-      }
+   /* no-op */
 }
 
+
 /**
  * Loop through debug group stack tearing down states for
  * filtering debug messages.
@@ -1015,13 +1109,16 @@ _mesa_init_errors(struct gl_context *ctx)
 void
 _mesa_free_errors_data(struct gl_context *ctx)
 {
-   GLint i;
+   if (ctx->Debug) {
+      GLint i;
 
-   for (i = 0; i <= ctx->Debug.GroupStackDepth; i++) {
-      free_errors_data(ctx, i);
+      for (i = 0; i <= ctx->Debug->GroupStackDepth; i++) {
+         free_errors_data(ctx, i);
+      }
    }
 }
 
+
 /**********************************************************************/
 /** \name Diagnostics */
 /*@{*/
@@ -1077,6 +1174,7 @@ output_if_debug(const char *prefixString, const char *outputString,
    }
 }
 
+
 /**
  * When a new type of error is recorded, print a message describing
  * previous errors which were accumulated.
@@ -1149,6 +1247,7 @@ _mesa_problem( const struct gl_context *ctx, const char *fmtString, ... )
    }
 }
 
+
 static GLboolean
 should_output(struct gl_context *ctx, GLenum error, const char *fmtString)
 {
@@ -1185,6 +1284,7 @@ should_output(struct gl_context *ctx, GLenum error, const char *fmtString)
    return GL_FALSE;
 }
 
+
 void
 _mesa_gl_debug(struct gl_context *ctx,
                GLuint *id,
@@ -1202,8 +1302,7 @@ _mesa_gl_debug(struct gl_context *ctx,
    len = _mesa_vsnprintf(s, MAX_DEBUG_MESSAGE_LENGTH, fmtString, args);
    va_end(args);
 
-   _mesa_log_msg(ctx, MESA_DEBUG_SOURCE_API, type,
-                 *id, severity, len, s);
+   log_msg(ctx, MESA_DEBUG_SOURCE_API, type, *id, severity, len, s);
 }
 
 
@@ -1248,7 +1347,8 @@ _mesa_error( struct gl_context *ctx, GLenum error, const char *fmtString, ... )
 
       if (len >= MAX_DEBUG_MESSAGE_LENGTH) {
          /* Too long error message. Whoever calls _mesa_error should use
-          * shorter strings. */
+          * shorter strings.
+          */
          ASSERT(0);
          return;
       }
@@ -1268,11 +1368,8 @@ _mesa_error( struct gl_context *ctx, GLenum error, const char *fmtString, ... )
 
       /* Log the error via ARB_debug_output if needed.*/
       if (do_log) {
-         _mesa_log_msg(ctx,
-                       MESA_DEBUG_SOURCE_API,
-                       MESA_DEBUG_TYPE_ERROR,
-                       error_msg_id,
-                       MESA_DEBUG_SEVERITY_HIGH, len, s2);
+         log_msg(ctx, MESA_DEBUG_SOURCE_API, MESA_DEBUG_TYPE_ERROR,
+                 error_msg_id, MESA_DEBUG_SEVERITY_HIGH, len, s2);
       }
    }
 
@@ -1329,7 +1426,7 @@ _mesa_shader_debug( struct gl_context *ctx, GLenum type, GLuint *id,
    if (len >= MAX_DEBUG_MESSAGE_LENGTH)
       len = MAX_DEBUG_MESSAGE_LENGTH - 1;
 
-   _mesa_log_msg(ctx, source, type, *id, severity, len, msg);
+   log_msg(ctx, source, type, *id, severity, len, msg);
 }
 
 /*@}*/
diff --git a/mesalib/src/mesa/main/errors.h b/mesalib/src/mesa/main/errors.h
index a837dc8e1..cd414e6b6 100644
--- a/mesalib/src/mesa/main/errors.h
+++ b/mesalib/src/mesa/main/errors.h
@@ -83,6 +83,9 @@ _mesa_gl_debug(struct gl_context *ctx,
    }                                                                      \
 } while (0)
 
+struct gl_debug_state *
+_mesa_get_debug_state(struct gl_context *ctx);
+
 extern void
 _mesa_shader_debug(struct gl_context *ctx, GLenum type, GLuint *id,
                    const char *msg, int len);
diff --git a/mesalib/src/mesa/main/execmem.c b/mesalib/src/mesa/main/execmem.c
index d63bb4a4e..7267cf85f 100644
--- a/mesalib/src/mesa/main/execmem.c
+++ b/mesalib/src/mesa/main/execmem.c
@@ -32,7 +32,6 @@
 
 
 #include "imports.h"
-#include "glapi/glthread.h"
 
 
 
@@ -59,7 +58,7 @@
 
 #define EXEC_HEAP_SIZE (10*1024*1024)
 
-_glthread_DECLARE_STATIC_MUTEX(exec_mutex);
+static mtx_t exec_mutex = _MTX_INITIALIZER_NP;
 
 static struct mem_block *exec_heap = NULL;
 static unsigned char *exec_mem = NULL;
@@ -93,7 +92,7 @@ _mesa_exec_malloc(GLuint size)
    struct mem_block *block = NULL;
    void *addr = NULL;
 
-   _glthread_LOCK_MUTEX(exec_mutex);
+   mtx_lock(&exec_mutex);
 
    if (!init_heap())
       goto bail;
@@ -109,7 +108,7 @@ _mesa_exec_malloc(GLuint size)
       printf("_mesa_exec_malloc failed\n");
 
 bail:
-   _glthread_UNLOCK_MUTEX(exec_mutex);
+   mtx_unlock(&exec_mutex);
    
    return addr;
 }
@@ -118,7 +117,7 @@ bail:
 void 
 _mesa_exec_free(void *addr)
 {
-   _glthread_LOCK_MUTEX(exec_mutex);
+   mtx_lock(&exec_mutex);
 
    if (exec_heap) {
       struct mem_block *block = mmFindBlock(exec_heap, (unsigned char *)addr - exec_mem);
@@ -127,7 +126,7 @@ _mesa_exec_free(void *addr)
 	 mmFreeMem(block);
    }
 
-   _glthread_UNLOCK_MUTEX(exec_mutex);
+   mtx_unlock(&exec_mutex);
 }
 
 
diff --git a/mesalib/src/mesa/main/extensions.c b/mesalib/src/mesa/main/extensions.c
index 5f741fbd2..c46d70b20 100644
--- a/mesalib/src/mesa/main/extensions.c
+++ b/mesalib/src/mesa/main/extensions.c
@@ -83,6 +83,7 @@ static const struct extension extension_table[] = {
    { "GL_ARB_arrays_of_arrays",                    o(ARB_arrays_of_arrays),                    GL,             2012 },
    { "GL_ARB_base_instance",                       o(ARB_base_instance),                       GL,             2011 },
    { "GL_ARB_blend_func_extended",                 o(ARB_blend_func_extended),                 GL,             2009 },
+   { "GL_ARB_buffer_storage",                      o(ARB_buffer_storage),                      GL,             2013 },
    { "GL_ARB_clear_buffer_object",                 o(dummy_true),                              GL,             2012 },
    { "GL_ARB_color_buffer_float",                  o(ARB_color_buffer_float),                  GL,             2004 },
    { "GL_ARB_compute_shader",                      o(ARB_compute_shader),                      GL,             2012 },
@@ -106,7 +107,7 @@ static const struct extension extension_table[] = {
    { "GL_ARB_framebuffer_sRGB",                    o(EXT_framebuffer_sRGB),                    GL,             1998 },
    { "GL_ARB_get_program_binary",                  o(dummy_true),                              GL,             2010 },
    { "GL_ARB_gpu_shader5",                         o(ARB_gpu_shader5),                         GL,             2010 },
-   { "GL_ARB_half_float_pixel",                    o(ARB_half_float_pixel),                    GL,             2003 },
+   { "GL_ARB_half_float_pixel",                    o(dummy_true),                              GL,             2003 },
    { "GL_ARB_half_float_vertex",                   o(ARB_half_float_vertex),                   GL,             2008 },
    { "GL_ARB_instanced_arrays",                    o(ARB_instanced_arrays),                    GL,             2008 },
    { "GL_ARB_internalformat_query",                o(ARB_internalformat_query),                GL,             2011 },
@@ -126,6 +127,7 @@ static const struct extension extension_table[] = {
    { "GL_ARB_sample_shading",                      o(ARB_sample_shading),                      GL,             2009 },
    { "GL_ARB_sampler_objects",                     o(dummy_true),                              GL,             2009 },
    { "GL_ARB_seamless_cube_map",                   o(ARB_seamless_cube_map),                   GL,             2009 },
+   { "GL_ARB_separate_shader_objects",             o(ARB_separate_shader_objects),             GL,             2010 },
    { "GL_ARB_shader_atomic_counters",              o(ARB_shader_atomic_counters),              GL,             2011 },
    { "GL_ARB_shader_bit_encoding",                 o(ARB_shader_bit_encoding),                 GL,             2010 },
    { "GL_ARB_shader_image_load_store",             o(ARB_shader_image_load_store),             GL,             2011 },
@@ -401,7 +403,6 @@ _mesa_enable_sw_extensions(struct gl_context *ctx)
    ctx->Extensions.ARB_fragment_program_shadow = GL_TRUE;
    ctx->Extensions.ARB_fragment_shader = GL_TRUE;
    ctx->Extensions.ARB_framebuffer_object = GL_TRUE;
-   ctx->Extensions.ARB_half_float_pixel = GL_TRUE;
    ctx->Extensions.ARB_half_float_vertex = GL_TRUE;
    ctx->Extensions.ARB_map_buffer_range = GL_TRUE;
    ctx->Extensions.ARB_occlusion_query = GL_TRUE;
diff --git a/mesalib/src/mesa/main/fbobject.c b/mesalib/src/mesa/main/fbobject.c
index e459e0c63..a9dcc5144 100644
--- a/mesalib/src/mesa/main/fbobject.c
+++ b/mesalib/src/mesa/main/fbobject.c
@@ -89,9 +89,9 @@ delete_dummy_framebuffer(struct gl_framebuffer *fb)
 void
 _mesa_init_fbobjects(struct gl_context *ctx)
 {
-   _glthread_INIT_MUTEX(DummyFramebuffer.Mutex);
-   _glthread_INIT_MUTEX(DummyRenderbuffer.Mutex);
-   _glthread_INIT_MUTEX(IncompleteFramebuffer.Mutex);
+   mtx_init(&DummyFramebuffer.Mutex, mtx_plain);
+   mtx_init(&DummyRenderbuffer.Mutex, mtx_plain);
+   mtx_init(&IncompleteFramebuffer.Mutex, mtx_plain);
    DummyFramebuffer.Delete = delete_dummy_framebuffer;
    DummyRenderbuffer.Delete = delete_dummy_renderbuffer;
    IncompleteFramebuffer.Delete = delete_dummy_framebuffer;
@@ -484,7 +484,7 @@ _mesa_framebuffer_renderbuffer(struct gl_context *ctx,
 {
    struct gl_renderbuffer_attachment *att;
 
-   _glthread_LOCK_MUTEX(fb->Mutex);
+   mtx_lock(&fb->Mutex);
 
    att = get_attachment(ctx, fb, attachment);
    ASSERT(att);
@@ -504,7 +504,7 @@ _mesa_framebuffer_renderbuffer(struct gl_context *ctx,
 
    invalidate_framebuffer(fb);
 
-   _glthread_UNLOCK_MUTEX(fb->Mutex);
+   mtx_unlock(&fb->Mutex);
 }
 
 
@@ -1352,9 +1352,9 @@ _mesa_GenRenderbuffers(GLsizei n, GLuint *renderbuffers)
       GLuint name = first + i;
       renderbuffers[i] = name;
       /* insert dummy placeholder into hash table */
-      _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_lock(&ctx->Shared->Mutex);
       _mesa_HashInsert(ctx->Shared->RenderBuffers, name, &DummyRenderbuffer);
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
    }
 }
 
@@ -2218,9 +2218,9 @@ _mesa_GenFramebuffers(GLsizei n, GLuint *framebuffers)
       GLuint name = first + i;
       framebuffers[i] = name;
       /* insert dummy placeholder into hash table */
-      _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_lock(&ctx->Shared->Mutex);
       _mesa_HashInsert(ctx->Shared->FrameBuffers, name, &DummyFramebuffer);
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
    }
 }
 
@@ -2433,7 +2433,7 @@ framebuffer_texture(struct gl_context *ctx, const char *caller, GLenum target,
 
    FLUSH_VERTICES(ctx, _NEW_BUFFERS);
 
-   _glthread_LOCK_MUTEX(fb->Mutex);
+   mtx_lock(&fb->Mutex);
    if (texObj) {
       if (attachment == GL_DEPTH_ATTACHMENT &&
           texObj == fb->Attachment[BUFFER_STENCIL].Texture &&
@@ -2491,7 +2491,7 @@ framebuffer_texture(struct gl_context *ctx, const char *caller, GLenum target,
 
    invalidate_framebuffer(fb);
 
-   _glthread_UNLOCK_MUTEX(fb->Mutex);
+   mtx_unlock(&fb->Mutex);
 }
 
 
diff --git a/mesalib/src/mesa/main/format_pack.c b/mesalib/src/mesa/main/format_pack.c
index dee253c40..2772ff2d3 100644
--- a/mesalib/src/mesa/main/format_pack.c
+++ b/mesalib/src/mesa/main/format_pack.c
@@ -1936,9 +1936,9 @@ _mesa_get_pack_ubyte_rgba_function(mesa_format format)
 
       /* should never convert RGBA to these formats */
       table[MESA_FORMAT_S8_UINT_Z24_UNORM] = NULL;
-      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = NULL;
-      table[MESA_FORMAT_Z_UNORM16] = NULL;
       table[MESA_FORMAT_Z24_UNORM_S8_UINT] = NULL;
+      table[MESA_FORMAT_Z_UNORM16] = NULL;
+      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = NULL;
       table[MESA_FORMAT_X8Z24_UNORM] = NULL;
       table[MESA_FORMAT_Z_UNORM32] = NULL;
       table[MESA_FORMAT_S_UINT8] = NULL;
@@ -2099,9 +2099,9 @@ _mesa_get_pack_float_rgba_function(mesa_format format)
 
       /* should never convert RGBA to these formats */
       table[MESA_FORMAT_S8_UINT_Z24_UNORM] = NULL;
-      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = NULL;
-      table[MESA_FORMAT_Z_UNORM16] = NULL;
       table[MESA_FORMAT_Z24_UNORM_S8_UINT] = NULL;
+      table[MESA_FORMAT_Z_UNORM16] = NULL;
+      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = NULL;
       table[MESA_FORMAT_X8Z24_UNORM] = NULL;
       table[MESA_FORMAT_Z_UNORM32] = NULL;
       table[MESA_FORMAT_S_UINT8] = NULL;
@@ -2433,8 +2433,8 @@ _mesa_get_pack_float_z_func(mesa_format format)
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
    case MESA_FORMAT_X8Z24_UNORM:
       return pack_float_z_Z24_S8;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       return pack_float_z_S8_Z24;
    case MESA_FORMAT_Z_UNORM16:
       return pack_float_z_Z16;
@@ -2518,8 +2518,8 @@ _mesa_get_pack_uint_z_func(mesa_format format)
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
    case MESA_FORMAT_X8Z24_UNORM:
       return pack_uint_z_Z24_S8;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       return pack_uint_z_S8_Z24;
    case MESA_FORMAT_Z_UNORM16:
       return pack_uint_z_Z16;
@@ -2581,7 +2581,7 @@ _mesa_get_pack_ubyte_stencil_func(mesa_format format)
    switch (format) {
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       return pack_ubyte_stencil_Z24_S8;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       return pack_ubyte_stencil_S8_Z24;
    case MESA_FORMAT_S_UINT8:
       return pack_ubyte_stencil_S8;
@@ -2616,8 +2616,8 @@ _mesa_pack_float_z_row(mesa_format format, GLuint n,
          }
       }
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       {
          /* don't disturb the stencil values */
          GLuint *d = ((GLuint *) dst);
@@ -2690,8 +2690,8 @@ _mesa_pack_uint_z_row(mesa_format format, GLuint n,
          }
       }
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       {
          /* don't disturb the stencil values */
          GLuint *d = ((GLuint *) dst);
@@ -2762,7 +2762,7 @@ _mesa_pack_ubyte_stencil_row(mesa_format format, GLuint n,
          }
       }
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       {
          /* don't disturb the Z values */
          GLuint *d = ((GLuint *) dst);
@@ -2803,7 +2803,7 @@ _mesa_pack_uint_24_8_depth_stencil_row(mesa_format format, GLuint n,
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       memcpy(dst, src, n * sizeof(GLuint));
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       {
          GLuint *d = ((GLuint *) dst);
          GLuint i;
diff --git a/mesalib/src/mesa/main/format_unpack.c b/mesalib/src/mesa/main/format_unpack.c
index 02ad00a9d..276ba556a 100644
--- a/mesalib/src/mesa/main/format_unpack.c
+++ b/mesalib/src/mesa/main/format_unpack.c
@@ -2366,9 +2366,9 @@ get_unpack_rgba_function(mesa_format format)
       table[MESA_FORMAT_B10G10R10A2_UINT] = unpack_ARGB2101010_UINT;
       table[MESA_FORMAT_R10G10B10A2_UINT] = unpack_ABGR2101010_UINT;
       table[MESA_FORMAT_S8_UINT_Z24_UNORM] = unpack_Z24_S8;
-      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = unpack_S8_Z24;
+      table[MESA_FORMAT_Z24_UNORM_S8_UINT] = unpack_S8_Z24;
       table[MESA_FORMAT_Z_UNORM16] = unpack_Z16;
-      table[MESA_FORMAT_Z24_UNORM_S8_UINT] = unpack_X8_Z24;
+      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = unpack_X8_Z24;
       table[MESA_FORMAT_X8Z24_UNORM] = unpack_Z24_X8;
       table[MESA_FORMAT_Z_UNORM32] = unpack_Z32;
       table[MESA_FORMAT_S_UINT8] = unpack_S8;
@@ -3986,8 +3986,8 @@ _mesa_unpack_float_z_row(mesa_format format, GLuint n,
    case MESA_FORMAT_X8Z24_UNORM:
       unpack = unpack_float_z_Z24_X8;
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       unpack = unpack_float_z_X8_Z24;
       break;
    case MESA_FORMAT_Z_UNORM16:
@@ -4091,8 +4091,8 @@ _mesa_unpack_uint_z_row(mesa_format format, GLuint n,
    case MESA_FORMAT_X8Z24_UNORM:
       unpack = unpack_uint_z_Z24_X8;
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       unpack = unpack_uint_z_X8_Z24;
       break;
    case MESA_FORMAT_Z_UNORM16:
@@ -4164,7 +4164,7 @@ _mesa_unpack_ubyte_stencil_row(mesa_format format, GLuint n,
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       unpack_ubyte_s_Z24_S8(src, dst, n);
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       unpack_ubyte_s_S8_Z24(src, dst, n);
       break;
    case MESA_FORMAT_Z32_FLOAT_S8X24_UINT:
@@ -4189,11 +4189,30 @@ unpack_uint_24_8_depth_stencil_S8_Z24(const GLuint *src, GLuint *dst, GLuint n)
 }
 
 static void
+unpack_uint_24_8_depth_stencil_Z32_S8X24(const GLuint *src,
+                                         GLuint *dst, GLuint n)
+{
+   GLuint i;
+
+   for (i = 0; i < n; i++) {
+      /* 8 bytes per pixel (float + uint32) */
+      GLfloat zf = ((GLfloat *) src)[i * 2 + 0];
+      GLuint z24 = (GLuint) (zf * (GLfloat) 0xffffff);
+      GLuint s = src[i * 2 + 1] & 0xff;
+      dst[i] = (z24 << 8) | s;
+   }
+}
+
+static void
 unpack_uint_24_8_depth_stencil_Z24_S8(const GLuint *src, GLuint *dst, GLuint n)
 {
    memcpy(dst, src, n * 4);
 }
 
+/**
+ * Unpack depth/stencil returning as GL_UNSIGNED_INT_24_8.
+ * \param format  the source data format
+ */
 void
 _mesa_unpack_uint_24_8_depth_stencil_row(mesa_format format, GLuint n,
 					 const void *src, GLuint *dst)
@@ -4202,9 +4221,12 @@ _mesa_unpack_uint_24_8_depth_stencil_row(mesa_format format, GLuint n,
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       unpack_uint_24_8_depth_stencil_Z24_S8(src, dst, n);
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       unpack_uint_24_8_depth_stencil_S8_Z24(src, dst, n);
       break;
+   case MESA_FORMAT_Z32_FLOAT_S8X24_UINT:
+      unpack_uint_24_8_depth_stencil_Z32_S8X24(src, dst, n);
+      break;
    default:
       _mesa_problem(NULL,
                     "bad format %s in _mesa_unpack_uint_24_8_depth_stencil_row",
diff --git a/mesalib/src/mesa/main/formats.c b/mesalib/src/mesa/main/formats.c
index 10731d5a4..f6c399ede 100644
--- a/mesalib/src/mesa/main/formats.c
+++ b/mesalib/src/mesa/main/formats.c
@@ -439,8 +439,8 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] =
       1, 1, 4                      /* BlockWidth/Height,Bytes */
    },
    {
-      MESA_FORMAT_Z24_UNORM_X8_UINT,          /* Name */
-      "MESA_FORMAT_Z24_UNORM_X8_UINT",        /* StrName */
+      MESA_FORMAT_Z24_UNORM_S8_UINT,          /* Name */
+      "MESA_FORMAT_Z24_UNORM_S8_UINT",        /* StrName */
       GL_DEPTH_STENCIL,            /* BaseFormat */
       GL_UNSIGNED_NORMALIZED,      /* DataType */
       0, 0, 0, 0,                  /* Red/Green/Blue/AlphaBits */
@@ -457,8 +457,8 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] =
       1, 1, 2                      /* BlockWidth/Height,Bytes */
    },
    {
-      MESA_FORMAT_Z24_UNORM_S8_UINT,          /* Name */
-      "MESA_FORMAT_Z24_UNORM_S8_UINT",        /* StrName */
+      MESA_FORMAT_Z24_UNORM_X8_UINT,          /* Name */
+      "MESA_FORMAT_Z24_UNORM_X8_UINT",        /* StrName */
       GL_DEPTH_COMPONENT,          /* BaseFormat */
       GL_UNSIGNED_NORMALIZED,      /* DataType */
       0, 0, 0, 0,                  /* Red/Green/Blue/AlphaBits */
@@ -2472,7 +2472,7 @@ _mesa_format_to_type_and_comps(mesa_format format,
       *comps = 2;
       return;
 
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       *datatype = GL_UNSIGNED_INT_8_24_REV_MESA;
       *comps = 2;
       return;
@@ -2482,7 +2482,7 @@ _mesa_format_to_type_and_comps(mesa_format format,
       *comps = 1;
       return;
 
-   case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       *datatype = GL_UNSIGNED_INT;
       *comps = 1;
       return;
@@ -3123,14 +3123,14 @@ _mesa_format_matches_format_and_type(mesa_format mesa_format,
       return format == GL_DEPTH_STENCIL && type == GL_UNSIGNED_INT_24_8 &&
          !swapBytes;
    case MESA_FORMAT_X8Z24_UNORM:
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       return GL_FALSE;
 
    case MESA_FORMAT_Z_UNORM16:
       return format == GL_DEPTH_COMPONENT && type == GL_UNSIGNED_SHORT &&
          !swapBytes;
 
-   case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       return GL_FALSE;
 
    case MESA_FORMAT_Z_UNORM32:
diff --git a/mesalib/src/mesa/main/formats.h b/mesalib/src/mesa/main/formats.h
index 63d9565e3..3102584b6 100644
--- a/mesalib/src/mesa/main/formats.h
+++ b/mesalib/src/mesa/main/formats.h
@@ -339,33 +339,33 @@ typedef enum
 
    /* Type P formats */          /* msb <------ TEXEL BITS -----------> lsb */
                                  /* ---- ---- ---- ---- ---- ---- ---- ---- */
-   MESA_FORMAT_A8B8G8R8_UNORM,   /* AAAA AAAA BBBB BBBB GGGG GGGG RRRR RRRR */
-   MESA_FORMAT_R8G8B8A8_UNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB AAAA AAAA */
-   MESA_FORMAT_B8G8R8A8_UNORM,   /* BBBB BBBB GGGG GGGG RRRR RRRR AAAA AAAA */
-   MESA_FORMAT_A8R8G8B8_UNORM,   /* AAAA AAAA RRRR RRRR GGGG GGGG BBBB BBBB */
-   MESA_FORMAT_X8B8G8R8_UNORM,   /* xxxx xxxx BBBB BBBB GGGG GGGG RRRR RRRR */
-   MESA_FORMAT_R8G8B8X8_UNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB xxxx xxxx */
-   MESA_FORMAT_B8G8R8X8_UNORM,   /* BBBB BBBB GGGG GGGG RRRR RRRR xxxx xxxx */
-   MESA_FORMAT_X8R8G8B8_UNORM,   /* xxxx xxxx RRRR RRRR GGGG GGGG BBBB BBBB */
+   MESA_FORMAT_A8B8G8R8_UNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB AAAA AAAA */
+   MESA_FORMAT_R8G8B8A8_UNORM,   /* AAAA AAAA BBBB BBBB GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_B8G8R8A8_UNORM,   /* AAAA AAAA RRRR RRRR GGGG GGGG BBBB BBBB */
+   MESA_FORMAT_A8R8G8B8_UNORM,   /* BBBB BBBB GGGG GGGG RRRR RRRR AAAA AAAA */
+   MESA_FORMAT_X8B8G8R8_UNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB xxxx xxxx */
+   MESA_FORMAT_R8G8B8X8_UNORM,   /* xxxx xxxx BBBB BBBB GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_B8G8R8X8_UNORM,   /* xxxx xxxx RRRR RRRR GGGG GGGG BBBB BBBB */
+   MESA_FORMAT_X8R8G8B8_UNORM,   /* BBBB BBBB GGGG GGGG RRRR RRRR xxxx xxxx */
 
    /* Type A formats */
    MESA_FORMAT_BGR_UNORM8,    /* uchar[i * 3] = B, [i * 3 + 1] = G, [i *3 + 2] = R */
    MESA_FORMAT_RGB_UNORM8,    /* uchar[i * 3] = R, [i * 3 + 1] = G, [i *3 + 2] = B */
 
    /* Type P formats */
-   MESA_FORMAT_B5G6R5_UNORM,                         /* BBBB BGGG GGGR RRRR */
-   MESA_FORMAT_R5G6B5_UNORM,                         /* RRRR RGGG GGGB BBBB */
-   MESA_FORMAT_B4G4R4A4_UNORM,                       /* BBBB GGGG RRRR AAAA */
-   MESA_FORMAT_A4R4G4B4_UNORM,                       /* AAAA RRRR GGGG BBBB */
-   MESA_FORMAT_A1B5G5R5_UNORM,                       /* ABBB BBGG GGGR RRRR */
-   MESA_FORMAT_B5G5R5A1_UNORM,                       /* BBBB BGGG GGRR RRRA */
-   MESA_FORMAT_A1R5G5B5_UNORM,                       /* ARRR RRGG GGGB BBBB */
-   MESA_FORMAT_L4A4_UNORM,                                     /* LLLL AAAA */
-   MESA_FORMAT_L8A8_UNORM,                           /* LLLL LLLL AAAA AAAA */
-   MESA_FORMAT_A8L8_UNORM,                           /* AAAA AAAA LLLL LLLL */
-   MESA_FORMAT_L16A16_UNORM,     /* LLLL LLLL LLLL LLLL AAAA AAAA AAAA AAAA */
-   MESA_FORMAT_A16L16_UNORM,     /* AAAA AAAA AAAA AAAA LLLL LLLL LLLL LLLL */
-   MESA_FORMAT_B2G3R3_UNORM,                                   /* BBGG GRRR */
+   MESA_FORMAT_B5G6R5_UNORM,                         /* RRRR RGGG GGGB BBBB */
+   MESA_FORMAT_R5G6B5_UNORM,                         /* BBBB BGGG GGGR RRRR */
+   MESA_FORMAT_B4G4R4A4_UNORM,                       /* AAAA RRRR GGGG BBBB */
+   MESA_FORMAT_A4R4G4B4_UNORM,                       /* BBBB GGGG RRRR AAAA */
+   MESA_FORMAT_A1B5G5R5_UNORM,                       /* RRRR RGGG GGBB BBBA */
+   MESA_FORMAT_B5G5R5A1_UNORM,                       /* ARRR RRGG GGGB BBBB */
+   MESA_FORMAT_A1R5G5B5_UNORM,                       /* BBBB BGGG GGRR RRRA */
+   MESA_FORMAT_L4A4_UNORM,                                     /* AAAA LLLL */
+   MESA_FORMAT_L8A8_UNORM,                           /* AAAA AAAA LLLL LLLL */
+   MESA_FORMAT_A8L8_UNORM,                           /* LLLL LLLL AAAA AAAA */
+   MESA_FORMAT_L16A16_UNORM,     /* AAAA AAAA AAAA AAAA LLLL LLLL LLLL LLLL */
+   MESA_FORMAT_A16L16_UNORM,     /* LLLL LLLL LLLL LLLL AAAA AAAA AAAA AAAA */
+   MESA_FORMAT_B2G3R3_UNORM,                                   /* RRRG GGBB */
 
    /* Type A formats */
    MESA_FORMAT_A_UNORM8,     /* uchar[i] = A */
@@ -383,25 +383,25 @@ typedef enum
    MESA_FORMAT_R_UNORM8,      /* uchar[i] = R */
 
    /* Type P formats */
-   MESA_FORMAT_R8G8_UNORM,                           /* RRRR RRRR GGGG GGGG */
-   MESA_FORMAT_G8R8_UNORM,                           /* GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_R8G8_UNORM,                           /* GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_G8R8_UNORM,                           /* RRRR RRRR GGGG GGGG */
 
    /* Type A format(s) */
    MESA_FORMAT_R_UNORM16,     /* ushort[i] = R */
 
    /* Type P formats */
-   MESA_FORMAT_R16G16_UNORM,     /* RRRR RRRR RRRR RRRR GGGG GGGG GGGG GGGG */
-   MESA_FORMAT_G16R16_UNORM,     /* GGGG GGGG GGGG GGGG RRRR RRRR RRRR RRRR */
-   MESA_FORMAT_B10G10R10A2_UNORM,/* BBBB BBBB BBGG GGGG GGGG RRRR RRRR RRAA */
-   MESA_FORMAT_S8_UINT_Z24_UNORM,/* SSSS SSSS ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ */
-   MESA_FORMAT_Z24_UNORM_X8_UINT,/* ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ xxxx xxxx */
+   MESA_FORMAT_R16G16_UNORM,     /* GGGG GGGG GGGG GGGG RRRR RRRR RRRR RRRR */
+   MESA_FORMAT_G16R16_UNORM,     /* RRRR RRRR RRRR RRRR GGGG GGGG GGGG GGGG */
+   MESA_FORMAT_B10G10R10A2_UNORM,/* AARR RRRR RRRR GGGG GGGG GGBB BBBB BBBB */
+   MESA_FORMAT_S8_UINT_Z24_UNORM,/* ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ SSSS SSSS */
+   MESA_FORMAT_Z24_UNORM_S8_UINT,/* SSSS SSSS ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ */
 
    /* Type A format(s) */
    MESA_FORMAT_Z_UNORM16,     /* ushort[i] = Z */
 
    /* Type P formats */
-   MESA_FORMAT_Z24_UNORM_S8_UINT,/* ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ SSSS SSSS */
-   MESA_FORMAT_X8Z24_UNORM,      /* xxxx xxxx ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ */
+   MESA_FORMAT_Z24_UNORM_X8_UINT,/* xxxx xxxx ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ */
+   MESA_FORMAT_X8Z24_UNORM,      /* ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ ZZZZ xxxx xxxx */
 
    /* Type A formats */
    MESA_FORMAT_Z_UNORM32,     /* uint[i] = Z */
@@ -416,14 +416,14 @@ typedef enum
    MESA_FORMAT_BGR_SRGB8,     /* uchar[i * 3] = B, [i * 3 + 1] = G, [i *3 + 2] = R */
 
    /* Type P formats */
-   MESA_FORMAT_A8B8G8R8_SRGB,    /* AAAA AAAA BBBB BBBB GGGG GGGG RRRR RRRR*/
-   MESA_FORMAT_B8G8R8A8_SRGB,    /* BBBB BBBB GGGG GGGG RRRR RRRR AAAA AAAA */
+   MESA_FORMAT_A8B8G8R8_SRGB,    /* RRRR RRRR GGGG GGGG BBBB BBBB AAAA AAAA */
+   MESA_FORMAT_B8G8R8A8_SRGB,    /* AAAA AAAA RRRR RRRR GGGG GGGG BBBB BBBB */
 
    /* Type A format(s) */
    MESA_FORMAT_L_SRGB8,       /* uchar[i] = L */
 
    /* Type P formats */
-   MESA_FORMAT_L8A8_SRGB,        /* LLLL LLLL AAAA AAAA */
+   MESA_FORMAT_L8A8_SRGB,                            /* AAAA AAAA LLLL LLLL */
 
    /* Type C formats */
    MESA_FORMAT_SRGB_DXT1,
@@ -548,16 +548,16 @@ typedef enum
    MESA_FORMAT_R_SNORM8,      /* char[i] = R */
 
    /* Type P formats */
-   MESA_FORMAT_R8G8_SNORM,                           /* RRRR RRRR GGGG GGGG */
-   MESA_FORMAT_X8B8G8R8_SNORM,   /* xxxx xxxx BBBB BBBB GGGG GGGG RRRR RRRR */
-   MESA_FORMAT_A8B8G8R8_SNORM,   /* AAAA AAAA BBBB BBBB GGGG GGGG RRRR RRRR */
-   MESA_FORMAT_R8G8B8A8_SNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB AAAA AAAA */
+   MESA_FORMAT_R8G8_SNORM,                           /* GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_X8B8G8R8_SNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB xxxx xxxx */
+   MESA_FORMAT_A8B8G8R8_SNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB AAAA AAAA */
+   MESA_FORMAT_R8G8B8A8_SNORM,   /* AAAA AAAA BBBB BBBB GGGG GGGG RRRR RRRR */
 
    /* Type A format(s) */
    MESA_FORMAT_R_SNORM16,     /* short[i] = R */
 
    /* Type P format(s) */
-   MESA_FORMAT_R16G16_SNORM,     /* RRRR RRRR RRRR RRRR GGGG GGGG GGGG GGGG */
+   MESA_FORMAT_R16G16_SNORM,     /* GGGG GGGG GGGG GGGG RRRR RRRR RRRR RRRR */
 
    /* Type A format(s) */
    MESA_FORMAT_RGB_SNORM16,   /* short[i * 3] = R, [i * 3 + 1] = G, [i *3 + 2] = B */
@@ -597,7 +597,7 @@ typedef enum
    MESA_FORMAT_L_SNORM8,      /* char[i] = L */
 
    /* Type P format(s) */
-   MESA_FORMAT_L8A8_SNORM,                           /* LLLL LLLL AAAA AAAA */
+   MESA_FORMAT_L8A8_SNORM,                           /* AAAA AAAA LLLL LLLL */
 
    /* Type A format(s) */
    MESA_FORMAT_I_SNORM8,      /* char[i] = I */
@@ -608,21 +608,21 @@ typedef enum
 
    /* Type P format(s) */
    MESA_FORMAT_R9G9B9E5_FLOAT,
-   MESA_FORMAT_R11G11B10_FLOAT,
+   MESA_FORMAT_R11G11B10_FLOAT,   /* BBBB BBBB BBGG GGGG GGGG GRRR RRRR RRRR */
 
    /* Type A format(s) */
    MESA_FORMAT_Z_FLOAT32,
 
    /* Type P formats */
-   MESA_FORMAT_Z32_FLOAT_S8X24_UINT,
+   MESA_FORMAT_Z32_FLOAT_S8X24_UINT, /* (float, x24s8) */
 
-   MESA_FORMAT_B10G10R10A2_UINT,
-   MESA_FORMAT_R10G10B10A2_UINT,
+   MESA_FORMAT_B10G10R10A2_UINT, /* AARR RRRR RRRR GGGG GGGG GGBB BBBB BBBB */
+   MESA_FORMAT_R10G10B10A2_UINT, /* AABB BBBB BBBB GGGG GGGG GGRR RRRR RRRR */
 
-   MESA_FORMAT_B4G4R4X4_UNORM,                       /* BBBB GGGG RRRR xxxx */
-   MESA_FORMAT_B5G5R5X1_UNORM,                       /* BBBB BGGG GGRR RRRx */
-   MESA_FORMAT_R8G8B8X8_SNORM,   /* RRRR RRRR GGGG GGGG BBBB BBBB xxxx xxxx */
-   MESA_FORMAT_R8G8B8X8_SRGB,    /* RRRR RRRR GGGG GGGG BBBB BBBB xxxx xxxx */
+   MESA_FORMAT_B4G4R4X4_UNORM,                       /* xxxx RRRR GGGG BBBB */
+   MESA_FORMAT_B5G5R5X1_UNORM,                       /* xRRR RRGG GGGB BBBB */
+   MESA_FORMAT_R8G8B8X8_SNORM,   /* xxxx xxxx BBBB BBBB GGGG GGGG RRRR RRRR */
+   MESA_FORMAT_R8G8B8X8_SRGB,    /* xxxx xxxx BBBB BBBB GGGG GGGG RRRR RRRR */
 
    /* Type A formats */
    MESA_FORMAT_RGBX_UINT8,    /* uchar[i * 4] = R, [i * 4 + 1] = G, [i * 4 + 2] = B, [i * 4 + 3] = x */
@@ -643,9 +643,9 @@ typedef enum
    MESA_FORMAT_RGBX_SINT32,   /* ... */
 
    /* Type P formats */
-   MESA_FORMAT_R10G10B10A2_UNORM,
-   MESA_FORMAT_G8R8_SNORM,
-   MESA_FORMAT_G16R16_SNORM,
+   MESA_FORMAT_R10G10B10A2_UNORM, /* AABB BBBB BBBB GGGG GGGG GGRR RRRR RRRR */
+   MESA_FORMAT_G8R8_SNORM,        /*                     RRRR RRRR GGGG GGGG */
+   MESA_FORMAT_G16R16_SNORM,      /* RRRR RRRR RRRR RRRR GGGG GGGG GGGG GGGG */
 
    MESA_FORMAT_COUNT
 } mesa_format;
diff --git a/mesalib/src/mesa/main/framebuffer.c b/mesalib/src/mesa/main/framebuffer.c
index 54eeda2b2..7416bb118 100644
--- a/mesalib/src/mesa/main/framebuffer.c
+++ b/mesalib/src/mesa/main/framebuffer.c
@@ -130,7 +130,7 @@ _mesa_initialize_window_framebuffer(struct gl_framebuffer *fb,
 
    memset(fb, 0, sizeof(struct gl_framebuffer));
 
-   _glthread_INIT_MUTEX(fb->Mutex);
+   mtx_init(&fb->Mutex, mtx_plain);
 
    fb->RefCount = 1;
 
@@ -182,7 +182,7 @@ _mesa_initialize_user_framebuffer(struct gl_framebuffer *fb, GLuint name)
    fb->ColorReadBuffer = GL_COLOR_ATTACHMENT0_EXT;
    fb->_ColorReadBufferIndex = BUFFER_COLOR0;
    fb->Delete = _mesa_destroy_framebuffer;
-   _glthread_INIT_MUTEX(fb->Mutex);
+   mtx_init(&fb->Mutex, mtx_plain);
 }
 
 
@@ -213,7 +213,7 @@ _mesa_free_framebuffer_data(struct gl_framebuffer *fb)
    assert(fb);
    assert(fb->RefCount == 0);
 
-   _glthread_DESTROY_MUTEX(fb->Mutex);
+   mtx_destroy(&fb->Mutex);
 
    for (i = 0; i < BUFFER_COUNT; i++) {
       struct gl_renderbuffer_attachment *att = &fb->Attachment[i];
@@ -244,11 +244,11 @@ _mesa_reference_framebuffer_(struct gl_framebuffer **ptr,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_framebuffer *oldFb = *ptr;
 
-      _glthread_LOCK_MUTEX(oldFb->Mutex);
+      mtx_lock(&oldFb->Mutex);
       ASSERT(oldFb->RefCount > 0);
       oldFb->RefCount--;
       deleteFlag = (oldFb->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(oldFb->Mutex);
+      mtx_unlock(&oldFb->Mutex);
       
       if (deleteFlag)
          oldFb->Delete(oldFb);
@@ -258,9 +258,9 @@ _mesa_reference_framebuffer_(struct gl_framebuffer **ptr,
    assert(!*ptr);
 
    if (fb) {
-      _glthread_LOCK_MUTEX(fb->Mutex);
+      mtx_lock(&fb->Mutex);
       fb->RefCount++;
-      _glthread_UNLOCK_MUTEX(fb->Mutex);
+      mtx_unlock(&fb->Mutex);
       *ptr = fb;
    }
 }
diff --git a/mesalib/src/mesa/main/get.c b/mesalib/src/mesa/main/get.c
index 54f7d7745..b1908515c 100644
--- a/mesalib/src/mesa/main/get.c
+++ b/mesalib/src/mesa/main/get.c
@@ -28,6 +28,7 @@
 #include "blend.h"
 #include "enable.h"
 #include "enums.h"
+#include "errors.h"
 #include "extensions.h"
 #include "get.h"
 #include "macros.h"
@@ -973,6 +974,26 @@ find_custom_value(struct gl_context *ctx, const struct value_desc *d, union valu
          _mesa_problem(ctx, "driver doesn't implement GetTimestamp");
       }
       break;
+   /* GL_KHR_DEBUG */
+   case GL_DEBUG_LOGGED_MESSAGES:
+      {
+         struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+         v->value_int = debug ? debug->NumMessages : 0;
+      }
+      break;
+   case GL_DEBUG_NEXT_LOGGED_MESSAGE_LENGTH:
+      {
+         struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+         v->value_int = debug ? debug->NextMsgLength : 0;
+      }
+      break;
+   case GL_DEBUG_GROUP_STACK_DEPTH:
+      {
+         struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+         v->value_int = debug ? debug->GroupStackDepth : 0;
+      }
+      break;
+
    /* GL_ARB_shader_atomic_counters */
    case GL_ATOMIC_COUNTER_BUFFER_BINDING:
       v->value_int = ctx->AtomicBuffer->Name;
diff --git a/mesalib/src/mesa/main/get_hash_params.py b/mesalib/src/mesa/main/get_hash_params.py
index f47cbd881..674d0032a 100644
--- a/mesalib/src/mesa/main/get_hash_params.py
+++ b/mesalib/src/mesa/main/get_hash_params.py
@@ -698,13 +698,13 @@ descriptor=[
   [ "RESET_NOTIFICATION_STRATEGY_ARB", "CONTEXT_ENUM(Const.ResetStrategy), NO_EXTRA" ],
 
 # GL_KHR_debug (GL 4.3)/ GL_ARB_debug_output
-  [ "DEBUG_LOGGED_MESSAGES", "CONTEXT_INT(Debug.NumMessages), NO_EXTRA" ],
-  [ "DEBUG_NEXT_LOGGED_MESSAGE_LENGTH", "CONTEXT_INT(Debug.NextMsgLength), NO_EXTRA" ],
+  [ "DEBUG_LOGGED_MESSAGES", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
+  [ "DEBUG_NEXT_LOGGED_MESSAGE_LENGTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
   [ "MAX_DEBUG_LOGGED_MESSAGES", "CONST(MAX_DEBUG_LOGGED_MESSAGES), NO_EXTRA" ],
   [ "MAX_DEBUG_MESSAGE_LENGTH", "CONST(MAX_DEBUG_MESSAGE_LENGTH), NO_EXTRA" ],
   [ "MAX_LABEL_LENGTH", "CONST(MAX_LABEL_LENGTH), NO_EXTRA" ],
   [ "MAX_DEBUG_GROUP_STACK_DEPTH", "CONST(MAX_DEBUG_GROUP_STACK_DEPTH), NO_EXTRA" ],
-  [ "DEBUG_GROUP_STACK_DEPTH", "CONTEXT_INT(Debug.GroupStackDepth), NO_EXTRA" ],
+  [ "DEBUG_GROUP_STACK_DEPTH", "LOC_CUSTOM, TYPE_INT, 0, NO_EXTRA" ],
 
   [ "MAX_DUAL_SOURCE_DRAW_BUFFERS", "CONTEXT_INT(Const.MaxDualSourceDrawBuffers), extra_ARB_blend_func_extended" ],
 
diff --git a/mesalib/src/mesa/main/getstring.c b/mesalib/src/mesa/main/getstring.c
index 674126702..3ac62d402 100644
--- a/mesalib/src/mesa/main/getstring.c
+++ b/mesalib/src/mesa/main/getstring.c
@@ -253,14 +253,22 @@ _mesa_GetPointerv( GLenum pname, GLvoid **params )
          *params = (GLvoid *) ctx->Array.VAO->VertexAttrib[VERT_ATTRIB_POINT_SIZE].Ptr;
          break;
       case GL_DEBUG_CALLBACK_FUNCTION_ARB:
-         if (!_mesa_is_desktop_gl(ctx))
+         if (!_mesa_is_desktop_gl(ctx)) {
             goto invalid_pname;
-         *params = (GLvoid *) ctx->Debug.Callback;
+         }
+         else {
+            struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+            *params = debug ? (void *) debug->Callback : NULL;
+         }
          break;
       case GL_DEBUG_CALLBACK_USER_PARAM_ARB:
-         if (!_mesa_is_desktop_gl(ctx))
+         if (!_mesa_is_desktop_gl(ctx)) {
             goto invalid_pname;
-         *params = (GLvoid *) ctx->Debug.CallbackData;
+         }
+         else {
+            struct gl_debug_state *debug = _mesa_get_debug_state(ctx);
+            *params = debug ? (void *) debug->CallbackData : NULL;
+         }
          break;
       default:
          goto invalid_pname;
@@ -326,7 +334,7 @@ _mesa_GetGraphicsResetStatusARB( void )
        */
       status = ctx->Driver.GetGraphicsResetStatus(ctx);
 
-      _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_lock(&ctx->Shared->Mutex);
 
       /* If this context has not been affected by a GPU reset, check to see if
        * some other context in the share group has been affected by a reset.
@@ -340,7 +348,7 @@ _mesa_GetGraphicsResetStatusARB( void )
       }
 
       ctx->ShareGroupReset = ctx->Shared->ShareGroupReset;
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
    }
 
    if (!ctx->Driver.GetGraphicsResetStatus && (MESA_VERBOSE & VERBOSE_API))
diff --git a/mesalib/src/mesa/main/glformats.c b/mesalib/src/mesa/main/glformats.c
index 7d4a31057..77cf26337 100644
--- a/mesalib/src/mesa/main/glformats.c
+++ b/mesalib/src/mesa/main/glformats.c
@@ -1323,10 +1323,8 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_INT:
             case GL_UNSIGNED_INT:
             case GL_FLOAT:
-               return GL_NO_ERROR;
             case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
+               return GL_NO_ERROR;
             default:
                return GL_INVALID_ENUM;
          }
@@ -1349,10 +1347,8 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_INT:
             case GL_UNSIGNED_INT:
             case GL_FLOAT:
-               return GL_NO_ERROR;
             case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
+               return GL_NO_ERROR;
             default:
                return GL_INVALID_ENUM;
          }
@@ -1368,10 +1364,8 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_INT:
             case GL_UNSIGNED_INT:
             case GL_FLOAT:
-               return GL_NO_ERROR;
             case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
+               return GL_NO_ERROR;
             default:
                return GL_INVALID_ENUM;
          }
@@ -1389,14 +1383,12 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_UNSIGNED_BYTE_2_3_3_REV:
             case GL_UNSIGNED_SHORT_5_6_5:
             case GL_UNSIGNED_SHORT_5_6_5_REV:
+            case GL_HALF_FLOAT:
                return GL_NO_ERROR;
             case GL_UNSIGNED_INT_2_10_10_10_REV:
                /* OK by GL_EXT_texture_type_2_10_10_10_REV */
                return (ctx->API == API_OPENGLES2)
                   ? GL_NO_ERROR : GL_INVALID_ENUM;
-            case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
             case GL_UNSIGNED_INT_5_9_9_9_REV:
                return ctx->Extensions.EXT_texture_shared_exponent
                   ? GL_NO_ERROR : GL_INVALID_ENUM;
@@ -1419,10 +1411,8 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_INT:
             case GL_UNSIGNED_INT:
             case GL_FLOAT:
-               return GL_NO_ERROR;
             case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
+               return GL_NO_ERROR;
             default:
                return GL_INVALID_ENUM;
          }
@@ -1446,10 +1436,8 @@ _mesa_error_check_format_and_type(const struct gl_context *ctx,
             case GL_UNSIGNED_INT_8_8_8_8_REV:
             case GL_UNSIGNED_INT_10_10_10_2:
             case GL_UNSIGNED_INT_2_10_10_10_REV:
-               return GL_NO_ERROR;
             case GL_HALF_FLOAT:
-               return ctx->Extensions.ARB_half_float_pixel
-                  ? GL_NO_ERROR : GL_INVALID_ENUM;
+               return GL_NO_ERROR;
             default:
                return GL_INVALID_ENUM;
          }
diff --git a/mesalib/src/mesa/main/hash.c b/mesalib/src/mesa/main/hash.c
index b31fd4839..4c92005e0 100644
--- a/mesalib/src/mesa/main/hash.c
+++ b/mesalib/src/mesa/main/hash.c
@@ -36,7 +36,6 @@
 
 #include "glheader.h"
 #include "imports.h"
-#include "glapi/glthread.h"
 #include "hash.h"
 #include "hash_table.h"
 
@@ -59,8 +58,8 @@
 struct _mesa_HashTable {
    struct hash_table *ht;
    GLuint MaxKey;                        /**< highest key inserted so far */
-   _glthread_Mutex Mutex;                /**< mutual exclusion lock */
-   _glthread_Mutex WalkMutex;            /**< for _mesa_HashWalk() */
+   mtx_t Mutex;                /**< mutual exclusion lock */
+   mtx_t WalkMutex;            /**< for _mesa_HashWalk() */
    GLboolean InDeleteAll;                /**< Debug check */
    /** Value that would be in the table for DELETED_KEY_VALUE. */
    void *deleted_key_data;
@@ -117,8 +116,8 @@ _mesa_NewHashTable(void)
    if (table) {
       table->ht = _mesa_hash_table_create(NULL, uint_key_compare);
       _mesa_hash_table_set_deleted_key(table->ht, uint_key(DELETED_KEY_VALUE));
-      _glthread_INIT_MUTEX(table->Mutex);
-      _glthread_INIT_MUTEX(table->WalkMutex);
+      mtx_init(&table->Mutex, mtx_plain);
+      mtx_init(&table->WalkMutex, mtx_plain);
    }
    return table;
 }
@@ -144,8 +143,8 @@ _mesa_DeleteHashTable(struct _mesa_HashTable *table)
 
    _mesa_hash_table_destroy(table->ht, NULL);
 
-   _glthread_DESTROY_MUTEX(table->Mutex);
-   _glthread_DESTROY_MUTEX(table->WalkMutex);
+   mtx_destroy(&table->Mutex);
+   mtx_destroy(&table->WalkMutex);
    free(table);
 }
 
@@ -187,9 +186,9 @@ _mesa_HashLookup(struct _mesa_HashTable *table, GLuint key)
 {
    void *res;
    assert(table);
-   _glthread_LOCK_MUTEX(table->Mutex);
+   mtx_lock(&table->Mutex);
    res = _mesa_HashLookup_unlocked(table, key);
-   _glthread_UNLOCK_MUTEX(table->Mutex);
+   mtx_unlock(&table->Mutex);
    return res;
 }
 
@@ -211,7 +210,7 @@ _mesa_HashInsert(struct _mesa_HashTable *table, GLuint key, void *data)
    assert(table);
    assert(key);
 
-   _glthread_LOCK_MUTEX(table->Mutex);
+   mtx_lock(&table->Mutex);
 
    if (key > table->MaxKey)
       table->MaxKey = key;
@@ -227,7 +226,7 @@ _mesa_HashInsert(struct _mesa_HashTable *table, GLuint key, void *data)
       }
    }
 
-   _glthread_UNLOCK_MUTEX(table->Mutex);
+   mtx_unlock(&table->Mutex);
 }
 
 
@@ -256,14 +255,14 @@ _mesa_HashRemove(struct _mesa_HashTable *table, GLuint key)
       return;
    }
 
-   _glthread_LOCK_MUTEX(table->Mutex);
+   mtx_lock(&table->Mutex);
    if (key == DELETED_KEY_VALUE) {
       table->deleted_key_data = NULL;
    } else {
       entry = _mesa_hash_table_search(table->ht, uint_hash(key), uint_key(key));
       _mesa_hash_table_remove(table->ht, entry);
    }
-   _glthread_UNLOCK_MUTEX(table->Mutex);
+   mtx_unlock(&table->Mutex);
 }
 
 
@@ -286,7 +285,7 @@ _mesa_HashDeleteAll(struct _mesa_HashTable *table,
 
    ASSERT(table);
    ASSERT(callback);
-   _glthread_LOCK_MUTEX(table->Mutex);
+   mtx_lock(&table->Mutex);
    table->InDeleteAll = GL_TRUE;
    hash_table_foreach(table->ht, entry) {
       callback((uintptr_t)entry->key, entry->data, userData);
@@ -297,7 +296,7 @@ _mesa_HashDeleteAll(struct _mesa_HashTable *table,
       table->deleted_key_data = NULL;
    }
    table->InDeleteAll = GL_FALSE;
-   _glthread_UNLOCK_MUTEX(table->Mutex);
+   mtx_unlock(&table->Mutex);
 }
 
 
@@ -315,7 +314,7 @@ _mesa_HashClone(const struct _mesa_HashTable *table)
    struct _mesa_HashTable *clonetable;
 
    ASSERT(table);
-   _glthread_LOCK_MUTEX(table2->Mutex);
+   mtx_lock(&table2->Mutex);
 
    clonetable = _mesa_NewHashTable();
    assert(clonetable);
@@ -323,7 +322,7 @@ _mesa_HashClone(const struct _mesa_HashTable *table)
       _mesa_HashInsert(clonetable, (GLint)(uintptr_t)entry->key, entry->data);
    }
 
-   _glthread_UNLOCK_MUTEX(table2->Mutex);
+   mtx_unlock(&table2->Mutex);
 
    return clonetable;
 }
@@ -352,13 +351,13 @@ _mesa_HashWalk(const struct _mesa_HashTable *table,
 
    ASSERT(table);
    ASSERT(callback);
-   _glthread_LOCK_MUTEX(table2->WalkMutex);
+   mtx_lock(&table2->WalkMutex);
    hash_table_foreach(table->ht, entry) {
       callback((uintptr_t)entry->key, entry->data, userData);
    }
    if (table->deleted_key_data)
       callback(DELETED_KEY_VALUE, table->deleted_key_data, userData);
-   _glthread_UNLOCK_MUTEX(table2->WalkMutex);
+   mtx_unlock(&table2->WalkMutex);
 }
 
 static void
@@ -398,10 +397,10 @@ GLuint
 _mesa_HashFindFreeKeyBlock(struct _mesa_HashTable *table, GLuint numKeys)
 {
    const GLuint maxKey = ~((GLuint) 0) - 1;
-   _glthread_LOCK_MUTEX(table->Mutex);
+   mtx_lock(&table->Mutex);
    if (maxKey - numKeys > table->MaxKey) {
       /* the quick solution */
-      _glthread_UNLOCK_MUTEX(table->Mutex);
+      mtx_unlock(&table->Mutex);
       return table->MaxKey + 1;
    }
    else {
@@ -419,13 +418,13 @@ _mesa_HashFindFreeKeyBlock(struct _mesa_HashTable *table, GLuint numKeys)
 	    /* this key not in use, check if we've found enough */
 	    freeCount++;
 	    if (freeCount == numKeys) {
-               _glthread_UNLOCK_MUTEX(table->Mutex);
+               mtx_unlock(&table->Mutex);
 	       return freeStart;
 	    }
 	 }
       }
       /* cannot allocate a block of numKeys consecutive keys */
-      _glthread_UNLOCK_MUTEX(table->Mutex);
+      mtx_unlock(&table->Mutex);
       return 0;
    }
 }
diff --git a/mesalib/src/mesa/main/imports.h b/mesalib/src/mesa/main/imports.h
index 53e40b445..29772be6e 100644
--- a/mesalib/src/mesa/main/imports.h
+++ b/mesalib/src/mesa/main/imports.h
@@ -132,7 +132,7 @@ typedef union { GLfloat f; GLint i; GLuint u; } fi_type;
 #define atanhf(f) ((float) atanh(f))
 #endif
 
-#if defined(_MSC_VER)
+#if defined(_MSC_VER) && (_MSC_VER < 1800)  /* Not req'd on VS2013 and above */
 static inline float truncf(float x) { return x < 0.0f ? ceilf(x) : floorf(x); }
 static inline float exp2f(float x) { return powf(2.0f, x); }
 static inline float log2f(float x) { return logf(x) * 1.442695041f; }
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index ceabd9df7..bbc377280 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -1190,7 +1190,7 @@ struct gl_sampler_object
  */
 struct gl_texture_object
 {
-   _glthread_Mutex Mutex;      /**< for thread safety */
+   mtx_t Mutex;      /**< for thread safety */
    GLint RefCount;             /**< reference count */
    GLuint Name;                /**< the user-visible texture object ID */
    GLchar *Label;               /**< GL_KHR_debug */
@@ -1439,28 +1439,44 @@ struct gl_viewport_attrib
 };
 
 
+typedef enum {
+   MAP_USER,
+   MAP_INTERNAL,
+
+   MAP_COUNT
+} gl_map_buffer_index;
+
+
+/**
+ * Fields describing a mapped buffer range.
+ */
+struct gl_buffer_mapping {
+   GLbitfield AccessFlags; /**< Mask of GL_MAP_x_BIT flags */
+   GLvoid *Pointer;     /**< User-space address of mapping */
+   GLintptr Offset;     /**< Mapped offset */
+   GLsizeiptr Length;   /**< Mapped length */
+};
+
+
 /**
  * GL_ARB_vertex/pixel_buffer_object buffer object
  */
 struct gl_buffer_object
 {
-   _glthread_Mutex Mutex;
+   mtx_t Mutex;
    GLint RefCount;
    GLuint Name;
    GLchar *Label;       /**< GL_KHR_debug */
    GLenum Usage;        /**< GL_STREAM_DRAW_ARB, GL_STREAM_READ_ARB, etc. */
+   GLbitfield StorageFlags; /**< GL_MAP_PERSISTENT_BIT, etc. */
    GLsizeiptrARB Size;  /**< Size of buffer storage in bytes */
    GLubyte *Data;       /**< Location of storage either in RAM or VRAM. */
-   /** Fields describing a mapped buffer */
-   /*@{*/
-   GLbitfield AccessFlags; /**< Mask of GL_MAP_x_BIT flags */
-   GLvoid *Pointer;     /**< User-space address of mapping */
-   GLintptr Offset;     /**< Mapped offset */
-   GLsizeiptr Length;   /**< Mapped length */
-   /*@}*/
    GLboolean DeletePending;   /**< true if buffer object is removed from the hash */
    GLboolean Written;   /**< Ever written to? (for debugging) */
    GLboolean Purgeable; /**< Is the buffer purgeable under memory pressure? */
+   GLboolean Immutable; /**< GL_ARB_buffer_storage */
+
+   struct gl_buffer_mapping Mappings[MAP_COUNT];
 };
 
 
@@ -1560,7 +1576,7 @@ struct gl_vertex_array_object
    GLchar *Label;       /**< GL_KHR_debug */
 
    GLint RefCount;
-   _glthread_Mutex Mutex;
+   mtx_t Mutex;
 
    /**
     * Does the VAO use ARB semantics or Apple semantics?
@@ -2013,6 +2029,7 @@ typedef enum
    SYSTEM_VALUE_SAMPLE_ID,      /**< Fragment shader only */
    SYSTEM_VALUE_SAMPLE_POS,     /**< Fragment shader only */
    SYSTEM_VALUE_SAMPLE_MASK_IN, /**< Fragment shader only */
+   SYSTEM_VALUE_INVOCATION_ID,  /**< Geometry shader only */
    SYSTEM_VALUE_MAX             /**< Number of values */
 } gl_system_value;
 
@@ -2143,6 +2160,7 @@ struct gl_geometry_program
 
    GLint VerticesIn;
    GLint VerticesOut;
+   GLint Invocations;
    GLenum InputType;  /**< GL_POINTS, GL_LINES, GL_LINES_ADJACENCY_ARB,
                            GL_TRIANGLES, or GL_TRIANGLES_ADJACENCY_ARB */
    GLenum OutputType; /**< GL_POINTS, GL_LINE_STRIP or GL_TRIANGLE_STRIP */
@@ -2419,6 +2437,11 @@ struct gl_shader
    struct {
       GLint VerticesOut;
       /**
+       * 0 - Invocations count not declared in shader, or
+       * 1 .. MAX_GEOMETRY_SHADER_INVOCATIONS
+       */
+      GLint Invocations;
+      /**
        * GL_POINTS, GL_LINES, GL_LINES_ADJACENCY, GL_TRIANGLES, or
        * GL_TRIANGLES_ADJACENCY, or PRIM_UNKNOWN if it's not set in this
        * shader.
@@ -2579,6 +2602,12 @@ struct gl_shader_program
     */
    GLboolean InternalSeparateShader;
 
+   /**
+    * Indicates whether program can be bound for individual pipeline stages
+    * using UseProgramStages after it is next linked.
+    */
+   GLboolean SeparateShader;
+
    GLuint NumShaders;          /**< number of attached shaders */
    struct gl_shader **Shaders; /**< List of attached the shaders */
 
@@ -2627,6 +2656,10 @@ struct gl_shader_program
    struct {
       GLint VerticesIn;
       GLint VerticesOut;
+      /**
+       * 1 .. MAX_GEOMETRY_SHADER_INVOCATIONS
+       */
+      GLint Invocations;
       GLenum InputType;  /**< GL_POINTS, GL_LINES, GL_LINES_ADJACENCY_ARB,
                               GL_TRIANGLES, or GL_TRIANGLES_ADJACENCY_ARB */
       GLenum OutputType; /**< GL_POINTS, GL_LINE_STRIP or GL_TRIANGLE_STRIP */
@@ -2746,9 +2779,19 @@ struct gl_shader_program
 
 /**
  * Context state for GLSL vertex/fragment shaders.
+ * Extended to support pipeline object
  */
-struct gl_shader_state
+struct gl_pipeline_object
 {
+   /** Name of the pipeline object as received from glGenProgramPipelines.
+    * It would be 0 for shaders without separate shader objects.
+    */
+   GLuint Name;
+
+   GLint RefCount;
+
+   mtx_t Mutex;
+
    /**
     * Programs used for rendering
     *
@@ -2768,8 +2811,21 @@ struct gl_shader_state
    struct gl_shader_program *ActiveProgram;
 
    GLbitfield Flags;                    /**< Mask of GLSL_x flags */
+
+   GLboolean EverBound;                 /**< Has the pipeline object been created */
 };
 
+/**
+ * Context state for GLSL pipeline shaders.
+ */
+struct gl_pipeline_shader_state
+{
+   /** Currently bound pipeline object. See _mesa_BindProgramPipeline() */
+   struct gl_pipeline_object *Current;
+
+   /** Pipeline objects */
+   struct _mesa_HashTable *Objects;
+};
 
 /**
  * Compiler options for a single GLSL shaders type
@@ -2871,7 +2927,7 @@ struct gl_sync_object
  */
 struct gl_shared_state
 {
-   _glthread_Mutex Mutex;		   /**< for thread safety */
+   mtx_t Mutex;		   /**< for thread safety */
    GLint RefCount;			   /**< Reference count */
    struct _mesa_HashTable *DisplayList;	   /**< Display lists hash table */
    struct _mesa_HashTable *TexObjects;	   /**< Texture objects hash table */
@@ -2889,7 +2945,7 @@ struct gl_shared_state
     * \todo Improve the granularity of locking.
     */
    /*@{*/
-   _glthread_Mutex TexMutex;		/**< texobj thread safety */
+   mtx_t TexMutex;		/**< texobj thread safety */
    GLuint TextureStateStamp;	        /**< state notification for shared tex */
    /*@}*/
 
@@ -2946,7 +3002,7 @@ struct gl_shared_state
  */
 struct gl_renderbuffer
 {
-   _glthread_Mutex Mutex; /**< for thread safety */
+   mtx_t Mutex; /**< for thread safety */
    GLuint ClassID;        /**< Useful for drivers */
    GLuint Name;
    GLchar *Label;         /**< GL_KHR_debug */
@@ -3024,7 +3080,7 @@ struct gl_renderbuffer_attachment
  */
 struct gl_framebuffer
 {
-   _glthread_Mutex Mutex;  /**< for thread safety */
+   mtx_t Mutex;  /**< for thread safety */
    /**
     * If zero, this is a window system framebuffer.  If non-zero, this
     * is a FBO framebuffer; note that for some devices (i.e. those with
@@ -3433,6 +3489,7 @@ struct gl_extensions
    GLboolean ARB_arrays_of_arrays;
    GLboolean ARB_base_instance;
    GLboolean ARB_blend_func_extended;
+   GLboolean ARB_buffer_storage;
    GLboolean ARB_color_buffer_float;
    GLboolean ARB_compute_shader;
    GLboolean ARB_conservative_depth;
@@ -3451,7 +3508,6 @@ struct gl_extensions
    GLboolean ARB_explicit_attrib_location;
    GLboolean ARB_geometry_shader4;
    GLboolean ARB_gpu_shader5;
-   GLboolean ARB_half_float_pixel;
    GLboolean ARB_half_float_vertex;
    GLboolean ARB_instanced_arrays;
    GLboolean ARB_internalformat_query;
@@ -3461,6 +3517,7 @@ struct gl_extensions
    GLboolean ARB_point_sprite;
    GLboolean ARB_sample_shading;
    GLboolean ARB_seamless_cube_map;
+   GLboolean ARB_separate_shader_objects;
    GLboolean ARB_shader_atomic_counters;
    GLboolean ARB_shader_bit_encoding;
    GLboolean ARB_shader_image_load_store;
@@ -4070,7 +4127,8 @@ struct gl_context
    struct gl_geometry_program_state GeometryProgram;
    struct gl_ati_fragment_shader_state ATIFragmentShader;
 
-   struct gl_shader_state Shader; /**< GLSL shader object state */
+   struct gl_pipeline_shader_state Pipeline; /**< GLSL pipeline shader object state */
+   struct gl_pipeline_object Shader; /**< GLSL shader object state */
    struct gl_shader_compiler_options ShaderCompilerOptions[MESA_SHADER_STAGES];
 
    struct gl_query_state Query;  /**< occlusion, timer queries */
@@ -4132,7 +4190,7 @@ struct gl_context
    GLuint ErrorDebugCount;
 
    /* GL_ARB_debug_output/GL_KHR_debug */
-   struct gl_debug_state Debug;
+   struct gl_debug_state *Debug;
 
    GLenum RenderMode;        /**< either GL_RENDER, GL_SELECT, GL_FEEDBACK */
    GLbitfield NewState;      /**< bitwise-or of _NEW_* flags */
diff --git a/mesalib/src/mesa/main/pbo.c b/mesalib/src/mesa/main/pbo.c
index 400cec3f0..a0d61a643 100644
--- a/mesalib/src/mesa/main/pbo.c
+++ b/mesalib/src/mesa/main/pbo.c
@@ -149,7 +149,8 @@ _mesa_map_pbo_source(struct gl_context *ctx,
       buf = (GLubyte *) ctx->Driver.MapBufferRange(ctx, 0,
 						   unpack->BufferObj->Size,
 						   GL_MAP_READ_BIT,
-						   unpack->BufferObj);
+						   unpack->BufferObj,
+                                                   MAP_INTERNAL);
       if (!buf)
          return NULL;
 
@@ -201,7 +202,7 @@ _mesa_map_validate_pbo_source(struct gl_context *ctx,
       return ptr;
    }
 
-   if (_mesa_bufferobj_mapped(unpack->BufferObj)) {
+   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;
@@ -221,7 +222,7 @@ _mesa_unmap_pbo_source(struct gl_context *ctx,
 {
    ASSERT(unpack != &ctx->Pack); /* catch pack/unpack mismatch */
    if (_mesa_is_bufferobj(unpack->BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj, MAP_INTERNAL);
    }
 }
 
@@ -246,7 +247,8 @@ _mesa_map_pbo_dest(struct gl_context *ctx,
       buf = (GLubyte *) ctx->Driver.MapBufferRange(ctx, 0,
 						   pack->BufferObj->Size,
 						   GL_MAP_WRITE_BIT,
-						   pack->BufferObj);
+						   pack->BufferObj,
+                                                   MAP_INTERNAL);
       if (!buf)
          return NULL;
 
@@ -297,7 +299,7 @@ _mesa_map_validate_pbo_dest(struct gl_context *ctx,
       return ptr;
    }
 
-   if (_mesa_bufferobj_mapped(unpack->BufferObj)) {
+   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;
@@ -317,7 +319,7 @@ _mesa_unmap_pbo_dest(struct gl_context *ctx,
 {
    ASSERT(pack != &ctx->Unpack); /* catch pack/unpack mismatch */
    if (_mesa_is_bufferobj(pack->BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, pack->BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, pack->BufferObj, MAP_INTERNAL);
    }
 }
 
@@ -351,7 +353,8 @@ _mesa_validate_pbo_teximage(struct gl_context *ctx, GLuint dimensions,
    buf = (GLubyte *) ctx->Driver.MapBufferRange(ctx, 0,
                                                 unpack->BufferObj->Size,
 						GL_MAP_READ_BIT,
-						unpack->BufferObj);
+						unpack->BufferObj,
+                                                MAP_INTERNAL);
    if (!buf) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(PBO is mapped)", funcName,
                   dimensions);
@@ -393,7 +396,8 @@ _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
    buf = (GLubyte*) ctx->Driver.MapBufferRange(ctx, 0,
 					       packing->BufferObj->Size,
 					       GL_MAP_READ_BIT,
-					       packing->BufferObj);
+					       packing->BufferObj,
+                                               MAP_INTERNAL);
    if (!buf) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(PBO is mapped)", funcName,
                   dimensions);
@@ -413,6 +417,6 @@ _mesa_unmap_teximage_pbo(struct gl_context *ctx,
                          const struct gl_pixelstore_attrib *unpack)
 {
    if (_mesa_is_bufferobj(unpack->BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj, MAP_INTERNAL);
    }
 }
diff --git a/mesalib/src/mesa/main/pipelineobj.c b/mesalib/src/mesa/main/pipelineobj.c
new file mode 100644
index 000000000..27012df72
--- /dev/null
+++ b/mesalib/src/mesa/main/pipelineobj.c
@@ -0,0 +1,445 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright © 2013 Gregory Hainaut <gregory.hainaut@gmail.com>
+ *
+ * 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.
+ */
+
+/**
+ * \file pipelineobj.c
+ * \author Hainaut Gregory <gregory.hainaut@gmail.com>
+ *
+ * Implementation of pipeline object related API functions. Based on
+ * GL_ARB_separate_shader_objects extension.
+ */
+
+#include "main/glheader.h"
+#include "main/context.h"
+#include "main/dispatch.h"
+#include "main/enums.h"
+#include "main/hash.h"
+#include "main/mtypes.h"
+#include "main/pipelineobj.h"
+#include "main/shaderapi.h"
+#include "main/shaderobj.h"
+#include "main/transformfeedback.h"
+#include "main/uniforms.h"
+#include "program/program.h"
+#include "program/prog_parameter.h"
+#include "ralloc.h"
+#include <stdbool.h>
+#include "../glsl/glsl_parser_extras.h"
+#include "../glsl/ir_uniform.h"
+
+/**
+ * Delete a pipeline object.
+ */
+void
+_mesa_delete_pipeline_object(struct gl_context *ctx,
+                             struct gl_pipeline_object *obj)
+{
+   unsigned i;
+
+   _mesa_reference_shader_program(ctx, &obj->_CurrentFragmentProgram, NULL);
+
+   for (i = 0; i < MESA_SHADER_STAGES; i++)
+      _mesa_reference_shader_program(ctx, &obj->CurrentProgram[i], NULL);
+
+   _mesa_reference_shader_program(ctx, &obj->ActiveProgram, NULL);
+   mtx_destroy(&obj->Mutex);
+   ralloc_free(obj);
+}
+
+/**
+ * Allocate and initialize a new pipeline object.
+ */
+static struct gl_pipeline_object *
+_mesa_new_pipeline_object(struct gl_context *ctx, GLuint name)
+{
+   struct gl_pipeline_object *obj = rzalloc(NULL, struct gl_pipeline_object);
+   if (obj) {
+      obj->Name = name;
+      mtx_init(&obj->Mutex, mtx_plain);
+      obj->RefCount = 1;
+      obj->Flags = _mesa_get_shader_flags();
+   }
+
+   return obj;
+}
+
+/**
+ * Initialize pipeline object state for given context.
+ */
+void
+_mesa_init_pipeline(struct gl_context *ctx)
+{
+   ctx->Pipeline.Objects = _mesa_NewHashTable();
+
+   ctx->Pipeline.Current = NULL;
+}
+
+
+/**
+ * Callback for deleting a pipeline object.  Called by _mesa_HashDeleteAll().
+ */
+static void
+delete_pipelineobj_cb(GLuint id, void *data, void *userData)
+{
+   struct gl_pipeline_object *obj = (struct gl_pipeline_object *) data;
+   struct gl_context *ctx = (struct gl_context *) userData;
+   _mesa_delete_pipeline_object(ctx, obj);
+}
+
+
+/**
+ * Free pipeline state for given context.
+ */
+void
+_mesa_free_pipeline_data(struct gl_context *ctx)
+{
+   _mesa_HashDeleteAll(ctx->Pipeline.Objects, delete_pipelineobj_cb, ctx);
+   _mesa_DeleteHashTable(ctx->Pipeline.Objects);
+}
+
+/**
+ * Look up the pipeline object for the given ID.
+ *
+ * \returns
+ * Either a pointer to the pipeline object with the specified ID or \c NULL for
+ * a non-existent ID.  The spec defines ID 0 as being technically
+ * non-existent.
+ */
+static inline struct gl_pipeline_object *
+lookup_pipeline_object(struct gl_context *ctx, GLuint id)
+{
+   if (id == 0)
+      return NULL;
+   else
+      return (struct gl_pipeline_object *)
+         _mesa_HashLookup(ctx->Pipeline.Objects, id);
+}
+
+/**
+ * Add the given pipeline object to the pipeline object pool.
+ */
+static void
+save_pipeline_object(struct gl_context *ctx, struct gl_pipeline_object *obj)
+{
+   if (obj->Name > 0) {
+      _mesa_HashInsert(ctx->Pipeline.Objects, obj->Name, obj);
+   }
+}
+
+/**
+ * Remove the given pipeline object from the pipeline object pool.
+ * Do not deallocate the pipeline object though.
+ */
+static void
+remove_pipeline_object(struct gl_context *ctx, struct gl_pipeline_object *obj)
+{
+   if (obj->Name > 0) {
+      _mesa_HashRemove(ctx->Pipeline.Objects, obj->Name);
+   }
+}
+
+/**
+ * Set ptr to obj w/ reference counting.
+ * Note: this should only be called from the _mesa_reference_pipeline_object()
+ * inline function.
+ */
+void
+_mesa_reference_pipeline_object_(struct gl_context *ctx,
+                                 struct gl_pipeline_object **ptr,
+                                 struct gl_pipeline_object *obj)
+{
+   assert(*ptr != obj);
+
+   if (*ptr) {
+      /* Unreference the old pipeline object */
+      GLboolean deleteFlag = GL_FALSE;
+      struct gl_pipeline_object *oldObj = *ptr;
+
+      mtx_lock(&oldObj->Mutex);
+      ASSERT(oldObj->RefCount > 0);
+      oldObj->RefCount--;
+      deleteFlag = (oldObj->RefCount == 0);
+      mtx_unlock(&oldObj->Mutex);
+
+      if (deleteFlag) {
+         _mesa_delete_pipeline_object(ctx, oldObj);
+      }
+
+      *ptr = NULL;
+   }
+   ASSERT(!*ptr);
+
+   if (obj) {
+      /* reference new pipeline object */
+      mtx_lock(&obj->Mutex);
+      if (obj->RefCount == 0) {
+         /* this pipeline's being deleted (look just above) */
+         /* Not sure this can ever really happen.  Warn if it does. */
+         _mesa_problem(NULL, "referencing deleted pipeline object");
+         *ptr = NULL;
+      }
+      else {
+         obj->RefCount++;
+         *ptr = obj;
+      }
+      mtx_unlock(&obj->Mutex);
+   }
+}
+
+/**
+ * Bound program to severals stages of the pipeline
+ */
+void GLAPIENTRY
+_mesa_UseProgramStages(GLuint pipeline, GLbitfield stages, GLuint program)
+{
+}
+
+/**
+ * Use the named shader program for subsequent glUniform calls (if pipeline
+ * bound)
+ */
+void GLAPIENTRY
+_mesa_ActiveShaderProgram(GLuint pipeline, GLuint program)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg = NULL;
+   struct gl_pipeline_object *pipe = lookup_pipeline_object(ctx, pipeline);
+
+   if (program != 0) {
+      shProg = _mesa_lookup_shader_program_err(ctx, program,
+                                               "glActiveShaderProgram(program)");
+      if (shProg == NULL)
+         return;
+   }
+
+   if (!pipe) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glActiveShaderProgram(pipeline)");
+      return;
+   }
+
+   /* Object is created by any Pipeline call but glGenProgramPipelines,
+    * glIsProgramPipeline and GetProgramPipelineInfoLog
+    */
+   pipe->EverBound = GL_TRUE;
+
+   if ((shProg != NULL) && !shProg->LinkStatus) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+            "glActiveShaderProgram(program %u not linked)", shProg->Name);
+      return;
+   }
+
+   _mesa_reference_shader_program(ctx, &pipe->ActiveProgram, shProg);
+}
+
+/**
+ * Make program of the pipeline current
+ */
+void GLAPIENTRY
+_mesa_BindProgramPipeline(GLuint pipeline)
+{
+}
+
+/**
+ * Delete a set of pipeline objects.
+ *
+ * \param n      Number of pipeline objects to delete.
+ * \param ids    pipeline of \c n pipeline object IDs.
+ */
+void GLAPIENTRY
+_mesa_DeleteProgramPipelines(GLsizei n, const GLuint *pipelines)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLsizei i;
+
+   if (n < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glDeleteProgramPipelines(n<0)");
+      return;
+   }
+
+   for (i = 0; i < n; i++) {
+      struct gl_pipeline_object *obj =
+         lookup_pipeline_object(ctx, pipelines[i]);
+
+      if (obj) {
+         ASSERT(obj->Name == pipelines[i]);
+
+         /* If the pipeline object is currently bound, the spec says "If an
+          * object that is currently bound is deleted, the binding for that
+          * object reverts to zero and no program pipeline object becomes
+          * current."
+          */
+         if (obj == ctx->Pipeline.Current) {
+            _mesa_BindProgramPipeline(0);
+         }
+
+         /* The ID is immediately freed for re-use */
+         remove_pipeline_object(ctx, obj);
+
+         /* Unreference the pipeline object.
+          * If refcount hits zero, the object will be deleted.
+          */
+         _mesa_reference_pipeline_object(ctx, &obj, NULL);
+      }
+   }
+}
+
+/**
+ * Generate a set of unique pipeline object IDs and store them in \c pipelines.
+ * \param n       Number of IDs to generate.
+ * \param pipelines  pipeline of \c n locations to store the IDs.
+ */
+void GLAPIENTRY
+_mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   GLuint first;
+   GLint i;
+
+   if (n < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glGenProgramPipelines(n<0)");
+      return;
+   }
+
+   if (!pipelines) {
+      return;
+   }
+
+   first = _mesa_HashFindFreeKeyBlock(ctx->Pipeline.Objects, n);
+
+   for (i = 0; i < n; i++) {
+      struct gl_pipeline_object *obj;
+      GLuint name = first + i;
+
+      obj = _mesa_new_pipeline_object(ctx, name);
+      if (!obj) {
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenProgramPipelines");
+         return;
+      }
+
+      save_pipeline_object(ctx, obj);
+      pipelines[i] = first + i;
+   }
+
+}
+
+/**
+ * Determine if ID is the name of an pipeline object.
+ *
+ * \param id  ID of the potential pipeline object.
+ * \return  \c GL_TRUE if \c id is the name of a pipeline object,
+ *          \c GL_FALSE otherwise.
+ */
+GLboolean GLAPIENTRY
+_mesa_IsProgramPipeline(GLuint pipeline)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   struct gl_pipeline_object *obj = lookup_pipeline_object(ctx, pipeline);
+   if (obj == NULL)
+      return GL_FALSE;
+
+   return obj->EverBound;
+}
+
+/**
+ * glGetProgramPipelineiv() - get pipeline shader state.
+ */
+void GLAPIENTRY
+_mesa_GetProgramPipelineiv(GLuint pipeline, GLenum pname, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_pipeline_object *pipe = lookup_pipeline_object(ctx, pipeline);
+
+   /* Are geometry shaders available in this context?
+    */
+   const bool has_gs = _mesa_has_geometry_shaders(ctx);
+
+   if (!pipe) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glGetProgramPipelineiv(pipeline)");
+      return;
+   }
+
+   /* Object is created by any Pipeline call but glGenProgramPipelines,
+    * glIsProgramPipeline and GetProgramPipelineInfoLog
+    */
+   pipe->EverBound = GL_TRUE;
+
+   switch (pname) {
+   case GL_ACTIVE_PROGRAM:
+      *params = pipe->ActiveProgram ? pipe->ActiveProgram->Name : 0;
+      return;
+   case GL_INFO_LOG_LENGTH:
+      /* FINISHME: Implement the info log.
+       */
+      *params = 0;
+      return;
+   case GL_VALIDATE_STATUS:
+      /* FINISHME: Implement validation status.
+       */
+      *params = 0;
+      return;
+   case GL_VERTEX_SHADER:
+      *params = pipe->CurrentProgram[MESA_SHADER_VERTEX]
+         ? pipe->CurrentProgram[MESA_SHADER_VERTEX]->Name : 0;
+      return;
+   case GL_TESS_EVALUATION_SHADER:
+      /* NOT YET SUPPORTED */
+      break;
+   case GL_TESS_CONTROL_SHADER:
+      /* NOT YET SUPPORTED */
+      break;
+   case GL_GEOMETRY_SHADER:
+      if (!has_gs)
+         break;
+      *params = pipe->CurrentProgram[MESA_SHADER_GEOMETRY]
+         ? pipe->CurrentProgram[MESA_SHADER_GEOMETRY]->Name : 0;
+      return;
+   case GL_FRAGMENT_SHADER:
+      *params = pipe->CurrentProgram[MESA_SHADER_FRAGMENT]
+         ? pipe->CurrentProgram[MESA_SHADER_FRAGMENT]->Name : 0;
+      return;
+   default:
+      break;
+   }
+
+   _mesa_error(ctx, GL_INVALID_ENUM, "glGetProgramPipelineiv(pname=%s)",
+               _mesa_lookup_enum_by_nr(pname));
+}
+
+/**
+ * Check compatibility of pipeline's program
+ */
+void GLAPIENTRY
+_mesa_ValidateProgramPipeline(GLuint pipeline)
+{
+}
+
+void GLAPIENTRY
+_mesa_GetProgramPipelineInfoLog(GLuint pipeline, GLsizei bufSize,
+                                GLsizei *length, GLchar *infoLog)
+{
+}
diff --git a/mesalib/src/mesa/main/pipelineobj.h b/mesalib/src/mesa/main/pipelineobj.h
new file mode 100644
index 000000000..46d5fab42
--- /dev/null
+++ b/mesalib/src/mesa/main/pipelineobj.h
@@ -0,0 +1,95 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright © 2013 Gregory Hainaut <gregory.hainaut@gmail.com>
+ *
+ * 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.
+ */
+
+#ifndef PIPELINEOBJ_H
+#define PIPELINEOBJ_H
+
+#include "glheader.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct _glapi_table;
+struct gl_context;
+struct gl_pipeline_object;
+
+extern void
+_mesa_delete_pipeline_object(struct gl_context *ctx, struct gl_pipeline_object *obj);
+
+extern void
+_mesa_init_pipeline(struct gl_context *ctx);
+
+extern void
+_mesa_free_pipeline_data(struct gl_context *ctx);
+
+extern void
+_mesa_reference_pipeline_object_(struct gl_context *ctx,
+                                 struct gl_pipeline_object **ptr,
+                                 struct gl_pipeline_object *obj);
+
+static inline void
+_mesa_reference_pipeline_object(struct gl_context *ctx,
+                                struct gl_pipeline_object **ptr,
+                                struct gl_pipeline_object *obj)
+{
+   if (*ptr != obj)
+      _mesa_reference_pipeline_object_(ctx, ptr, obj);
+}
+
+
+extern void GLAPIENTRY
+_mesa_UseProgramStages(GLuint pipeline, GLbitfield stages, GLuint program);
+
+extern void GLAPIENTRY
+_mesa_ActiveShaderProgram(GLuint pipeline, GLuint program);
+
+extern void GLAPIENTRY
+_mesa_BindProgramPipeline(GLuint pipeline);
+
+extern void GLAPIENTRY
+_mesa_DeleteProgramPipelines(GLsizei n, const GLuint *pipelines);
+
+extern void GLAPIENTRY
+_mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines);
+
+extern GLboolean GLAPIENTRY
+_mesa_IsProgramPipeline(GLuint pipeline);
+
+extern void GLAPIENTRY
+_mesa_GetProgramPipelineiv(GLuint pipeline, GLenum pname, GLint *params);
+
+extern void GLAPIENTRY
+_mesa_ValidateProgramPipeline(GLuint pipeline);
+
+extern void GLAPIENTRY
+_mesa_GetProgramPipelineInfoLog(GLuint pipeline, GLsizei bufSize,
+                                GLsizei *length, GLchar *infoLog);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* PIPELINEOBJ_H */
diff --git a/mesalib/src/mesa/main/readpix.c b/mesalib/src/mesa/main/readpix.c
index e16346896..b09cf5499 100644
--- a/mesalib/src/mesa/main/readpix.c
+++ b/mesalib/src/mesa/main/readpix.c
@@ -595,7 +595,7 @@ fast_read_depth_stencil_pixels(struct gl_context *ctx,
       return GL_FALSE;
 
    if (rb->Format != MESA_FORMAT_S8_UINT_Z24_UNORM &&
-       rb->Format != MESA_FORMAT_Z24_UNORM_X8_UINT)
+       rb->Format != MESA_FORMAT_Z24_UNORM_S8_UINT)
       return GL_FALSE;
 
    ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
@@ -1033,7 +1033,7 @@ _mesa_ReadnPixelsARB( GLint x, GLint y, GLsizei width, GLsizei height,
    }
 
    if (_mesa_is_bufferobj(ctx->Pack.BufferObj) &&
-       _mesa_bufferobj_mapped(ctx->Pack.BufferObj)) {
+       _mesa_check_disallowed_mapping(ctx->Pack.BufferObj)) {
       /* buffer is mapped - that's an error */
       _mesa_error(ctx, GL_INVALID_OPERATION, "glReadPixels(PBO is mapped)");
       return;
diff --git a/mesalib/src/mesa/main/renderbuffer.c b/mesalib/src/mesa/main/renderbuffer.c
index 2ff96e548..0bc7f2b96 100644
--- a/mesalib/src/mesa/main/renderbuffer.c
+++ b/mesalib/src/mesa/main/renderbuffer.c
@@ -38,7 +38,7 @@
 void
 _mesa_init_renderbuffer(struct gl_renderbuffer *rb, GLuint name)
 {
-   _glthread_INIT_MUTEX(rb->Mutex);
+   mtx_init(&rb->Mutex, mtx_plain);
 
    rb->ClassID = 0;
    rb->Name = name;
@@ -83,7 +83,7 @@ _mesa_new_renderbuffer(struct gl_context *ctx, GLuint name)
 void
 _mesa_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
-   _glthread_DESTROY_MUTEX(rb->Mutex);
+   mtx_destroy(&rb->Mutex);
    free(rb->Label);
    free(rb);
 }
@@ -153,12 +153,12 @@ _mesa_reference_renderbuffer_(struct gl_renderbuffer **ptr,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_renderbuffer *oldRb = *ptr;
 
-      _glthread_LOCK_MUTEX(oldRb->Mutex);
+      mtx_lock(&oldRb->Mutex);
       ASSERT(oldRb->RefCount > 0);
       oldRb->RefCount--;
       /*printf("RB DECR %p (%d) to %d\n", (void*) oldRb, oldRb->Name, oldRb->RefCount);*/
       deleteFlag = (oldRb->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(oldRb->Mutex);
+      mtx_unlock(&oldRb->Mutex);
 
       if (deleteFlag) {
          GET_CURRENT_CONTEXT(ctx);
@@ -171,10 +171,10 @@ _mesa_reference_renderbuffer_(struct gl_renderbuffer **ptr,
 
    if (rb) {
       /* reference new renderbuffer */
-      _glthread_LOCK_MUTEX(rb->Mutex);
+      mtx_lock(&rb->Mutex);
       rb->RefCount++;
       /*printf("RB INCR %p (%d) to %d\n", (void*) rb, rb->Name, rb->RefCount);*/
-      _glthread_UNLOCK_MUTEX(rb->Mutex);
+      mtx_unlock(&rb->Mutex);
       *ptr = rb;
    }
 }
diff --git a/mesalib/src/mesa/main/samplerobj.c b/mesalib/src/mesa/main/samplerobj.c
index 7285ef5ed..4900d5256 100644
--- a/mesalib/src/mesa/main/samplerobj.c
+++ b/mesalib/src/mesa/main/samplerobj.c
@@ -66,7 +66,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_sampler_object *oldSamp = *ptr;
 
-      /*_glthread_LOCK_MUTEX(oldSamp->Mutex);*/
+      /*mtx_lock(&oldSamp->Mutex);*/
       ASSERT(oldSamp->RefCount > 0);
       oldSamp->RefCount--;
 #if 0
@@ -74,7 +74,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
              (void *) oldSamp, oldSamp->Name, oldSamp->RefCount);
 #endif
       deleteFlag = (oldSamp->RefCount == 0);
-      /*_glthread_UNLOCK_MUTEX(oldSamp->Mutex);*/
+      /*mtx_unlock(&oldSamp->Mutex);*/
 
       if (deleteFlag) {
 	 ASSERT(ctx->Driver.DeleteSamplerObject);
@@ -87,7 +87,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
 
    if (samp) {
       /* reference new sampler */
-      /*_glthread_LOCK_MUTEX(samp->Mutex);*/
+      /*mtx_lock(&samp->Mutex);*/
       if (samp->RefCount == 0) {
          /* this sampler's being deleted (look just above) */
          /* Not sure this can every really happen.  Warn if it does. */
@@ -102,7 +102,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
 #endif
          *ptr = samp;
       }
-      /*_glthread_UNLOCK_MUTEX(samp->Mutex);*/
+      /*mtx_unlock(&samp->Mutex);*/
    }
 }
 
@@ -203,7 +203,7 @@ _mesa_DeleteSamplers(GLsizei count, const GLuint *samplers)
       return;
    }
 
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
 
    for (i = 0; i < count; i++) {
       if (samplers[i]) {
@@ -228,7 +228,7 @@ _mesa_DeleteSamplers(GLsizei count, const GLuint *samplers)
       }
    }
 
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 }
 
 
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index 65b6b16bb..5060cbb0b 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -65,8 +65,8 @@
 /**
  * Return mask of GLSL_x flags by examining the MESA_GLSL env var.
  */
-static GLbitfield
-get_shader_flags(void)
+GLbitfield
+_mesa_get_shader_flags(void)
 {
    GLbitfield flags = 0x0;
    const char *env = _mesa_getenv("MESA_GLSL");
@@ -120,7 +120,11 @@ _mesa_init_shader_state(struct gl_context *ctx)
    for (sh = 0; sh < MESA_SHADER_STAGES; ++sh)
       memcpy(&ctx->ShaderCompilerOptions[sh], &options, sizeof(options));
 
-   ctx->Shader.Flags = get_shader_flags();
+   ctx->Shader.Flags = _mesa_get_shader_flags();
+
+   /* Extended for ARB_separate_shader_objects */
+   ctx->Shader.RefCount = 1;
+   mtx_init(&ctx->Shader.Mutex, mtx_plain);
 }
 
 
@@ -138,6 +142,10 @@ _mesa_free_shader_state(struct gl_context *ctx)
    _mesa_reference_shader_program(ctx, &ctx->Shader._CurrentFragmentProgram,
 				  NULL);
    _mesa_reference_shader_program(ctx, &ctx->Shader.ActiveProgram, NULL);
+
+   /* Extended for ARB_separate_shader_objects */
+   assert(ctx->Shader.RefCount == 1);
+   mtx_destroy(&ctx->Shader.Mutex);
 }
 
 
@@ -383,30 +391,31 @@ detach_shader(struct gl_context *ctx, GLuint program, GLuint shader)
          _mesa_reference_shader(ctx, &shProg->Shaders[i], NULL);
 
          /* alloc new, smaller array */
-         newList =
-            malloc((n - 1) * sizeof(struct gl_shader *));
+         newList = malloc((n - 1) * sizeof(struct gl_shader *));
          if (!newList) {
             _mesa_error(ctx, GL_OUT_OF_MEMORY, "glDetachShader");
             return;
          }
+         /* Copy old list entries to new list, skipping removed entry at [i] */
          for (j = 0; j < i; j++) {
             newList[j] = shProg->Shaders[j];
          }
-         while (++i < n)
+         while (++i < n) {
             newList[j++] = shProg->Shaders[i];
-         free(shProg->Shaders);
+         }
 
+         /* Free old list and install new one */
+         free(shProg->Shaders);
          shProg->Shaders = newList;
          shProg->NumShaders = n - 1;
 
 #ifdef DEBUG
-         /* sanity check */
-         {
-            for (j = 0; j < shProg->NumShaders; j++) {
-               assert(shProg->Shaders[j]->Type == GL_VERTEX_SHADER ||
-                      shProg->Shaders[j]->Type == GL_FRAGMENT_SHADER);
-               assert(shProg->Shaders[j]->RefCount > 0);
-            }
+         /* sanity check - make sure the new list's entries are sensible */
+         for (j = 0; j < shProg->NumShaders; j++) {
+            assert(shProg->Shaders[j]->Type == GL_VERTEX_SHADER ||
+                   shProg->Shaders[j]->Type == GL_GEOMETRY_SHADER ||
+                   shProg->Shaders[j]->Type == GL_FRAGMENT_SHADER);
+            assert(shProg->Shaders[j]->RefCount > 0);
          }
 #endif
 
@@ -605,6 +614,12 @@ get_programiv(struct gl_context *ctx, GLuint program, GLenum pname, GLint *param
       if (check_gs_query(ctx, shProg))
          *params = shProg->Geom.VerticesOut;
       return;
+   case GL_GEOMETRY_SHADER_INVOCATIONS:
+      if (!has_core_gs || !ctx->Extensions.ARB_gpu_shader5)
+         break;
+      if (check_gs_query(ctx, shProg))
+         *params = shProg->Geom.Invocations;
+      return;
    case GL_GEOMETRY_INPUT_TYPE:
       if (!has_core_gs)
          break;
@@ -681,6 +696,12 @@ get_programiv(struct gl_context *ctx, GLuint program, GLenum pname, GLint *param
          params[i] = shProg->Comp.LocalSize[i];
       return;
    }
+   case GL_PROGRAM_SEPARABLE:
+      if (!ctx->Extensions.ARB_separate_shader_objects)
+         break;
+
+      *params = shProg->SeparateShader;
+      return;
    default:
       break;
    }
@@ -817,9 +838,10 @@ compile_shader(struct gl_context *ctx, GLuint shaderObj)
       sh->CompileStatus = GL_FALSE;
    } else {
       if (ctx->Shader.Flags & GLSL_DUMP) {
-         printf("GLSL source for %s shader %d:\n",
-                _mesa_shader_stage_to_string(sh->Stage), sh->Name);
-         printf("%s\n", sh->Source);
+         fprintf(stderr, "GLSL source for %s shader %d:\n",
+                 _mesa_shader_stage_to_string(sh->Stage), sh->Name);
+         fprintf(stderr, "%s\n", sh->Source);
+         fflush(stderr);
       }
 
       /* this call will set the shader->CompileStatus field to indicate if
@@ -833,16 +855,17 @@ compile_shader(struct gl_context *ctx, GLuint shaderObj)
 
       if (ctx->Shader.Flags & GLSL_DUMP) {
          if (sh->CompileStatus) {
-            printf("GLSL IR for shader %d:\n", sh->Name);
-            _mesa_print_ir(sh->ir, NULL);
-            printf("\n\n");
+            fprintf(stderr, "GLSL IR for shader %d:\n", sh->Name);
+            _mesa_print_ir(stderr, sh->ir, NULL);
+            fprintf(stderr, "\n\n");
          } else {
-            printf("GLSL shader %d failed to compile.\n", sh->Name);
+            fprintf(stderr, "GLSL shader %d failed to compile.\n", sh->Name);
          }
          if (sh->InfoLog && sh->InfoLog[0] != 0) {
-            printf("GLSL shader %d info log:\n", sh->Name);
-            printf("%s\n", sh->InfoLog);
+            fprintf(stderr, "GLSL shader %d info log:\n", sh->Name);
+            fprintf(stderr, "%s\n", sh->InfoLog);
          }
+         fflush(stderr);
       }
 
    }
@@ -1721,6 +1744,25 @@ _mesa_ProgramParameteri(GLuint program, GLenum pname, GLint value)
        */
       shProg->BinaryRetreivableHint = value;
       return;
+
+   case GL_PROGRAM_SEPARABLE:
+      if (!ctx->Extensions.ARB_separate_shader_objects)
+         break;
+
+      /* Spec imply that the behavior is the same as ARB_get_program_binary
+       * Chapter 7.3 Program Objects
+       */
+      if (value != GL_TRUE && value != GL_FALSE) {
+         _mesa_error(ctx, GL_INVALID_VALUE,
+                     "glProgramParameteri(pname=%s, value=%d): "
+                     "value must be 0 or 1.",
+                     _mesa_lookup_enum_by_nr(pname),
+                     value);
+         return;
+      }
+      shProg->SeparateShader = value;
+      return;
+
    default:
       break;
    }
@@ -1792,19 +1834,16 @@ _mesa_ActiveProgramEXT(GLuint program)
    return;
 }
 
-
-/**
- * For GL_EXT_separate_shader_objects
- */
-GLuint GLAPIENTRY
-_mesa_CreateShaderProgramEXT(GLenum type, const GLchar *string)
+static GLuint
+_mesa_create_shader_program(struct gl_context* ctx, GLboolean separate,
+                            GLenum type, GLsizei count, const GLchar* const *strings)
 {
-   GET_CURRENT_CONTEXT(ctx);
    const GLuint shader = create_shader(ctx, type);
    GLuint program = 0;
 
    if (shader) {
-      shader_source(ctx, shader, _mesa_strdup(string));
+      _mesa_ShaderSource(shader, count, strings, NULL);
+
       compile_shader(ctx, shader);
 
       program = create_shader_program(ctx);
@@ -1816,6 +1855,8 @@ _mesa_CreateShaderProgramEXT(GLenum type, const GLchar *string)
 	 shProg = _mesa_lookup_shader_program(ctx, program);
 	 sh = _mesa_lookup_shader(ctx, shader);
 
+	 shProg->SeparateShader = separate;
+
 	 get_shaderiv(ctx, shader, GL_COMPILE_STATUS, &compiled);
 	 if (compiled) {
 	    attach_shader(ctx, program, shader);
@@ -1858,6 +1899,7 @@ _mesa_copy_linked_program_data(gl_shader_stage type,
       struct gl_geometry_program *dst_gp = (struct gl_geometry_program *) dst;
       dst_gp->VerticesIn = src->Geom.VerticesIn;
       dst_gp->VerticesOut = src->Geom.VerticesOut;
+      dst_gp->Invocations = src->Geom.Invocations;
       dst_gp->InputType = src->Geom.InputType;
       dst_gp->OutputType = src->Geom.OutputType;
       dst->UsesClipDistanceOut = src->Geom.UsesClipDistance;
@@ -1875,3 +1917,30 @@ _mesa_copy_linked_program_data(gl_shader_stage type,
       break;
    }
 }
+
+
+/**
+ * For GL_EXT_separate_shader_objects
+ */
+GLuint GLAPIENTRY
+_mesa_CreateShaderProgramEXT(GLenum type, const GLchar *string)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   return _mesa_create_shader_program(ctx, GL_FALSE, type, 1, &string);
+}
+
+/**
+ * ARB_separate_shader_objects: Compile & Link Program
+ *
+ * Basically the same as _mesa_CreateShaderProgramEXT but with support of
+ * multiple strings and sets the SeparateShader flag to true.
+ */
+GLuint GLAPIENTRY
+_mesa_CreateShaderProgramv(GLenum type, GLsizei count,
+                           const GLchar* const *strings)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   return _mesa_create_shader_program(ctx, GL_TRUE, type, count, strings);
+}
diff --git a/mesalib/src/mesa/main/shaderapi.h b/mesalib/src/mesa/main/shaderapi.h
index 10f810caf..44426cc65 100644
--- a/mesalib/src/mesa/main/shaderapi.h
+++ b/mesalib/src/mesa/main/shaderapi.h
@@ -40,6 +40,9 @@ struct _glapi_table;
 struct gl_context;
 struct gl_shader_program;
 
+extern GLbitfield
+_mesa_get_shader_flags(void);
+
 extern void
 _mesa_copy_string(GLchar *dst, GLsizei maxLength,
                   GLsizei *length, const GLchar *src);
@@ -219,6 +222,11 @@ extern bool
 _mesa_validate_shader_target(const struct gl_context *ctx, GLenum type);
 
 
+/* GL_ARB_separate_shader_objects */
+extern GLuint GLAPIENTRY
+_mesa_CreateShaderProgramv(GLenum type, GLsizei count,
+                           const GLchar* const *strings);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/mesalib/src/mesa/main/shaderobj.c b/mesalib/src/mesa/main/shaderobj.c
index 4f4bb69a8..d5c3d8099 100644
--- a/mesalib/src/mesa/main/shaderobj.c
+++ b/mesalib/src/mesa/main/shaderobj.c
@@ -355,6 +355,7 @@ _mesa_free_shader_program_data(struct gl_context *ctx,
    }
 
    free(shProg->Label);
+   shProg->Label = NULL;
 }
 
 
diff --git a/mesalib/src/mesa/main/shared.c b/mesalib/src/mesa/main/shared.c
index c11c7f9e9..dc22025c2 100644
--- a/mesalib/src/mesa/main/shared.c
+++ b/mesalib/src/mesa/main/shared.c
@@ -62,7 +62,7 @@ _mesa_alloc_shared_state(struct gl_context *ctx)
    if (!shared)
       return NULL;
 
-   _glthread_INIT_MUTEX(shared->Mutex);
+   mtx_init(&shared->Mutex, mtx_plain);
 
    shared->DisplayList = _mesa_NewHashTable();
    shared->TexObjects = _mesa_NewHashTable();
@@ -113,7 +113,7 @@ _mesa_alloc_shared_state(struct gl_context *ctx)
    assert(shared->DefaultTex[TEXTURE_1D_INDEX]->RefCount == 1);
 
    /* Mutex and timestamp for texobj state validation */
-   _glthread_INIT_MUTEX(shared->TexMutex);
+   mtx_init(&shared->TexMutex, mtx_plain);
    shared->TextureStateStamp = 0;
 
    shared->FrameBuffers = _mesa_NewHashTable();
@@ -186,10 +186,8 @@ delete_bufferobj_cb(GLuint id, void *data, void *userData)
 {
    struct gl_buffer_object *bufObj = (struct gl_buffer_object *) data;
    struct gl_context *ctx = (struct gl_context *) userData;
-   if (_mesa_bufferobj_mapped(bufObj)) {
-      ctx->Driver.UnmapBuffer(ctx, bufObj);
-      bufObj->Pointer = NULL;
-   }
+
+   _mesa_buffer_unmap_all_mappings(ctx, bufObj);
    _mesa_reference_buffer_object(ctx, &bufObj, NULL);
 }
 
@@ -356,8 +354,8 @@ free_shared_state(struct gl_context *ctx, struct gl_shared_state *shared)
    _mesa_HashDeleteAll(shared->TexObjects, delete_texture_cb, ctx);
    _mesa_DeleteHashTable(shared->TexObjects);
 
-   _glthread_DESTROY_MUTEX(shared->Mutex);
-   _glthread_DESTROY_MUTEX(shared->TexMutex);
+   mtx_destroy(&shared->Mutex);
+   mtx_destroy(&shared->TexMutex);
 
    free(shared);
 }
@@ -380,11 +378,11 @@ _mesa_reference_shared_state(struct gl_context *ctx,
       struct gl_shared_state *old = *ptr;
       GLboolean delete;
 
-      _glthread_LOCK_MUTEX(old->Mutex);
+      mtx_lock(&old->Mutex);
       assert(old->RefCount >= 1);
       old->RefCount--;
       delete = (old->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(old->Mutex);
+      mtx_unlock(&old->Mutex);
 
       if (delete) {
          free_shared_state(ctx, old);
@@ -395,9 +393,9 @@ _mesa_reference_shared_state(struct gl_context *ctx,
 
    if (state) {
       /* reference new state */
-      _glthread_LOCK_MUTEX(state->Mutex);
+      mtx_lock(&state->Mutex);
       state->RefCount++;
       *ptr = state;
-      _glthread_UNLOCK_MUTEX(state->Mutex);
+      mtx_unlock(&state->Mutex);
    }
 }
diff --git a/mesalib/src/mesa/main/syncobj.c b/mesalib/src/mesa/main/syncobj.c
index ae49db2e4..9721b14b4 100644
--- a/mesalib/src/mesa/main/syncobj.c
+++ b/mesalib/src/mesa/main/syncobj.c
@@ -184,9 +184,9 @@ _mesa_validate_sync(struct gl_context *ctx,
 void GLAPIENTRY
 _mesa_ref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj)
 {
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
    syncObj->RefCount++;
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 }
 
 
@@ -195,7 +195,7 @@ _mesa_unref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj)
 {
    struct set_entry *entry;
 
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
    syncObj->RefCount--;
    if (syncObj->RefCount == 0) {
       entry = _mesa_set_search(ctx->Shared->SyncObjects,
@@ -203,11 +203,11 @@ _mesa_unref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj)
                                syncObj);
       assert (entry != NULL);
       _mesa_set_remove(ctx->Shared->SyncObjects, entry);
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
 
       ctx->Driver.DeleteSyncObject(ctx, syncObj);
    } else {
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
    }
 }
 
@@ -288,11 +288,11 @@ _mesa_FenceSync(GLenum condition, GLbitfield flags)
 
       ctx->Driver.FenceSync(ctx, syncObj, condition, flags);
 
-      _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_lock(&ctx->Shared->Mutex);
       _mesa_set_add(ctx->Shared->SyncObjects,
                     _mesa_hash_pointer(syncObj),
                     syncObj);
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
 
       return (GLsync) syncObj;
    }
diff --git a/mesalib/src/mesa/main/texcompress_etc.c b/mesalib/src/mesa/main/texcompress_etc.c
index e3862be45..cbda68940 100644
--- a/mesalib/src/mesa/main/texcompress_etc.c
+++ b/mesalib/src/mesa/main/texcompress_etc.c
@@ -429,8 +429,7 @@ etc2_rgb8_parse_block(struct etc2_block *block,
       block->is_planar_mode = true;
 
       /* opaque bit must be set in planar mode */
-      if (!block->opaque)
-         block->opaque = true;
+      block->opaque = true;
 
       for (i = 0; i < 3; i++) {
          block->base_colors[0][i] = etc2_base_color_o_planar(src, i);
diff --git a/mesalib/src/mesa/main/texformat.c b/mesalib/src/mesa/main/texformat.c
index bda2d8e17..004e7ebac 100644
--- a/mesalib/src/mesa/main/texformat.c
+++ b/mesalib/src/mesa/main/texformat.c
@@ -209,13 +209,13 @@ _mesa_choose_tex_format(struct gl_context *ctx, GLenum target,
    case GL_DEPTH_COMPONENT24:
    case GL_DEPTH_COMPONENT32:
       RETURN_IF_SUPPORTED(MESA_FORMAT_Z_UNORM32);
-      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_S8_UINT);
       RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_X8_UINT);
+      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_S8_UINT);
       break;
    case GL_DEPTH_COMPONENT16:
       RETURN_IF_SUPPORTED(MESA_FORMAT_Z_UNORM16);
-      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_S8_UINT);
       RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_X8_UINT);
+      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_S8_UINT);
       break;
 
    case GL_COMPRESSED_ALPHA_ARB:
@@ -426,7 +426,7 @@ _mesa_choose_tex_format(struct gl_context *ctx, GLenum target,
    case GL_DEPTH_STENCIL_EXT:
    case GL_DEPTH24_STENCIL8_EXT:
       RETURN_IF_SUPPORTED(MESA_FORMAT_S8_UINT_Z24_UNORM);
-      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_X8_UINT);
+      RETURN_IF_SUPPORTED(MESA_FORMAT_Z24_UNORM_S8_UINT);
       break;
 
    case GL_DEPTH_COMPONENT32F:
diff --git a/mesalib/src/mesa/main/texgetimage.c b/mesalib/src/mesa/main/texgetimage.c
index 133fa5376..daabf2e81 100644
--- a/mesalib/src/mesa/main/texgetimage.c
+++ b/mesalib/src/mesa/main/texgetimage.c
@@ -130,6 +130,10 @@ get_tex_depth_stencil(struct gl_context *ctx, GLuint dimensions,
    const GLint depth = texImage->Depth;
    GLint img, row;
 
+   assert(format == GL_DEPTH_STENCIL);
+   assert(type == GL_UNSIGNED_INT_24_8);
+   /* XXX type == GL_FLOAT_32_UNSIGNED_INT_24_8_REV is not handled yet */
+
    for (img = 0; img < depth; img++) {
       GLubyte *srcMap;
       GLint rowstride;
@@ -145,8 +149,11 @@ get_tex_depth_stencil(struct gl_context *ctx, GLuint dimensions,
             void *dest = _mesa_image_address(dimensions, &ctx->Pack, pixels,
                                              width, height, format, type,
                                              img, row, 0);
-            /* XXX Z24_S8 vs. S8_Z24??? */
-            memcpy(dest, src, width * sizeof(GLuint));
+            /* Unpack from texture's format to GL's z24_s8 layout */
+            _mesa_unpack_uint_24_8_depth_stencil_row(texImage->TexFormat,
+                                                     width,
+                                                     (const GLuint *) src,
+                                                     dest);
             if (ctx->Pack.SwapBytes) {
                _mesa_swap4((GLuint *) dest, width);
             }
@@ -635,7 +642,8 @@ _mesa_get_teximage(struct gl_context *ctx,
        */
       GLubyte *buf = (GLubyte *)
          ctx->Driver.MapBufferRange(ctx, 0, ctx->Pack.BufferObj->Size,
-				    GL_MAP_WRITE_BIT, ctx->Pack.BufferObj);
+				    GL_MAP_WRITE_BIT, ctx->Pack.BufferObj,
+                                    MAP_INTERNAL);
       if (!buf) {
          /* out of memory or other unexpected error */
          _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGetTexImage(map PBO failed)");
@@ -664,7 +672,7 @@ _mesa_get_teximage(struct gl_context *ctx,
    }
 
    if (_mesa_is_bufferobj(ctx->Pack.BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, ctx->Pack.BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, ctx->Pack.BufferObj, MAP_INTERNAL);
    }
 }
 
@@ -689,7 +697,8 @@ _mesa_get_compressed_teximage(struct gl_context *ctx,
       /* pack texture image into a PBO */
       GLubyte *buf = (GLubyte *)
          ctx->Driver.MapBufferRange(ctx, 0, ctx->Pack.BufferObj->Size,
-				    GL_MAP_WRITE_BIT, ctx->Pack.BufferObj);
+				    GL_MAP_WRITE_BIT, ctx->Pack.BufferObj,
+                                    MAP_INTERNAL);
       if (!buf) {
          /* out of memory or other unexpected error */
          _mesa_error(ctx, GL_OUT_OF_MEMORY,
@@ -731,7 +740,7 @@ _mesa_get_compressed_teximage(struct gl_context *ctx,
    }
 
    if (_mesa_is_bufferobj(ctx->Pack.BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, ctx->Pack.BufferObj);
+      ctx->Driver.UnmapBuffer(ctx, ctx->Pack.BufferObj, MAP_INTERNAL);
    }
 }
 
@@ -861,7 +870,7 @@ getteximage_error_check(struct gl_context *ctx, GLenum target, GLint level,
 
    if (_mesa_is_bufferobj(ctx->Pack.BufferObj)) {
       /* PBO should not be mapped */
-      if (_mesa_bufferobj_mapped(ctx->Pack.BufferObj)) {
+      if (_mesa_check_disallowed_mapping(ctx->Pack.BufferObj)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
                      "glGetTexImage(PBO is mapped)");
          return GL_TRUE;
@@ -1004,7 +1013,7 @@ getcompressedteximage_error_check(struct gl_context *ctx, GLenum target,
       }
 
       /* make sure PBO is not mapped */
-      if (_mesa_bufferobj_mapped(ctx->Pack.BufferObj)) {
+      if (_mesa_check_disallowed_mapping(ctx->Pack.BufferObj)) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
                      "glGetCompressedTexImage(PBO is mapped)");
          return GL_TRUE;
diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c
index 4d635fe7f..0519d221f 100644
--- a/mesalib/src/mesa/main/teximage.c
+++ b/mesalib/src/mesa/main/teximage.c
@@ -397,11 +397,6 @@ _mesa_base_tex_format( struct gl_context *ctx, GLint internalFormat )
    if (ctx->Extensions.ARB_texture_rg) {
       switch (internalFormat) {
       case GL_R16F:
-	 /* R16F depends on both ARB_half_float_pixel and ARB_texture_float.
-	  */
-	 if (!ctx->Extensions.ARB_half_float_pixel)
-	    break;
-	 /* FALLTHROUGH */
       case GL_R32F:
 	 if (!ctx->Extensions.ARB_texture_float)
 	    break;
@@ -422,11 +417,6 @@ _mesa_base_tex_format( struct gl_context *ctx, GLint internalFormat )
          return GL_RED;
 
       case GL_RG16F:
-	 /* RG16F depends on both ARB_half_float_pixel and ARB_texture_float.
-	  */
-	 if (!ctx->Extensions.ARB_half_float_pixel)
-	    break;
-	 /* FALLTHROUGH */
       case GL_RG32F:
 	 if (!ctx->Extensions.ARB_texture_float)
 	    break;
@@ -2254,9 +2244,10 @@ compressed_texture_error_check(struct gl_context *ctx, GLint dimensions,
 
    /* This will detect any invalid internalFormat value */
    if (!_mesa_is_compressed_format(ctx, internalFormat)) {
-      reason = "internalFormat";
-      error = GL_INVALID_ENUM;
-      goto error;
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCompressedTexImage%dD(internalFormat=%s)",
+                  dimensions, _mesa_lookup_enum_by_nr(internalFormat));
+      return GL_TRUE;
    }
 
    switch (internalFormat) {
@@ -2348,6 +2339,7 @@ compressed_texture_error_check(struct gl_context *ctx, GLint dimensions,
    return GL_FALSE;
 
 error:
+   /* Note: not all error paths exit through here. */
    _mesa_error(ctx, error, "glCompressedTexImage%dD(%s)", dimensions, reason);
    return GL_TRUE;
 }
@@ -2563,7 +2555,8 @@ copytexture_error_check( struct gl_context *ctx, GLuint dimensions,
          break;
       default:
          _mesa_error(ctx, GL_INVALID_VALUE,
-                     "glCopyTexImage%dD(internalFormat)", dimensions);
+                     "glCopyTexImage%dD(internalFormat=%s)", dimensions,
+                     _mesa_lookup_enum_by_nr(internalFormat));
          return GL_TRUE;
       }
    }
@@ -2571,7 +2564,8 @@ copytexture_error_check( struct gl_context *ctx, GLuint dimensions,
    baseFormat = _mesa_base_tex_format(ctx, internalFormat);
    if (baseFormat < 0) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCopyTexImage%dD(internalFormat)", dimensions);
+                  "glCopyTexImage%dD(internalFormat=%s)", dimensions,
+                  _mesa_lookup_enum_by_nr(internalFormat));
       return GL_TRUE;
    }
 
@@ -2580,7 +2574,8 @@ copytexture_error_check( struct gl_context *ctx, GLuint dimensions,
    if (_mesa_is_color_format(internalFormat)) {
       if (rb_base_format < 0) {
          _mesa_error(ctx, GL_INVALID_VALUE,
-                     "glCopyTexImage%dD(internalFormat)", dimensions);
+                     "glCopyTexImage%dD(internalFormat=%s)", dimensions,
+                     _mesa_lookup_enum_by_nr(internalFormat));
          return GL_TRUE;
       }
    }
@@ -2606,7 +2601,8 @@ copytexture_error_check( struct gl_context *ctx, GLuint dimensions,
       }
       if (!valid) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
-                     "glCopyTexImage%dD(internalFormat)", dimensions);
+                     "glCopyTexImage%dD(internalFormat=%s)", dimensions,
+                     _mesa_lookup_enum_by_nr(internalFormat));
          return GL_TRUE;
       }
    }
@@ -4183,10 +4179,18 @@ _mesa_validate_texbuffer_format(const struct gl_context *ctx,
       return MESA_FORMAT_NONE;
 
    datatype = _mesa_get_format_datatype(format);
-   if (datatype == GL_FLOAT && !ctx->Extensions.ARB_texture_float)
-      return MESA_FORMAT_NONE;
 
-   if (datatype == GL_HALF_FLOAT && !ctx->Extensions.ARB_half_float_pixel)
+   /* The GL_ARB_texture_buffer_object spec says:
+    *
+    *     "If ARB_texture_float is not supported, references to the
+    *     floating-point internal formats provided by that extension should be
+    *     removed, and such formats may not be passed to TexBufferARB."
+    *
+    * As a result, GL_HALF_FLOAT internal format depends on both
+    * GL_ARB_texture_float and GL_ARB_half_float_pixel.
+    */
+   if ((datatype == GL_FLOAT || datatype == GL_HALF_FLOAT) &&
+       !ctx->Extensions.ARB_texture_float)
       return MESA_FORMAT_NONE;
 
    if (!ctx->Extensions.ARB_texture_rg) {
diff --git a/mesalib/src/mesa/main/teximage.h b/mesalib/src/mesa/main/teximage.h
index 5f8a47776..51d94d17e 100644
--- a/mesalib/src/mesa/main/teximage.h
+++ b/mesalib/src/mesa/main/teximage.h
@@ -163,7 +163,7 @@ _mesa_legal_texture_base_format_for_target(struct gl_context *ctx,
 static inline void
 _mesa_lock_texture(struct gl_context *ctx, struct gl_texture_object *texObj)
 {
-   _glthread_LOCK_MUTEX(ctx->Shared->TexMutex);
+   mtx_lock(&ctx->Shared->TexMutex);
    ctx->Shared->TextureStateStamp++;
    (void) texObj;
 }
@@ -172,7 +172,7 @@ static inline void
 _mesa_unlock_texture(struct gl_context *ctx, struct gl_texture_object *texObj)
 {
    (void) texObj;
-   _glthread_UNLOCK_MUTEX(ctx->Shared->TexMutex);
+   mtx_unlock(&ctx->Shared->TexMutex);
 }
 
 /*@}*/
diff --git a/mesalib/src/mesa/main/texobj.c b/mesalib/src/mesa/main/texobj.c
index 6adc0ae84..3375fe36b 100644
--- a/mesalib/src/mesa/main/texobj.c
+++ b/mesalib/src/mesa/main/texobj.c
@@ -115,7 +115,7 @@ _mesa_initialize_texture_object( struct gl_context *ctx,
 
    memset(obj, 0, sizeof(*obj));
    /* init the non-zero fields */
-   _glthread_INIT_MUTEX(obj->Mutex);
+   mtx_init(&obj->Mutex, mtx_plain);
    obj->RefCount = 1;
    obj->Name = name;
    obj->Target = target;
@@ -237,7 +237,7 @@ _mesa_delete_texture_object(struct gl_context *ctx,
    _mesa_reference_buffer_object(ctx, &texObj->BufferObject, NULL);
 
    /* destroy the mutex -- it may have allocated memory (eg on bsd) */
-   _glthread_DESTROY_MUTEX(texObj->Mutex);
+   mtx_destroy(&texObj->Mutex);
 
    free(texObj->Label);
 
@@ -374,12 +374,12 @@ _mesa_reference_texobj_(struct gl_texture_object **ptr,
       ASSERT(valid_texture_object(oldTex));
       (void) valid_texture_object; /* silence warning in release builds */
 
-      _glthread_LOCK_MUTEX(oldTex->Mutex);
+      mtx_lock(&oldTex->Mutex);
       ASSERT(oldTex->RefCount > 0);
       oldTex->RefCount--;
 
       deleteFlag = (oldTex->RefCount == 0);
-      _glthread_UNLOCK_MUTEX(oldTex->Mutex);
+      mtx_unlock(&oldTex->Mutex);
 
       if (deleteFlag) {
          GET_CURRENT_CONTEXT(ctx);
@@ -396,7 +396,7 @@ _mesa_reference_texobj_(struct gl_texture_object **ptr,
    if (tex) {
       /* reference new texture */
       ASSERT(valid_texture_object(tex));
-      _glthread_LOCK_MUTEX(tex->Mutex);
+      mtx_lock(&tex->Mutex);
       if (tex->RefCount == 0) {
          /* this texture's being deleted (look just above) */
          /* Not sure this can every really happen.  Warn if it does. */
@@ -407,7 +407,7 @@ _mesa_reference_texobj_(struct gl_texture_object **ptr,
          tex->RefCount++;
          *ptr = tex;
       }
-      _glthread_UNLOCK_MUTEX(tex->Mutex);
+      mtx_unlock(&tex->Mutex);
    }
 }
 
@@ -1009,7 +1009,7 @@ _mesa_GenTextures( GLsizei n, GLuint *textures )
    /*
     * This must be atomic (generation and allocation of texture IDs)
     */
-   _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_lock(&ctx->Shared->Mutex);
 
    first = _mesa_HashFindFreeKeyBlock(ctx->Shared->TexObjects, n);
 
@@ -1020,7 +1020,7 @@ _mesa_GenTextures( GLsizei n, GLuint *textures )
       GLenum target = 0;
       texObj = ctx->Driver.NewTextureObject(ctx, name, target);
       if (!texObj) {
-         _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+         mtx_unlock(&ctx->Shared->Mutex);
          _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenTextures");
          return;
       }
@@ -1031,7 +1031,7 @@ _mesa_GenTextures( GLsizei n, GLuint *textures )
       textures[i] = name;
    }
 
-   _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+   mtx_unlock(&ctx->Shared->Mutex);
 }
 
 
@@ -1178,9 +1178,9 @@ _mesa_DeleteTextures( GLsizei n, const GLuint *textures)
             /* The texture _name_ is now free for re-use.
              * Remove it from the hash table now.
              */
-            _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+            mtx_lock(&ctx->Shared->Mutex);
             _mesa_HashRemove(ctx->Shared->TexObjects, delObj->Name);
-            _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+            mtx_unlock(&ctx->Shared->Mutex);
 
             /* Unreference the texobj.  If refcount hits zero, the texture
              * will be deleted.
@@ -1313,9 +1313,9 @@ _mesa_BindTexture( GLenum target, GLuint texName )
          }
 
          /* and insert it into hash table */
-         _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+         mtx_lock(&ctx->Shared->Mutex);
          _mesa_HashInsert(ctx->Shared->TexObjects, texName, newTexObj);
-         _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+         mtx_unlock(&ctx->Shared->Mutex);
       }
       newTexObj->Target = target;
    }
@@ -1327,10 +1327,10 @@ _mesa_BindTexture( GLenum target, GLuint texName )
     */
    {
       GLboolean early_out;
-      _glthread_LOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_lock(&ctx->Shared->Mutex);
       early_out = ((ctx->Shared->RefCount == 1)
                    && (newTexObj == texUnit->CurrentTex[targetIndex]));
-      _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex);
+      mtx_unlock(&ctx->Shared->Mutex);
       if (early_out) {
          return;
       }
@@ -1493,7 +1493,7 @@ _mesa_IsTexture( GLuint texture )
 void
 _mesa_lock_context_textures( struct gl_context *ctx )
 {
-   _glthread_LOCK_MUTEX(ctx->Shared->TexMutex);
+   mtx_lock(&ctx->Shared->TexMutex);
 
    if (ctx->Shared->TextureStateStamp != ctx->TextureStateTimestamp) {
       ctx->NewState |= _NEW_TEXTURE;
@@ -1506,7 +1506,7 @@ void
 _mesa_unlock_context_textures( struct gl_context *ctx )
 {
    assert(ctx->Shared->TextureStateStamp == ctx->TextureStateTimestamp);
-   _glthread_UNLOCK_MUTEX(ctx->Shared->TexMutex);
+   mtx_unlock(&ctx->Shared->TexMutex);
 }
 
 void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/texparam.c b/mesalib/src/mesa/main/texparam.c
index b7ed50dd5..bbdbc2763 100644
--- a/mesalib/src/mesa/main/texparam.c
+++ b/mesalib/src/mesa/main/texparam.c
@@ -986,6 +986,9 @@ legal_get_tex_level_parameter_target(struct gl_context *ctx, GLenum target)
    case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_ARB:
    case GL_PROXY_TEXTURE_CUBE_MAP_ARB:
       return ctx->Extensions.ARB_texture_cube_map;
+   case GL_TEXTURE_CUBE_MAP_ARRAY_ARB:
+   case GL_PROXY_TEXTURE_CUBE_MAP_ARRAY_ARB:
+      return ctx->Extensions.ARB_texture_cube_map_array;
    case GL_TEXTURE_RECTANGLE_NV:
    case GL_PROXY_TEXTURE_RECTANGLE_NV:
       return ctx->Extensions.NV_texture_rectangle;
diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c
index 1e4308a04..3e22a0b46 100644
--- a/mesalib/src/mesa/main/texstore.c
+++ b/mesalib/src/mesa/main/texstore.c
@@ -1049,7 +1049,7 @@ _mesa_texstore_x8_z24(TEXSTORE_PARAMS)
    const GLuint depthScale = 0xffffff;
 
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_S8_UINT);
+   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_X8_UINT);
 
    {
       /* general path */
@@ -2558,7 +2558,7 @@ _mesa_texstore_s8_z24(TEXSTORE_PARAMS)
    GLuint *depth;
    GLubyte *stencil;
 
-   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_X8_UINT);
+   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_S8_UINT);
    ASSERT(srcFormat == GL_DEPTH_STENCIL_EXT ||
           srcFormat == GL_DEPTH_COMPONENT ||
           srcFormat == GL_STENCIL_INDEX);
@@ -3699,9 +3699,9 @@ _mesa_get_texstore_func(mesa_format format)
       table[MESA_FORMAT_G16R16_UNORM] = _mesa_texstore_unorm1616;
       table[MESA_FORMAT_B10G10R10A2_UNORM] = _mesa_texstore_argb2101010;
       table[MESA_FORMAT_S8_UINT_Z24_UNORM] = _mesa_texstore_z24_s8;
-      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = _mesa_texstore_s8_z24;
+      table[MESA_FORMAT_Z24_UNORM_S8_UINT] = _mesa_texstore_s8_z24;
       table[MESA_FORMAT_Z_UNORM16] = _mesa_texstore_z16;
-      table[MESA_FORMAT_Z24_UNORM_S8_UINT] = _mesa_texstore_x8_z24;
+      table[MESA_FORMAT_Z24_UNORM_X8_UINT] = _mesa_texstore_x8_z24;
       table[MESA_FORMAT_X8Z24_UNORM] = _mesa_texstore_z24_x8;
       table[MESA_FORMAT_Z_UNORM32] = _mesa_texstore_z32;
       table[MESA_FORMAT_S_UINT8] = _mesa_texstore_s8;
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index 82d7628e8..8cc5da752 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -684,6 +684,7 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
       match = true;
       break;
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
       match = (basicType == GLSL_TYPE_INT);
       break;
    default:
@@ -735,6 +736,22 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
       }
    }
 
+   if (uni->type->is_image()) {
+      int i;
+
+      for (i = 0; i < count; i++) {
+         const int unit = ((GLint *) values)[i];
+
+         /* check that the image unit is legal */
+         if (unit < 0 || unit >= (int)ctx->Const.MaxImageUnits) {
+            _mesa_error(ctx, GL_INVALID_VALUE,
+                        "glUniform1i(invalid image unit index for uniform %d)",
+                        location);
+            return;
+         }
+      }
+   }
+
    /* Page 82 (page 96 of the PDF) of the OpenGL 2.1 spec says:
     *
     *     "When loading N elements starting at an arbitrary position k in a
@@ -830,6 +847,25 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
 	 }
       }
    }
+
+   /* If the uniform is an image, update the mapping from image
+    * uniforms to image units present in the shader data structure.
+    */
+   if (uni->type->is_image()) {
+      int i, j;
+
+      for (i = 0; i < MESA_SHADER_STAGES; i++) {
+	 if (uni->image[i].active) {
+            struct gl_shader *sh = shProg->_LinkedShaders[i];
+
+            for (j = 0; j < count; j++)
+               sh->ImageUnits[uni->image[i].index + offset + j] =
+                  ((GLint *) values)[j];
+         }
+      }
+
+      ctx->NewDriverState |= ctx->DriverFlags.NewImageUnits;
+   }
 }
 
 /**
diff --git a/mesalib/src/mesa/main/uniforms.c b/mesalib/src/mesa/main/uniforms.c
index 071d668f3..c25c2df7e 100644
--- a/mesalib/src/mesa/main/uniforms.c
+++ b/mesalib/src/mesa/main/uniforms.c
@@ -270,6 +270,197 @@ _mesa_Uniform4iv(GLint location, GLsizei count, const GLint * value)
    _mesa_uniform(ctx, ctx->Shader.ActiveProgram, location, count, value, GL_INT_VEC4);
 }
 
+/** Same as above with direct state access **/
+void GLAPIENTRY
+_mesa_ProgramUniform1f(GLuint program, GLint location, GLfloat v0)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1f");
+   _mesa_uniform(ctx, shProg, location, 1, &v0, GL_FLOAT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2f(GLuint program, GLint location, GLfloat v0, GLfloat v1)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLfloat v[2];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform2f");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_FLOAT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3f(GLuint program, GLint location, GLfloat v0, GLfloat v1,
+                       GLfloat v2)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLfloat v[3];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform3f");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_FLOAT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4f(GLuint program, GLint location, GLfloat v0, GLfloat v1,
+                       GLfloat v2, GLfloat v3)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLfloat v[4];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   v[3] = v3;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform4f");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_FLOAT_VEC4);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform1i(GLuint program, GLint location, GLint v0)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1i");
+   _mesa_uniform(ctx, shProg, location, 1, &v0, GL_INT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2i(GLuint program, GLint location, GLint v0, GLint v1)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint v[2];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform2i");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_INT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3i(GLuint program, GLint location, GLint v0, GLint v1,
+                       GLint v2)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint v[3];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform3i");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_INT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4i(GLuint program, GLint location, GLint v0, GLint v1,
+                       GLint v2, GLint v3)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint v[4];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   v[3] = v3;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform4i");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_INT_VEC4);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform1fv(GLuint program, GLint location, GLsizei count,
+                        const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1fv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_FLOAT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2fv(GLuint program, GLint location, GLsizei count,
+                        const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform2fv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_FLOAT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3fv(GLuint program, GLint location, GLsizei count,
+                        const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform3fv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_FLOAT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4fv(GLuint program, GLint location, GLsizei count,
+                        const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform4fv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_FLOAT_VEC4);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform1iv(GLuint program, GLint location, GLsizei count,
+                        const GLint * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1iv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_INT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2iv(GLuint program, GLint location, GLsizei count,
+                        const GLint * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform2iv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_INT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3iv(GLuint program, GLint location, GLsizei count,
+                        const GLint * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform3iv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_INT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4iv(GLuint program, GLint location, GLsizei count,
+                        const GLint * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform4iv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_INT_VEC4);
+}
+
 
 /** OpenGL 3.0 GLuint-valued functions **/
 void GLAPIENTRY
@@ -369,6 +560,140 @@ _mesa_UniformMatrix4fv(GLint location, GLsizei count, GLboolean transpose,
 			4, 4, location, count, transpose, value);
 }
 
+/** Same as above with direct state access **/
+
+void GLAPIENTRY
+_mesa_ProgramUniform1ui(GLuint program, GLint location, GLuint v0)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1ui");
+   _mesa_uniform(ctx, shProg, location, 1, &v0, GL_UNSIGNED_INT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2ui(GLuint program, GLint location, GLuint v0, GLuint v1)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint v[2];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   shProg = _mesa_lookup_shader_program_err(ctx, program,
+                                            "glProgramUniform2ui");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_UNSIGNED_INT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3ui(GLuint program, GLint location, GLuint v0, GLuint v1,
+                        GLuint v2)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint v[3];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   shProg = _mesa_lookup_shader_program_err(ctx, program,
+                                            "glProgramUniform3ui");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_UNSIGNED_INT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4ui(GLuint program, GLint location, GLuint v0, GLuint v1,
+                        GLuint v2, GLuint v3)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLuint v[4];
+   struct gl_shader_program *shProg;
+   v[0] = v0;
+   v[1] = v1;
+   v[2] = v2;
+   v[3] = v3;
+   shProg = _mesa_lookup_shader_program_err(ctx, program, "glProgramUniform4ui");
+   _mesa_uniform(ctx, shProg, location, 1, v, GL_UNSIGNED_INT_VEC4);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform1uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform1uiv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_UNSIGNED_INT);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform2uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform2uiv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_UNSIGNED_INT_VEC2);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform3uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform3uiv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_UNSIGNED_INT_VEC3);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniform4uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniform4uiv");
+   _mesa_uniform(ctx, shProg, location, count, value, GL_UNSIGNED_INT_VEC4);
+}
+
+
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2fv(GLuint program, GLint location, GLsizei count,
+                              GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix2fv");
+   _mesa_uniform_matrix(ctx, shProg, 2, 2, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3fv(GLuint program, GLint location, GLsizei count,
+                              GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix3fv");
+   _mesa_uniform_matrix(ctx, shProg, 3, 3, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4fv(GLuint program, GLint location, GLsizei count,
+                              GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix4fv");
+   _mesa_uniform_matrix(ctx, shProg, 4, 4, location, count, transpose, value);
+}
+
 
 /**
  * Non-square UniformMatrix are OpenGL 2.1
@@ -427,6 +752,74 @@ _mesa_UniformMatrix4x3fv(GLint location, GLsizei count, GLboolean transpose,
 			4, 3, location, count, transpose, value);
 }
 
+/** Same as above with direct state access **/
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2x3fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix2x3fv");
+   _mesa_uniform_matrix(ctx, shProg, 2, 3, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3x2fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix3x2fv");
+   _mesa_uniform_matrix(ctx, shProg, 3, 2, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2x4fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix2x4fv");
+   _mesa_uniform_matrix(ctx, shProg, 2, 4, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4x2fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix4x2fv");
+   _mesa_uniform_matrix(ctx, shProg, 4, 2, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3x4fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix3x4fv");
+   _mesa_uniform_matrix(ctx, shProg, 3, 4, location, count, transpose, value);
+}
+
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4x3fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat * value)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+            "glProgramUniformMatrix4x3fv");
+   _mesa_uniform_matrix(ctx, shProg, 4, 3, location, count, transpose, value);
+}
+
 
 void GLAPIENTRY
 _mesa_GetnUniformfvARB(GLhandleARB program, GLint location,
diff --git a/mesalib/src/mesa/main/uniforms.h b/mesalib/src/mesa/main/uniforms.h
index f7cac6328..bd50fd9b8 100644
--- a/mesalib/src/mesa/main/uniforms.h
+++ b/mesalib/src/mesa/main/uniforms.h
@@ -111,6 +111,89 @@ _mesa_UniformMatrix3x4fv(GLint location, GLsizei count, GLboolean transpose,
 void GLAPIENTRY
 _mesa_UniformMatrix4x3fv(GLint location, GLsizei count, GLboolean transpose,
                          const GLfloat *value);
+
+void GLAPIENTRY
+_mesa_ProgramUniform1f(GLuint program, GLint, GLfloat);
+void GLAPIENTRY
+_mesa_ProgramUniform2f(GLuint program, GLint, GLfloat, GLfloat);
+void GLAPIENTRY
+_mesa_ProgramUniform3f(GLuint program, GLint, GLfloat, GLfloat, GLfloat);
+void GLAPIENTRY
+_mesa_ProgramUniform4f(GLuint program, GLint, GLfloat, GLfloat, GLfloat, GLfloat);
+void GLAPIENTRY
+_mesa_ProgramUniform1i(GLuint program, GLint, GLint);
+void GLAPIENTRY
+_mesa_ProgramUniform2i(GLuint program, GLint, GLint, GLint);
+void GLAPIENTRY
+_mesa_ProgramUniform3i(GLuint program, GLint, GLint, GLint, GLint);
+void GLAPIENTRY
+_mesa_ProgramUniform4i(GLuint program, GLint, GLint, GLint, GLint, GLint);
+void GLAPIENTRY
+_mesa_ProgramUniform1fv(GLuint program, GLint, GLsizei, const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniform2fv(GLuint program, GLint, GLsizei, const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniform3fv(GLuint program, GLint, GLsizei, const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniform4fv(GLuint program, GLint, GLsizei, const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniform1iv(GLuint program, GLint, GLsizei, const GLint *);
+void GLAPIENTRY
+_mesa_ProgramUniform2iv(GLuint program, GLint, GLsizei, const GLint *);
+void GLAPIENTRY
+_mesa_ProgramUniform3iv(GLuint program, GLint, GLsizei, const GLint *);
+void GLAPIENTRY
+_mesa_ProgramUniform4iv(GLuint program, GLint, GLsizei, const GLint *);
+void GLAPIENTRY
+_mesa_ProgramUniform1ui(GLuint program, GLint location, GLuint v0);
+void GLAPIENTRY
+_mesa_ProgramUniform2ui(GLuint program, GLint location, GLuint v0, GLuint v1);
+void GLAPIENTRY
+_mesa_ProgramUniform3ui(GLuint program, GLint location, GLuint v0, GLuint v1,
+                        GLuint v2);
+void GLAPIENTRY
+_mesa_ProgramUniform4ui(GLuint program, GLint location, GLuint v0, GLuint v1,
+                        GLuint v2, GLuint v3);
+void GLAPIENTRY
+_mesa_ProgramUniform1uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value);
+void GLAPIENTRY
+_mesa_ProgramUniform2uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value);
+void GLAPIENTRY
+_mesa_ProgramUniform3uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value);
+void GLAPIENTRY
+_mesa_ProgramUniform4uiv(GLuint program, GLint location, GLsizei count,
+                         const GLuint *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2fv(GLuint program, GLint, GLsizei, GLboolean,
+                              const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3fv(GLuint program, GLint, GLsizei, GLboolean,
+                              const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4fv(GLuint program, GLint, GLsizei, GLboolean,
+                              const GLfloat *);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2x3fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3x2fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix2x4fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4x2fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix3x4fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+void GLAPIENTRY
+_mesa_ProgramUniformMatrix4x3fv(GLuint program, GLint location, GLsizei count,
+                                GLboolean transpose, const GLfloat *value);
+
 void GLAPIENTRY
 _mesa_GetnUniformfvARB(GLhandleARB, GLint, GLsizei, GLfloat *);
 void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/vdpau.c b/mesalib/src/mesa/main/vdpau.c
index 359757607..c2cf20664 100644
--- a/mesalib/src/mesa/main/vdpau.c
+++ b/mesalib/src/mesa/main/vdpau.c
@@ -205,7 +205,7 @@ _mesa_VDPAURegisterOutputSurfaceNV(const GLvoid *vdpSurface, GLenum target,
                            numTextureNames, textureNames);
 }
 
-void GLAPIENTRY
+GLboolean GLAPIENTRY
 _mesa_VDPAUIsSurfaceNV(GLintptr surface)
 {
    struct vdp_surface *surf = (struct vdp_surface *)surface;
@@ -213,13 +213,14 @@ _mesa_VDPAUIsSurfaceNV(GLintptr surface)
 
    if (!ctx->vdpDevice || !ctx->vdpGetProcAddress || !ctx->vdpSurfaces) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "VDPAUIsSurfaceNV");
-      return;
+      return false;
    }
 
    if (!_mesa_set_search(ctx->vdpSurfaces, _mesa_hash_pointer(surf), surf)) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "VDPAUIsSurfaceNV");
-      return;
+      return false;
    }
+
+   return true;
 }
 
 void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/vdpau.h b/mesalib/src/mesa/main/vdpau.h
index f32d6dacb..627609c50 100644
--- a/mesalib/src/mesa/main/vdpau.h
+++ b/mesalib/src/mesa/main/vdpau.h
@@ -50,7 +50,7 @@ _mesa_VDPAURegisterOutputSurfaceNV(const GLvoid *vdpSurface, GLenum target,
                                    GLsizei numTextureNames,
                                    const GLuint *textureNames);
 
-extern void GLAPIENTRY
+extern GLboolean GLAPIENTRY
 _mesa_VDPAUIsSurfaceNV(GLintptr surface);
 
 extern void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/version.c b/mesalib/src/mesa/main/version.c
index 587cb9ad5..63cba26da 100644
--- a/mesalib/src/mesa/main/version.c
+++ b/mesalib/src/mesa/main/version.c
@@ -232,7 +232,6 @@ compute_version(struct gl_context *ctx)
                               (ctx->API == API_OPENGL_CORE ||
                                ctx->Extensions.ARB_color_buffer_float) &&
                               ctx->Extensions.ARB_depth_buffer_float &&
-                              ctx->Extensions.ARB_half_float_pixel &&
                               ctx->Extensions.ARB_half_float_vertex &&
                               ctx->Extensions.ARB_map_buffer_range &&
                               ctx->Extensions.ARB_shader_texture_lod &&
diff --git a/mesalib/src/mesa/program/ir_to_mesa.cpp b/mesalib/src/mesa/program/ir_to_mesa.cpp
index 74c512b33..9919874a6 100644
--- a/mesalib/src/mesa/program/ir_to_mesa.cpp
+++ b/mesalib/src/mesa/program/ir_to_mesa.cpp
@@ -618,6 +618,7 @@ type_size(const struct glsl_type *type)
       }
       return size;
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
       /* Samplers take up one slot in UNIFORMS[], but they're baked in
        * at link time.
        */
@@ -2559,6 +2560,7 @@ _mesa_associate_uniform_storage(struct gl_context *ctx,
 	    columns = 1;
 	    break;
 	 case GLSL_TYPE_SAMPLER:
+	 case GLSL_TYPE_IMAGE:
 	    format = uniform_native;
 	    columns = 1;
 	    break;
@@ -2916,16 +2918,17 @@ get_mesa_program(struct gl_context *ctx,
    set_branchtargets(&v, mesa_instructions, num_instructions);
 
    if (ctx->Shader.Flags & GLSL_DUMP) {
-      printf("\n");
-      printf("GLSL IR for linked %s program %d:\n", target_string,
-	     shader_program->Name);
-      _mesa_print_ir(shader->ir, NULL);
-      printf("\n");
-      printf("\n");
-      printf("Mesa IR for linked %s program %d:\n", target_string,
-	     shader_program->Name);
+      fprintf(stderr, "\n");
+      fprintf(stderr, "GLSL IR for linked %s program %d:\n", target_string,
+	      shader_program->Name);
+      _mesa_print_ir(stderr, shader->ir, NULL);
+      fprintf(stderr, "\n");
+      fprintf(stderr, "\n");
+      fprintf(stderr, "Mesa IR for linked %s program %d:\n", target_string,
+	      shader_program->Name);
       print_program(mesa_instructions, mesa_instruction_annotation,
 		    num_instructions);
+      fflush(stderr);
    }
 
    prog->Instructions = mesa_instructions;
@@ -3095,12 +3098,12 @@ _mesa_glsl_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
 
    if (ctx->Shader.Flags & GLSL_DUMP) {
       if (!prog->LinkStatus) {
-	 printf("GLSL shader program %d failed to link\n", prog->Name);
+	 fprintf(stderr, "GLSL shader program %d failed to link\n", prog->Name);
       }
 
       if (prog->InfoLog && prog->InfoLog[0] != 0) {
-	 printf("GLSL shader program %d info log:\n", prog->Name);
-	 printf("%s\n", prog->InfoLog);
+	 fprintf(stderr, "GLSL shader program %d info log:\n", prog->Name);
+	 fprintf(stderr, "%s\n", prog->InfoLog);
       }
    }
 }
diff --git a/mesalib/src/mesa/program/program.c b/mesalib/src/mesa/program/program.c
index fa80bc58c..b7332fccf 100644
--- a/mesalib/src/mesa/program/program.c
+++ b/mesalib/src/mesa/program/program.c
@@ -424,7 +424,7 @@ _mesa_reference_program_(struct gl_context *ctx,
    if (*ptr) {
       GLboolean deleteFlag;
 
-      /*_glthread_LOCK_MUTEX((*ptr)->Mutex);*/
+      /*mtx_lock(&(*ptr)->Mutex);*/
 #if 0
       printf("Program %p ID=%u Target=%s  Refcount-- to %d\n",
              *ptr, (*ptr)->Id,
@@ -436,7 +436,7 @@ _mesa_reference_program_(struct gl_context *ctx,
       (*ptr)->RefCount--;
 
       deleteFlag = ((*ptr)->RefCount == 0);
-      /*_glthread_UNLOCK_MUTEX((*ptr)->Mutex);*/
+      /*mtx_lock(&(*ptr)->Mutex);*/
 
       if (deleteFlag) {
          ASSERT(ctx);
@@ -448,7 +448,7 @@ _mesa_reference_program_(struct gl_context *ctx,
 
    assert(!*ptr);
    if (prog) {
-      /*_glthread_LOCK_MUTEX(prog->Mutex);*/
+      /*mtx_lock(&prog->Mutex);*/
       prog->RefCount++;
 #if 0
       printf("Program %p ID=%u Target=%s  Refcount++ to %d\n",
@@ -457,7 +457,7 @@ _mesa_reference_program_(struct gl_context *ctx,
               (prog->Target == MESA_GEOMETRY_PROGRAM ? "GP" : "FP")),
              prog->RefCount);
 #endif
-      /*_glthread_UNLOCK_MUTEX(prog->Mutex);*/
+      /*mtx_unlock(&prog->Mutex);*/
    }
 
    *ptr = prog;
@@ -550,6 +550,7 @@ _mesa_clone_program(struct gl_context *ctx, const struct gl_program *prog)
          struct gl_geometry_program *gpc = gl_geometry_program(clone);
          gpc->VerticesOut = gp->VerticesOut;
          gpc->InputType = gp->InputType;
+         gpc->Invocations = gp->Invocations;
          gpc->OutputType = gp->OutputType;
       }
       break;
diff --git a/mesalib/src/mesa/state_tracker/st_atom_scissor.c b/mesalib/src/mesa/state_tracker/st_atom_scissor.c
index a1f72da47..a19ade1fa 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_scissor.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_scissor.c
@@ -43,51 +43,56 @@
 static void
 update_scissor( struct st_context *st )
 {
-   struct pipe_scissor_state scissor;
+   struct pipe_scissor_state scissor[PIPE_MAX_VIEWPORTS];
    const struct gl_context *ctx = st->ctx;
    const struct gl_framebuffer *fb = ctx->DrawBuffer;
    GLint miny, maxy;
+   int i;
+   bool changed = false;
+   for (i = 0 ; i < ctx->Const.MaxViewports; i++) {
+      scissor[i].minx = 0;
+      scissor[i].miny = 0;
+      scissor[i].maxx = fb->Width;
+      scissor[i].maxy = fb->Height;
 
-   scissor.minx = 0;
-   scissor.miny = 0;
-   scissor.maxx = fb->Width;
-   scissor.maxy = fb->Height;
+      if (ctx->Scissor.EnableFlags & (1 << i)) {
+         /* need to be careful here with xmax or ymax < 0 */
+         GLint xmax = MAX2(0, ctx->Scissor.ScissorArray[i].X + ctx->Scissor.ScissorArray[i].Width);
+         GLint ymax = MAX2(0, ctx->Scissor.ScissorArray[i].Y + ctx->Scissor.ScissorArray[i].Height);
 
-   if (ctx->Scissor.EnableFlags & 1) {
-      /* need to be careful here with xmax or ymax < 0 */
-      GLint xmax = MAX2(0, ctx->Scissor.ScissorArray[0].X + ctx->Scissor.ScissorArray[0].Width);
-      GLint ymax = MAX2(0, ctx->Scissor.ScissorArray[0].Y + ctx->Scissor.ScissorArray[0].Height);
+         if (ctx->Scissor.ScissorArray[i].X > (GLint)scissor[i].minx)
+            scissor[i].minx = ctx->Scissor.ScissorArray[i].X;
+         if (ctx->Scissor.ScissorArray[i].Y > (GLint)scissor[i].miny)
+            scissor[i].miny = ctx->Scissor.ScissorArray[i].Y;
 
-      if (ctx->Scissor.ScissorArray[0].X > (GLint)scissor.minx)
-         scissor.minx = ctx->Scissor.ScissorArray[0].X;
-      if (ctx->Scissor.ScissorArray[0].Y > (GLint)scissor.miny)
-         scissor.miny = ctx->Scissor.ScissorArray[0].Y;
+         if (xmax < (GLint) scissor[i].maxx)
+            scissor[i].maxx = xmax;
+         if (ymax < (GLint) scissor[i].maxy)
+            scissor[i].maxy = ymax;
+         
+         /* check for null space */
+         if (scissor[i].minx >= scissor[i].maxx || scissor[i].miny >= scissor[i].maxy)
+            scissor[i].minx = scissor[i].miny = scissor[i].maxx = scissor[i].maxy = 0;
+      }
 
-      if (xmax < (GLint) scissor.maxx)
-         scissor.maxx = xmax;
-      if (ymax < (GLint) scissor.maxy)
-         scissor.maxy = ymax;
+      /* Now invert Y if needed.
+       * Gallium drivers use the convention Y=0=top for surfaces.
+       */
+      if (st_fb_orientation(fb) == Y_0_TOP) {
+         miny = fb->Height - scissor[i].maxy;
+         maxy = fb->Height - scissor[i].miny;
+         scissor[i].miny = miny;
+         scissor[i].maxy = maxy;
+      }
 
-      /* check for null space */
-      if (scissor.minx >= scissor.maxx || scissor.miny >= scissor.maxy)
-         scissor.minx = scissor.miny = scissor.maxx = scissor.maxy = 0;
-   }
-
-   /* Now invert Y if needed.
-    * Gallium drivers use the convention Y=0=top for surfaces.
-    */
-   if (st_fb_orientation(fb) == Y_0_TOP) {
-      miny = fb->Height - scissor.maxy;
-      maxy = fb->Height - scissor.miny;
-      scissor.miny = miny;
-      scissor.maxy = maxy;
-   }
-
-   if (memcmp(&scissor, &st->state.scissor, sizeof(scissor)) != 0) {
-      /* state has changed */
-      st->state.scissor = scissor;  /* struct copy */
-      st->pipe->set_scissor_states(st->pipe, 0, 1, &scissor); /* activate */
+      if (memcmp(&scissor[i], &st->state.scissor[i], sizeof(scissor)) != 0) {
+         /* state has changed */
+         st->state.scissor[i] = scissor[i];  /* struct copy */
+         changed = true;
+      }
    }
+   if (changed)
+      st->pipe->set_scissor_states(st->pipe, 0, ctx->Const.MaxViewports, scissor); /* activate */
 }
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_atom_viewport.c b/mesalib/src/mesa/state_tracker/st_atom_viewport.c
index 8c6d679a0..7584f9b3d 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_viewport.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_viewport.c
@@ -43,7 +43,7 @@ update_viewport( struct st_context *st )
 {
    struct gl_context *ctx = st->ctx;
    GLfloat yScale, yBias;
-
+   int i;
    /* _NEW_BUFFERS
     */
    if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) {
@@ -61,26 +61,29 @@ update_viewport( struct st_context *st )
 
    /* _NEW_VIEWPORT 
     */
+   for (i = 0; i < ctx->Const.MaxViewports; i++)
    {
-      GLfloat x = ctx->ViewportArray[0].X;
-      GLfloat y = ctx->ViewportArray[0].Y;
-      GLfloat z = ctx->ViewportArray[0].Near;
-      GLfloat half_width = ctx->ViewportArray[0].Width * 0.5f;
-      GLfloat half_height = ctx->ViewportArray[0].Height * 0.5f;
-      GLfloat half_depth = (GLfloat)(ctx->ViewportArray[0].Far - ctx->ViewportArray[0].Near) * 0.5f;
+      GLfloat x = ctx->ViewportArray[i].X;
+      GLfloat y = ctx->ViewportArray[i].Y;
+      GLfloat z = ctx->ViewportArray[i].Near;
+      GLfloat half_width = ctx->ViewportArray[i].Width * 0.5f;
+      GLfloat half_height = ctx->ViewportArray[i].Height * 0.5f;
+      GLfloat half_depth = (GLfloat)(ctx->ViewportArray[i].Far - ctx->ViewportArray[i].Near) * 0.5f;
       
-      st->state.viewport.scale[0] = half_width;
-      st->state.viewport.scale[1] = half_height * yScale;
-      st->state.viewport.scale[2] = half_depth;
-      st->state.viewport.scale[3] = 1.0;
-
-      st->state.viewport.translate[0] = half_width + x;
-      st->state.viewport.translate[1] = (half_height + y) * yScale + yBias;
-      st->state.viewport.translate[2] = half_depth + z;
-      st->state.viewport.translate[3] = 0.0;
+      st->state.viewport[i].scale[0] = half_width;
+      st->state.viewport[i].scale[1] = half_height * yScale;
+      st->state.viewport[i].scale[2] = half_depth;
+      st->state.viewport[i].scale[3] = 1.0;
 
-      cso_set_viewport(st->cso_context, &st->state.viewport);
+      st->state.viewport[i].translate[0] = half_width + x;
+      st->state.viewport[i].translate[1] = (half_height + y) * yScale + yBias;
+      st->state.viewport[i].translate[2] = half_depth + z;
+      st->state.viewport[i].translate[3] = 0.0;
    }
+
+   cso_set_viewport(st->cso_context, &st->state.viewport[0]);
+   if (ctx->Const.MaxViewports > 1)
+      st->pipe->set_viewport_states(st->pipe, 1, ctx->Const.MaxViewports - 1, &st->state.viewport[1]);
 }
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
index c8f088dc8..49c4b903e 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
@@ -76,7 +76,7 @@ st_bufferobj_free(struct gl_context *ctx, struct gl_buffer_object *obj)
    struct st_buffer_object *st_obj = st_buffer_object(obj);
 
    assert(obj->RefCount == 0);
-   assert(st_obj->transfer == NULL);
+   _mesa_buffer_unmap_all_mappings(ctx, obj);
 
    if (st_obj->buffer)
       pipe_resource_reference(&st_obj->buffer, NULL);
@@ -176,15 +176,18 @@ st_bufferobj_data(struct gl_context *ctx,
 		  GLsizeiptrARB size,
 		  const GLvoid * data,
 		  GLenum usage,
+                  GLbitfield storageFlags,
 		  struct gl_buffer_object *obj)
 {
    struct st_context *st = st_context(ctx);
    struct pipe_context *pipe = st->pipe;
    struct st_buffer_object *st_obj = st_buffer_object(obj);
-   unsigned bind, pipe_usage;
+   unsigned bind, pipe_usage, pipe_flags = 0;
 
    if (size && data && st_obj->buffer &&
-       st_obj->Base.Size == size && st_obj->Base.Usage == usage) {
+       st_obj->Base.Size == size &&
+       st_obj->Base.Usage == usage &&
+       st_obj->Base.StorageFlags == storageFlags) {
       /* Just discard the old contents and write new data.
        * This should be the same as creating a new buffer, but we avoid
        * a lot of validation in Mesa.
@@ -200,6 +203,7 @@ st_bufferobj_data(struct gl_context *ctx,
 
    st_obj->Base.Size = size;
    st_obj->Base.Usage = usage;
+   st_obj->Base.StorageFlags = storageFlags;
 
    switch (target) {
    case GL_PIXEL_PACK_BUFFER_ARB:
@@ -225,24 +229,41 @@ st_bufferobj_data(struct gl_context *ctx,
       bind = 0;
    }
 
-   switch (usage) {
-   case GL_STATIC_DRAW:
-   case GL_STATIC_READ:
-   case GL_STATIC_COPY:
-   default:
-      pipe_usage = PIPE_USAGE_DEFAULT;
-      break;
-   case GL_DYNAMIC_DRAW:
-   case GL_DYNAMIC_READ:
-   case GL_DYNAMIC_COPY:
-      pipe_usage = PIPE_USAGE_DYNAMIC;
-      break;
-   case GL_STREAM_DRAW:
-   case GL_STREAM_READ:
-   case GL_STREAM_COPY:
-      pipe_usage = PIPE_USAGE_STREAM;
-      break;
+   /* Set usage. */
+   if (st_obj->Base.Immutable) {
+      /* BufferStorage */
+      if (storageFlags & GL_CLIENT_STORAGE_BIT)
+         pipe_usage = PIPE_USAGE_STAGING;
+      else
+         pipe_usage = PIPE_USAGE_DEFAULT;
    }
+   else {
+      /* BufferData */
+      switch (usage) {
+      case GL_STATIC_DRAW:
+      case GL_STATIC_READ:
+      case GL_STATIC_COPY:
+      default:
+	 pipe_usage = PIPE_USAGE_DEFAULT;
+         break;
+      case GL_DYNAMIC_DRAW:
+      case GL_DYNAMIC_READ:
+      case GL_DYNAMIC_COPY:
+         pipe_usage = PIPE_USAGE_DYNAMIC;
+         break;
+      case GL_STREAM_DRAW:
+      case GL_STREAM_READ:
+      case GL_STREAM_COPY:
+         pipe_usage = PIPE_USAGE_STREAM;
+         break;
+      }
+   }
+
+   /* Set flags. */
+   if (storageFlags & GL_MAP_PERSISTENT_BIT)
+      pipe_flags |= PIPE_RESOURCE_FLAG_MAP_PERSISTENT;
+   if (storageFlags & GL_MAP_COHERENT_BIT)
+      pipe_flags |= PIPE_RESOURCE_FLAG_MAP_COHERENT;
 
    pipe_resource_reference( &st_obj->buffer, NULL );
 
@@ -251,8 +272,20 @@ st_bufferobj_data(struct gl_context *ctx,
    }
 
    if (size != 0) {
-      st_obj->buffer = pipe_buffer_create(pipe->screen, bind,
-                                          pipe_usage, size);
+      struct pipe_resource buffer;
+
+      memset(&buffer, 0, sizeof buffer);
+      buffer.target = PIPE_BUFFER;
+      buffer.format = PIPE_FORMAT_R8_UNORM; /* want TYPELESS or similar */
+      buffer.bind = bind;
+      buffer.usage = pipe_usage;
+      buffer.flags = pipe_flags;
+      buffer.width0 = size;
+      buffer.height0 = 1;
+      buffer.depth0 = 1;
+      buffer.array_size = 1;
+
+      st_obj->buffer = pipe->screen->resource_create(pipe->screen, &buffer);
 
       if (!st_obj->buffer) {
          /* out of memory */
@@ -277,7 +310,8 @@ st_bufferobj_data(struct gl_context *ctx,
 static void *
 st_bufferobj_map_range(struct gl_context *ctx,
                        GLintptr offset, GLsizeiptr length, GLbitfield access,
-                       struct gl_buffer_object *obj)
+                       struct gl_buffer_object *obj,
+                       gl_map_buffer_index index)
 {
    struct pipe_context *pipe = st_context(ctx)->pipe;
    struct st_buffer_object *st_obj = st_buffer_object(obj);
@@ -305,6 +339,12 @@ st_bufferobj_map_range(struct gl_context *ctx,
    if (access & GL_MAP_UNSYNCHRONIZED_BIT)
       flags |= PIPE_TRANSFER_UNSYNCHRONIZED;
 
+   if (access & GL_MAP_PERSISTENT_BIT)
+      flags |= PIPE_TRANSFER_PERSISTENT;
+
+   if (access & GL_MAP_COHERENT_BIT)
+      flags |= PIPE_TRANSFER_COHERENT;
+
    /* ... other flags ...
     */
 
@@ -316,28 +356,29 @@ st_bufferobj_map_range(struct gl_context *ctx,
    assert(offset < obj->Size);
    assert(offset + length <= obj->Size);
 
-   obj->Pointer = pipe_buffer_map_range(pipe,
+   obj->Mappings[index].Pointer = pipe_buffer_map_range(pipe,
                                         st_obj->buffer,
                                         offset, length,
                                         flags,
-                                        &st_obj->transfer);
-   if (obj->Pointer) {
-      obj->Offset = offset;
-      obj->Length = length;
-      obj->AccessFlags = access;
+                                        &st_obj->transfer[index]);
+   if (obj->Mappings[index].Pointer) {
+      obj->Mappings[index].Offset = offset;
+      obj->Mappings[index].Length = length;
+      obj->Mappings[index].AccessFlags = access;
    }
    else {
-      st_obj->transfer = NULL;
+      st_obj->transfer[index] = NULL;
    }
 
-   return obj->Pointer;
+   return obj->Mappings[index].Pointer;
 }
 
 
 static void
 st_bufferobj_flush_mapped_range(struct gl_context *ctx,
                                 GLintptr offset, GLsizeiptr length,
-                                struct gl_buffer_object *obj)
+                                struct gl_buffer_object *obj,
+                                gl_map_buffer_index index)
 {
    struct pipe_context *pipe = st_context(ctx)->pipe;
    struct st_buffer_object *st_obj = st_buffer_object(obj);
@@ -345,14 +386,15 @@ st_bufferobj_flush_mapped_range(struct gl_context *ctx,
    /* Subrange is relative to mapped range */
    assert(offset >= 0);
    assert(length >= 0);
-   assert(offset + length <= obj->Length);
-   assert(obj->Pointer);
+   assert(offset + length <= obj->Mappings[index].Length);
+   assert(obj->Mappings[index].Pointer);
 
    if (!length)
       return;
 
-   pipe_buffer_flush_mapped_range(pipe, st_obj->transfer,
-                                  obj->Offset + offset, length);
+   pipe_buffer_flush_mapped_range(pipe, st_obj->transfer[index],
+                                  obj->Mappings[index].Offset + offset,
+                                  length);
 }
 
 
@@ -360,18 +402,19 @@ st_bufferobj_flush_mapped_range(struct gl_context *ctx,
  * Called via glUnmapBufferARB().
  */
 static GLboolean
-st_bufferobj_unmap(struct gl_context *ctx, struct gl_buffer_object *obj)
+st_bufferobj_unmap(struct gl_context *ctx, struct gl_buffer_object *obj,
+                   gl_map_buffer_index index)
 {
    struct pipe_context *pipe = st_context(ctx)->pipe;
    struct st_buffer_object *st_obj = st_buffer_object(obj);
 
-   if (obj->Length)
-      pipe_buffer_unmap(pipe, st_obj->transfer);
+   if (obj->Mappings[index].Length)
+      pipe_buffer_unmap(pipe, st_obj->transfer[index]);
 
-   st_obj->transfer = NULL;
-   obj->Pointer = NULL;
-   obj->Offset = 0;
-   obj->Length = 0;
+   st_obj->transfer[index] = NULL;
+   obj->Mappings[index].Pointer = NULL;
+   obj->Mappings[index].Offset = 0;
+   obj->Mappings[index].Length = 0;
    return GL_TRUE;
 }
 
@@ -395,8 +438,8 @@ st_copy_buffer_subdata(struct gl_context *ctx,
       return;
 
    /* buffer should not already be mapped */
-   assert(!src->Pointer);
-   assert(!dst->Pointer);
+   assert(!_mesa_check_disallowed_mapping(src));
+   assert(!_mesa_check_disallowed_mapping(dst));
 
    u_box_1d(readOffset, size, &box);
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
index a039c0a50..a5ab12219 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
@@ -43,7 +43,7 @@ struct st_buffer_object
 {
    struct gl_buffer_object Base;
    struct pipe_resource *buffer;     /* GPU storage */
-   struct pipe_transfer *transfer; /* In-progress map information */
+   struct pipe_transfer *transfer[MAP_COUNT];
 };
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c
index 2f1a22b58..dd4dde74c 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c
@@ -54,7 +54,24 @@ st_TextureBarrier(struct gl_context *ctx)
 }
 
 
+/**
+ * Called via ctx->Driver.MemoryBarrier()
+ */
+static void
+st_MemoryBarrier(struct gl_context *ctx, GLbitfield barriers)
+{
+   struct pipe_context *pipe = st_context(ctx)->pipe;
+   unsigned flags = 0;
+
+   if (barriers & GL_CLIENT_MAPPED_BUFFER_BARRIER_BIT)
+      flags |= PIPE_BARRIER_MAPPED_BUFFER;
+
+   if (flags && pipe->memory_barrier)
+      pipe->memory_barrier(pipe, flags);
+}
+
 void st_init_texture_barrier_functions(struct dd_function_table *functions)
 {
    functions->TextureBarrier = st_TextureBarrier;
+   functions->MemoryBarrier = st_MemoryBarrier;
 }
diff --git a/mesalib/src/mesa/state_tracker/st_context.h b/mesalib/src/mesa/state_tracker/st_context.h
index 996e0c6bc..9c699a015 100644
--- a/mesalib/src/mesa/state_tracker/st_context.h
+++ b/mesalib/src/mesa/state_tracker/st_context.h
@@ -115,8 +115,8 @@ struct st_context
          unsigned size;
       } constants[PIPE_SHADER_TYPES];
       struct pipe_framebuffer_state framebuffer;
-      struct pipe_scissor_state scissor;
-      struct pipe_viewport_state viewport;
+      struct pipe_scissor_state scissor[PIPE_MAX_VIEWPORTS];
+      struct pipe_viewport_state viewport[PIPE_MAX_VIEWPORTS];
       unsigned sample_mask;
 
       GLuint poly_stipple[32];  /**< In OpenGL's bottom-to-top order */
diff --git a/mesalib/src/mesa/state_tracker/st_draw_feedback.c b/mesalib/src/mesa/state_tracker/st_draw_feedback.c
index 09cd9511d..177f6b5ae 100644
--- a/mesalib/src/mesa/state_tracker/st_draw_feedback.c
+++ b/mesalib/src/mesa/state_tracker/st_draw_feedback.c
@@ -156,7 +156,7 @@ st_feedback_draw_vbo(struct gl_context *ctx,
     * code sends state updates to the pipe, not to our private draw module.
     */
    assert(draw);
-   draw_set_viewport_states(draw, 0, 1, &st->state.viewport);
+   draw_set_viewport_states(draw, 0, 1, &st->state.viewport[0]);
    draw_set_clip_state(draw, &st->state.clip);
    draw_set_rasterizer_state(draw, &st->state.rasterizer, NULL);
    draw_bind_vertex_shader(draw, st->vp_variant->draw_shader);
diff --git a/mesalib/src/mesa/state_tracker/st_extensions.c b/mesalib/src/mesa/state_tracker/st_extensions.c
index 9829eb596..18ddd4e47 100644
--- a/mesalib/src/mesa/state_tracker/st_extensions.c
+++ b/mesalib/src/mesa/state_tracker/st_extensions.c
@@ -39,6 +39,12 @@
 #include "st_extensions.h"
 #include "st_format.h"
 
+
+/*
+ * Note: we use these function rather than the MIN2, MAX2, CLAMP macros to
+ * avoid evaluating arguments (which are often function calls) more than once.
+ */
+
 static unsigned _min(unsigned a, unsigned b)
 {
    return (a < b) ? a : b;
@@ -262,10 +268,13 @@ void st_init_limits(struct st_context *st)
    c->Program[MESA_SHADER_VERTEX].MaxOutputComponents = c->MaxVarying * 4;
    c->Program[MESA_SHADER_GEOMETRY].MaxInputComponents = c->MaxVarying * 4;
    c->Program[MESA_SHADER_GEOMETRY].MaxOutputComponents = c->MaxVarying * 4;
+   c->MaxGeometryOutputVertices = screen->get_param(screen, PIPE_CAP_MAX_GEOMETRY_OUTPUT_VERTICES);
+   c->MaxGeometryTotalOutputComponents = screen->get_param(screen, PIPE_CAP_MAX_GEOMETRY_TOTAL_OUTPUT_COMPONENTS);
 
    c->MinProgramTexelOffset = screen->get_param(screen, PIPE_CAP_MIN_TEXEL_OFFSET);
    c->MaxProgramTexelOffset = screen->get_param(screen, PIPE_CAP_MAX_TEXEL_OFFSET);
 
+   c->MaxProgramTextureGatherComponents = screen->get_param(screen, PIPE_CAP_MAX_TEXTURE_GATHER_COMPONENTS);
    c->UniformBooleanTrue = ~0;
 
    c->MaxTransformFeedbackBuffers =
@@ -373,6 +382,7 @@ void st_init_extensions(struct st_context *st)
 
    static const struct st_extension_cap_mapping cap_mapping[] = {
       { o(ARB_base_instance),                PIPE_CAP_START_INSTANCE                   },
+      { o(ARB_buffer_storage),               PIPE_CAP_BUFFER_MAP_PERSISTENT_COHERENT },
       { o(ARB_depth_clamp),                  PIPE_CAP_DEPTH_CLIP_DISABLE               },
       { o(ARB_depth_texture),                PIPE_CAP_TEXTURE_SHADOW_MAP               },
       { o(ARB_draw_buffers_blend),           PIPE_CAP_INDEP_BLEND_FUNC                 },
@@ -526,7 +536,6 @@ void st_init_extensions(struct st_context *st)
    ctx->Extensions.ARB_fragment_coord_conventions = GL_TRUE;
    ctx->Extensions.ARB_fragment_program = GL_TRUE;
    ctx->Extensions.ARB_fragment_shader = GL_TRUE;
-   ctx->Extensions.ARB_half_float_pixel = GL_TRUE;
    ctx->Extensions.ARB_half_float_vertex = GL_TRUE;
    ctx->Extensions.ARB_internalformat_query = GL_TRUE;
    ctx->Extensions.ARB_map_buffer_range = GL_TRUE;
@@ -777,4 +786,15 @@ void st_init_extensions(struct st_context *st)
       if (!ctx->Extensions.EXT_transform_feedback)
          ctx->Const.DisableVaryingPacking = GL_TRUE;
    }
+
+   if (ctx->API == API_OPENGL_CORE) {
+      ctx->Const.MaxViewports = screen->get_param(screen, PIPE_CAP_MAX_VIEWPORTS);
+      if (ctx->Const.MaxViewports >= 16) {
+         ctx->Const.ViewportBounds.Min = -16384.0;
+         ctx->Const.ViewportBounds.Max = 16384.0;
+         ctx->Extensions.ARB_viewport_array = GL_TRUE;
+      }
+   }
+   if (ctx->Const.MaxProgramTextureGatherComponents > 0)
+      ctx->Extensions.ARB_texture_gather = GL_TRUE;
 }
diff --git a/mesalib/src/mesa/state_tracker/st_format.c b/mesalib/src/mesa/state_tracker/st_format.c
index 90e9f2f78..5f951eb02 100644
--- a/mesalib/src/mesa/state_tracker/st_format.c
+++ b/mesalib/src/mesa/state_tracker/st_format.c
@@ -107,11 +107,11 @@ st_mesa_format_to_pipe_format(mesa_format mesaFormat)
       return PIPE_FORMAT_Z32_UNORM;
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       return PIPE_FORMAT_S8_UINT_Z24_UNORM;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       return PIPE_FORMAT_Z24_UNORM_S8_UINT;
    case MESA_FORMAT_X8Z24_UNORM:
       return PIPE_FORMAT_X8Z24_UNORM;
-   case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
       return PIPE_FORMAT_Z24X8_UNORM;
    case MESA_FORMAT_S_UINT8:
       return PIPE_FORMAT_S8_UINT;
@@ -465,9 +465,9 @@ st_pipe_format_to_mesa_format(enum pipe_format format)
    case PIPE_FORMAT_X8Z24_UNORM:
       return MESA_FORMAT_X8Z24_UNORM;
    case PIPE_FORMAT_Z24X8_UNORM:
-      return MESA_FORMAT_Z24_UNORM_S8_UINT;
-   case PIPE_FORMAT_Z24_UNORM_S8_UINT:
       return MESA_FORMAT_Z24_UNORM_X8_UINT;
+   case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+      return MESA_FORMAT_Z24_UNORM_S8_UINT;
    case PIPE_FORMAT_Z32_FLOAT:
       return MESA_FORMAT_Z_FLOAT32;
    case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
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 610fc68a8..a9e75d8a6 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -87,7 +87,7 @@ extern "C" {
  */
 #define MAX_ARRAYS        256
 
-/* will be 4 for GLSL 4.00 */
+/* if we support a native gallium TG4 with the ability to take 4 texoffsets then bump this */
 #define MAX_GLSL_TEXTURE_OFFSET 1
 
 class st_src_reg;
@@ -249,7 +249,8 @@ public:
    int sampler; /**< sampler index */
    int tex_target; /**< One of TEXTURE_*_INDEX */
    GLboolean tex_shadow;
-   struct tgsi_texture_offset tex_offsets[MAX_GLSL_TEXTURE_OFFSET];
+
+   st_src_reg tex_offsets[MAX_GLSL_TEXTURE_OFFSET];
    unsigned tex_offset_num_offset;
    int dead_mask; /**< Used in dead code elimination */
 
@@ -983,6 +984,7 @@ type_size(const struct glsl_type *type)
       }
       return size;
    case GLSL_TYPE_SAMPLER:
+   case GLSL_TYPE_IMAGE:
       /* Samplers take up one slot in UNIFORMS[], but they're baked in
        * at link time.
        */
@@ -2685,7 +2687,7 @@ glsl_to_tgsi_visitor::visit(ir_call *ir)
 void
 glsl_to_tgsi_visitor::visit(ir_texture *ir)
 {
-   st_src_reg result_src, coord, cube_sc, lod_info, projector, dx, dy, offset, sample_index;
+   st_src_reg result_src, coord, cube_sc, lod_info, projector, dx, dy, offset, sample_index, component;
    st_dst_reg result_dst, coord_dst, cube_sc_dst;
    glsl_to_tgsi_instruction *inst = NULL;
    unsigned opcode = TGSI_OPCODE_NOP;
@@ -2779,12 +2781,20 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
       ir->lod_info.sample_index->accept(this);
       sample_index = this->result;
       break;
+   case ir_tg4:
+      opcode = TGSI_OPCODE_TG4;
+      ir->lod_info.component->accept(this);
+      component = this->result;
+      if (ir->offset) {
+         ir->offset->accept(this);
+         /* this should have been lowered */
+         assert(ir->offset->type->base_type != GLSL_TYPE_ARRAY);
+         offset = this->result;
+      }
+      break;
    case ir_lod:
       assert(!"Unexpected ir_lod opcode");
       break;
-   case ir_tg4:
-      assert(!"Unexpected ir_tg4 opcode");
-      break;
    case ir_query_levels:
       assert(!"Unexpected ir_query_levels opcode");
       break;
@@ -2892,7 +2902,13 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
       inst = emit(ir, opcode, result_dst, coord, lod_info);
    } else if (opcode == TGSI_OPCODE_TEX2) {
       inst = emit(ir, opcode, result_dst, coord, cube_sc);
-   } else 
+   } else if (opcode == TGSI_OPCODE_TG4) {
+      if (is_cube_array && ir->shadow_comparitor) {
+         inst = emit(ir, opcode, result_dst, coord, cube_sc);
+      } else {
+         inst = emit(ir, opcode, result_dst, coord, component);
+      }
+   } else
       inst = emit(ir, opcode, result_dst, coord);
 
    if (ir->shadow_comparitor)
@@ -2903,12 +2919,8 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
         					   this->prog);
 
    if (ir->offset) {
-       inst->tex_offset_num_offset = 1;
-       inst->tex_offsets[0].Index = offset.index;
-       inst->tex_offsets[0].File = offset.file;
-       inst->tex_offsets[0].SwizzleX = GET_SWZ(offset.swizzle, 0);
-       inst->tex_offsets[0].SwizzleY = GET_SWZ(offset.swizzle, 1);
-       inst->tex_offsets[0].SwizzleZ = GET_SWZ(offset.swizzle, 2);
+      inst->tex_offset_num_offset = 1;
+      inst->tex_offsets[0] = offset;
    }
 
    switch (sampler_type->sampler_dimensionality) {
@@ -3266,6 +3278,13 @@ glsl_to_tgsi_visitor::rename_temp_register(int index, int new_index)
             inst->src[j].index = new_index;
          }
       }
+
+      for (j=0; j < inst->tex_offset_num_offset; j++) {
+         if (inst->tex_offsets[j].file == PROGRAM_TEMPORARY && 
+             inst->tex_offsets[j].index == index) {
+            inst->tex_offsets[j].index = new_index;
+         }
+      }
       
       if (inst->dst.file == PROGRAM_TEMPORARY && inst->dst.index == index) {
          inst->dst.index = new_index;
@@ -3289,6 +3308,12 @@ glsl_to_tgsi_visitor::get_first_temp_read(int index)
             return (depth == 0) ? i : loop_start;
          }
       }
+      for (j=0; j < inst->tex_offset_num_offset; j++) {
+         if (inst->tex_offsets[j].file == PROGRAM_TEMPORARY && 
+             inst->tex_offsets[j].index == index) {
+            return (depth == 0) ? i : loop_start;
+         }
+      }
       
       if (inst->op == TGSI_OPCODE_BGNLOOP) {
          if(depth++ == 0)
@@ -3350,6 +3375,11 @@ glsl_to_tgsi_visitor::get_last_temp_read(int index)
             last = (depth == 0) ? i : -2;
          }
       }
+      for (j=0; j < inst->tex_offset_num_offset; j++) {
+          if (inst->tex_offsets[j].file == PROGRAM_TEMPORARY &&
+              inst->tex_offsets[j].index == index)
+              last = (depth == 0) ? i : -2;
+      }
       
       if (inst->op == TGSI_OPCODE_BGNLOOP)
          depth++;
@@ -3726,6 +3756,26 @@ glsl_to_tgsi_visitor::eliminate_dead_code_advanced(void)
                }
             }
          }
+         for (unsigned i = 0; i < inst->tex_offset_num_offset; i++) {
+            if (inst->tex_offsets[i].file == PROGRAM_TEMPORARY && inst->tex_offsets[i].reladdr){
+               /* Any temporary might be read, so no dead code elimination 
+                * across this instruction.
+                */
+               memset(writes, 0, sizeof(*writes) * this->next_temp * 4);
+            } else if (inst->tex_offsets[i].file == PROGRAM_TEMPORARY) {
+               /* Clear where it's used as src. */
+               int src_chans = 1 << GET_SWZ(inst->tex_offsets[i].swizzle, 0);
+               src_chans |= 1 << GET_SWZ(inst->tex_offsets[i].swizzle, 1);
+               src_chans |= 1 << GET_SWZ(inst->tex_offsets[i].swizzle, 2);
+               src_chans |= 1 << GET_SWZ(inst->tex_offsets[i].swizzle, 3);
+               
+               for (int c = 0; c < 4; c++) {
+              	   if (src_chans & (1 << c)) {
+              	      writes[4 * inst->tex_offsets[i].index + c] = NULL;
+              	   }
+               }
+            }
+         }
          break;
       }
 
@@ -4079,7 +4129,7 @@ struct st_translate {
    struct ureg_dst address[2];
    struct ureg_src samplers[PIPE_MAX_SAMPLERS];
    struct ureg_src systemValues[SYSTEM_VALUE_MAX];
-
+   struct tgsi_texture_offset tex_offsets[MAX_GLSL_TEXTURE_OFFSET];
    unsigned array_sizes[MAX_ARRAYS];
 
    const GLuint *inputMapping;
@@ -4379,22 +4429,34 @@ translate_src(struct st_translate *t, const st_src_reg *src_reg)
 
 static struct tgsi_texture_offset
 translate_tex_offset(struct st_translate *t,
-                     const struct tgsi_texture_offset *in_offset)
+                     const st_src_reg *in_offset, int idx)
 {
    struct tgsi_texture_offset offset;
    struct ureg_src imm_src;
 
-   assert(in_offset->File == PROGRAM_IMMEDIATE);
-   imm_src = t->immediates[in_offset->Index];
-
-   offset.File = imm_src.File;
-   offset.Index = imm_src.Index;
-   offset.SwizzleX = imm_src.SwizzleX;
-   offset.SwizzleY = imm_src.SwizzleY;
-   offset.SwizzleZ = imm_src.SwizzleZ;
-   offset.File = TGSI_FILE_IMMEDIATE;
-   offset.Padding = 0;
-
+   switch (in_offset->file) {
+   case PROGRAM_IMMEDIATE:
+      imm_src = t->immediates[in_offset->index];
+
+      offset.File = imm_src.File;
+      offset.Index = imm_src.Index;
+      offset.SwizzleX = imm_src.SwizzleX;
+      offset.SwizzleY = imm_src.SwizzleY;
+      offset.SwizzleZ = imm_src.SwizzleZ;
+      offset.Padding = 0;
+      break;
+   case PROGRAM_TEMPORARY:
+      imm_src = ureg_src(t->temps[in_offset->index]);
+      offset.File = imm_src.File;
+      offset.Index = imm_src.Index;
+      offset.SwizzleX = GET_SWZ(in_offset->swizzle, 0);
+      offset.SwizzleY = GET_SWZ(in_offset->swizzle, 1);
+      offset.SwizzleZ = GET_SWZ(in_offset->swizzle, 2);
+      offset.Padding = 0;
+      break;
+   default:
+      break;
+   }
    return offset;
 }
 
@@ -4450,9 +4512,10 @@ compile_tgsi_instruction(struct st_translate *t,
    case TGSI_OPCODE_TEX2:
    case TGSI_OPCODE_TXB2:
    case TGSI_OPCODE_TXL2:
+   case TGSI_OPCODE_TG4:
       src[num_src++] = t->samplers[inst->sampler];
       for (i = 0; i < inst->tex_offset_num_offset; i++) {
-         texoffsets[i] = translate_tex_offset(t, &inst->tex_offsets[i]);
+         texoffsets[i] = translate_tex_offset(t, &inst->tex_offsets[i], i);
       }
       tex_target = st_translate_texture_target(inst->tex_target, inst->tex_shadow);
 
@@ -5140,7 +5203,7 @@ get_mesa_program(struct gl_context *ctx,
       printf("GLSL IR for linked %s program %d:\n",
              _mesa_shader_stage_to_string(shader->Stage),
              shader_program->Name);
-      _mesa_print_ir(shader->ir, NULL);
+      _mesa_print_ir(stdout, shader->ir, NULL);
       printf("\n");
       printf("\n");
       fflush(stdout);
@@ -5182,6 +5245,7 @@ get_mesa_program(struct gl_context *ctx,
       stgp->Base.InputType = shader_program->Geom.InputType;
       stgp->Base.OutputType = shader_program->Geom.OutputType;
       stgp->Base.VerticesOut = shader_program->Geom.VerticesOut;
+      stgp->Base.Invocations = shader_program->Geom.Invocations;
       break;
    default:
       assert(!"should not be reached");
@@ -5268,6 +5332,7 @@ st_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
          lower_packing_builtins(ir, lower_inst);
       }
 
+      lower_offset_arrays(ir);
       do_mat_op_to_vec(ir);
       lower_instructions(ir,
                          MOD_TO_FRACT |
diff --git a/mesalib/src/mesa/state_tracker/st_manager.c b/mesalib/src/mesa/state_tracker/st_manager.c
index 215b27634..6a776a8a2 100644
--- a/mesalib/src/mesa/state_tracker/st_manager.c
+++ b/mesalib/src/mesa/state_tracker/st_manager.c
@@ -30,6 +30,7 @@
 #include "main/texobj.h"
 #include "main/teximage.h"
 #include "main/texstate.h"
+#include "main/errors.h"
 #include "main/framebuffer.h"
 #include "main/fbobject.h"
 #include "main/renderbuffer.h"
@@ -626,10 +627,14 @@ st_api_create_context(struct st_api *stapi, struct st_manager *smapi,
       return NULL;
    }
 
-   st->ctx->Debug.DebugOutput = GL_FALSE;
    if (attribs->flags & ST_CONTEXT_FLAG_DEBUG){
+      struct gl_debug_state *debug = _mesa_get_debug_state(st->ctx);
+      if (!debug) {
+         *error = ST_CONTEXT_ERROR_NO_MEMORY;
+         return NULL;
+      }
       st->ctx->Const.ContextFlags |= GL_CONTEXT_FLAG_DEBUG_BIT;
-      st->ctx->Debug.DebugOutput = GL_TRUE;
+      debug->DebugOutput = GL_TRUE;
    }
 
    if (attribs->flags & ST_CONTEXT_FLAG_FORWARD_COMPATIBLE)
diff --git a/mesalib/src/mesa/state_tracker/st_program.c b/mesalib/src/mesa/state_tracker/st_program.c
index cadbe1705..e9074ac97 100644
--- a/mesalib/src/mesa/state_tracker/st_program.c
+++ b/mesalib/src/mesa/state_tracker/st_program.c
@@ -572,6 +572,11 @@ st_translate_fragment_program(struct st_context *st,
             input_semantic_index[slot] = 0;
             interpMode[slot] = TGSI_INTERPOLATE_CONSTANT;
             break;
+         case VARYING_SLOT_VIEWPORT:
+            input_semantic_name[slot] = TGSI_SEMANTIC_VIEWPORT_INDEX;
+            input_semantic_index[slot] = 0;
+            interpMode[slot] = TGSI_INTERPOLATE_CONSTANT;
+            break;
          case VARYING_SLOT_CLIP_DIST0:
             input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
             input_semantic_index[slot] = 0;
@@ -1027,6 +1032,10 @@ st_translate_geometry_program(struct st_context *st,
             gs_output_semantic_name[slot] = TGSI_SEMANTIC_PRIMID;
             gs_output_semantic_index[slot] = 0;
             break;
+         case VARYING_SLOT_VIEWPORT:
+            gs_output_semantic_name[slot] = TGSI_SEMANTIC_VIEWPORT_INDEX;
+            gs_output_semantic_index[slot] = 0;
+            break;
          case VARYING_SLOT_TEX0:
          case VARYING_SLOT_TEX1:
          case VARYING_SLOT_TEX2:
@@ -1087,6 +1096,7 @@ st_translate_geometry_program(struct st_context *st,
    ureg_property_gs_input_prim(ureg, stgp->Base.InputType);
    ureg_property_gs_output_prim(ureg, stgp->Base.OutputType);
    ureg_property_gs_max_vertices(ureg, stgp->Base.VerticesOut);
+   ureg_property_gs_invocations(ureg, stgp->Base.Invocations);
 
    if (stgp->glsl_to_tgsi)
       st_translate_program(st->ctx,
diff --git a/mesalib/src/mesa/swrast/s_depth.c b/mesalib/src/mesa/swrast/s_depth.c
index 93aaffc57..8c225e111 100644
--- a/mesalib/src/mesa/swrast/s_depth.c
+++ b/mesalib/src/mesa/swrast/s_depth.c
@@ -552,8 +552,8 @@ _swrast_clear_depth_buffer(struct gl_context *ctx)
    height = ctx->DrawBuffer->_Ymax - ctx->DrawBuffer->_Ymin;
 
    mapMode = GL_MAP_WRITE_BIT;
-   if (rb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT ||
-       rb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT ||
+   if (rb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT ||
+       rb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT ||
        rb->Format == MESA_FORMAT_S8_UINT_Z24_UNORM ||
        rb->Format == MESA_FORMAT_X8Z24_UNORM) {
       mapMode |= GL_MAP_READ_BIT;
@@ -602,8 +602,8 @@ _swrast_clear_depth_buffer(struct gl_context *ctx)
          }
       }
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_Z24_UNORM_S8_UINT:
+   case MESA_FORMAT_Z24_UNORM_X8_UINT:
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
    case MESA_FORMAT_X8Z24_UNORM:
       {
@@ -611,8 +611,8 @@ _swrast_clear_depth_buffer(struct gl_context *ctx)
          GLuint clearVal = 0;
          GLuint mask;
 
-         if (rb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT ||
-             rb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT)
+         if (rb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT ||
+             rb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT)
             mask = 0xff000000;
          else
             mask = 0xff;
@@ -692,7 +692,7 @@ _swrast_clear_depth_stencil_buffer(struct gl_context *ctx)
    }
 
    switch (rb->Format) {
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
    case MESA_FORMAT_S8_UINT_Z24_UNORM:
       {
          GLfloat zClear = (GLfloat) ctx->Depth.Clear;
@@ -700,7 +700,7 @@ _swrast_clear_depth_stencil_buffer(struct gl_context *ctx)
 
          _mesa_pack_float_z_row(rb->Format, 1, &zClear, &clear);
 
-         if (rb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT) {
+         if (rb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT) {
             mask = ((~writeMask) & 0xff) << 24;
             clear |= (ctx->Stencil.Clear & writeMask & 0xff) << 24;
          }
diff --git a/mesalib/src/mesa/swrast/s_drawpix.c b/mesalib/src/mesa/swrast/s_drawpix.c
index ed4487f3e..f7926e426 100644
--- a/mesalib/src/mesa/swrast/s_drawpix.c
+++ b/mesalib/src/mesa/swrast/s_drawpix.c
@@ -500,7 +500,7 @@ draw_rgba_pixels( struct gl_context *ctx, GLint x, GLint y,
 
 
 /**
- * Draw depth+stencil values into a MESA_FORAMT_Z24_S8 or MESA_FORMAT_Z24_UNORM_X8_UINT
+ * Draw depth+stencil values into a MESA_FORAMT_Z24_S8 or MESA_FORMAT_Z24_UNORM_S8_UINT
  * renderbuffer.  No masking, zooming, scaling, etc.
  */
 static void
@@ -573,7 +573,7 @@ draw_depth_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
 
    if (depthRb == stencilRb &&
        (depthRb->Format == MESA_FORMAT_S8_UINT_Z24_UNORM ||
-        depthRb->Format == MESA_FORMAT_Z24_UNORM_X8_UINT) &&
+        depthRb->Format == MESA_FORMAT_Z24_UNORM_S8_UINT) &&
        type == GL_UNSIGNED_INT_24_8 &&
        !scaleOrBias &&
        !zoom &&
diff --git a/mesalib/src/mesa/swrast/s_renderbuffer.c b/mesalib/src/mesa/swrast/s_renderbuffer.c
index cf5e2230d..dfd3a6057 100644
--- a/mesalib/src/mesa/swrast/s_renderbuffer.c
+++ b/mesalib/src/mesa/swrast/s_renderbuffer.c
@@ -102,7 +102,7 @@ soft_renderbuffer_storage(struct gl_context *ctx, struct gl_renderbuffer *rb,
       rb->Format = MESA_FORMAT_Z_UNORM16;
       break;
    case GL_DEPTH_COMPONENT24:
-      rb->Format = MESA_FORMAT_Z24_UNORM_S8_UINT;
+      rb->Format = MESA_FORMAT_Z24_UNORM_X8_UINT;
       break;
    case GL_DEPTH_COMPONENT32:
       rb->Format = MESA_FORMAT_Z_UNORM32;
diff --git a/mesalib/src/mesa/swrast/s_stencil.c b/mesalib/src/mesa/swrast/s_stencil.c
index e7c35ff69..eba9da863 100644
--- a/mesalib/src/mesa/swrast/s_stencil.c
+++ b/mesalib/src/mesa/swrast/s_stencil.c
@@ -614,7 +614,7 @@ _swrast_clear_stencil_buffer(struct gl_context *ctx)
          }
       }
       break;
-   case MESA_FORMAT_Z24_UNORM_X8_UINT:
+   case MESA_FORMAT_Z24_UNORM_S8_UINT:
       {
          GLuint clear = (ctx->Stencil.Clear & writeMask & 0xff) << 24;
          GLuint mask = (((~writeMask) & 0xff) << 24) | 0xffffff;
diff --git a/mesalib/src/mesa/swrast/s_texfetch.c b/mesalib/src/mesa/swrast/s_texfetch.c
index 7a3e76f85..b620748c9 100644
--- a/mesalib/src/mesa/swrast/s_texfetch.c
+++ b/mesalib/src/mesa/swrast/s_texfetch.c
@@ -383,7 +383,7 @@ texfetch_funcs[] =
       fetch_texel_3d_f_z24_s8
    },
    {
-      MESA_FORMAT_Z24_UNORM_X8_UINT,
+      MESA_FORMAT_Z24_UNORM_S8_UINT,
       fetch_texel_1d_f_s8_z24,
       fetch_texel_2d_f_s8_z24,
       fetch_texel_3d_f_s8_z24
@@ -395,7 +395,7 @@ texfetch_funcs[] =
       fetch_texel_3d_f_z16
    },
    {
-      MESA_FORMAT_Z24_UNORM_S8_UINT,
+      MESA_FORMAT_Z24_UNORM_X8_UINT,
       fetch_texel_1d_f_s8_z24,
       fetch_texel_2d_f_s8_z24,
       fetch_texel_3d_f_s8_z24
diff --git a/mesalib/src/mesa/swrast/s_texfetch_tmp.h b/mesalib/src/mesa/swrast/s_texfetch_tmp.h
index f3048894e..806f0fd1a 100644
--- a/mesalib/src/mesa/swrast/s_texfetch_tmp.h
+++ b/mesalib/src/mesa/swrast/s_texfetch_tmp.h
@@ -1556,8 +1556,8 @@ static void FETCH(f_s8_z24)( const struct swrast_texture_image *texImage,
    const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
    const GLdouble scale = 1.0 / (GLdouble) 0xffffff;
    texel[0] = (GLfloat) (((*src) & 0x00ffffff) * scale);
-   ASSERT(texImage->Base.TexFormat == MESA_FORMAT_Z24_UNORM_X8_UINT ||
-	  texImage->Base.TexFormat == MESA_FORMAT_Z24_UNORM_S8_UINT);
+   ASSERT(texImage->Base.TexFormat == MESA_FORMAT_Z24_UNORM_S8_UINT ||
+	  texImage->Base.TexFormat == MESA_FORMAT_Z24_UNORM_X8_UINT);
    ASSERT(texel[0] >= 0.0F);
    ASSERT(texel[0] <= 1.0F);
 }
diff --git a/mesalib/src/mesa/tnl/t_draw.c b/mesalib/src/mesa/tnl/t_draw.c
index 8e5af9627..2755ae62d 100644
--- a/mesalib/src/mesa/tnl/t_draw.c
+++ b/mesalib/src/mesa/tnl/t_draw.c
@@ -279,17 +279,18 @@ static void bind_inputs( struct gl_context *ctx,
       const void *ptr;
 
       if (inputs[i]->BufferObj->Name) { 
-	 if (!inputs[i]->BufferObj->Pointer) {
+	 if (!inputs[i]->BufferObj->Mappings[MAP_INTERNAL].Pointer) {
 	    bo[*nr_bo] = inputs[i]->BufferObj;
 	    (*nr_bo)++;
 	    ctx->Driver.MapBufferRange(ctx, 0, inputs[i]->BufferObj->Size,
 				       GL_MAP_READ_BIT,
-				       inputs[i]->BufferObj);
+				       inputs[i]->BufferObj,
+                                       MAP_INTERNAL);
 	    
-	    assert(inputs[i]->BufferObj->Pointer);
+	    assert(inputs[i]->BufferObj->Mappings[MAP_INTERNAL].Pointer);
 	 }
 	 
-	 ptr = ADD_POINTERS(inputs[i]->BufferObj->Pointer,
+	 ptr = ADD_POINTERS(inputs[i]->BufferObj->Mappings[MAP_INTERNAL].Pointer,
 			    inputs[i]->Ptr);
       }
       else
@@ -348,17 +349,19 @@ static void bind_indices( struct gl_context *ctx,
       return;
    }
 
-   if (_mesa_is_bufferobj(ib->obj) && !_mesa_bufferobj_mapped(ib->obj)) {
+   if (_mesa_is_bufferobj(ib->obj) &&
+       !_mesa_bufferobj_mapped(ib->obj, MAP_INTERNAL)) {
       /* if the buffer object isn't mapped yet, map it now */
       bo[*nr_bo] = ib->obj;
       (*nr_bo)++;
       ptr = ctx->Driver.MapBufferRange(ctx, (GLsizeiptr) ib->ptr,
                                        ib->count * vbo_sizeof_ib_type(ib->type),
-				       GL_MAP_READ_BIT, ib->obj);
-      assert(ib->obj->Pointer);
+				       GL_MAP_READ_BIT, ib->obj,
+                                       MAP_INTERNAL);
+      assert(ib->obj->Mappings[MAP_INTERNAL].Pointer);
    } else {
       /* user-space elements, or buffer already mapped */
-      ptr = ADD_POINTERS(ib->obj->Pointer, ib->ptr);
+      ptr = ADD_POINTERS(ib->obj->Mappings[MAP_INTERNAL].Pointer, ib->ptr);
    }
 
    if (ib->type == GL_UNSIGNED_INT && VB->Primitive[0].basevertex == 0) {
@@ -403,7 +406,7 @@ static void unmap_vbos( struct gl_context *ctx,
 {
    GLuint i;
    for (i = 0; i < nr_bo; i++) { 
-      ctx->Driver.UnmapBuffer(ctx, bo[i]);
+      ctx->Driver.UnmapBuffer(ctx, bo[i], MAP_INTERNAL);
    }
 }
 
diff --git a/mesalib/src/mesa/vbo/vbo_exec_api.c b/mesalib/src/mesa/vbo/vbo_exec_api.c
index 810913645..74aec127c 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_api.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_api.c
@@ -998,7 +998,11 @@ void vbo_use_buffer_objects(struct gl_context *ctx)
    /* Allocate a real buffer object now */
    _mesa_reference_buffer_object(ctx, &exec->vtx.bufferobj, NULL);
    exec->vtx.bufferobj = ctx->Driver.NewBufferObject(ctx, bufName, target);
-   if (!ctx->Driver.BufferData(ctx, target, size, NULL, usage, exec->vtx.bufferobj)) {
+   if (!ctx->Driver.BufferData(ctx, target, size, NULL, usage,
+                               GL_MAP_WRITE_BIT |
+                               GL_DYNAMIC_STORAGE_BIT |
+                               GL_CLIENT_STORAGE_BIT,
+                               exec->vtx.bufferobj)) {
       _mesa_error(ctx, GL_OUT_OF_MEMORY, "VBO allocation");
    }
 }
@@ -1114,8 +1118,8 @@ void vbo_exec_vtx_destroy( struct vbo_exec_context *exec )
 
    /* Free the vertex buffer.  Unmap first if needed.
     */
-   if (_mesa_bufferobj_mapped(exec->vtx.bufferobj)) {
-      ctx->Driver.UnmapBuffer(ctx, exec->vtx.bufferobj);
+   if (_mesa_bufferobj_mapped(exec->vtx.bufferobj, MAP_INTERNAL)) {
+      ctx->Driver.UnmapBuffer(ctx, exec->vtx.bufferobj, MAP_INTERNAL);
    }
    _mesa_reference_buffer_object(ctx, &exec->vtx.bufferobj, NULL);
 }
diff --git a/mesalib/src/mesa/vbo/vbo_exec_array.c b/mesalib/src/mesa/vbo/vbo_exec_array.c
index b316f097d..07ce08fef 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_array.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_array.c
@@ -53,7 +53,7 @@ check_buffers_are_unmapped(const struct gl_client_array **inputs)
    for (i = 0; i < VERT_ATTRIB_MAX; i++) {
       if (inputs[i]) {
          struct gl_buffer_object *obj = inputs[i]->BufferObj;
-         assert(!_mesa_bufferobj_mapped(obj));
+         assert(!_mesa_check_disallowed_mapping(obj));
          (void) obj;
       }
    }
@@ -73,7 +73,7 @@ vbo_check_buffers_are_unmapped(struct gl_context *ctx)
    /* check the current vertex arrays */
    check_buffers_are_unmapped(exec->array.inputs);
    /* check the current glBegin/glVertex/glEnd-style VBO */
-   assert(!_mesa_bufferobj_mapped(exec->vtx.bufferobj));
+   assert(!_mesa_check_disallowed_mapping(exec->vtx.bufferobj));
 }
 
 
@@ -101,7 +101,8 @@ vbo_get_minmax_index(struct gl_context *ctx,
    if (_mesa_is_bufferobj(ib->obj)) {
       GLsizeiptr size = MIN2(count * index_size, ib->obj->Size);
       indices = ctx->Driver.MapBufferRange(ctx, (GLintptr) indices, size,
-                                           GL_MAP_READ_BIT, ib->obj);
+                                           GL_MAP_READ_BIT, ib->obj,
+                                           MAP_INTERNAL);
    }
 
    switch (ib->type) {
@@ -177,7 +178,7 @@ vbo_get_minmax_index(struct gl_context *ctx,
    }
 
    if (_mesa_is_bufferobj(ib->obj)) {
-      ctx->Driver.UnmapBuffer(ctx, ib->obj);
+      ctx->Driver.UnmapBuffer(ctx, ib->obj, MAP_INTERNAL);
    }
 }
 
@@ -229,13 +230,15 @@ check_array_data(struct gl_context *ctx, struct gl_client_array *array,
    if (array->Enabled) {
       const void *data = array->Ptr;
       if (_mesa_is_bufferobj(array->BufferObj)) {
-         if (!array->BufferObj->Pointer) {
+         if (!array->BufferObj->Mappings[MAP_INTERNAL].Pointer) {
             /* need to map now */
-            array->BufferObj->Pointer =
+            array->BufferObj->Mappings[MAP_INTERNAL].Pointer =
                ctx->Driver.MapBufferRange(ctx, 0, array->BufferObj->Size,
-					  GL_MAP_READ_BIT, array->BufferObj);
+					  GL_MAP_READ_BIT, array->BufferObj,
+                                          MAP_INTERNAL);
          }
-         data = ADD_POINTERS(data, array->BufferObj->Pointer);
+         data = ADD_POINTERS(data,
+                             array->BufferObj->Mappings[MAP_INTERNAL].Pointer);
       }
       switch (array->Type) {
       case GL_FLOAT:
@@ -273,8 +276,8 @@ unmap_array_buffer(struct gl_context *ctx, struct gl_client_array *array)
 {
    if (array->Enabled &&
        _mesa_is_bufferobj(array->BufferObj) &&
-       _mesa_bufferobj_mapped(array->BufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, array->BufferObj);
+       _mesa_bufferobj_mapped(array->BufferObj, MAP_INTERNAL)) {
+      ctx->Driver.UnmapBuffer(ctx, array->BufferObj, MAP_INTERNAL);
    }
 }
 
@@ -295,7 +298,8 @@ check_draw_elements_data(struct gl_context *ctx, GLsizei count, GLenum elemType,
       elemMap = ctx->Driver.MapBufferRange(ctx, 0,
 					   ctx->Array.VAO->IndexBufferObj->Size,
 					   GL_MAP_READ_BIT,
-					   ctx->Array.VAO->IndexBufferObj);
+					   ctx->Array.VAO->IndexBufferObj,
+                                           MAP_INTERNAL);
       elements = ADD_POINTERS(elements, elemMap);
    }
 
@@ -324,7 +328,8 @@ check_draw_elements_data(struct gl_context *ctx, GLsizei count, GLenum elemType,
    }
 
    if (_mesa_is_bufferobj(vao->IndexBufferObj)) {
-      ctx->Driver.UnmapBuffer(ctx, ctx->Array.VAO->IndexBufferObj);
+      ctx->Driver.UnmapBuffer(ctx, ctx->Array.VAO->IndexBufferObj,
+                              MAP_INTERNAL);
    }
 
    for (k = 0; k < Elements(vao->_VertexAttrib); k++) {
@@ -374,7 +379,8 @@ print_draw_arrays(struct gl_context *ctx,
 
       if (bufName) {
          GLubyte *p = ctx->Driver.MapBufferRange(ctx, 0, bufObj->Size,
-						 GL_MAP_READ_BIT, bufObj);
+						 GL_MAP_READ_BIT, bufObj,
+                                                 MAP_INTERNAL);
          int offset = (int) (GLintptr) exec->array.inputs[i]->Ptr;
          float *f = (float *) (p + offset);
          int *k = (int *) f;
@@ -386,7 +392,7 @@ print_draw_arrays(struct gl_context *ctx,
          for (i = 0; i < n; i++) {
             printf("    float[%d] = 0x%08x %f\n", i, k[i], f[i]);
          }
-         ctx->Driver.UnmapBuffer(ctx, bufObj);
+         ctx->Driver.UnmapBuffer(ctx, bufObj, MAP_INTERNAL);
       }
    }
 }
@@ -885,7 +891,8 @@ dump_element_buffer(struct gl_context *ctx, GLenum type)
       ctx->Driver.MapBufferRange(ctx, 0,
 				 ctx->Array.VAO->IndexBufferObj->Size,
 				 GL_MAP_READ_BIT,
-				 ctx->Array.VAO->IndexBufferObj);
+                                 ctx->Array.VAO->IndexBufferObj,
+                                 MAP_INTERNAL);
    switch (type) {
    case GL_UNSIGNED_BYTE:
       {
@@ -927,7 +934,8 @@ dump_element_buffer(struct gl_context *ctx, GLenum type)
       ;
    }
 
-   ctx->Driver.UnmapBuffer(ctx, ctx->Array.VAO->IndexBufferObj);
+   ctx->Driver.UnmapBuffer(ctx, ctx->Array.VAO->IndexBufferObj,
+                           MAP_INTERNAL);
 }
 #endif
 
diff --git a/mesalib/src/mesa/vbo/vbo_exec_draw.c b/mesalib/src/mesa/vbo/vbo_exec_draw.c
index 75115ce66..c454c6943 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_draw.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_draw.c
@@ -220,9 +220,10 @@ vbo_exec_bind_arrays( struct gl_context *ctx )
 
          if (_mesa_is_bufferobj(exec->vtx.bufferobj)) {
             /* a real buffer obj: Ptr is an offset, not a pointer*/
-            assert(exec->vtx.bufferobj->Pointer);  /* buf should be mapped */
+            assert(exec->vtx.bufferobj->Mappings[MAP_INTERNAL].Pointer);
             assert(offset >= 0);
-            arrays[attr].Ptr = (GLubyte *)exec->vtx.bufferobj->Offset + offset;
+            arrays[attr].Ptr = (GLubyte *)
+               exec->vtx.bufferobj->Mappings[MAP_INTERNAL].Offset + offset;
          }
          else {
             /* Ptr into ordinary app memory */
@@ -261,12 +262,15 @@ vbo_exec_vtx_unmap( struct vbo_exec_context *exec )
       struct gl_context *ctx = exec->ctx;
       
       if (ctx->Driver.FlushMappedBufferRange) {
-         GLintptr offset = exec->vtx.buffer_used - exec->vtx.bufferobj->Offset;
-         GLsizeiptr length = (exec->vtx.buffer_ptr - exec->vtx.buffer_map) * sizeof(float);
+         GLintptr offset = exec->vtx.buffer_used -
+                           exec->vtx.bufferobj->Mappings[MAP_INTERNAL].Offset;
+         GLsizeiptr length = (exec->vtx.buffer_ptr - exec->vtx.buffer_map) *
+                             sizeof(float);
 
          if (length)
             ctx->Driver.FlushMappedBufferRange(ctx, offset, length,
-                                               exec->vtx.bufferobj);
+                                               exec->vtx.bufferobj,
+                                               MAP_INTERNAL);
       }
 
       exec->vtx.buffer_used += (exec->vtx.buffer_ptr -
@@ -275,7 +279,7 @@ vbo_exec_vtx_unmap( struct vbo_exec_context *exec )
       assert(exec->vtx.buffer_used <= VBO_VERT_BUFFER_SIZE);
       assert(exec->vtx.buffer_ptr != NULL);
       
-      ctx->Driver.UnmapBuffer(ctx, exec->vtx.bufferobj);
+      ctx->Driver.UnmapBuffer(ctx, exec->vtx.bufferobj, MAP_INTERNAL);
       exec->vtx.buffer_map = NULL;
       exec->vtx.buffer_ptr = NULL;
       exec->vtx.max_vert = 0;
@@ -312,7 +316,8 @@ vbo_exec_vtx_map( struct vbo_exec_context *exec )
                                                   (VBO_VERT_BUFFER_SIZE - 
                                                    exec->vtx.buffer_used),
                                                   accessRange,
-                                                  exec->vtx.bufferobj);
+                                                  exec->vtx.bufferobj,
+                                                  MAP_INTERNAL);
          exec->vtx.buffer_ptr = exec->vtx.buffer_map;
       }
       else {
@@ -325,14 +330,19 @@ vbo_exec_vtx_map( struct vbo_exec_context *exec )
       exec->vtx.buffer_used = 0;
 
       if (ctx->Driver.BufferData(ctx, GL_ARRAY_BUFFER_ARB,
-                                  VBO_VERT_BUFFER_SIZE, 
-                                  NULL, usage, exec->vtx.bufferobj)) {
+                                 VBO_VERT_BUFFER_SIZE,
+                                 NULL, usage,
+                                 GL_MAP_WRITE_BIT |
+                                 GL_DYNAMIC_STORAGE_BIT |
+                                 GL_CLIENT_STORAGE_BIT,
+                                 exec->vtx.bufferobj)) {
          /* buffer allocation worked, now map the buffer */
          exec->vtx.buffer_map =
             (GLfloat *)ctx->Driver.MapBufferRange(ctx,
                                                   0, VBO_VERT_BUFFER_SIZE,
                                                   accessRange,
-                                                  exec->vtx.bufferobj);
+                                                  exec->vtx.bufferobj,
+                                                  MAP_INTERNAL);
       }
       else {
          _mesa_error(ctx, GL_OUT_OF_MEMORY, "VBO allocation");
diff --git a/mesalib/src/mesa/vbo/vbo_primitive_restart.c b/mesalib/src/mesa/vbo/vbo_primitive_restart.c
index 9c3055431..25c89669c 100644
--- a/mesalib/src/mesa/vbo/vbo_primitive_restart.c
+++ b/mesalib/src/mesa/vbo/vbo_primitive_restart.c
@@ -177,7 +177,7 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
    struct _mesa_prim temp_prim;
    struct vbo_context *vbo = vbo_context(ctx);
    vbo_draw_func draw_prims_func = vbo->draw_prims;
-   GLboolean map_ib = ib->obj->Name && !ib->obj->Pointer;
+   GLboolean map_ib = ib->obj->Name && !ib->obj->Mappings[MAP_INTERNAL].Pointer;
    void *ptr;
 
    /* If there is an indirect buffer, map it and extract the draw params */
@@ -186,7 +186,7 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
       struct _mesa_index_buffer new_ib = *ib;
       const uint32_t *indirect_params;
       if (!ctx->Driver.MapBufferRange(ctx, 0, indirect->Size, GL_MAP_READ_BIT,
-            indirect)) {
+                                      indirect, MAP_INTERNAL)) {
 
          /* something went wrong with mapping, give up */
          _mesa_error(ctx, GL_OUT_OF_MEMORY,
@@ -195,8 +195,9 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
       }
 
       assert(nr_prims == 1);
-      indirect_params = (const uint32_t *) ADD_POINTERS(indirect->Pointer,
-            new_prim.indirect_offset);
+      indirect_params = (const uint32_t *)
+                        ADD_POINTERS(indirect->Mappings[MAP_INTERNAL].Pointer,
+                                     new_prim.indirect_offset);
 
       new_prim.is_indirect = 0;
       new_prim.count = indirect_params[0];
@@ -210,7 +211,7 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
       prims = &new_prim;
       ib = &new_ib;
 
-      ctx->Driver.UnmapBuffer(ctx, indirect);
+      ctx->Driver.UnmapBuffer(ctx, indirect, MAP_INTERNAL);
    }
 
    /* Find the sub-primitives. These are regions in the index buffer which
@@ -218,17 +219,17 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
     */
    if (map_ib) {
       ctx->Driver.MapBufferRange(ctx, 0, ib->obj->Size, GL_MAP_READ_BIT,
-                                 ib->obj);
+                                 ib->obj, MAP_INTERNAL);
    }
 
-   ptr = ADD_POINTERS(ib->obj->Pointer, ib->ptr);
+   ptr = ADD_POINTERS(ib->obj->Mappings[MAP_INTERNAL].Pointer, ib->ptr);
 
    sub_prims = find_sub_primitives(ptr, vbo_sizeof_ib_type(ib->type),
                                    0, ib->count, restart_index,
                                    &num_sub_prims);
 
    if (map_ib) {
-      ctx->Driver.UnmapBuffer(ctx, ib->obj);
+      ctx->Driver.UnmapBuffer(ctx, ib->obj, MAP_INTERNAL);
    }
 
    /* Loop over the primitives, and use the located sub-primitives to draw
diff --git a/mesalib/src/mesa/vbo/vbo_rebase.c b/mesalib/src/mesa/vbo/vbo_rebase.c
index c700621ce..f3fe5f757 100644
--- a/mesalib/src/mesa/vbo/vbo_rebase.c
+++ b/mesalib/src/mesa/vbo/vbo_rebase.c
@@ -157,15 +157,16 @@ void vbo_rebase_prims( struct gl_context *ctx,
    } else if (ib) {
       /* Unfortunately need to adjust each index individually.
        */
-      GLboolean map_ib = ib->obj->Name && !ib->obj->Pointer;
+      GLboolean map_ib = ib->obj->Name &&
+                         !ib->obj->Mappings[MAP_INTERNAL].Pointer;
       void *ptr;
 
       if (map_ib) 
 	 ctx->Driver.MapBufferRange(ctx, 0, ib->obj->Size, GL_MAP_READ_BIT,
-				    ib->obj);
+				    ib->obj, MAP_INTERNAL);
 
 
-      ptr = ADD_POINTERS(ib->obj->Pointer, ib->ptr);
+      ptr = ADD_POINTERS(ib->obj->Mappings[MAP_INTERNAL].Pointer, ib->ptr);
 
       /* Some users might prefer it if we translated elements to
        * GLuints here.  Others wouldn't...
@@ -183,7 +184,7 @@ void vbo_rebase_prims( struct gl_context *ctx,
       }      
 
       if (map_ib) 
-	 ctx->Driver.UnmapBuffer(ctx, ib->obj);
+	 ctx->Driver.UnmapBuffer(ctx, ib->obj, MAP_INTERNAL);
 
       tmp_ib.obj = ctx->Shared->NullBufferObj;
       tmp_ib.ptr = tmp_indices;
diff --git a/mesalib/src/mesa/vbo/vbo_save_api.c b/mesalib/src/mesa/vbo/vbo_save_api.c
index 98ab68623..1c6b91942 100644
--- a/mesalib/src/mesa/vbo/vbo_save_api.c
+++ b/mesalib/src/mesa/vbo/vbo_save_api.c
@@ -200,6 +200,8 @@ alloc_vertex_store(struct gl_context *ctx)
                                  GL_ARRAY_BUFFER_ARB,
                                  VBO_SAVE_BUFFER_SIZE * sizeof(GLfloat),
                                  NULL, GL_STATIC_DRAW_ARB,
+                                 GL_MAP_WRITE_BIT |
+                                 GL_DYNAMIC_STORAGE_BIT,
                                  vertex_store->bufferobj);
    }
    else {
@@ -251,7 +253,8 @@ vbo_save_map_vertex_store(struct gl_context *ctx,
       GLsizeiptr size = vertex_store->bufferobj->Size - offset;
       GLfloat *range = (GLfloat *)
          ctx->Driver.MapBufferRange(ctx, offset, size, access,
-                                    vertex_store->bufferobj);
+                                    vertex_store->bufferobj,
+                                    MAP_INTERNAL);
       if (range) {
          /* compute address of start of whole buffer (needed elsewhere) */
          vertex_store->buffer = range - vertex_store->used;
@@ -277,13 +280,14 @@ vbo_save_unmap_vertex_store(struct gl_context *ctx,
    if (vertex_store->bufferobj->Size > 0) {
       GLintptr offset = 0;
       GLsizeiptr length = vertex_store->used * sizeof(GLfloat)
-         - vertex_store->bufferobj->Offset;
+         - vertex_store->bufferobj->Mappings[MAP_INTERNAL].Offset;
 
       /* Explicitly flush the region we wrote to */
       ctx->Driver.FlushMappedBufferRange(ctx, offset, length,
-                                         vertex_store->bufferobj);
+                                         vertex_store->bufferobj,
+                                         MAP_INTERNAL);
 
-      ctx->Driver.UnmapBuffer(ctx, vertex_store->bufferobj);
+      ctx->Driver.UnmapBuffer(ctx, vertex_store->bufferobj, MAP_INTERNAL);
    }
    vertex_store->buffer = NULL;
 }
@@ -1116,6 +1120,7 @@ _save_OBE_DrawElements(GLenum mode, GLsizei count, GLenum type,
 {
    GET_CURRENT_CONTEXT(ctx);
    struct vbo_save_context *save = &vbo_context(ctx)->save;
+   struct gl_buffer_object *indexbuf = ctx->Array.VAO->IndexBufferObj;
    GLint i;
 
    if (!_mesa_is_valid_prim_mode(ctx, mode)) {
@@ -1138,9 +1143,9 @@ _save_OBE_DrawElements(GLenum mode, GLsizei count, GLenum type,
 
    _ae_map_vbos(ctx);
 
-   if (_mesa_is_bufferobj(ctx->Array.VAO->IndexBufferObj))
+   if (_mesa_is_bufferobj(indexbuf))
       indices =
-         ADD_POINTERS(ctx->Array.VAO->IndexBufferObj->Pointer, indices);
+         ADD_POINTERS(indexbuf->Mappings[MAP_INTERNAL].Pointer, indices);
 
    vbo_save_NotifyBegin(ctx, (mode | VBO_SAVE_PRIM_WEAK |
                               VBO_SAVE_PRIM_NO_CURRENT_UPDATE));
diff --git a/mesalib/src/mesa/vbo/vbo_save_draw.c b/mesalib/src/mesa/vbo/vbo_save_draw.c
index 4b13cc4e5..89fd30e0f 100644
--- a/mesalib/src/mesa/vbo/vbo_save_draw.c
+++ b/mesalib/src/mesa/vbo/vbo_save_draw.c
@@ -225,7 +225,8 @@ vbo_save_loopback_vertex_list(struct gl_context *ctx,
       ctx->Driver.MapBufferRange(ctx, 0,
 				 list->vertex_store->bufferobj->Size,
 				 GL_MAP_READ_BIT, /* ? */
-				 list->vertex_store->bufferobj);
+				 list->vertex_store->bufferobj,
+                                 MAP_INTERNAL);
 
    vbo_loopback_vertex_list(ctx,
                             (const GLfloat *)(buffer + list->buffer_offset),
@@ -235,7 +236,8 @@ vbo_save_loopback_vertex_list(struct gl_context *ctx,
                             list->wrap_count,
                             list->vertex_size);
 
-   ctx->Driver.UnmapBuffer(ctx, list->vertex_store->bufferobj);
+   ctx->Driver.UnmapBuffer(ctx, list->vertex_store->bufferobj,
+                           MAP_INTERNAL);
 }
 
 
diff --git a/mesalib/src/mesa/vbo/vbo_split_copy.c b/mesalib/src/mesa/vbo/vbo_split_copy.c
index 4332ef2e6..719ad651a 100644
--- a/mesalib/src/mesa/vbo/vbo_split_copy.c
+++ b/mesalib/src/mesa/vbo/vbo_split_copy.c
@@ -451,11 +451,14 @@ replay_init( struct copy_context *copy )
 	 copy->varying[j].size = attr_size(copy->array[i]);
 	 copy->vertex_size += attr_size(copy->array[i]);
       
-	 if (_mesa_is_bufferobj(vbo) && !_mesa_bufferobj_mapped(vbo)) 
-	    ctx->Driver.MapBufferRange(ctx, 0, vbo->Size, GL_MAP_READ_BIT, vbo);
+	 if (_mesa_is_bufferobj(vbo) &&
+             !_mesa_bufferobj_mapped(vbo, MAP_INTERNAL))
+	    ctx->Driver.MapBufferRange(ctx, 0, vbo->Size, GL_MAP_READ_BIT, vbo,
+                                       MAP_INTERNAL);
 
-	 copy->varying[j].src_ptr = ADD_POINTERS(vbo->Pointer,
-						 copy->array[i]->Ptr);
+	 copy->varying[j].src_ptr =
+               ADD_POINTERS(vbo->Mappings[MAP_INTERNAL].Pointer,
+                            copy->array[i]->Ptr);
 
 	 copy->dstarray_ptr[i] = &copy->varying[j].dstarray;
       }
@@ -466,12 +469,13 @@ replay_init( struct copy_context *copy )
     * do it internally.
     */
    if (_mesa_is_bufferobj(copy->ib->obj) &&
-       !_mesa_bufferobj_mapped(copy->ib->obj)) 
+       !_mesa_bufferobj_mapped(copy->ib->obj, MAP_INTERNAL))
       ctx->Driver.MapBufferRange(ctx, 0, copy->ib->obj->Size, GL_MAP_READ_BIT,
-				 copy->ib->obj);
+				 copy->ib->obj, MAP_INTERNAL);
 
-   srcptr = (const GLubyte *) ADD_POINTERS(copy->ib->obj->Pointer,
-                                           copy->ib->ptr);
+   srcptr = (const GLubyte *)
+            ADD_POINTERS(copy->ib->obj->Mappings[MAP_INTERNAL].Pointer,
+                         copy->ib->ptr);
 
    switch (copy->ib->type) {
    case GL_UNSIGNED_BYTE:
@@ -572,15 +576,15 @@ replay_finish( struct copy_context *copy )
     */
    for (i = 0; i < copy->nr_varying; i++) {
       struct gl_buffer_object *vbo = copy->varying[i].array->BufferObj;
-      if (_mesa_is_bufferobj(vbo) && _mesa_bufferobj_mapped(vbo)) 
-	 ctx->Driver.UnmapBuffer(ctx, vbo);
+      if (_mesa_is_bufferobj(vbo) && _mesa_bufferobj_mapped(vbo, MAP_INTERNAL))
+	 ctx->Driver.UnmapBuffer(ctx, vbo, MAP_INTERNAL);
    }
 
    /* Unmap index buffer:
     */
    if (_mesa_is_bufferobj(copy->ib->obj) &&
-       _mesa_bufferobj_mapped(copy->ib->obj)) {
-      ctx->Driver.UnmapBuffer(ctx, copy->ib->obj);
+       _mesa_bufferobj_mapped(copy->ib->obj, MAP_INTERNAL)) {
+      ctx->Driver.UnmapBuffer(ctx, copy->ib->obj, MAP_INTERNAL);
    }
 }