Merge remote-tracking branch 'origin/released'

* origin/released:
  pixman xkbcomp libX11 libXau mesa fontconfig xserver xkeyboard-config git update 7 jan 2013

Conflicts:
	fontconfig/src/fcatomic.c
	pixman/pixman/pixman-mmx.c
	pixman/pixman/pixman-sse2.c
	xorg-server/dix/dispatch.c
	xorg-server/hw/xwin/wincursor.c
	xorg-server/hw/xwin/winmsg.c
	xorg-server/hw/xwin/winscrinit.c
	xorg-server/hw/xwin/winsetsp.c
	xorg-server/hw/xwin/winwin32rootless.c
	xorg-server/xfixes/cursor.c

46 files changed, 946 insertions, 405 deletions

diff --git a/mesalib/src/.gitignore b/mesalib/src/.gitignore
index 5fc607b9e..ddfb40586 100644
--- a/mesalib/src/.gitignore
+++ b/mesalib/src/.gitignore
@@ -1 +1,3 @@
 /Makefile
+*.trs
+*.log
diff --git a/mesalib/src/gallium/auxiliary/util/u_blitter.c b/mesalib/src/gallium/auxiliary/util/u_blitter.c
index 7c7e06219..95224020c 100644
--- a/mesalib/src/gallium/auxiliary/util/u_blitter.c
+++ b/mesalib/src/gallium/auxiliary/util/u_blitter.c
@@ -127,6 +127,10 @@ struct blitter_context_priv
    void   (*delete_fs_state)(struct pipe_context *, void *);
 };
 
+static struct pipe_surface *
+util_blitter_get_next_surface_layer(struct pipe_context *pipe,
+                                    struct pipe_surface *surf);
+
 struct blitter_context *util_blitter_create(struct pipe_context *pipe)
 {
    struct blitter_context_priv *ctx;
@@ -143,6 +147,7 @@ struct blitter_context *util_blitter_create(struct pipe_context *pipe)
 
    ctx->base.pipe = pipe;
    ctx->base.draw_rectangle = util_blitter_draw_rectangle;
+   ctx->base.get_next_surface_layer = util_blitter_get_next_surface_layer;
 
    ctx->bind_fs_state = pipe->bind_fs_state;
    ctx->delete_fs_state = pipe->delete_fs_state;
@@ -681,6 +686,11 @@ static void blitter_set_texcoords(struct blitter_context_priv *ctx,
       }
       break;
 
+   case PIPE_TEXTURE_CUBE_ARRAY:
+      for (i = 0; i < 4; i++)
+         ctx->vertices[i][1][3] = (float) (layer / 6); /*w*/
+      break;
+
    case PIPE_TEXTURE_2D:
       for (i = 0; i < 4; i++) {
          ctx->vertices[i][1][2] = (float) sample; /*r*/
@@ -1056,14 +1066,28 @@ void util_blitter_custom_clear_depth(struct blitter_context *blitter,
 void util_blitter_default_dst_texture(struct pipe_surface *dst_templ,
                                       struct pipe_resource *dst,
                                       unsigned dstlevel,
-                                      unsigned dstz,
-                                      const struct pipe_box *srcbox)
+                                      unsigned dstz)
 {
     memset(dst_templ, 0, sizeof(*dst_templ));
     dst_templ->format = util_format_linear(dst->format);
     dst_templ->u.tex.level = dstlevel;
     dst_templ->u.tex.first_layer = dstz;
-    dst_templ->u.tex.last_layer = dstz + srcbox->depth - 1;
+    dst_templ->u.tex.last_layer = dstz;
+}
+
+static struct pipe_surface *
+util_blitter_get_next_surface_layer(struct pipe_context *pipe,
+                                    struct pipe_surface *surf)
+{
+   struct pipe_surface dst_templ;
+
+   memset(&dst_templ, 0, sizeof(dst_templ));
+   dst_templ.format = surf->format;
+   dst_templ.u.tex.level = surf->u.tex.level;
+   dst_templ.u.tex.first_layer = surf->u.tex.first_layer + 1;
+   dst_templ.u.tex.last_layer = surf->u.tex.last_layer + 1;
+
+   return pipe->create_surface(pipe, surf->texture, &dst_templ);
 }
 
 void util_blitter_default_src_texture(struct pipe_sampler_view *src_templ,
@@ -1176,12 +1200,16 @@ void util_blitter_copy_texture(struct blitter_context *blitter,
    struct pipe_context *pipe = ctx->base.pipe;
    struct pipe_surface *dst_view, dst_templ;
    struct pipe_sampler_view src_templ, *src_view;
+   struct pipe_box dstbox;
 
    assert(dst && src);
    assert(src->target < PIPE_MAX_TEXTURE_TYPES);
 
+   u_box_3d(dstx, dsty, dstz, abs(srcbox->width), abs(srcbox->height),
+            abs(srcbox->depth), &dstbox);
+
    /* Initialize the surface. */
-   util_blitter_default_dst_texture(&dst_templ, dst, dst_level, dstz, srcbox);
+   util_blitter_default_dst_texture(&dst_templ, dst, dst_level, dstz);
    dst_view = pipe->create_surface(pipe, dst, &dst_templ);
 
    /* Initialize the sampler view. */
@@ -1189,8 +1217,7 @@ void util_blitter_copy_texture(struct blitter_context *blitter,
    src_view = pipe->create_sampler_view(pipe, src, &src_templ);
 
    /* Copy. */
-   util_blitter_blit_generic(blitter, dst_view, dstx, dsty,
-                             abs(srcbox->width), abs(srcbox->height),
+   util_blitter_blit_generic(blitter, dst_view, &dstbox,
                              src_view, srcbox, src->width0, src->height0,
                              mask, PIPE_TEX_FILTER_NEAREST, NULL,
                              copy_all_samples);
@@ -1201,8 +1228,7 @@ void util_blitter_copy_texture(struct blitter_context *blitter,
 
 void util_blitter_blit_generic(struct blitter_context *blitter,
                                struct pipe_surface *dst,
-                               int dstx, int dsty,
-                               unsigned dst_width, unsigned dst_height,
+                               const struct pipe_box *dstbox,
                                struct pipe_sampler_view *src,
                                const struct pipe_box *srcbox,
                                unsigned src_width0, unsigned src_height0,
@@ -1214,6 +1240,7 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
    struct pipe_context *pipe = ctx->base.pipe;
    struct pipe_framebuffer_state fb_state;
    enum pipe_texture_target src_target = src->texture->target;
+   unsigned src_samples = src->texture->nr_samples;
    boolean has_depth, has_stencil, has_color;
    boolean blit_stencil, blit_depth, blit_color;
    void *sampler_state;
@@ -1238,9 +1265,6 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
       return;
    }
 
-   /* XXX should handle 3d regions */
-   assert(srcbox->depth == 1);
-
    /* Check whether the states are properly saved. */
    blitter_set_running_flag(ctx);
    blitter_check_saved_vertex_states(ctx);
@@ -1252,6 +1276,9 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
    /* Initialize framebuffer state. */
    fb_state.width = dst->width;
    fb_state.height = dst->height;
+   fb_state.nr_cbufs = blit_depth || blit_stencil ? 0 : 1;
+   fb_state.cbufs[0] = NULL;
+   fb_state.zsbuf = NULL;
 
    if (blit_depth || blit_stencil) {
       pipe->bind_blend_state(pipe, ctx->blend[0]);
@@ -1260,41 +1287,36 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
          pipe->bind_depth_stencil_alpha_state(pipe,
                                               ctx->dsa_write_depth_stencil);
          ctx->bind_fs_state(pipe,
-               blitter_get_fs_texfetch_depthstencil(ctx, src->texture->target,
-                                                    src->texture->nr_samples));
+               blitter_get_fs_texfetch_depthstencil(ctx, src_target,
+                                                    src_samples));
       } else if (blit_depth) {
          pipe->bind_depth_stencil_alpha_state(pipe,
                                               ctx->dsa_write_depth_keep_stencil);
          ctx->bind_fs_state(pipe,
-               blitter_get_fs_texfetch_depth(ctx, src->texture->target,
-                                             src->texture->nr_samples));
+               blitter_get_fs_texfetch_depth(ctx, src_target,
+                                             src_samples));
       } else { /* is_stencil */
          pipe->bind_depth_stencil_alpha_state(pipe,
                                               ctx->dsa_keep_depth_write_stencil);
          ctx->bind_fs_state(pipe,
-               blitter_get_fs_texfetch_stencil(ctx, src->texture->target,
-                                               src->texture->nr_samples));
+               blitter_get_fs_texfetch_stencil(ctx, src_target,
+                                               src_samples));
       }
 
-      fb_state.nr_cbufs = 0;
-      fb_state.zsbuf = dst;
    } else {
       pipe->bind_blend_state(pipe, ctx->blend[mask & PIPE_MASK_RGBA]);
       pipe->bind_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
       ctx->bind_fs_state(pipe,
-            blitter_get_fs_texfetch_col(ctx, src->texture->target,
-                                        src->texture->nr_samples));
-
-      fb_state.nr_cbufs = 1;
-      fb_state.cbufs[0] = dst;
-      fb_state.zsbuf = 0;
+            blitter_get_fs_texfetch_col(ctx, src_target,
+                                        src_samples));
    }
 
    /* Set the linear filter only for scaled color non-MSAA blits. */
    if (filter == PIPE_TEX_FILTER_LINEAR &&
        !blit_depth && !blit_stencil &&
-       src->texture->nr_samples <= 1 &&
-       (dst_width != abs(srcbox->width) || dst_height != abs(srcbox->height))) {
+       src_samples <= 1 &&
+       (dstbox->width != abs(srcbox->width) ||
+        dstbox->height != abs(srcbox->height))) {
       sampler_state = ctx->sampler_state_linear;
    } else {
       sampler_state = ctx->sampler_state;
@@ -1339,8 +1361,6 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
    }
 
    pipe->bind_vertex_elements_state(pipe, ctx->velem_state);
-   pipe->set_framebuffer_state(pipe, &fb_state);
-
    if (scissor) {
       pipe->set_scissor_state(pipe, scissor);
    }
@@ -1351,7 +1371,7 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
    if ((src_target == PIPE_TEXTURE_1D ||
         src_target == PIPE_TEXTURE_2D ||
         src_target == PIPE_TEXTURE_RECT) &&
-       src->texture->nr_samples <= 1) {
+       src_samples <= 1) {
       /* Draw the quad with the draw_rectangle callback. */
 
       /* Set texture coordinates. - use a pipe color union
@@ -1363,35 +1383,72 @@ void util_blitter_blit_generic(struct blitter_context *blitter,
       get_texcoords(src, src_width0, src_height0, srcbox->x, srcbox->y,
                     srcbox->x+srcbox->width, srcbox->y+srcbox->height, coord.f);
 
+      /* Set framebuffer state. */
+      if (blit_depth || blit_stencil) {
+         fb_state.zsbuf = dst;
+      } else {
+         fb_state.cbufs[0] = dst;
+      }
+      pipe->set_framebuffer_state(pipe, &fb_state);
+
       /* Draw. */
       pipe->set_sample_mask(pipe, ~0);
-      blitter->draw_rectangle(blitter, dstx, dsty,
-                              dstx+dst_width, dsty+dst_height, 0,
+      blitter->draw_rectangle(blitter, dstbox->x, dstbox->y,
+                              dstbox->x + dstbox->width,
+                              dstbox->y + dstbox->height, 0,
                               UTIL_BLITTER_ATTRIB_TEXCOORD, &coord);
    } else {
       /* Draw the quad with the generic codepath. */
-      if (copy_all_samples &&
-          src->texture->nr_samples == dst->texture->nr_samples &&
-          dst->texture->nr_samples > 1) {
-         /* MSAA copy. */
-         unsigned i, max_sample = MAX2(dst->texture->nr_samples, 1) - 1;
-
-         for (i = 0; i <= max_sample; i++) {
-            pipe->set_sample_mask(pipe, 1 << i);
-            blitter_set_texcoords(ctx, src, src_width0, src_height0, srcbox->z,
-                                  i, srcbox->x, srcbox->y,
+      int z;
+      for (z = 0; z < dstbox->depth; z++) {
+         struct pipe_surface *old;
+
+         /* Set framebuffer state. */
+         if (blit_depth || blit_stencil) {
+            fb_state.zsbuf = dst;
+         } else {
+            fb_state.cbufs[0] = dst;
+         }
+         pipe->set_framebuffer_state(pipe, &fb_state);
+
+         /* See if we need to blit a multisample or singlesample buffer. */
+         if (copy_all_samples &&
+             src_samples == dst->texture->nr_samples &&
+             dst->texture->nr_samples > 1) {
+            unsigned i, max_sample = MAX2(dst->texture->nr_samples, 1) - 1;
+
+            for (i = 0; i <= max_sample; i++) {
+               pipe->set_sample_mask(pipe, 1 << i);
+               blitter_set_texcoords(ctx, src, src_width0, src_height0,
+                                     srcbox->z + z,
+                                     i, srcbox->x, srcbox->y,
+                                     srcbox->x + srcbox->width,
+                                     srcbox->y + srcbox->height);
+               blitter_draw(ctx, dstbox->x, dstbox->y,
+                            dstbox->x + dstbox->width,
+                            dstbox->y + dstbox->height, 0);
+            }
+         } else {
+            pipe->set_sample_mask(pipe, ~0);
+            blitter_set_texcoords(ctx, src, src_width0, src_height0,
+                                  srcbox->z + z, 0,
+                                  srcbox->x, srcbox->y,
                                   srcbox->x + srcbox->width,
                                   srcbox->y + srcbox->height);
-            blitter_draw(ctx, dstx, dsty,
-                         dstx+dst_width, dsty+dst_height, 0);
+            blitter_draw(ctx, dstbox->x, dstbox->y,
+                         dstbox->x + dstbox->width,
+                         dstbox->y + dstbox->height, 0);
+         }
+
+         /* Get the next surface or (if this is the last iteration)
+          * just unreference the last one. */
+         old = dst;
+         if (z < dstbox->depth-1) {
+            dst = ctx->base.get_next_surface_layer(ctx->base.pipe, dst);
+         }
+         if (z) {
+            pipe_surface_reference(&old, NULL);
          }
-      } else {
-         pipe->set_sample_mask(pipe, ~0);
-         blitter_set_texcoords(ctx, src, src_width0, src_height0, srcbox->z, 0,
-                               srcbox->x, srcbox->y,
-                               srcbox->x + srcbox->width,
-                               srcbox->y + srcbox->height);
-         blitter_draw(ctx, dstx, dsty, dstx+dst_width, dsty+dst_height, 0);
       }
    }
 
@@ -1419,7 +1476,7 @@ util_blitter_blit(struct blitter_context *blitter,
 
    /* Initialize the surface. */
    util_blitter_default_dst_texture(&dst_templ, dst, info->dst.level,
-                                    info->dst.box.z, &info->src.box);
+                                    info->dst.box.z);
    dst_templ.format = info->dst.format;
    dst_view = pipe->create_surface(pipe, dst, &dst_templ);
 
@@ -1429,9 +1486,7 @@ util_blitter_blit(struct blitter_context *blitter,
    src_view = pipe->create_sampler_view(pipe, src, &src_templ);
 
    /* Copy. */
-   util_blitter_blit_generic(blitter, dst_view,
-                             info->dst.box.x, info->dst.box.y,
-                             info->dst.box.width, info->dst.box.height,
+   util_blitter_blit_generic(blitter, dst_view, &info->dst.box,
                              src_view, &info->src.box, src->width0, src->height0,
                              info->mask, info->filter,
                              info->scissor_enable ? &info->scissor : NULL, TRUE);
@@ -1761,7 +1816,8 @@ void util_blitter_custom_color(struct blitter_context *blitter,
    blitter_disable_render_cond(ctx);
 
    /* bind states */
-   pipe->bind_blend_state(pipe, custom_blend);
+   pipe->bind_blend_state(pipe, custom_blend ? custom_blend
+                                             : ctx->blend[PIPE_MASK_RGBA]);
    pipe->bind_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
    ctx->bind_fs_state(pipe, blitter_get_fs_col(ctx, 1, FALSE));
    pipe->bind_vertex_elements_state(pipe, ctx->velem_state);
diff --git a/mesalib/src/gallium/auxiliary/util/u_blitter.h b/mesalib/src/gallium/auxiliary/util/u_blitter.h
index 0b5e4aa45..20e69ed5b 100644
--- a/mesalib/src/gallium/auxiliary/util/u_blitter.h
+++ b/mesalib/src/gallium/auxiliary/util/u_blitter.h
@@ -83,6 +83,15 @@ struct blitter_context
                           enum blitter_attrib_type type,
                           const union pipe_color_union *color);
 
+   /**
+    * Get the next surface layer for the pipe surface, i.e. make a copy
+    * of the surface and increment the first and last layer by 1.
+    *
+    * This callback is exposed, so that drivers can override it if needed.
+    */
+   struct pipe_surface *(*get_next_surface_layer)(struct pipe_context *pipe,
+                                                  struct pipe_surface *surf);
+
    /* Whether the blitter is running. */
    boolean running;
 
@@ -217,14 +226,15 @@ void util_blitter_copy_texture(struct blitter_context *blitter,
                                boolean copy_all_samples);
 
 /**
- * Same as util_blitter_copy_texture with the capabilities of util_blitter_blit,
- * but dst and src are pipe_surface and pipe_sampler_view, respectively.
- * The mipmap level and dstz are part of the views.
+ * This is a generic implementation of pipe->blit, which accepts
+ * sampler/surface views instead of resources.
+ *
+ * The layer and mipmap level are specified by the views.
  *
  * Drivers can use this to change resource properties (like format, width,
  * height) by changing how the views interpret them, instead of changing
- * pipe_resource directly. This is usually needed to accelerate copying of
- * compressed formats.
+ * pipe_resource directly. This is used to blit resources of formats which
+ * are not renderable.
  *
  * src_width0 and src_height0 are sampler_view-private properties that
  * override pipe_resource. The blitter uses them for computation of texture
@@ -236,8 +246,7 @@ void util_blitter_copy_texture(struct blitter_context *blitter,
  */
 void util_blitter_blit_generic(struct blitter_context *blitter,
                                struct pipe_surface *dst,
-                               int dstx, int dsty,
-                               unsigned dst_width, unsigned dst_height,
+                               const struct pipe_box *dstbox,
                                struct pipe_sampler_view *src,
                                const struct pipe_box *srcbox,
                                unsigned src_width0, unsigned src_height0,
@@ -255,8 +264,7 @@ void util_blitter_blit(struct blitter_context *blitter,
 void util_blitter_default_dst_texture(struct pipe_surface *dst_templ,
                                       struct pipe_resource *dst,
                                       unsigned dstlevel,
-                                      unsigned dstz,
-                                      const struct pipe_box *srcbox);
+                                      unsigned dstz);
 
 /**
  * Helper function to initialize a view for copy_texture_view.
diff --git a/mesalib/src/gallium/auxiliary/util/u_dump_state.c b/mesalib/src/gallium/auxiliary/util/u_dump_state.c
index 09faffe73..d7df5b495 100644
--- a/mesalib/src/gallium/auxiliary/util/u_dump_state.c
+++ b/mesalib/src/gallium/auxiliary/util/u_dump_state.c
@@ -681,13 +681,14 @@ util_dump_transfer(FILE *stream, const struct pipe_transfer *state)
    util_dump_struct_begin(stream, "pipe_transfer");
 
    util_dump_member(stream, ptr, state, resource);
-   /*util_dump_member(stream, uint, state, box);*/
-
+   util_dump_member(stream, uint, state, level);
+   util_dump_member(stream, uint, state, usage);
+   util_dump_member_begin(stream, "box");
+   util_dump_box(stream, &state->box);
+   util_dump_member_end(stream);
    util_dump_member(stream, uint, state, stride);
    util_dump_member(stream, uint, state, layer_stride);
 
-   /*util_dump_member(stream, ptr, state, data);*/
-
    util_dump_struct_end(stream);
 }
 
diff --git a/mesalib/src/gallium/auxiliary/util/u_format_zs.c b/mesalib/src/gallium/auxiliary/util/u_format_zs.c
index 816ef1420..ed45c52f9 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format_zs.c
+++ b/mesalib/src/gallium/auxiliary/util/u_format_zs.c
@@ -72,7 +72,7 @@ static INLINE uint16_t
 z32_float_to_z16_unorm(float z)
 {
    const float scale = 0xffff;
-   return (uint16_t)(z * scale);
+   return (uint16_t)(z * scale + 0.5f);
 }
 
 static INLINE float
diff --git a/mesalib/src/gallium/auxiliary/util/u_inlines.h b/mesalib/src/gallium/auxiliary/util/u_inlines.h
index 582aacdca..2ff90c982 100644
--- a/mesalib/src/gallium/auxiliary/util/u_inlines.h
+++ b/mesalib/src/gallium/auxiliary/util/u_inlines.h
@@ -421,6 +421,24 @@ pipe_transfer_map(struct pipe_context *context,
                                 &box, transfer);
 }
 
+static INLINE void *
+pipe_transfer_map_3d(struct pipe_context *context,
+                     struct pipe_resource *resource,
+                     unsigned level,
+                     enum pipe_transfer_usage usage,
+                     unsigned x, unsigned y, unsigned z,
+                     unsigned w, unsigned h, unsigned d,
+                     struct pipe_transfer **transfer)
+{
+   struct pipe_box box;
+   u_box_3d(x, y, z, w, h, d, &box);
+   return context->transfer_map(context,
+                                resource,
+                                level,
+                                usage,
+                                &box, transfer);
+}
+
 static INLINE void
 pipe_transfer_unmap( struct pipe_context *context,
                      struct pipe_transfer *transfer )
diff --git a/mesalib/src/gallium/auxiliary/util/u_surface.c b/mesalib/src/gallium/auxiliary/util/u_surface.c
index 5e771c950..37f48154f 100644
--- a/mesalib/src/gallium/auxiliary/util/u_surface.c
+++ b/mesalib/src/gallium/auxiliary/util/u_surface.c
@@ -56,61 +56,6 @@ u_surface_default_template(struct pipe_surface *surf,
    surf->format = texture->format;
 }
 
-/**
- * Helper to quickly create an RGBA rendering surface of a certain size.
- * \param textureOut  returns the new texture
- * \param surfaceOut  returns the new surface
- * \return TRUE for success, FALSE if failure
- */
-boolean
-util_create_rgba_texture(struct pipe_context *pipe,
-                         uint width, uint height, uint bind,
-                         struct pipe_resource **textureOut)
-{
-   static const enum pipe_format rgbaFormats[] = {
-      PIPE_FORMAT_B8G8R8A8_UNORM,
-      PIPE_FORMAT_A8R8G8B8_UNORM,
-      PIPE_FORMAT_A8B8G8R8_UNORM,
-      PIPE_FORMAT_NONE
-   };
-   const uint target = PIPE_TEXTURE_2D;
-   enum pipe_format format = PIPE_FORMAT_NONE;
-   struct pipe_resource templ;
-   struct pipe_surface surf_templ;
-   struct pipe_screen *screen = pipe->screen;
-   uint i;
-
-   /* Choose surface format */
-   for (i = 0; rgbaFormats[i]; i++) {
-      if (screen->is_format_supported(screen, rgbaFormats[i],
-                                      target, 0, bind)) {
-         format = rgbaFormats[i];
-         break;
-      }
-   }
-   if (format == PIPE_FORMAT_NONE)
-      return FALSE;  /* unable to get an rgba format!?! */
-
-   /* create texture */
-   memset(&templ, 0, sizeof(templ));
-   templ.target = target;
-   templ.format = format;
-   templ.last_level = 0;
-   templ.width0 = width;
-   templ.height0 = height;
-   templ.depth0 = 1;
-   templ.array_size = 1;
-   templ.bind = bind;
-
-   *textureOut = screen->resource_create(screen, &templ);
-   if (!*textureOut)
-      return FALSE;
-
-   /* create surface */
-   u_surface_default_template(&surf_templ, *textureOut);
-   return TRUE;
-}
-
 
 /**
  * Copy 2D rect from one place to another.
diff --git a/mesalib/src/gallium/auxiliary/util/u_surface.h b/mesalib/src/gallium/auxiliary/util/u_surface.h
index fe950c818..d6184acbd 100644
--- a/mesalib/src/gallium/auxiliary/util/u_surface.h
+++ b/mesalib/src/gallium/auxiliary/util/u_surface.h
@@ -44,12 +44,6 @@ extern void
 u_surface_default_template(struct pipe_surface *view,
                            const struct pipe_resource *texture);
 
-extern boolean
-util_create_rgba_texture(struct pipe_context *ctx,
-                         uint width, uint height, uint bind,
-                         struct pipe_resource **textureOut);
-
-
 extern void
 util_copy_rect(ubyte * dst, enum pipe_format format,
                unsigned dst_stride, unsigned dst_x, unsigned dst_y,
diff --git a/mesalib/src/gallium/auxiliary/util/u_tile.c b/mesalib/src/gallium/auxiliary/util/u_tile.c
index f4b5cad0e..6c618a674 100644
--- a/mesalib/src/gallium/auxiliary/util/u_tile.c
+++ b/mesalib/src/gallium/auxiliary/util/u_tile.c
@@ -678,6 +678,28 @@ pipe_get_tile_z(struct pipe_transfer *pt,
          }
       }
       break;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+      {
+         const float *ptrc = (const float *)(map + y * pt->stride + x*8);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert float Z to 32-bit Z */
+               if (ptrc[j] <= 0.0) {
+                  pDest[j*2] = 0;
+               }
+               else if (ptrc[j] >= 1.0) {
+                  pDest[j*2] = 0xffffffff;
+               }
+               else {
+                  double z = ptrc[j] * 0xffffffff;
+                  pDest[j*2] = (uint) z;
+               }
+            }
+            pDest += dstStride;
+            ptrc += pt->stride/4;
+         }
+      }
+      break;
    default:
       assert(0);
    }
@@ -779,7 +801,7 @@ pipe_put_tile_z(struct pipe_transfer *pt,
       break;
    case PIPE_FORMAT_Z32_FLOAT:
       {
-         float *pDest = (float *) (map + y * pt->stride + x*2);
+         float *pDest = (float *) (map + y * pt->stride + x*4);
          for (i = 0; i < h; i++) {
             for (j = 0; j < w; j++) {
                /* convert 32-bit integer Z to float Z */
@@ -791,6 +813,20 @@ pipe_put_tile_z(struct pipe_transfer *pt,
          }
       }
       break;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+      {
+         float *pDest = (float *) (map + y * pt->stride + x*8);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit integer Z to float Z */
+               const double scale = 1.0 / 0xffffffffU;
+               pDest[j*2] = ptrc[j] * scale;
+            }
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
    default:
       assert(0);
    }
diff --git a/mesalib/src/glsl/glcpp/Makefile.am b/mesalib/src/glsl/glcpp/Makefile.am
index d8d39d295..1efc9dc1a 100644
--- a/mesalib/src/glsl/glcpp/Makefile.am
+++ b/mesalib/src/glsl/glcpp/Makefile.am
@@ -28,6 +28,7 @@ AM_CFLAGS = \
 	-I$(top_srcdir)/include \
 	-I$(top_srcdir)/src/mapi \
 	-I$(top_srcdir)/src/mesa/ \
+	$(VISIBILITY_CFLAGS) \
 	$(LIBRARY_INCLUDES) \
 	$(DEFINES) \
 	$(LIBRARY_DEFINES) \
diff --git a/mesalib/src/glsl/ir_uniform.h b/mesalib/src/glsl/ir_uniform.h
index 913c53773..30e6f260d 100644
--- a/mesalib/src/glsl/ir_uniform.h
+++ b/mesalib/src/glsl/ir_uniform.h
@@ -80,6 +80,10 @@ struct gl_uniform_driver_storage {
 
 struct gl_uniform_storage {
    char *name;
+   /** Type of this uniform data stored.
+    *
+    * In the case of an array, it's the type of a single array element.
+    */
    const struct glsl_type *type;
 
    /**
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index ac036eaac..d211fda9d 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -474,8 +474,7 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
     * able to change other state.
     */
    save->TransformFeedbackNeedsResume =
-      ctx->TransformFeedback.CurrentObject->Active &&
-      !ctx->TransformFeedback.CurrentObject->Paused;
+      _mesa_is_xfb_active_and_unpaused(ctx);
    if (save->TransformFeedbackNeedsResume)
       _mesa_PauseTransformFeedback();
 
diff --git a/mesalib/src/mesa/main/api_validate.c b/mesalib/src/mesa/main/api_validate.c
index e47db23e0..133fb6370 100644
--- a/mesalib/src/mesa/main/api_validate.c
+++ b/mesalib/src/mesa/main/api_validate.c
@@ -32,6 +32,7 @@
 #include "mtypes.h"
 #include "enums.h"
 #include "vbo/vbo.h"
+#include "transformfeedback.h"
 #include <stdbool.h>
 
 
@@ -252,8 +253,7 @@ _mesa_valid_prim_mode(struct gl_context *ctx, GLenum mode, const char *name)
     *        current transform feedback state as given by table X.1.
     *
     */
-   if (ctx->TransformFeedback.CurrentObject->Active &&
-       !ctx->TransformFeedback.CurrentObject->Paused) {
+   if (_mesa_is_xfb_active_and_unpaused(ctx)) {
       GLboolean pass = GL_TRUE;
 
       switch (mode) {
@@ -316,6 +316,19 @@ _mesa_validate_DrawElements(struct gl_context *ctx,
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
    FLUSH_CURRENT(ctx, 0);
 
+   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
+    * Primitive Capture):
+    *
+    *   The error INVALID_OPERATION is also generated by DrawElements,
+    *   DrawElementsInstanced, and DrawRangeElements while transform feedback
+    *   is active and not paused, regardless of mode.
+    */
+   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glDrawElements(transform feedback active)");
+      return GL_FALSE;
+   }
+
    if (count <= 0) {
       if (count < 0)
 	 _mesa_error(ctx, GL_INVALID_VALUE, "glDrawElements(count)" );
@@ -434,6 +447,19 @@ _mesa_validate_DrawRangeElements(struct gl_context *ctx, GLenum mode,
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
    FLUSH_CURRENT(ctx, 0);
 
+   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
+    * Primitive Capture):
+    *
+    *   The error INVALID_OPERATION is also generated by DrawElements,
+    *   DrawElementsInstanced, and DrawRangeElements while transform feedback
+    *   is active and not paused, regardless of mode.
+    */
+   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glDrawElements(transform feedback active)");
+      return GL_FALSE;
+   }
+
    if (count <= 0) {
       if (count < 0)
 	 _mesa_error(ctx, GL_INVALID_VALUE, "glDrawRangeElements(count)" );
@@ -486,6 +512,8 @@ GLboolean
 _mesa_validate_DrawArrays(struct gl_context *ctx,
 			  GLenum mode, GLint start, GLsizei count)
 {
+   struct gl_transform_feedback_object *xfb_obj
+      = ctx->TransformFeedback.CurrentObject;
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
    FLUSH_CURRENT(ctx, 0);
 
@@ -507,6 +535,29 @@ _mesa_validate_DrawArrays(struct gl_context *ctx,
          return GL_FALSE;
    }
 
+   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
+    * Primitive Capture):
+    *
+    *   The error INVALID_OPERATION is generated by DrawArrays and
+    *   DrawArraysInstanced if recording the vertices of a primitive to the
+    *   buffer objects being used for transform feedback purposes would result
+    *   in either exceeding the limits of any buffer object’s size, or in
+    *   exceeding the end position offset + size − 1, as set by
+    *   BindBufferRange.
+    *
+    * This is in contrast to the behaviour of desktop GL, where the extra
+    * primitives are silently dropped from the transform feedback buffer.
+    */
+   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx)) {
+      size_t prim_count = vbo_count_tessellated_primitives(mode, count, 1);
+      if (xfb_obj->GlesRemainingPrims < prim_count) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glDrawArrays(exceeds transform feedback size)");
+         return GL_FALSE;
+      }
+      xfb_obj->GlesRemainingPrims -= prim_count;
+   }
+
    return GL_TRUE;
 }
 
@@ -515,6 +566,8 @@ GLboolean
 _mesa_validate_DrawArraysInstanced(struct gl_context *ctx, GLenum mode, GLint first,
                                    GLsizei count, GLsizei numInstances)
 {
+   struct gl_transform_feedback_object *xfb_obj
+      = ctx->TransformFeedback.CurrentObject;
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
    FLUSH_CURRENT(ctx, 0);
 
@@ -550,6 +603,30 @@ _mesa_validate_DrawArraysInstanced(struct gl_context *ctx, GLenum mode, GLint fi
          return GL_FALSE;
    }
 
+   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
+    * Primitive Capture):
+    *
+    *   The error INVALID_OPERATION is generated by DrawArrays and
+    *   DrawArraysInstanced if recording the vertices of a primitive to the
+    *   buffer objects being used for transform feedback purposes would result
+    *   in either exceeding the limits of any buffer object’s size, or in
+    *   exceeding the end position offset + size − 1, as set by
+    *   BindBufferRange.
+    *
+    * This is in contrast to the behaviour of desktop GL, where the extra
+    * primitives are silently dropped from the transform feedback buffer.
+    */
+   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx)) {
+      size_t prim_count
+         = vbo_count_tessellated_primitives(mode, count, numInstances);
+      if (xfb_obj->GlesRemainingPrims < prim_count) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glDrawArraysInstanced(exceeds transform feedback size)");
+         return GL_FALSE;
+      }
+      xfb_obj->GlesRemainingPrims -= prim_count;
+   }
+
    return GL_TRUE;
 }
 
@@ -563,6 +640,19 @@ _mesa_validate_DrawElementsInstanced(struct gl_context *ctx,
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
    FLUSH_CURRENT(ctx, 0);
 
+   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
+    * Primitive Capture):
+    *
+    *   The error INVALID_OPERATION is also generated by DrawElements,
+    *   DrawElementsInstanced, and DrawRangeElements while transform feedback
+    *   is active and not paused, regardless of mode.
+    */
+   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glDrawElements(transform feedback active)");
+      return GL_FALSE;
+   }
+
    if (count <= 0) {
       if (count < 0)
 	 _mesa_error(ctx, GL_INVALID_VALUE,
diff --git a/mesalib/src/mesa/main/arrayobj.c b/mesalib/src/mesa/main/arrayobj.c
index a909e0eff..032af43b5 100644
--- a/mesalib/src/mesa/main/arrayobj.c
+++ b/mesalib/src/mesa/main/arrayobj.c
@@ -373,7 +373,7 @@ bind_vertex_array(struct gl_context *ctx, GLuint id, GLboolean genRequired)
          save_array_object(ctx, newObj);
       }
 
-      if (!newObj->_Used) {
+      if (!newObj->EverBound) {
          /* The "Interactions with APPLE_vertex_array_object" section of the
           * GL_ARB_vertex_array_object spec says:
           *
@@ -381,7 +381,7 @@ bind_vertex_array(struct gl_context *ctx, GLuint id, GLboolean genRequired)
           *     BindVertexArrayAPPLE, determines the semantic of the object."
           */
          newObj->ARBsemantics = genRequired;
-         newObj->_Used = GL_TRUE;
+         newObj->EverBound = GL_TRUE;
       }
    }
 
@@ -551,6 +551,8 @@ _mesa_IsVertexArray( GLuint id )
       return GL_FALSE;
 
    obj = lookup_arrayobj(ctx, id);
+   if (obj == NULL)
+      return GL_FALSE;
 
-   return (obj != NULL) ? GL_TRUE : GL_FALSE;
+   return obj->EverBound;
 }
diff --git a/mesalib/src/mesa/main/buffers.c b/mesalib/src/mesa/main/buffers.c
index 76f0d4671..d10a57394 100644
--- a/mesalib/src/mesa/main/buffers.c
+++ b/mesalib/src/mesa/main/buffers.c
@@ -532,8 +532,8 @@ _mesa_ReadBuffer(GLenum buffer)
    if (MESA_VERBOSE & VERBOSE_API)
       _mesa_debug(ctx, "glReadBuffer %s\n", _mesa_lookup_enum_by_nr(buffer));
 
-   if (_mesa_is_user_fbo(fb) && buffer == GL_NONE) {
-      /* This is legal for user-created framebuffer objects */
+   if (buffer == GL_NONE) {
+      /* This is legal--it means that no buffer should be bound for reading. */
       srcBuffer = -1;
    }
    else {
diff --git a/mesalib/src/mesa/main/fbobject.c b/mesalib/src/mesa/main/fbobject.c
index ce77b9f77..281cdd05c 100644
--- a/mesalib/src/mesa/main/fbobject.c
+++ b/mesalib/src/mesa/main/fbobject.c
@@ -1169,35 +1169,23 @@ _mesa_base_fbo_format(struct gl_context *ctx, GLenum internalFormat)
       return ctx->API != API_OPENGLES && ctx->Extensions.ARB_texture_rg
          ? GL_RG : 0;
    /* signed normalized texture formats */
-   case GL_R8_SNORM:
-      return ctx->Version >= 30
-         || (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.EXT_texture_snorm)
-         ? GL_RED : 0;
    case GL_RED_SNORM:
+   case GL_R8_SNORM:
    case GL_R16_SNORM:
       return _mesa_is_desktop_gl(ctx) && ctx->Extensions.EXT_texture_snorm
          ? GL_RED : 0;
-   case GL_RG8_SNORM:
-      return ctx->Version >= 30
-         || (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.EXT_texture_snorm)
-         ? GL_RG : 0;
    case GL_RG_SNORM:
+   case GL_RG8_SNORM:
    case GL_RG16_SNORM:
       return _mesa_is_desktop_gl(ctx) && ctx->Extensions.EXT_texture_snorm
          ? GL_RG : 0;
-   case GL_RGB8_SNORM:
-      return ctx->Version >= 30
-         || (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.EXT_texture_snorm)
-         ? GL_RGB : 0;
    case GL_RGB_SNORM:
+   case GL_RGB8_SNORM:
    case GL_RGB16_SNORM:
       return _mesa_is_desktop_gl(ctx) && ctx->Extensions.EXT_texture_snorm
          ? GL_RGB : 0;
-   case GL_RGBA8_SNORM:
-      return ctx->Version >= 30
-         || (ctx->API == API_OPENGL_COMPAT && ctx->Extensions.EXT_texture_snorm)
-         ? GL_RGBA : 0;
    case GL_RGBA_SNORM:
+   case GL_RGBA8_SNORM:
    case GL_RGBA16_SNORM:
       return _mesa_is_desktop_gl(ctx) && ctx->Extensions.EXT_texture_snorm
          ? GL_RGBA : 0;
@@ -1207,24 +1195,6 @@ _mesa_base_fbo_format(struct gl_context *ctx, GLenum internalFormat)
       return ctx->API == API_OPENGL_COMPAT &&
              ctx->Extensions.EXT_texture_snorm &&
              ctx->Extensions.ARB_framebuffer_object ? GL_ALPHA : 0;
-   case GL_LUMINANCE_SNORM:
-   case GL_LUMINANCE8_SNORM:
-   case GL_LUMINANCE16_SNORM:
-      return ctx->API == API_OPENGL_COMPAT &&
-             ctx->Extensions.EXT_texture_snorm &&
-             ctx->Extensions.ARB_framebuffer_object ? GL_LUMINANCE : 0;
-   case GL_LUMINANCE_ALPHA_SNORM:
-   case GL_LUMINANCE8_ALPHA8_SNORM:
-   case GL_LUMINANCE16_ALPHA16_SNORM:
-      return ctx->API == API_OPENGL_COMPAT &&
-             ctx->Extensions.EXT_texture_snorm &&
-             ctx->Extensions.ARB_framebuffer_object ? GL_LUMINANCE_ALPHA : 0;
-   case GL_INTENSITY_SNORM:
-   case GL_INTENSITY8_SNORM:
-   case GL_INTENSITY16_SNORM:
-      return ctx->API == API_OPENGL_COMPAT &&
-             ctx->Extensions.EXT_texture_snorm &&
-             ctx->Extensions.ARB_framebuffer_object ? GL_INTENSITY : 0;
    case GL_R16F:
    case GL_R32F:
       return ctx->Version >= 30
@@ -1291,10 +1261,8 @@ _mesa_base_fbo_format(struct gl_context *ctx, GLenum internalFormat)
    case GL_RGB8I_EXT:
    case GL_RGB16I_EXT:
    case GL_RGB32I_EXT:
-      return ctx->Version >= 30
-         || (_mesa_is_desktop_gl(ctx) &&
-             ctx->Extensions.EXT_texture_integer) ? GL_RGB : 0;
-
+      return _mesa_is_desktop_gl(ctx) && ctx->Extensions.EXT_texture_integer
+         ? GL_RGB : 0;
    case GL_R8UI:
    case GL_R8I:
    case GL_R16UI:
diff --git a/mesalib/src/mesa/main/get.c b/mesalib/src/mesa/main/get.c
index f3dbda2d3..273a79f7f 100644
--- a/mesalib/src/mesa/main/get.c
+++ b/mesalib/src/mesa/main/get.c
@@ -1574,7 +1574,8 @@ find_value_indexed(const char *func, GLenum pname, GLuint index, union value *v)
 	 goto invalid_value;
       if (!ctx->Extensions.EXT_transform_feedback)
 	 goto invalid_enum;
-      v->value_int64 = ctx->TransformFeedback.CurrentObject->Size[index];
+      v->value_int64
+         = ctx->TransformFeedback.CurrentObject->RequestedSize[index];
       return TYPE_INT64;
 
    case GL_TRANSFORM_FEEDBACK_BUFFER_BINDING:
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index 67eaadd9f..7c513a1c5 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -1595,7 +1595,7 @@ struct gl_array_object
    /**
     * Has this array object been bound?
     */
-   GLboolean _Used;
+   GLboolean EverBound;
 
    /** Vertex attribute arrays */
    struct gl_client_array VertexAttrib[VERT_ATTRIB_MAX];
@@ -1807,6 +1807,18 @@ struct gl_transform_feedback_object
    GLboolean Paused;  /**< Is transform feedback paused? */
    GLboolean EndedAnytime; /**< Has EndTransformFeedback been called
                                 at least once? */
+   GLboolean EverBound; /**< Has this object been bound? */
+
+   /**
+    * GLES: if Active is true, remaining number of primitives which can be
+    * rendered without overflow.  This is necessary to track because GLES
+    * requires us to generate INVALID_OPERATION if a call to glDrawArrays or
+    * glDrawArraysInstanced would overflow transform feedback buffers.
+    * Undefined if Active is false.
+    *
+    * Not tracked for desktop GL since it's unnecessary.
+    */
+   unsigned GlesRemainingPrims;
 
    /** The feedback buffers */
    GLuint BufferNames[MAX_FEEDBACK_BUFFERS];
@@ -1814,8 +1826,19 @@ struct gl_transform_feedback_object
 
    /** Start of feedback data in dest buffer */
    GLintptr Offset[MAX_FEEDBACK_BUFFERS];
-   /** Max data to put into dest buffer (in bytes) */
+
+   /**
+    * Max data to put into dest buffer (in bytes).  Computed based on
+    * RequestedSize and the actual size of the buffer.
+    */
    GLsizeiptr Size[MAX_FEEDBACK_BUFFERS];
+
+   /**
+    * Size that was specified when the buffer was bound.  If the buffer was
+    * bound with glBindBufferBase() or glBindBufferOffsetEXT(), this value is
+    * zero.
+    */
+   GLsizeiptr RequestedSize[MAX_FEEDBACK_BUFFERS];
 };
 
 
diff --git a/mesalib/src/mesa/main/pbo.c b/mesalib/src/mesa/main/pbo.c
index c73d7492f..0dc4c04eb 100644
--- a/mesalib/src/mesa/main/pbo.c
+++ b/mesalib/src/mesa/main/pbo.c
@@ -342,7 +342,8 @@ _mesa_validate_pbo_teximage(struct gl_context *ctx, GLuint dimensions,
    }
    if (!_mesa_validate_pbo_access(dimensions, unpack, width, height, depth,
                                   format, type, INT_MAX, pixels)) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, funcName, "(invalid PBO access)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(invalid PBO access)",
+                  funcName, dimensions);
       return NULL;
    }
 
@@ -351,7 +352,8 @@ _mesa_validate_pbo_teximage(struct gl_context *ctx, GLuint dimensions,
 						GL_MAP_READ_BIT,
 						unpack->BufferObj);
    if (!buf) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, funcName, "(PBO is mapped)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(PBO is mapped)", funcName,
+                  dimensions);
       return NULL;
    }
 
@@ -368,7 +370,8 @@ _mesa_validate_pbo_teximage(struct gl_context *ctx, GLuint dimensions,
  */
 const GLvoid *
 _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
-                                 GLsizei imageSize, const GLvoid *pixels,
+                                 GLuint dimensions, GLsizei imageSize,
+                                 const GLvoid *pixels,
                                  const struct gl_pixelstore_attrib *packing,
                                  const char *funcName)
 {
@@ -381,7 +384,8 @@ _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
    if ((const GLubyte *) pixels + imageSize >
        ((const GLubyte *) 0) + packing->BufferObj->Size) {
       /* out of bounds read! */
-      _mesa_error(ctx, GL_INVALID_OPERATION, funcName, "(invalid PBO access)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(invalid PBO access)",
+                  funcName, dimensions);
       return NULL;
    }
 
@@ -390,7 +394,8 @@ _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
 					       GL_MAP_READ_BIT,
 					       packing->BufferObj);
    if (!buf) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, funcName, "(PBO is mapped");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s%uD(PBO is mapped)", funcName,
+                  dimensions);
       return NULL;
    }
 
diff --git a/mesalib/src/mesa/main/pbo.h b/mesalib/src/mesa/main/pbo.h
index 00a6e617f..9eba33514 100644
--- a/mesalib/src/mesa/main/pbo.h
+++ b/mesalib/src/mesa/main/pbo.h
@@ -81,7 +81,8 @@ _mesa_validate_pbo_teximage(struct gl_context *ctx, GLuint dimensions,
 
 extern const GLvoid *
 _mesa_validate_pbo_compressed_teximage(struct gl_context *ctx,
-                                    GLsizei imageSize, const GLvoid *pixels,
+                                    GLuint dimensions, GLsizei imageSize,
+                                    const GLvoid *pixels,
                                     const struct gl_pixelstore_attrib *packing,
                                     const char *funcName);
 
diff --git a/mesalib/src/mesa/main/samplerobj.c b/mesalib/src/mesa/main/samplerobj.c
index 3c3bfffd4..773bebf26 100644
--- a/mesalib/src/mesa/main/samplerobj.c
+++ b/mesalib/src/mesa/main/samplerobj.c
@@ -110,7 +110,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
 /**
  * Initialize the fields of the given sampler object.
  */
-void
+static void
 _mesa_init_sampler_object(struct gl_sampler_object *sampObj, GLuint name)
 {
    sampObj->Name = name;
@@ -151,7 +151,7 @@ _mesa_new_sampler_object(struct gl_context *ctx, GLuint name)
 /**
  * Fallback for ctx->Driver.DeleteSamplerObject();
  */
-void
+static void
 _mesa_delete_sampler_object(struct gl_context *ctx,
                             struct gl_sampler_object *sampObj)
 {
@@ -606,6 +606,8 @@ _mesa_SamplerParameteri(GLuint sampler, GLenum pname, GLint param)
    GLuint res;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glSamplerParameteri(sampler %u)",
@@ -777,6 +779,8 @@ _mesa_SamplerParameteriv(GLuint sampler, GLenum pname, const GLint *params)
    GLuint res;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glSamplerParameteriv(sampler %u)",
@@ -956,6 +960,8 @@ _mesa_SamplerParameterIiv(GLuint sampler, GLenum pname, const GLint *params)
    GLuint res;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glSamplerParameterIiv(sampler %u)",
@@ -1042,6 +1048,8 @@ _mesa_SamplerParameterIuiv(GLuint sampler, GLenum pname, const GLuint *params)
    GLuint res;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glSamplerParameterIuiv(sampler %u)",
@@ -1127,6 +1135,8 @@ _mesa_GetSamplerParameteriv(GLuint sampler, GLenum pname, GLint *params)
    struct gl_sampler_object *sampObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glGetSamplerParameteriv(sampler %u)",
@@ -1205,6 +1215,8 @@ _mesa_GetSamplerParameterfv(GLuint sampler, GLenum pname, GLfloat *params)
    struct gl_sampler_object *sampObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glGetSamplerParameterfv(sampler %u)",
@@ -1283,6 +1295,8 @@ _mesa_GetSamplerParameterIiv(GLuint sampler, GLenum pname, GLint *params)
    struct gl_sampler_object *sampObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE,
@@ -1362,6 +1376,8 @@ _mesa_GetSamplerParameterIuiv(GLuint sampler, GLenum pname, GLuint *params)
    struct gl_sampler_object *sampObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   ASSERT_OUTSIDE_BEGIN_END(ctx);
+
    sampObj = _mesa_lookup_samplerobj(ctx, sampler);
    if (!sampObj) {
       _mesa_error(ctx, GL_INVALID_VALUE,
diff --git a/mesalib/src/mesa/main/samplerobj.h b/mesalib/src/mesa/main/samplerobj.h
index ecff032c9..dce7c6c30 100644
--- a/mesalib/src/mesa/main/samplerobj.h
+++ b/mesalib/src/mesa/main/samplerobj.h
@@ -63,17 +63,10 @@ _mesa_reference_sampler_object(struct gl_context *ctx,
 }
 
 
-extern void
-_mesa_init_sampler_object(struct gl_sampler_object *sampObj, GLuint name);
-
 extern struct gl_sampler_object *
 _mesa_new_sampler_object(struct gl_context *ctx, GLuint name);
 
 extern void
-_mesa_delete_sampler_object(struct gl_context *ctx,
-                            struct gl_sampler_object *sampObj);
-
-extern void
 _mesa_init_sampler_object_functions(struct dd_function_table *driver);
 
 extern void
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index 53e9d0486..1f0002f6f 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -45,6 +45,7 @@
 #include "main/mtypes.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"
@@ -1383,12 +1384,10 @@ _mesa_UseProgram(GLhandleARB program)
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_shader_program *shProg;
-   struct gl_transform_feedback_object *obj =
-      ctx->TransformFeedback.CurrentObject;
 
    ASSERT_OUTSIDE_BEGIN_END(ctx);
 
-   if (obj->Active && !obj->Paused) {
+   if (_mesa_is_xfb_active_and_unpaused(ctx)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glUseProgram(transform feedback active)");
       return;
@@ -1590,8 +1589,7 @@ _mesa_UseShaderProgramEXT(GLenum type, GLuint program)
       return;
    }
 
-   if (ctx->TransformFeedback.CurrentObject->Active &&
-       !ctx->TransformFeedback.CurrentObject->Paused) {
+   if (_mesa_is_xfb_active_and_unpaused(ctx)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glUseShaderProgramEXT(transform feedback is active)");
       return;
diff --git a/mesalib/src/mesa/main/texcompress.c b/mesalib/src/mesa/main/texcompress.c
index 33c580a88..077a967da 100644
--- a/mesalib/src/mesa/main/texcompress.c
+++ b/mesalib/src/mesa/main/texcompress.c
@@ -533,6 +533,10 @@ _mesa_get_compressed_fetch_func(gl_format format)
    case MESA_FORMAT_RGBA_DXT1:
    case MESA_FORMAT_RGBA_DXT3:
    case MESA_FORMAT_RGBA_DXT5:
+   case MESA_FORMAT_SRGB_DXT1:
+   case MESA_FORMAT_SRGBA_DXT1:
+   case MESA_FORMAT_SRGBA_DXT3:
+   case MESA_FORMAT_SRGBA_DXT5:
       return _mesa_get_dxt_fetch_func(format);
    case MESA_FORMAT_RGB_FXT1:
    case MESA_FORMAT_RGBA_FXT1:
diff --git a/mesalib/src/mesa/main/texcompress_rgtc.c b/mesalib/src/mesa/main/texcompress_rgtc.c
index fa9172aa2..1f8de25ec 100644
--- a/mesalib/src/mesa/main/texcompress_rgtc.c
+++ b/mesalib/src/mesa/main/texcompress_rgtc.c
@@ -364,9 +364,10 @@ fetch_signed_l_latc1(const GLubyte *map, const GLuint imageOffsets[],
                      GLint rowStride, GLint i, GLint j, GLint k,
                      GLfloat *texel)
 {
-   GLubyte red;
+   GLbyte red;
    GLuint sliceOffset = k ? imageOffsets[k] / 2 : 0;
-   unsigned_fetch_texel_rgtc(rowStride,  map + sliceOffset, i, j, &red, 1);
+   signed_fetch_texel_rgtc(rowStride, (GLbyte *) map + sliceOffset,
+                           i, j, &red, 1);
    texel[RCOMP] =
    texel[GCOMP] =
    texel[BCOMP] = BYTE_TO_FLOAT(red);
diff --git a/mesalib/src/mesa/main/texcompress_s3tc.c b/mesalib/src/mesa/main/texcompress_s3tc.c
index 23a5a0868..4a81ad803 100644
--- a/mesalib/src/mesa/main/texcompress_s3tc.c
+++ b/mesalib/src/mesa/main/texcompress_s3tc.c
@@ -417,6 +417,80 @@ fetch_rgba_dxt5(const GLubyte *map, const GLuint imageOffsets[],
 }
 
 
+static void
+fetch_srgb_dxt1(const GLubyte *map, const GLuint imageOffsets[],
+                GLint rowStride, GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   if (fetch_ext_rgb_dxt1) {
+      GLuint sliceOffset = k ? imageOffsets[k] / 2 : 0;
+      GLubyte tex[4];
+      fetch_ext_rgb_dxt1(rowStride, map + sliceOffset, i, j, tex);
+      texel[RCOMP] = _mesa_nonlinear_to_linear(tex[RCOMP]);
+      texel[GCOMP] = _mesa_nonlinear_to_linear(tex[GCOMP]);
+      texel[BCOMP] = _mesa_nonlinear_to_linear(tex[BCOMP]);
+      texel[ACOMP] = UBYTE_TO_FLOAT(tex[ACOMP]);
+   }
+   else {
+      problem("srgb_dxt1");
+   }
+}
+
+static void
+fetch_srgba_dxt1(const GLubyte *map, const GLuint imageOffsets[],
+                 GLint rowStride, GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   if (fetch_ext_rgba_dxt1) {
+      GLuint sliceOffset = k ? imageOffsets[k] / 2 : 0;
+      GLubyte tex[4];
+      fetch_ext_rgba_dxt1(rowStride, map + sliceOffset, i, j, tex);
+      texel[RCOMP] = _mesa_nonlinear_to_linear(tex[RCOMP]);
+      texel[GCOMP] = _mesa_nonlinear_to_linear(tex[GCOMP]);
+      texel[BCOMP] = _mesa_nonlinear_to_linear(tex[BCOMP]);
+      texel[ACOMP] = UBYTE_TO_FLOAT(tex[ACOMP]);
+   }
+   else {
+      problem("srgba_dxt1");
+   }
+}
+
+static void
+fetch_srgba_dxt3(const GLubyte *map, const GLuint imageOffsets[],
+                 GLint rowStride, GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   if (fetch_ext_rgba_dxt3) {
+      GLuint sliceOffset = k ? imageOffsets[k] : 0;
+      GLubyte tex[4];
+      fetch_ext_rgba_dxt3(rowStride, map + sliceOffset, i, j, tex);
+      texel[RCOMP] = _mesa_nonlinear_to_linear(tex[RCOMP]);
+      texel[GCOMP] = _mesa_nonlinear_to_linear(tex[GCOMP]);
+      texel[BCOMP] = _mesa_nonlinear_to_linear(tex[BCOMP]);
+      texel[ACOMP] = UBYTE_TO_FLOAT(tex[ACOMP]);
+   }
+   else {
+      problem("srgba_dxt3");
+   }
+}
+
+static void
+fetch_srgba_dxt5(const GLubyte *map, const GLuint imageOffsets[],
+                 GLint rowStride, GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   if (fetch_ext_rgba_dxt5) {
+      GLuint sliceOffset = k ? imageOffsets[k] : 0;
+      GLubyte tex[4];
+      fetch_ext_rgba_dxt5(rowStride, map + sliceOffset, i, j, tex);
+      texel[RCOMP] = _mesa_nonlinear_to_linear(tex[RCOMP]);
+      texel[GCOMP] = _mesa_nonlinear_to_linear(tex[GCOMP]);
+      texel[BCOMP] = _mesa_nonlinear_to_linear(tex[BCOMP]);
+      texel[ACOMP] = UBYTE_TO_FLOAT(tex[ACOMP]);
+   }
+   else {
+      problem("srgba_dxt5");
+   }
+}
+
+
+
 compressed_fetch_func
 _mesa_get_dxt_fetch_func(gl_format format)
 {
@@ -429,6 +503,14 @@ _mesa_get_dxt_fetch_func(gl_format format)
       return fetch_rgba_dxt3;
    case MESA_FORMAT_RGBA_DXT5:
       return fetch_rgba_dxt5;
+   case MESA_FORMAT_SRGB_DXT1:
+      return fetch_srgb_dxt1;
+   case MESA_FORMAT_SRGBA_DXT1:
+      return fetch_srgba_dxt1;
+   case MESA_FORMAT_SRGBA_DXT3:
+      return fetch_srgba_dxt3;
+   case MESA_FORMAT_SRGBA_DXT5:
+      return fetch_srgba_dxt5;
    default:
       return NULL;
    }
diff --git a/mesalib/src/mesa/main/texgetimage.c b/mesalib/src/mesa/main/texgetimage.c
index 69067df9f..d435c8269 100644
--- a/mesalib/src/mesa/main/texgetimage.c
+++ b/mesalib/src/mesa/main/texgetimage.c
@@ -683,6 +683,8 @@ legal_getteximage_target(struct gl_context *ctx, GLenum target)
    case GL_TEXTURE_2D_ARRAY_EXT:
       return (ctx->Extensions.MESA_texture_array ||
               ctx->Extensions.EXT_texture_array);
+   case GL_TEXTURE_CUBE_MAP_ARRAY:
+      return ctx->Extensions.ARB_texture_cube_map_array;
    default:
       return GL_FALSE;
    }
diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c
index 7a0d944fd..33f81a25c 100644
--- a/mesalib/src/mesa/main/teximage.c
+++ b/mesalib/src/mesa/main/teximage.c
@@ -3507,7 +3507,8 @@ compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
    GLint expectedSize;
    GLboolean targetOK;
 
-   if (dims == 2) {
+   switch (dims) {
+   case 2:
       switch (target) {
       case GL_TEXTURE_2D:
       case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
@@ -3520,12 +3521,17 @@ compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
          break;
       default:
          targetOK = GL_FALSE;
+         break;
       }
-   }
-   else {
-      assert(dims == 1 || dims == 3);
-      /* no 1D or 3D compressed textures at this time */
+      break;
+   case 3:
+      targetOK = (target == GL_TEXTURE_2D_ARRAY);
+      break;
+   default:
+      assert(dims == 1);
+      /* no 1D compressed textures at this time */
       targetOK = GL_FALSE;
+      break;
    }
  
    if (!targetOK) {
diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c
index 26c5b6703..751150937 100644
--- a/mesalib/src/mesa/main/texstore.c
+++ b/mesalib/src/mesa/main/texstore.c
@@ -4441,9 +4441,9 @@ _mesa_store_compressed_teximage(struct gl_context *ctx, GLuint dims,
                                 struct gl_texture_image *texImage,
                                 GLsizei imageSize, const GLvoid *data)
 {
-   /* only 2D compressed images are supported at this time */
-   if (dims != 2) {
-      _mesa_problem(ctx, "Unexpected glCompressedTexImage1D/3D call");
+   /* only 2D and 3D compressed images are supported at this time */
+   if (dims == 1) {
+      _mesa_problem(ctx, "Unexpected glCompressedTexImage1D call");
       return;
    }
 
@@ -4454,11 +4454,11 @@ _mesa_store_compressed_teximage(struct gl_context *ctx, GLuint dims,
    ASSERT(texImage);
    ASSERT(texImage->Width > 0);
    ASSERT(texImage->Height > 0);
-   ASSERT(texImage->Depth == 1);
+   ASSERT(texImage->Depth > 0);
 
    /* allocate storage for texture data */
    if (!ctx->Driver.AllocTextureImageBuffer(ctx, texImage)) {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCompressedTexImage2D");
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCompressedTexImage%uD", dims);
       return;
    }
 
@@ -4487,45 +4487,49 @@ _mesa_store_compressed_texsubimage(struct gl_context *ctx, GLuint dims,
    const GLubyte *src;
    const gl_format texFormat = texImage->TexFormat;
    GLuint bw, bh;
+   GLuint slice;
 
-   if (dims != 2) {
-      _mesa_problem(ctx, "Unexpected 1D/3D compressed texsubimage call");
+   if (dims == 1) {
+      _mesa_problem(ctx, "Unexpected 1D compressed texsubimage call");
       return;
    }
 
    _mesa_get_format_block_size(texFormat, &bw, &bh);
 
    /* get pointer to src pixels (may be in a pbo which we'll map here) */
-   data = _mesa_validate_pbo_compressed_teximage(ctx, imageSize, data,
+   data = _mesa_validate_pbo_compressed_teximage(ctx, dims, imageSize, data,
                                                  &ctx->Unpack,
-                                                 "glCompressedTexSubImage2D");
+                                                 "glCompressedTexSubImage");
    if (!data)
       return;
 
    srcRowStride = _mesa_format_row_stride(texFormat, width);
    src = (const GLubyte *) data;
 
-   /* Map dest texture buffer */
-   ctx->Driver.MapTextureImage(ctx, texImage, 0,
-                               xoffset, yoffset, width, height,
-                               GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_RANGE_BIT,
-                               &dstMap, &dstRowStride);
+   for (slice = 0; slice < depth; slice++) {
+      /* Map dest texture buffer */
+      ctx->Driver.MapTextureImage(ctx, texImage, slice + zoffset,
+                                  xoffset, yoffset, width, height,
+                                  GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_RANGE_BIT,
+                                  &dstMap, &dstRowStride);
 
-   if (dstMap) {
-      bytesPerRow = srcRowStride;  /* bytes per row of blocks */
-      rows = (height + bh - 1) / bh;  /* rows in blocks */
+      if (dstMap) {
+         bytesPerRow = srcRowStride;  /* bytes per row of blocks */
+         rows = (height + bh - 1) / bh;  /* rows in blocks */
 
-      /* copy rows of blocks */
-      for (i = 0; i < rows; i++) {
-         memcpy(dstMap, src, bytesPerRow);
-         dstMap += dstRowStride;
-         src += srcRowStride;
-      }
+         /* copy rows of blocks */
+         for (i = 0; i < rows; i++) {
+            memcpy(dstMap, src, bytesPerRow);
+            dstMap += dstRowStride;
+            src += srcRowStride;
+         }
 
-      ctx->Driver.UnmapTextureImage(ctx, texImage, 0);
-   }
-   else {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCompressedTexSubImage2D");
+         ctx->Driver.UnmapTextureImage(ctx, texImage, slice + zoffset);
+      }
+      else {
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glCompressedTexSubImage%uD",
+                     dims);
+      }
    }
 
    _mesa_unmap_teximage_pbo(ctx, &ctx->Unpack);
diff --git a/mesalib/src/mesa/main/transformfeedback.c b/mesalib/src/mesa/main/transformfeedback.c
index 22060c34c..1bf4beaeb 100644
--- a/mesalib/src/mesa/main/transformfeedback.c
+++ b/mesalib/src/mesa/main/transformfeedback.c
@@ -34,6 +34,7 @@
 #include "bufferobj.h"
 #include "context.h"
 #include "hash.h"
+#include "macros.h"
 #include "mfeatures.h"
 #include "mtypes.h"
 #include "transformfeedback.h"
@@ -79,6 +80,7 @@ reference_transform_feedback_object(struct gl_transform_feedback_object **ptr,
       }
       else {
          obj->RefCount++;
+         obj->EverBound = GL_TRUE;
          *ptr = obj;
       }
    }
@@ -177,6 +179,7 @@ new_transform_feedback(struct gl_context *ctx, GLuint name)
    if (obj) {
       obj->Name = name;
       obj->RefCount = 1;
+      obj->EverBound = GL_FALSE;
    }
    return obj;
 }
@@ -246,6 +249,83 @@ _mesa_init_transform_feedback_functions(struct dd_function_table *driver)
 
 
 /**
+ * Fill in the correct Size value for each buffer in \c obj.
+ *
+ * From the GL 4.3 spec, section 6.1.1 ("Binding Buffer Objects to Indexed
+ * Targets"):
+ *
+ *   BindBufferBase binds the entire buffer, even when the size of the buffer
+ *   is changed after the binding is established. It is equivalent to calling
+ *   BindBufferRange with offset zero, while size is determined by the size of
+ *   the bound buffer at the time the binding is used.
+ *
+ *   Regardless of the size specified with BindBufferRange, or indirectly with
+ *   BindBufferBase, the GL will never read or write beyond the end of a bound
+ *   buffer. In some cases this constraint may result in visibly different
+ *   behavior when a buffer overflow would otherwise result, such as described
+ *   for transform feedback operations in section 13.2.2.
+ */
+static void
+compute_transform_feedback_buffer_sizes(
+      struct gl_transform_feedback_object *obj)
+{
+   unsigned i = 0;
+   for (i = 0; i < MAX_FEEDBACK_BUFFERS; ++i) {
+      GLintptr offset = obj->Offset[i];
+      GLsizeiptr buffer_size
+         = obj->Buffers[i] == NULL ? 0 : obj->Buffers[i]->Size;
+      GLsizeiptr available_space
+         = buffer_size <= offset ? 0 : buffer_size - offset;
+      GLsizeiptr computed_size;
+      if (obj->RequestedSize[i] == 0) {
+         /* No size was specified at the time the buffer was bound, so allow
+          * writing to all available space in the buffer.
+          */
+         computed_size = available_space;
+      } else {
+         /* A size was specified at the time the buffer was bound, however
+          * it's possible that the buffer has shrunk since then.  So only
+          * allow writing to the minimum of the specified size and the space
+          * available.
+          */
+         computed_size = MIN2(available_space, obj->RequestedSize[i]);
+      }
+
+      /* Legal sizes must be multiples of four, so round down if necessary. */
+      obj->Size[i] = computed_size & ~0x3;
+   }
+}
+
+
+/**
+ * Compute the maximum number of vertices that can be written to the currently
+ * enabled transform feedback buffers without overflowing any of them.
+ */
+unsigned
+_mesa_compute_max_transform_feedback_vertices(
+      const struct gl_transform_feedback_object *obj,
+      const struct gl_transform_feedback_info *info)
+{
+   unsigned max_index = 0xffffffff;
+   unsigned i;
+
+   for (i = 0; i < info->NumBuffers; ++i) {
+      unsigned stride = info->BufferStride[i];
+      unsigned max_for_this_buffer;
+
+      /* Skip any inactive buffers, which have a stride of 0. */
+      if (stride == 0)
+	 continue;
+
+      max_for_this_buffer = obj->Size[i] / (4 * stride);
+      max_index = MIN2(max_index, max_for_this_buffer);
+   }
+
+   return max_index;
+}
+
+
+/**
  ** Begin API functions
  **/
 
@@ -256,6 +336,7 @@ _mesa_BeginTransformFeedback(GLenum mode)
    struct gl_transform_feedback_object *obj;
    struct gl_transform_feedback_info *info;
    GLuint i;
+   unsigned vertices_per_prim;
    GET_CURRENT_CONTEXT(ctx);
 
    obj = ctx->TransformFeedback.CurrentObject;
@@ -276,9 +357,13 @@ _mesa_BeginTransformFeedback(GLenum mode)
 
    switch (mode) {
    case GL_POINTS:
+      vertices_per_prim = 1;
+      break;
    case GL_LINES:
+      vertices_per_prim = 2;
+      break;
    case GL_TRIANGLES:
-      /* legal */
+      vertices_per_prim = 3;
       break;
    default:
       _mesa_error(ctx, GL_INVALID_ENUM, "glBeginTransformFeedback(mode)");
@@ -304,6 +389,20 @@ _mesa_BeginTransformFeedback(GLenum mode)
    obj->Active = GL_TRUE;
    ctx->TransformFeedback.Mode = mode;
 
+   compute_transform_feedback_buffer_sizes(obj);
+
+   if (_mesa_is_gles3(ctx)) {
+      /* In GLES3, we are required to track the usage of the transform
+       * feedback buffer and report INVALID_OPERATION if a draw call tries to
+       * exceed it.  So compute the maximum number of vertices that we can
+       * write without overflowing any of the buffers currently being used for
+       * feedback.
+       */
+      unsigned max_vertices
+         = _mesa_compute_max_transform_feedback_vertices(obj, info);
+      obj->GlesRemainingPrims = max_vertices / vertices_per_prim;
+   }
+
    assert(ctx->Driver.BeginTransformFeedback);
    ctx->Driver.BeginTransformFeedback(ctx, mode, obj);
 }
@@ -362,7 +461,7 @@ bind_buffer_range(struct gl_context *ctx, GLuint index,
    obj->BufferNames[index] = bufObj->Name;
 
    obj->Offset[index] = offset;
-   obj->Size[index] = size;
+   obj->RequestedSize[index] = size;
 }
 
 
@@ -421,7 +520,6 @@ _mesa_bind_buffer_base_transform_feedback(struct gl_context *ctx,
 					  struct gl_buffer_object *bufObj)
 {
    struct gl_transform_feedback_object *obj;
-   GLsizeiptr size;
 
    obj = ctx->TransformFeedback.CurrentObject;
 
@@ -436,12 +534,7 @@ _mesa_bind_buffer_base_transform_feedback(struct gl_context *ctx,
       return;
    }
 
-   /* default size is the buffer size rounded down to nearest
-    * multiple of four.
-    */
-   size = bufObj->Size & ~0x3;
-
-   bind_buffer_range(ctx, index, bufObj, 0, size);
+   bind_buffer_range(ctx, index, bufObj, 0, 0);
 }
 
 
@@ -457,7 +550,6 @@ _mesa_BindBufferOffsetEXT(GLenum target, GLuint index, GLuint buffer,
    struct gl_transform_feedback_object *obj;
    struct gl_buffer_object *bufObj;
    GET_CURRENT_CONTEXT(ctx);
-   GLsizeiptr size;
 
    if (target != GL_TRANSFORM_FEEDBACK_BUFFER) {
       _mesa_error(ctx, GL_INVALID_ENUM, "glBindBufferOffsetEXT(target)");
@@ -497,12 +589,7 @@ _mesa_BindBufferOffsetEXT(GLenum target, GLuint index, GLuint buffer,
       return;
    }
 
-   /* default size is the buffer size rounded down to nearest
-    * multiple of four.
-    */
-   size = (bufObj->Size - offset) & ~0x3;
-
-   bind_buffer_range(ctx, index, bufObj, offset, size);
+   bind_buffer_range(ctx, index, bufObj, offset, 0);
 }
 
 
@@ -708,14 +795,19 @@ _mesa_GenTransformFeedbacks(GLsizei n, GLuint *names)
 GLboolean GLAPIENTRY
 _mesa_IsTransformFeedback(GLuint name)
 {
+   struct gl_transform_feedback_object *obj;
    GET_CURRENT_CONTEXT(ctx);
 
    ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE);
 
-   if (name && _mesa_lookup_transform_feedback_object(ctx, name))
-      return GL_TRUE;
-   else
+   if (name == 0)
       return GL_FALSE;
+
+   obj = _mesa_lookup_transform_feedback_object(ctx, name);
+   if (obj == NULL)
+      return GL_FALSE;
+
+   return obj->EverBound;
 }
 
 
@@ -734,8 +826,7 @@ _mesa_BindTransformFeedback(GLenum target, GLuint name)
       return;
    }
 
-   if (ctx->TransformFeedback.CurrentObject->Active &&
-       !ctx->TransformFeedback.CurrentObject->Paused) {
+   if (_mesa_is_xfb_active_and_unpaused(ctx)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
               "glBindTransformFeedback(transform is active, or not paused)");
       return;
@@ -805,7 +896,7 @@ _mesa_PauseTransformFeedback(void)
 
    obj = ctx->TransformFeedback.CurrentObject;
 
-   if (!obj->Active || obj->Paused) {
+   if (!_mesa_is_xfb_active_and_unpaused(ctx)) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
            "glPauseTransformFeedback(feedback not active or already paused)");
       return;
diff --git a/mesalib/src/mesa/main/transformfeedback.h b/mesalib/src/mesa/main/transformfeedback.h
index 01c2af359..c128746b3 100644
--- a/mesalib/src/mesa/main/transformfeedback.h
+++ b/mesalib/src/mesa/main/transformfeedback.h
@@ -25,9 +25,11 @@
 #ifndef TRANSFORM_FEEDBACK_H
 #define TRANSFORM_FEEDBACK_H
 
+#include <stdbool.h>
 #include "compiler.h"
 #include "glheader.h"
 #include "mfeatures.h"
+#include "mtypes.h"
 
 struct _glapi_table;
 struct dd_function_table;
@@ -46,6 +48,11 @@ _mesa_validate_transform_feedback_buffers(struct gl_context *ctx);
 extern void
 _mesa_init_transform_feedback_functions(struct dd_function_table *driver);
 
+extern unsigned
+_mesa_compute_max_transform_feedback_vertices(
+      const struct gl_transform_feedback_object *obj,
+      const struct gl_transform_feedback_info *info);
+
 
 /*** GL_EXT_transform_feedback ***/
 
@@ -106,4 +113,11 @@ _mesa_PauseTransformFeedback(void);
 extern void GLAPIENTRY
 _mesa_ResumeTransformFeedback(void);
 
+static inline bool
+_mesa_is_xfb_active_and_unpaused(const struct gl_context *ctx)
+{
+   return ctx->TransformFeedback.CurrentObject->Active &&
+      !ctx->TransformFeedback.CurrentObject->Paused;
+}
+
 #endif /* TRANSFORM_FEEDBACK_H */
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index b6b73d16f..afe9a08c1 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -241,11 +241,14 @@ validate_uniform_parameters(struct gl_context *ctx,
       return false;
    }
 
-   /* This case should be impossible.  The implication is that a call like
-    * glGetUniformLocation(prog, "foo[8]") was successful but "foo" is not an
-    * array.
+   /* If the uniform is an array, check that array_index is in bounds.
+    * If not an array, check that array_index is zero.
+    * array_index is unsigned so no need to check for less than zero.
     */
-   if (*array_index != 0 && shProg->UniformStorage[*loc].array_elements == 0) {
+   unsigned limit = shProg->UniformStorage[*loc].array_elements;
+   if (limit == 0)
+      limit = 1;
+   if (*array_index >= limit) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "%s(location=%d)",
 		  caller, location);
       return false;
@@ -732,9 +735,6 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
     * will have already generated an error.
     */
    if (uni->array_elements != 0) {
-      if (offset >= uni->array_elements)
-	 return;
-
       count = MIN2(count, (int) (uni->array_elements - offset));
    }
 
@@ -889,9 +889,6 @@ _mesa_uniform_matrix(struct gl_context *ctx, struct gl_shader_program *shProg,
     * will have already generated an error.
     */
    if (uni->array_elements != 0) {
-      if (offset >= uni->array_elements)
-	 return;
-
       count = MIN2(count, (int) (uni->array_elements - offset));
    }
 
@@ -1025,10 +1022,13 @@ _mesa_get_uniform_location(struct gl_context *ctx,
    if (!found)
       return GL_INVALID_INDEX;
 
-   /* Since array_elements is 0 for non-arrays, this causes look-ups of 'a[0]'
-    * to (correctly) fail if 'a' is not an array.
+   /* If the uniform is an array, fail if the index is out of bounds.
+    * (A negative index is caught above.)  This also fails if the uniform
+    * is not an array, but the user is trying to index it, because
+    * array_elements is zero and offset >= 0.
     */
-   if (array_lookup && shProg->UniformStorage[location].array_elements == 0) {
+   if (array_lookup
+       && offset >= (long) shProg->UniformStorage[location].array_elements) {
       return GL_INVALID_INDEX;
    }
 
diff --git a/mesalib/src/mesa/state_tracker/st_atom_array.c b/mesalib/src/mesa/state_tracker/st_atom_array.c
index 15f5d1c95..2437245bd 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_array.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_array.c
@@ -403,9 +403,9 @@ setup_interleaved_attribs(const struct st_vertex_program *vp,
       const GLuint mesaAttr = vp->index_to_input[attr];
       const struct gl_client_array *array = arrays[mesaAttr];
       unsigned src_offset = (unsigned) (array->Ptr - low_addr);
-      GLuint element_size = array->_ElementSize;
 
-      assert(element_size == array->Size * _mesa_sizeof_type(array->Type));
+      assert(array->_ElementSize ==
+             _mesa_bytes_per_vertex_attrib(array->Size, array->Type));
 
       velements[attr].src_offset = src_offset;
       velements[attr].instance_divisor = array->InstanceDivisor;
@@ -474,7 +474,8 @@ setup_non_interleaved_attribs(struct st_context *st,
       struct gl_buffer_object *bufobj = array->BufferObj;
       GLsizei stride = array->StrideB;
 
-      assert(array->_ElementSize == array->Size * _mesa_sizeof_type(array->Type));
+      assert(array->_ElementSize ==
+             _mesa_bytes_per_vertex_attrib(array->Size, array->Type));
 
       if (_mesa_is_bufferobj(bufobj)) {
          /* Attribute data is in a VBO.
diff --git a/mesalib/src/mesa/state_tracker/st_cb_flush.c b/mesalib/src/mesa/state_tracker/st_cb_flush.c
index 7c9f91f1b..b569e3b3c 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_flush.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_flush.c
@@ -76,15 +76,16 @@ display_front_buffer(struct st_context *st)
 }
 
 
-void st_flush( struct st_context *st,
-               struct pipe_fence_handle **fence )
+void st_flush(struct st_context *st,
+              struct pipe_fence_handle **fence,
+              enum pipe_flush_flags flags)
 {
    FLUSH_VERTICES(st->ctx, 0);
    FLUSH_CURRENT(st->ctx, 0);
 
    st_flush_bitmap_cache(st);
 
-   st->pipe->flush( st->pipe, fence );
+   st->pipe->flush(st->pipe, fence, flags);
 }
 
 
@@ -95,7 +96,7 @@ void st_finish( struct st_context *st )
 {
    struct pipe_fence_handle *fence = NULL;
 
-   st_flush(st, &fence);
+   st_flush(st, &fence, 0);
 
    if(fence) {
       st->pipe->screen->fence_finish(st->pipe->screen, fence,
@@ -118,7 +119,7 @@ static void st_glFlush(struct gl_context *ctx)
     * synchronization issues.  Calling finish() here will just hide
     * problems that need to be fixed elsewhere.
     */
-   st_flush(st, NULL);
+   st_flush(st, NULL, 0);
 
    if (is_front_buffer_dirty(st)) {
       display_front_buffer(st);
diff --git a/mesalib/src/mesa/state_tracker/st_cb_flush.h b/mesalib/src/mesa/state_tracker/st_cb_flush.h
index 598536ba0..003e2a2a9 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_flush.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_flush.h
@@ -41,7 +41,8 @@ st_init_flush_functions(struct dd_function_table *functions);
 
 extern void
 st_flush(struct st_context *st,
-         struct pipe_fence_handle **fence);
+         struct pipe_fence_handle **fence,
+         enum pipe_flush_flags flags);
 
 extern void
 st_finish(struct st_context *st);
diff --git a/mesalib/src/mesa/state_tracker/st_cb_queryobj.c b/mesalib/src/mesa/state_tracker/st_cb_queryobj.c
index b10fd9335..98b61f576 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_queryobj.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_queryobj.c
@@ -72,6 +72,11 @@ st_DeleteQuery(struct gl_context *ctx, struct gl_query_object *q)
       stq->pq = NULL;
    }
 
+   if (stq->pq_begin) {
+      pipe->destroy_query(pipe, stq->pq_begin);
+      stq->pq_begin = NULL;
+   }
+
    free(stq);
 }
 
@@ -79,7 +84,8 @@ st_DeleteQuery(struct gl_context *ctx, struct gl_query_object *q)
 static void
 st_BeginQuery(struct gl_context *ctx, struct gl_query_object *q)
 {
-   struct pipe_context *pipe = st_context(ctx)->pipe;
+   struct st_context *st = st_context(ctx);
+   struct pipe_context *pipe = st->pipe;
    struct st_query_object *stq = st_query_object(q);
    unsigned type;
 
@@ -98,29 +104,46 @@ st_BeginQuery(struct gl_context *ctx, struct gl_query_object *q)
    case GL_TRANSFORM_FEEDBACK_PRIMITIVES_WRITTEN:
       type = PIPE_QUERY_PRIMITIVES_EMITTED;
       break;
-   case GL_TIME_ELAPSED_EXT:
-      type = PIPE_QUERY_TIME_ELAPSED;
+   case GL_TIME_ELAPSED:
+      if (st->has_time_elapsed)
+         type = PIPE_QUERY_TIME_ELAPSED;
+      else
+         type = PIPE_QUERY_TIMESTAMP;
       break;
    default:
       assert(0 && "unexpected query target in st_BeginQuery()");
       return;
    }
 
-   if (stq->pq && stq->type != type) {
+   if (stq->type != type) {
       /* free old query of different type */
-      pipe->destroy_query(pipe, stq->pq);
-      stq->pq = NULL;
+      if (stq->pq) {
+         pipe->destroy_query(pipe, stq->pq);
+         stq->pq = NULL;
+      }
+      if (stq->pq_begin) {
+         pipe->destroy_query(pipe, stq->pq_begin);
+         stq->pq_begin = NULL;
+      }
       stq->type = PIPE_QUERY_TYPES; /* an invalid value */
    }
 
-   if (!stq->pq) {
-      stq->pq = pipe->create_query(pipe, type);
-      stq->type = type;
+   if (q->Target == GL_TIME_ELAPSED &&
+       type == PIPE_QUERY_TIMESTAMP) {
+      /* Determine time elapsed by emitting two timestamp queries. */
+      if (!stq->pq_begin) {
+         stq->pq_begin = pipe->create_query(pipe, type);
+         stq->type = type;
+      }
+      pipe->end_query(pipe, stq->pq_begin);
+   } else {
+      if (!stq->pq) {
+         stq->pq = pipe->create_query(pipe, type);
+         stq->type = type;
+      }
+      pipe->begin_query(pipe, stq->pq);
    }
-
    assert(stq->type == type);
-
-   pipe->begin_query(pipe, stq->pq);
 }
 
 
@@ -132,7 +155,9 @@ st_EndQuery(struct gl_context *ctx, struct gl_query_object *q)
 
    st_flush_bitmap_cache(st_context(ctx));
 
-   if (q->Target == GL_TIMESTAMP && !stq->pq) {
+   if ((q->Target == GL_TIMESTAMP ||
+        q->Target == GL_TIME_ELAPSED) &&
+       !stq->pq) {
       stq->pq = pipe->create_query(pipe, PIPE_QUERY_TIMESTAMP);
       stq->type = PIPE_QUERY_TIMESTAMP;
    }
@@ -141,6 +166,33 @@ st_EndQuery(struct gl_context *ctx, struct gl_query_object *q)
 }
 
 
+static boolean
+get_query_result(struct pipe_context *pipe,
+                 struct st_query_object *stq,
+                 boolean wait)
+{
+   if (!pipe->get_query_result(pipe,
+                               stq->pq,
+                               wait,
+                               (void *)&stq->base.Result)) {
+      return FALSE;
+   }
+
+   if (stq->base.Target == GL_TIME_ELAPSED &&
+       stq->type == PIPE_QUERY_TIMESTAMP) {
+      /* Calculate the elapsed time from the two timestamp queries */
+      GLuint64EXT Result0 = 0;
+      assert(stq->pq_begin);
+      pipe->get_query_result(pipe, stq->pq_begin, TRUE, (void *)&Result0);
+      stq->base.Result -= Result0;
+   } else {
+      assert(!stq->pq_begin);
+   }
+
+   return TRUE;
+}
+
+
 static void
 st_WaitQuery(struct gl_context *ctx, struct gl_query_object *q)
 {
@@ -151,10 +203,7 @@ st_WaitQuery(struct gl_context *ctx, struct gl_query_object *q)
    assert(!stq->base.Ready);
 
    while (!stq->base.Ready &&
-	  !pipe->get_query_result(pipe, 
-				  stq->pq,
-				  TRUE,
-				  (void*)&q->Result))
+	  !get_query_result(pipe, stq, TRUE))
    {
       /* nothing */
    }
@@ -169,7 +218,7 @@ st_CheckQuery(struct gl_context *ctx, struct gl_query_object *q)
    struct pipe_context *pipe = st_context(ctx)->pipe;
    struct st_query_object *stq = st_query_object(q);
    assert(!q->Ready);   /* we should not get called if Ready is TRUE */
-   q->Ready = pipe->get_query_result(pipe, stq->pq, FALSE, (void*)&q->Result);
+   q->Ready = get_query_result(pipe, stq, FALSE);
 }
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
index 03487b153..d5b107532 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
@@ -39,6 +39,10 @@ struct st_query_object
 {
    struct gl_query_object base;
    struct pipe_query *pq;
+
+   /* Begin TIMESTAMP query for GL_TIME_ELAPSED_EXT queries */
+   struct pipe_query *pq_begin;
+
    unsigned type;  /**< PIPE_QUERY_x */
 };
 
diff --git a/mesalib/src/mesa/state_tracker/st_cb_syncobj.c b/mesalib/src/mesa/state_tracker/st_cb_syncobj.c
index f9f2348a3..94bf4861d 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_syncobj.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_syncobj.c
@@ -72,7 +72,7 @@ static void st_fence_sync(struct gl_context *ctx, struct gl_sync_object *obj,
    assert(condition == GL_SYNC_GPU_COMMANDS_COMPLETE && flags == 0);
    assert(so->fence == NULL);
 
-   pipe->flush(pipe, &so->fence);
+   pipe->flush(pipe, &so->fence, 0);
 }
 
 static void st_check_sync(struct gl_context *ctx, struct gl_sync_object *obj)
diff --git a/mesalib/src/mesa/state_tracker/st_cb_texture.c b/mesalib/src/mesa/state_tracker/st_cb_texture.c
index bf13526d2..02fc675ae 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_texture.c
@@ -201,7 +201,7 @@ st_MapTextureImage(struct gl_context *ctx,
    if (mode & GL_MAP_INVALIDATE_RANGE_BIT)
       pipeMode |= PIPE_TRANSFER_DISCARD_RANGE;
 
-   map = st_texture_image_map(st, stImage, slice, pipeMode, x, y, w, h);
+   map = st_texture_image_map(st, stImage, pipeMode, x, y, slice, w, h, 1);
    if (map) {
       *mapOut = map;
       *rowStrideOut = stImage->transfer->stride;
@@ -572,37 +572,68 @@ decompress_with_blit(struct gl_context * ctx,
 {
    struct st_context *st = st_context(ctx);
    struct pipe_context *pipe = st->pipe;
-   struct st_texture_image *stImage = st_texture_image(texImage);
-   struct st_texture_object *stObj = st_texture_object(texImage->TexObject);
+   struct pipe_screen *screen = pipe->screen;
    const GLuint width = texImage->Width;
    const GLuint height = texImage->Height;
-   struct pipe_resource *dst_texture;
+   const GLuint depth = texImage->Depth;
+   struct pipe_resource *src = st_texture_object(texImage->TexObject)->pt;
+   struct pipe_resource *dst;
+   struct pipe_resource dst_templ;
+   enum pipe_format pipe_format;
+   gl_format mesa_format;
+   GLenum gl_target = texImage->TexObject->Target;
+   enum pipe_texture_target pipe_target;
    struct pipe_blit_info blit;
    unsigned bind = (PIPE_BIND_RENDER_TARGET | PIPE_BIND_TRANSFER_READ);
    struct pipe_transfer *tex_xfer;
    ubyte *map;
 
-   /* create temp / dest surface */
-   if (!util_create_rgba_texture(pipe, width, height, bind,
-                                 &dst_texture)) {
-      _mesa_problem(ctx, "util_create_rgba_texture() failed "
-                    "in decompress_with_blit()");
+   /* GetTexImage only returns a single face for cubemaps. */
+   if (gl_target == GL_TEXTURE_CUBE_MAP) {
+      gl_target = GL_TEXTURE_2D;
+   }
+
+   pipe_target = gl_target_to_pipe(gl_target);
+
+   /* Find the best match for the format+type combo. */
+   pipe_format = st_choose_format(pipe->screen, GL_RGBA8, format, type,
+                                  pipe_target, 0, bind);
+   if (pipe_format == PIPE_FORMAT_NONE) {
+      /* unable to get an rgba format!?! */
+      _mesa_problem(ctx, "%s: cannot find a supported format", __func__);
       return;
    }
 
-   blit.src.resource = stObj->pt;
+   /* create the destination texture */
+   memset(&dst_templ, 0, sizeof(dst_templ));
+   dst_templ.target = pipe_target;
+   dst_templ.format = pipe_format;
+   dst_templ.bind = bind;
+   dst_templ.usage = PIPE_USAGE_STAGING;
+
+   st_gl_texture_dims_to_pipe_dims(gl_target, width, height, depth,
+                                   &dst_templ.width0, &dst_templ.height0,
+                                   &dst_templ.depth0, &dst_templ.array_size);
+
+   dst = screen->resource_create(screen, &dst_templ);
+   if (!dst) {
+      _mesa_problem(ctx, "%s: cannot create a temporary texture", __func__);
+      return;
+   }
+
+   blit.src.resource = src;
    blit.src.level = texImage->Level;
-   blit.src.format = util_format_linear(stObj->pt->format);
-   blit.dst.resource = dst_texture;
+   blit.src.format = util_format_linear(src->format);
+   blit.dst.resource = dst;
    blit.dst.level = 0;
-   blit.dst.format = dst_texture->format;
+   blit.dst.format = dst->format;
    blit.src.box.x = blit.dst.box.x = 0;
    blit.src.box.y = blit.dst.box.y = 0;
-   blit.src.box.z = 0; /* XXX compressed array textures? */
+   blit.src.box.z = texImage->Face;
    blit.dst.box.z = 0;
    blit.src.box.width = blit.dst.box.width = width;
    blit.src.box.height = blit.dst.box.height = height;
-   blit.src.box.depth = blit.dst.box.depth = 1;
+   blit.src.box.depth = blit.dst.box.depth = depth;
    blit.mask = PIPE_MASK_RGBA;
    blit.filter = PIPE_TEX_FILTER_NEAREST;
    blit.scissor_enable = FALSE;
@@ -612,33 +643,38 @@ decompress_with_blit(struct gl_context * ctx,
 
    pixels = _mesa_map_pbo_dest(ctx, &ctx->Pack, pixels);
 
-   map = pipe_transfer_map(pipe, dst_texture, 0, 0,
-                           PIPE_TRANSFER_READ,
-                           0, 0, width, height, &tex_xfer);
+   map = pipe_transfer_map_3d(pipe, dst, 0, PIPE_TRANSFER_READ,
+                              0, 0, 0, width, height, depth, &tex_xfer);
    if (!map) {
       goto end;
    }
 
+   mesa_format = st_pipe_format_to_mesa_format(pipe_format);
+
    /* copy/pack data into user buffer */
-   if (_mesa_format_matches_format_and_type(stImage->base.TexFormat,
-                                            format, type,
+   if (_mesa_format_matches_format_and_type(mesa_format, format, type,
                                             ctx->Pack.SwapBytes)) {
       /* memcpy */
-      const uint bytesPerRow = width * util_format_get_blocksize(stImage->pt->format);
-      /* map the dst_surface so we can read from it */
-      GLuint row;
-      for (row = 0; row < height; row++) {
-         GLvoid *dest = _mesa_image_address2d(&ctx->Pack, pixels, width,
-                                              height, format, type, row, 0);
-         memcpy(dest, map, bytesPerRow);
-         map += tex_xfer->stride;
+      const uint bytesPerRow = width * util_format_get_blocksize(pipe_format);
+      GLuint row, slice;
+
+      for (slice = 0; slice < depth; slice++) {
+         ubyte *slice_map = map;
+
+         for (row = 0; row < height; row++) {
+            GLvoid *dest = _mesa_image_address3d(&ctx->Pack, pixels,
+                                                 width, height, format,
+                                                 type, slice, row, 0);
+            memcpy(dest, slice_map, bytesPerRow);
+            slice_map += tex_xfer->stride;
+         }
+         map += tex_xfer->layer_stride;
       }
-      pipe_transfer_unmap(pipe, tex_xfer);
    }
    else {
       /* format translation via floats */
-      GLuint row;
-      enum pipe_format pformat = util_format_linear(dst_texture->format);
+      GLuint row, slice;
+      enum pipe_format pformat = util_format_linear(dst->format);
       GLfloat *rgba;
 
       rgba = malloc(width * 4 * sizeof(GLfloat));
@@ -647,20 +683,24 @@ decompress_with_blit(struct gl_context * ctx,
          goto end;
       }
 
-      for (row = 0; row < height; row++) {
-         const GLbitfield transferOps = 0x0; /* bypassed for glGetTexImage() */
-         GLvoid *dest = _mesa_image_address2d(&ctx->Pack, pixels, width,
-                                              height, format, type, row, 0);
+      for (slice = 0; slice < depth; slice++) {
+         for (row = 0; row < height; row++) {
+            const GLbitfield transferOps = 0x0; /* bypassed for glGetTexImage() */
+            GLvoid *dest = _mesa_image_address3d(&ctx->Pack, pixels,
+                                                 width, height, format,
+                                                 type, slice, row, 0);
 
-         if (ST_DEBUG & DEBUG_FALLBACK)
-            debug_printf("%s: fallback format translation\n", __FUNCTION__);
+            if (ST_DEBUG & DEBUG_FALLBACK)
+               debug_printf("%s: fallback format translation\n", __FUNCTION__);
 
-         /* get float[4] rgba row from surface */
-         pipe_get_tile_rgba_format(tex_xfer, map, 0, row, width, 1,
-                                   pformat, rgba);
+            /* get float[4] rgba row from surface */
+            pipe_get_tile_rgba_format(tex_xfer, map, 0, row, width, 1,
+                                      pformat, rgba);
 
-         _mesa_pack_rgba_span_float(ctx, width, (GLfloat (*)[4]) rgba, format,
-                                    type, dest, &ctx->Pack, transferOps);
+            _mesa_pack_rgba_span_float(ctx, width, (GLfloat (*)[4]) rgba, format,
+                                       type, dest, &ctx->Pack, transferOps);
+         }
+         map += tex_xfer->layer_stride;
       }
 
       free(rgba);
@@ -671,7 +711,7 @@ end:
       pipe_transfer_unmap(pipe, tex_xfer);
 
    _mesa_unmap_pbo_dest(ctx, &ctx->Pack);
-   pipe_resource_reference(&dst_texture, NULL);
+   pipe_resource_reference(&dst, NULL);
 }
 
 
@@ -722,6 +762,9 @@ fallback_copy_texsubimage(struct gl_context *ctx,
    GLvoid *texDest;
    enum pipe_transfer_usage transfer_usage;
    void *map;
+   unsigned dst_width = width;
+   unsigned dst_height = height;
+   unsigned dst_depth = 1;
 
    if (ST_DEBUG & DEBUG_FALLBACK)
       debug_printf("%s: fallback processing\n", __FUNCTION__);
@@ -730,6 +773,14 @@ fallback_copy_texsubimage(struct gl_context *ctx,
       srcY = strb->Base.Height - srcY - height;
    }
 
+   if (stImage->pt->target == PIPE_TEXTURE_1D_ARRAY) {
+      /* Move y/height to z/depth for 1D array textures.  */
+      destZ = destY;
+      destY = 0;
+      dst_depth = dst_height;
+      dst_height = 1;
+   }
+
    map = pipe_transfer_map(pipe,
                            strb->texture,
                            strb->rtt_level,
@@ -745,9 +796,9 @@ fallback_copy_texsubimage(struct gl_context *ctx,
    else
       transfer_usage = PIPE_TRANSFER_WRITE;
 
-   /* XXX this used to ignore destZ param */
-   texDest = st_texture_image_map(st, stImage, destZ, transfer_usage,
-                                  destX, destY, width, height);
+   texDest = st_texture_image_map(st, stImage, transfer_usage,
+                                  destX, destY, destZ,
+                                  dst_width, dst_height, dst_depth);
 
    if (baseFormat == GL_DEPTH_COMPONENT ||
        baseFormat == GL_DEPTH_STENCIL) {
@@ -775,8 +826,16 @@ fallback_copy_texsubimage(struct gl_context *ctx,
             if (scaleOrBias) {
                _mesa_scale_and_bias_depth_uint(ctx, width, data);
             }
-            pipe_put_tile_z(stImage->transfer, texDest, 0, row, width, 1,
-                            data);
+
+            if (stImage->pt->target == PIPE_TEXTURE_1D_ARRAY) {
+               pipe_put_tile_z(stImage->transfer,
+                               texDest + row*stImage->transfer->layer_stride,
+                               0, 0, width, 1, data);
+            }
+            else {
+               pipe_put_tile_z(stImage->transfer, texDest, 0, row, width, 1,
+                               data);
+            }
          }
       }
       else {
@@ -792,7 +851,7 @@ fallback_copy_texsubimage(struct gl_context *ctx,
 
       if (tempSrc && texDest) {
          const GLint dims = 2;
-         const GLint dstRowStride = stImage->transfer->stride;
+         GLint dstRowStride;
          struct gl_texture_image *texImage = &stImage->base;
          struct gl_pixelstore_attrib unpack = ctx->DefaultPacking;
 
@@ -800,6 +859,13 @@ fallback_copy_texsubimage(struct gl_context *ctx,
             unpack.Invert = GL_TRUE;
          }
 
+         if (stImage->pt->target == PIPE_TEXTURE_1D_ARRAY) {
+            dstRowStride = stImage->transfer->layer_stride;
+         }
+         else {
+            dstRowStride = stImage->transfer->stride;
+         }
+
          /* get float/RGBA image from framebuffer */
          /* XXX this usually involves a lot of int/float conversion.
           * try to avoid that someday.
@@ -901,6 +967,43 @@ compatible_src_dst_formats(struct gl_context *ctx,
 }
 
 
+/**
+ * Do pipe->blit. Return FALSE if the blitting is unsupported
+ * for the given formats.
+ */
+static GLboolean
+st_pipe_blit(struct pipe_context *pipe, struct pipe_blit_info *blit)
+{
+   struct pipe_screen *screen = pipe->screen;
+   unsigned dst_usage;
+
+   if (util_format_is_depth_or_stencil(blit->dst.format)) {
+      dst_usage = PIPE_BIND_DEPTH_STENCIL;
+   }
+   else {
+      dst_usage = PIPE_BIND_RENDER_TARGET;
+   }
+
+   /* try resource_copy_region in case the format is not supported
+    * for rendering */
+   if (util_try_blit_via_copy_region(pipe, blit)) {
+      return GL_TRUE; /* done */
+   }
+
+   /* check the format support */
+   if (!screen->is_format_supported(screen, blit->src.format,
+                                    PIPE_TEXTURE_2D, 0,
+                                    PIPE_BIND_SAMPLER_VIEW) ||
+       !screen->is_format_supported(screen, blit->dst.format,
+                                    PIPE_TEXTURE_2D, 0,
+                                    dst_usage)) {
+      return GL_FALSE;
+   }
+
+   pipe->blit(pipe, blit);
+   return GL_TRUE;
+}
+
 
 /**
  * Do a CopyTex[Sub]Image1/2/3D() using a hardware (blit) path if possible.
@@ -929,7 +1032,7 @@ st_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
    struct pipe_surface surf_tmpl;
    unsigned dst_usage;
    unsigned blit_mask;
-   GLint srcY0, srcY1;
+   GLint srcY0, srcY1, yStep;
 
    /* make sure finalize_textures has been called? 
     */
@@ -950,10 +1053,12 @@ st_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
    if (do_flip) {
       srcY1 = strb->Base.Height - srcY - height;
       srcY0 = srcY1 + height;
+      yStep = -1;
    }
    else {
       srcY0 = srcY;
       srcY1 = srcY0 + height;
+      yStep = 1;
    }
 
    if (ctx->_ImageTransferState) {
@@ -965,16 +1070,6 @@ st_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
       goto fallback;
    }
 
-   if (texImage->TexObject->Target == GL_TEXTURE_1D_ARRAY) {
-      /* 1D arrays might be thought of as 2D images but the actual layout
-       * might not be that way.  At some points, we convert OpenGL's 1D
-       * array 'height' into gallium 'layers' and that prevents the blit
-       * utility code from doing the right thing.  Simpy use the memcpy-based
-       * fallback.
-       */
-      goto fallback;
-   }
-
    /* Set the blit writemask. */
    switch (texBaseFormat) {
    case GL_DEPTH_STENCIL:
@@ -1055,27 +1150,38 @@ st_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
       blit.mask = blit_mask;
       blit.filter = PIPE_TEX_FILTER_NEAREST;
 
-      /* try resource_copy_region in case the format is not supported
-       * for rendering */
-      if (util_try_blit_via_copy_region(pipe, &blit)) {
-         return; /* done */
-      }
+      /* 1D array textures need special treatment.
+       * Blit rows from the source to layers in the destination. */
+      if (texImage->TexObject->Target == GL_TEXTURE_1D_ARRAY) {
+         int y, layer;
 
-      /* check the format support */
-      if (!screen->is_format_supported(screen, src_format,
-                                       PIPE_TEXTURE_2D, 0,
-                                       PIPE_BIND_SAMPLER_VIEW) ||
-          !screen->is_format_supported(screen, dest_format,
-                                       PIPE_TEXTURE_2D, 0,
-                                       dst_usage)) {
-         goto fallback;
-      }
+         for (y = srcY0, layer = 0; layer < height; y += yStep, layer++) {
+            blit.src.box.y = y;
+            blit.src.box.height = 1;
+            blit.dst.box.y = 0;
+            blit.dst.box.height = 1;
+            blit.dst.box.z = destY + layer;
 
-      pipe->blit(pipe, &blit);
+            if (!st_pipe_blit(pipe, &blit)) {
+               goto fallback;
+            }
+         }
+      }
+      else {
+         /* All the other texture targets. */
+         if (!st_pipe_blit(pipe, &blit)) {
+            goto fallback;
+         }
+      }
       return;
    }
 
    /* try u_blit */
+   if (texImage->TexObject->Target == GL_TEXTURE_1D_ARRAY) {
+      /* u_blit cannot copy 1D array textures as required by CopyTexSubImage */
+      goto fallback;
+   }
+
    color_writemask = compatible_src_dst_formats(ctx, &strb->Base, texImage);
 
    if (!color_writemask ||
diff --git a/mesalib/src/mesa/state_tracker/st_context.h b/mesalib/src/mesa/state_tracker/st_context.h
index eeb7f0c4f..2be74b452 100644
--- a/mesalib/src/mesa/state_tracker/st_context.h
+++ b/mesalib/src/mesa/state_tracker/st_context.h
@@ -83,6 +83,7 @@ struct st_context
    GLboolean clamp_frag_color_in_shader;
    GLboolean clamp_vert_color_in_shader;
    boolean has_stencil_export; /**< can do shader stencil export? */
+   boolean has_time_elapsed;
 
    /* On old libGL's for linux we need to invalidate the drawables
     * on glViewpport calls, this is set via a option.
diff --git a/mesalib/src/mesa/state_tracker/st_extensions.c b/mesalib/src/mesa/state_tracker/st_extensions.c
index 262cc3d6c..e7cf4f3bb 100644
--- a/mesalib/src/mesa/state_tracker/st_extensions.c
+++ b/mesalib/src/mesa/state_tracker/st_extensions.c
@@ -371,6 +371,7 @@ void st_init_extensions(struct st_context *st)
       { o(ARB_shader_texture_lod),           PIPE_CAP_SM3                              },
       { o(ARB_shadow),                       PIPE_CAP_TEXTURE_SHADOW_MAP               },
       { o(ARB_texture_non_power_of_two),     PIPE_CAP_NPOT_TEXTURES                    },
+      { o(ARB_timer_query),                  PIPE_CAP_QUERY_TIMESTAMP                  },
       { o(ARB_transform_feedback2),          PIPE_CAP_STREAM_OUTPUT_PAUSE_RESUME       },
       { o(ARB_transform_feedback3),          PIPE_CAP_STREAM_OUTPUT_PAUSE_RESUME       },
 
@@ -382,7 +383,6 @@ void st_init_extensions(struct st_context *st)
       { o(EXT_texture_filter_anisotropic),   PIPE_CAP_ANISOTROPIC_FILTER               },
       { o(EXT_texture_mirror_clamp),         PIPE_CAP_TEXTURE_MIRROR_CLAMP             },
       { o(EXT_texture_swizzle),              PIPE_CAP_TEXTURE_SWIZZLE                  },
-      { o(EXT_timer_query),                  PIPE_CAP_TIMER_QUERY                      },
       { o(EXT_transform_feedback),           PIPE_CAP_MAX_STREAM_OUTPUT_BUFFERS        },
 
       { o(AMD_seamless_cubemap_per_texture), PIPE_CAP_SEAMLESS_CUBE_MAP_PER_TEXTURE    },
@@ -652,9 +652,12 @@ void st_init_extensions(struct st_context *st)
    if (ctx->Const.MaxDualSourceDrawBuffers > 0)
       ctx->Extensions.ARB_blend_func_extended = GL_TRUE;
 
-   if (screen->get_param(screen, PIPE_CAP_TIMER_QUERY) &&
-       screen->get_param(screen, PIPE_CAP_QUERY_TIMESTAMP)) {
-      ctx->Extensions.ARB_timer_query = GL_TRUE;
+   st->has_time_elapsed =
+      screen->get_param(screen, PIPE_CAP_QUERY_TIME_ELAPSED);
+
+   if (st->has_time_elapsed ||
+       ctx->Extensions.ARB_timer_query) {
+      ctx->Extensions.EXT_timer_query = GL_TRUE;
    }
 
    if (ctx->Extensions.ARB_transform_feedback2 &&
diff --git a/mesalib/src/mesa/state_tracker/st_manager.c b/mesalib/src/mesa/state_tracker/st_manager.c
index 30f5ca5ed..a3a67712a 100644
--- a/mesalib/src/mesa/state_tracker/st_manager.c
+++ b/mesalib/src/mesa/state_tracker/st_manager.c
@@ -455,7 +455,13 @@ st_context_flush(struct st_context_iface *stctxi, unsigned flags,
                  struct pipe_fence_handle **fence)
 {
    struct st_context *st = (struct st_context *) stctxi;
-   st_flush(st, fence);
+   enum pipe_flush_flags pipe_flags = 0;
+
+   if (flags & ST_FLUSH_END_OF_FRAME) {
+      pipe_flags |= PIPE_FLUSH_END_OF_FRAME;
+   }
+
+   st_flush(st, fence, pipe_flags);
    if (flags & ST_FLUSH_FRONT)
       st_manager_flush_frontbuffer(st);
 }
diff --git a/mesalib/src/mesa/state_tracker/st_texture.c b/mesalib/src/mesa/state_tracker/st_texture.c
index 5a4dcaab2..ee4d7622d 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_texture.c
@@ -232,8 +232,9 @@ st_texture_match_image(const struct pipe_resource *pt,
  */
 GLubyte *
 st_texture_image_map(struct st_context *st, struct st_texture_image *stImage,
-                     GLuint zoffset, enum pipe_transfer_usage usage,
-                     GLuint x, GLuint y, GLuint w, GLuint h)
+                     enum pipe_transfer_usage usage,
+                     GLuint x, GLuint y, GLuint z,
+                     GLuint w, GLuint h, GLuint d)
 {
    struct st_texture_object *stObj =
       st_texture_object(stImage->base.TexObject);
@@ -249,9 +250,9 @@ st_texture_image_map(struct st_context *st, struct st_texture_image *stImage,
    else
       level = stImage->base.Level;
 
-   return pipe_transfer_map(st->pipe, stImage->pt, level,
-                            stImage->base.Face + zoffset,
-                            usage, x, y, w, h, &stImage->transfer);
+   return pipe_transfer_map_3d(st->pipe, stImage->pt, level, usage,
+                               x, y, z + stImage->base.Face,
+                               w, h, d, &stImage->transfer);
 }
 
 
diff --git a/mesalib/src/mesa/state_tracker/st_texture.h b/mesalib/src/mesa/state_tracker/st_texture.h
index 62e975bb4..8a701009a 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.h
+++ b/mesalib/src/mesa/state_tracker/st_texture.h
@@ -176,12 +176,10 @@ st_texture_match_image(const struct pipe_resource *pt,
  * well.
  */
 extern GLubyte *
-st_texture_image_map(struct st_context *st,
-                     struct st_texture_image *stImage,
-		     GLuint zoffset,
+st_texture_image_map(struct st_context *st, struct st_texture_image *stImage,
                      enum pipe_transfer_usage usage,
-                     unsigned x, unsigned y,
-                     unsigned w, unsigned h);
+                     GLuint x, GLuint y, GLuint z,
+                     GLuint w, GLuint h, GLuint d);
 
 extern void
 st_texture_image_unmap(struct st_context *st,
diff --git a/mesalib/src/mesa/vbo/vbo.h b/mesalib/src/mesa/vbo/vbo.h
index 04930b971..49dab2314 100644
--- a/mesalib/src/mesa/vbo/vbo.h
+++ b/mesalib/src/mesa/vbo/vbo.h
@@ -156,7 +156,8 @@ void vbo_check_buffers_are_unmapped(struct gl_context *ctx);
 void vbo_bind_arrays(struct gl_context *ctx);
 
 size_t
-vbo_count_tessellated_primitives(const struct _mesa_prim *prim);
+vbo_count_tessellated_primitives(GLenum mode, GLuint count,
+                                 GLuint num_instances);
 
 void
 vbo_sw_primitive_restart(struct gl_context *ctx,
diff --git a/mesalib/src/mesa/vbo/vbo_exec.c b/mesalib/src/mesa/vbo/vbo_exec.c
index 361d7a3e4..5827f90c3 100644
--- a/mesalib/src/mesa/vbo/vbo_exec.c
+++ b/mesalib/src/mesa/vbo/vbo_exec.c
@@ -94,47 +94,51 @@ void vbo_exec_invalidate_state( struct gl_context *ctx, GLuint new_state )
 
 /**
  * Figure out the number of transform feedback primitives that will be output
- * by the given _mesa_prim command, assuming that no geometry shading is done
- * and primitive restart is not used.
+ * considering the drawing mode, number of vertices, and instance count,
+ * assuming that no geometry shading is done and primitive restart is not
+ * used.
  *
- * This is intended for use by driver back-ends in implementing the
- * PRIMITIVES_GENERATED and TRANSFORM_FEEDBACK_PRIMITIVES_WRITTEN queries.
+ * This is used by driver back-ends in implementing the PRIMITIVES_GENERATED
+ * and TRANSFORM_FEEDBACK_PRIMITIVES_WRITTEN queries.  It is also used to
+ * pre-validate draw calls in GLES3 (where draw calls only succeed if there is
+ * enough room in the transform feedback buffer for the result).
  */
 size_t
-vbo_count_tessellated_primitives(const struct _mesa_prim *prim)
+vbo_count_tessellated_primitives(GLenum mode, GLuint count,
+                                 GLuint num_instances)
 {
    size_t num_primitives;
-   switch (prim->mode) {
+   switch (mode) {
    case GL_POINTS:
-      num_primitives = prim->count;
+      num_primitives = count;
       break;
    case GL_LINE_STRIP:
-      num_primitives = prim->count >= 2 ? prim->count - 1 : 0;
+      num_primitives = count >= 2 ? count - 1 : 0;
       break;
    case GL_LINE_LOOP:
-      num_primitives = prim->count >= 2 ? prim->count : 0;
+      num_primitives = count >= 2 ? count : 0;
       break;
    case GL_LINES:
-      num_primitives = prim->count / 2;
+      num_primitives = count / 2;
       break;
    case GL_TRIANGLE_STRIP:
    case GL_TRIANGLE_FAN:
    case GL_POLYGON:
-      num_primitives = prim->count >= 3 ? prim->count - 2 : 0;
+      num_primitives = count >= 3 ? count - 2 : 0;
       break;
    case GL_TRIANGLES:
-      num_primitives = prim->count / 3;
+      num_primitives = count / 3;
       break;
    case GL_QUAD_STRIP:
-      num_primitives = prim->count >= 4 ? ((prim->count / 2) - 1) * 2 : 0;
+      num_primitives = count >= 4 ? ((count / 2) - 1) * 2 : 0;
       break;
    case GL_QUADS:
-      num_primitives = (prim->count / 4) * 2;
+      num_primitives = (count / 4) * 2;
       break;
    default:
       assert(!"Unexpected primitive type in count_tessellated_primitives");
       num_primitives = 0;
       break;
    }
-   return num_primitives * prim->num_instances;
+   return num_primitives * num_instances;
 }