Merge remote-tracking branch 'origin/released'

Conflicts:
	mesalib/src/mesa/main/.gitignore
	mesalib/src/mesa/main/dlopen.h
	xorg-server/hw/xwin/glx/gen_gl_wrappers.py
	xorg-server/hw/xwin/win.h
	xorg-server/hw/xwin/winengine.c
	xorg-server/hw/xwin/winglobals.c
	xorg-server/hw/xwin/winscrinit.c
	xorg-server/hw/xwin/winshaddd.c
	xorg-server/randr/rrxinerama.c

198 files changed, 7807 insertions, 2140 deletions

diff --git a/mesalib/src/Makefile.am b/mesalib/src/Makefile.am
index 8edf33373..18cb4ce76 100644
--- a/mesalib/src/Makefile.am
+++ b/mesalib/src/Makefile.am
@@ -53,7 +53,13 @@ if HAVE_GALLIUM
 SUBDIRS += gallium
 endif
 
-EXTRA_DIST = egl/docs getopt hgl SConscript
+EXTRA_DIST = \
+	egl/drivers/haiku \
+	egl/docs \
+	getopt hgl SConscript
+
+AM_CFLAGS = $(VISIBILITY_CFLAGS)
+AM_CXXFLAGS = $(VISIBILITY_CXXFLAGS)
 
 AM_CPPFLAGS = \
 	-I$(top_srcdir)/include/ \
diff --git a/mesalib/src/SConscript b/mesalib/src/SConscript
index 188ab08b6..b0578e892 100644
--- a/mesalib/src/SConscript
+++ b/mesalib/src/SConscript
@@ -12,7 +12,8 @@ if env['hostonly']:
     # compilation
     Return()
 
-SConscript('loader/SConscript')
+if env['platform'] != 'windows':
+    SConscript('loader/SConscript')
 
 # When env['gles'] is set, the targets defined in mapi/glapi/SConscript are not
 # used.  libgl-xlib and libgl-gdi adapt themselves to use the targets defined
@@ -30,7 +31,7 @@ SConscript('mesa/SConscript')
 if not env['embedded']:
     if env['platform'] not in ('cygwin', 'darwin', 'freebsd', 'haiku', 'windows'):
         SConscript('glx/SConscript')
-    if env['platform'] not in ['darwin', 'haiku', 'sunos']:
+    if env['platform'] not in ['darwin', 'haiku', 'sunos', 'windows']:
         if env['dri']:
             SConscript('egl/drivers/dri2/SConscript')
         SConscript('egl/main/SConscript')
diff --git a/mesalib/src/gallium/auxiliary/Makefile.am b/mesalib/src/gallium/auxiliary/Makefile.am
index 27a8b3fe4..89c7a13e9 100644
--- a/mesalib/src/gallium/auxiliary/Makefile.am
+++ b/mesalib/src/gallium/auxiliary/Makefile.am
@@ -10,6 +10,7 @@ include $(top_srcdir)/src/gallium/Automake.inc
 noinst_LTLIBRARIES = libgallium.la
 
 AM_CFLAGS = \
+	-I$(top_builddir)/src/glsl/nir \
 	-I$(top_srcdir)/src/gallium/auxiliary/util \
 	$(GALLIUM_CFLAGS) \
 	$(VISIBILITY_CFLAGS) \
@@ -21,6 +22,7 @@ AM_CXXFLAGS = \
 
 libgallium_la_SOURCES = \
 	$(C_SOURCES) \
+	$(NIR_SOURCES) \
 	$(GENERATED_SOURCES)
 
 if HAVE_MESA_LLVM
diff --git a/mesalib/src/gallium/auxiliary/Makefile.sources b/mesalib/src/gallium/auxiliary/Makefile.sources
index b7174d6e7..ec7547cad 100644
--- a/mesalib/src/gallium/auxiliary/Makefile.sources
+++ b/mesalib/src/gallium/auxiliary/Makefile.sources
@@ -251,8 +251,6 @@ C_SOURCES := \
 	util/u_keymap.h \
 	util/u_linear.c \
 	util/u_linear.h \
-	util/u_linkage.c \
-	util/u_linkage.h \
 	util/u_math.c \
 	util/u_math.h \
 	util/u_memory.h \
@@ -304,6 +302,10 @@ C_SOURCES := \
 	util/u_vbuf.h \
 	util/u_video.h
 
+NIR_SOURCES := \
+	nir/tgsi_to_nir.c \
+	nir/tgsi_to_nir.h
+
 VL_SOURCES := \
 	vl/vl_compositor.c \
 	vl/vl_compositor.h \
diff --git a/mesalib/src/gallium/auxiliary/hud/hud_driver_query.c b/mesalib/src/gallium/auxiliary/hud/hud_driver_query.c
index b48708c2f..53771fcae 100644
--- a/mesalib/src/gallium/auxiliary/hud/hud_driver_query.c
+++ b/mesalib/src/gallium/auxiliary/hud/hud_driver_query.c
@@ -157,7 +157,8 @@ hud_pipe_query_install(struct hud_pane *pane, struct pipe_context *pipe,
    if (!gr)
       return;
 
-   strcpy(gr->name, name);
+   strncpy(gr->name, name, sizeof(gr->name));
+   gr->name[sizeof(gr->name) - 1] = '\0';
    gr->query_data = CALLOC_STRUCT(query_info);
    if (!gr->query_data) {
       FREE(gr);
diff --git a/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.c b/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.c
new file mode 100644
index 000000000..59aaf6778
--- /dev/null
+++ b/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.c
@@ -0,0 +1,1750 @@
+/*
+ * Copyright © 2014-2015 Broadcom
+ * Copyright (C) 2014 Rob Clark <robclark@freedesktop.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "util/ralloc.h"
+#include "glsl/nir/nir.h"
+#include "glsl/nir/nir_builder.h"
+#include "glsl/list.h"
+#include "glsl/shader_enums.h"
+
+#include "nir/tgsi_to_nir.h"
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_dump.h"
+#include "tgsi/tgsi_info.h"
+#include "tgsi/tgsi_scan.h"
+
+#define SWIZ(X, Y, Z, W) (unsigned[4]){      \
+      TGSI_SWIZZLE_##X,                      \
+      TGSI_SWIZZLE_##Y,                      \
+      TGSI_SWIZZLE_##Z,                      \
+      TGSI_SWIZZLE_##W,                      \
+   }
+
+struct ttn_reg_info {
+   /** nir register containing this TGSI index. */
+   nir_register *reg;
+   nir_variable *var;
+   /** Offset (in vec4s) from the start of var for this TGSI index. */
+   int offset;
+};
+
+struct ttn_compile {
+   union tgsi_full_token *token;
+   nir_builder build;
+   struct tgsi_shader_info *scan;
+
+   struct ttn_reg_info *output_regs;
+   struct ttn_reg_info *temp_regs;
+   nir_ssa_def **imm_defs;
+
+   nir_register *addr_reg;
+
+   /**
+    * Stack of cf_node_lists where instructions should be pushed as we pop
+    * back out of the control flow stack.
+    *
+    * For each IF/ELSE/ENDIF block, if_stack[if_stack_pos] has where the else
+    * instructions should be placed, and if_stack[if_stack_pos - 1] has where
+    * the next instructions outside of the if/then/else block go.
+    */
+   struct exec_list **if_stack;
+   unsigned if_stack_pos;
+
+   /**
+    * Stack of cf_node_lists where instructions should be pushed as we pop
+    * back out of the control flow stack.
+    *
+    * loop_stack[loop_stack_pos - 1] contains the cf_node_list for the outside
+    * of the loop.
+    */
+   struct exec_list **loop_stack;
+   unsigned loop_stack_pos;
+
+   /* How many TGSI_FILE_IMMEDIATE vec4s have been parsed so far. */
+   unsigned next_imm;
+};
+
+#define ttn_swizzle(b, src, x, y, z, w) \
+   nir_swizzle(b, src, SWIZ(x, y, z, w), 4, false)
+#define ttn_channel(b, src, swiz) \
+   nir_swizzle(b, src, SWIZ(swiz, swiz, swiz, swiz), 1, false)
+
+static nir_ssa_def *
+ttn_src_for_dest(nir_builder *b, nir_alu_dest *dest)
+{
+   nir_alu_src src;
+   memset(&src, 0, sizeof(src));
+
+   if (dest->dest.is_ssa)
+      src.src = nir_src_for_ssa(&dest->dest.ssa);
+   else {
+      assert(!dest->dest.reg.indirect);
+      src.src = nir_src_for_reg(dest->dest.reg.reg);
+      src.src.reg.base_offset = dest->dest.reg.base_offset;
+   }
+
+   for (int i = 0; i < 4; i++)
+      src.swizzle[i] = i;
+
+   return nir_fmov_alu(b, src, 4);
+}
+
+static void
+ttn_emit_declaration(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_full_declaration *decl = &c->token->FullDeclaration;
+   unsigned array_size = decl->Range.Last - decl->Range.First + 1;
+   unsigned file = decl->Declaration.File;
+   unsigned i;
+
+   if (file == TGSI_FILE_TEMPORARY) {
+      if (decl->Declaration.Array) {
+         /* for arrays, we create variables instead of registers: */
+         nir_variable *var = rzalloc(b->shader, nir_variable);
+
+         var->type = glsl_array_type(glsl_vec4_type(), array_size);
+         var->data.mode = nir_var_global;
+         var->name = ralloc_asprintf(var, "arr_%d", decl->Array.ArrayID);
+
+         exec_list_push_tail(&b->shader->globals, &var->node);
+
+         for (i = 0; i < array_size; i++) {
+            /* point all the matching slots to the same var,
+             * with appropriate offset set, mostly just so
+             * we know what to do when tgsi does a non-indirect
+             * access
+             */
+            c->temp_regs[decl->Range.First + i].reg = NULL;
+            c->temp_regs[decl->Range.First + i].var = var;
+            c->temp_regs[decl->Range.First + i].offset = i;
+         }
+      } else {
+         for (i = 0; i < array_size; i++) {
+            nir_register *reg = nir_local_reg_create(b->impl);
+            reg->num_components = 4;
+            c->temp_regs[decl->Range.First + i].reg = reg;
+            c->temp_regs[decl->Range.First + i].var = NULL;
+            c->temp_regs[decl->Range.First + i].offset = 0;
+         }
+      }
+   } else if (file == TGSI_FILE_ADDRESS) {
+      c->addr_reg = nir_local_reg_create(b->impl);
+      c->addr_reg->num_components = 4;
+   } else if (file == TGSI_FILE_SYSTEM_VALUE) {
+      /* Nothing to record for system values. */
+   } else if (file == TGSI_FILE_SAMPLER) {
+      /* Nothing to record for samplers. */
+   } else {
+      nir_variable *var;
+      assert(file == TGSI_FILE_INPUT ||
+             file == TGSI_FILE_OUTPUT ||
+             file == TGSI_FILE_CONSTANT);
+
+      /* nothing to do for UBOs: */
+      if ((file == TGSI_FILE_CONSTANT) && decl->Declaration.Dimension)
+         return;
+
+      var = rzalloc(b->shader, nir_variable);
+      var->data.driver_location = decl->Range.First;
+
+      var->type = glsl_vec4_type();
+      if (array_size > 1)
+         var->type = glsl_array_type(var->type, array_size);
+
+      switch (file) {
+      case TGSI_FILE_INPUT:
+         var->data.read_only = true;
+         var->data.mode = nir_var_shader_in;
+         var->name = ralloc_asprintf(var, "in_%d", decl->Range.First);
+
+         /* We should probably translate to a VERT_ATTRIB_* or VARYING_SLOT_*
+          * instead, but nothing in NIR core is looking at the value
+          * currently, and this is less change to drivers.
+          */
+         var->data.location = decl->Semantic.Name;
+         var->data.index = decl->Semantic.Index;
+
+         /* We definitely need to translate the interpolation field, because
+          * nir_print will decode it.
+          */
+         switch (decl->Interp.Interpolate) {
+         case TGSI_INTERPOLATE_CONSTANT:
+            var->data.interpolation = INTERP_QUALIFIER_FLAT;
+            break;
+         case TGSI_INTERPOLATE_LINEAR:
+            var->data.interpolation = INTERP_QUALIFIER_NOPERSPECTIVE;
+            break;
+         case TGSI_INTERPOLATE_PERSPECTIVE:
+            var->data.interpolation = INTERP_QUALIFIER_SMOOTH;
+            break;
+         }
+
+         exec_list_push_tail(&b->shader->inputs, &var->node);
+         break;
+      case TGSI_FILE_OUTPUT: {
+         /* Since we can't load from outputs in the IR, we make temporaries
+          * for the outputs and emit stores to the real outputs at the end of
+          * the shader.
+          */
+         nir_register *reg = nir_local_reg_create(b->impl);
+         reg->num_components = 4;
+         if (array_size > 1)
+            reg->num_array_elems = array_size;
+
+         var->data.mode = nir_var_shader_out;
+         var->name = ralloc_asprintf(var, "out_%d", decl->Range.First);
+
+         var->data.location = decl->Semantic.Name;
+         var->data.index = decl->Semantic.Index;
+
+         for (i = 0; i < array_size; i++) {
+            c->output_regs[decl->Range.First + i].offset = i;
+            c->output_regs[decl->Range.First + i].reg = reg;
+         }
+
+         exec_list_push_tail(&b->shader->outputs, &var->node);
+      }
+         break;
+      case TGSI_FILE_CONSTANT:
+         var->data.mode = nir_var_uniform;
+         var->name = ralloc_asprintf(var, "uniform_%d", decl->Range.First);
+
+         exec_list_push_tail(&b->shader->uniforms, &var->node);
+         break;
+      default:
+         unreachable("bad declaration file");
+         return;
+      }
+
+   }
+}
+
+static void
+ttn_emit_immediate(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_full_immediate *tgsi_imm = &c->token->FullImmediate;
+   nir_load_const_instr *load_const;
+   int i;
+
+   load_const = nir_load_const_instr_create(b->shader, 4);
+   c->imm_defs[c->next_imm] = &load_const->def;
+   c->next_imm++;
+
+   for (i = 0; i < 4; i++)
+      load_const->value.u[i] = tgsi_imm->u[i].Uint;
+
+   nir_instr_insert_after_cf_list(b->cf_node_list, &load_const->instr);
+}
+
+static nir_src
+ttn_src_for_indirect(struct ttn_compile *c, struct tgsi_ind_register *indirect);
+
+/* generate either a constant or indirect deref chain for accessing an
+ * array variable.
+ */
+static nir_deref_var *
+ttn_array_deref(struct ttn_compile *c, nir_intrinsic_instr *instr,
+                nir_variable *var, unsigned offset,
+                struct tgsi_ind_register *indirect)
+{
+   nir_deref_var *deref = nir_deref_var_create(instr, var);
+   nir_deref_array *arr = nir_deref_array_create(deref);
+
+   arr->base_offset = offset;
+   arr->deref.type = glsl_get_array_element(var->type);
+
+   if (indirect) {
+      arr->deref_array_type = nir_deref_array_type_indirect;
+      arr->indirect = ttn_src_for_indirect(c, indirect);
+   } else {
+      arr->deref_array_type = nir_deref_array_type_direct;
+   }
+
+   deref->deref.child = &arr->deref;
+
+   return deref;
+}
+
+static nir_src
+ttn_src_for_file_and_index(struct ttn_compile *c, unsigned file, unsigned index,
+                           struct tgsi_ind_register *indirect,
+                           struct tgsi_dimension *dim,
+                           struct tgsi_ind_register *dimind)
+{
+   nir_builder *b = &c->build;
+   nir_src src;
+
+   memset(&src, 0, sizeof(src));
+
+   switch (file) {
+   case TGSI_FILE_TEMPORARY:
+      if (c->temp_regs[index].var) {
+         unsigned offset = c->temp_regs[index].offset;
+         nir_variable *var = c->temp_regs[index].var;
+         nir_intrinsic_instr *load;
+
+         load = nir_intrinsic_instr_create(b->shader,
+                                           nir_intrinsic_load_var);
+         load->num_components = 4;
+         load->variables[0] = ttn_array_deref(c, load, var, offset, indirect);
+
+         nir_ssa_dest_init(&load->instr, &load->dest, 4, NULL);
+         nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+
+         src = nir_src_for_ssa(&load->dest.ssa);
+
+      } else {
+         assert(!indirect);
+         src.reg.reg = c->temp_regs[index].reg;
+      }
+      assert(!dim);
+      break;
+
+   case TGSI_FILE_ADDRESS:
+      src.reg.reg = c->addr_reg;
+      assert(!dim);
+      break;
+
+   case TGSI_FILE_IMMEDIATE:
+      src = nir_src_for_ssa(c->imm_defs[index]);
+      assert(!indirect);
+      assert(!dim);
+      break;
+
+   case TGSI_FILE_SYSTEM_VALUE: {
+      nir_intrinsic_instr *load;
+      nir_intrinsic_op op;
+      unsigned ncomp = 1;
+
+      assert(!indirect);
+      assert(!dim);
+
+      switch (c->scan->system_value_semantic_name[index]) {
+      case TGSI_SEMANTIC_VERTEXID_NOBASE:
+         op = nir_intrinsic_load_vertex_id_zero_base;
+         break;
+      case TGSI_SEMANTIC_VERTEXID:
+         op = nir_intrinsic_load_vertex_id;
+         break;
+      case TGSI_SEMANTIC_BASEVERTEX:
+         op = nir_intrinsic_load_base_vertex;
+         break;
+      case TGSI_SEMANTIC_INSTANCEID:
+         op = nir_intrinsic_load_instance_id;
+         break;
+      default:
+         unreachable("bad system value");
+      }
+
+      load = nir_intrinsic_instr_create(b->shader, op);
+      load->num_components = ncomp;
+
+      nir_ssa_dest_init(&load->instr, &load->dest, ncomp, NULL);
+      nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+
+      src = nir_src_for_ssa(&load->dest.ssa);
+      break;
+   }
+
+   case TGSI_FILE_INPUT:
+   case TGSI_FILE_CONSTANT: {
+      nir_intrinsic_instr *load;
+      nir_intrinsic_op op;
+      unsigned srcn = 0;
+
+      switch (file) {
+      case TGSI_FILE_INPUT:
+         op = indirect ? nir_intrinsic_load_input_indirect :
+                         nir_intrinsic_load_input;
+         assert(!dim);
+         break;
+      case TGSI_FILE_CONSTANT:
+         if (dim) {
+            op = indirect ? nir_intrinsic_load_ubo_indirect :
+                            nir_intrinsic_load_ubo;
+            /* convert index from vec4 to byte: */
+            index *= 16;
+         } else {
+            op = indirect ? nir_intrinsic_load_uniform_indirect :
+                            nir_intrinsic_load_uniform;
+         }
+         break;
+      default:
+         unreachable("No other load files supported");
+         break;
+      }
+
+      load = nir_intrinsic_instr_create(b->shader, op);
+
+      load->num_components = 4;
+      load->const_index[0] = index;
+      load->const_index[1] = 1;
+      if (dim) {
+         if (dimind) {
+            load->src[srcn] =
+               ttn_src_for_file_and_index(c, dimind->File, dimind->Index,
+                                          NULL, NULL, NULL);
+         } else {
+            /* UBOs start at index 1 in TGSI: */
+            load->src[srcn] =
+               nir_src_for_ssa(nir_imm_int(b, dim->Index - 1));
+         }
+         srcn++;
+      }
+      if (indirect) {
+         load->src[srcn] = ttn_src_for_indirect(c, indirect);
+         if (dim) {
+            assert(load->src[srcn].is_ssa);
+            /* we also need to covert vec4 to byte here too: */
+            load->src[srcn] =
+               nir_src_for_ssa(nir_ishl(b, load->src[srcn].ssa,
+                                        nir_imm_int(b, 4)));
+         }
+         srcn++;
+      }
+      nir_ssa_dest_init(&load->instr, &load->dest, 4, NULL);
+      nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+
+      src = nir_src_for_ssa(&load->dest.ssa);
+      break;
+   }
+
+   default:
+      unreachable("bad src file");
+   }
+
+
+   return src;
+}
+
+static nir_src
+ttn_src_for_indirect(struct ttn_compile *c, struct tgsi_ind_register *indirect)
+{
+   nir_builder *b = &c->build;
+   nir_alu_src src;
+   memset(&src, 0, sizeof(src));
+   for (int i = 0; i < 4; i++)
+      src.swizzle[i] = indirect->Swizzle;
+   src.src = ttn_src_for_file_and_index(c,
+                                        indirect->File,
+                                        indirect->Index,
+                                        NULL, NULL, NULL);
+   return nir_src_for_ssa(nir_imov_alu(b, src, 1));
+}
+
+static nir_alu_dest
+ttn_get_dest(struct ttn_compile *c, struct tgsi_full_dst_register *tgsi_fdst)
+{
+   struct tgsi_dst_register *tgsi_dst = &tgsi_fdst->Register;
+   nir_alu_dest dest;
+   unsigned index = tgsi_dst->Index;
+
+   memset(&dest, 0, sizeof(dest));
+
+   if (tgsi_dst->File == TGSI_FILE_TEMPORARY) {
+      if (c->temp_regs[index].var) {
+          nir_builder *b = &c->build;
+          nir_intrinsic_instr *load;
+          struct tgsi_ind_register *indirect =
+                tgsi_dst->Indirect ? &tgsi_fdst->Indirect : NULL;
+          nir_register *reg;
+
+         /* this works, because TGSI will give us a base offset
+          * (in case of indirect index) that points back into
+          * the array.  Access can be direct or indirect, we
+          * don't really care.  Just create a one-shot dst reg
+          * that will get store_var'd back into the array var
+          * at the end of ttn_emit_instruction()
+          */
+         reg = nir_local_reg_create(c->build.impl);
+         reg->num_components = 4;
+         dest.dest.reg.reg = reg;
+         dest.dest.reg.base_offset = 0;
+
+         /* since the alu op might not write to all components
+          * of the temporary, we must first do a load_var to
+          * get the previous array elements into the register.
+          * This is one area that NIR could use a bit of
+          * improvement (or opt pass to clean up the mess
+          * once things are scalarized)
+          */
+
+         load = nir_intrinsic_instr_create(c->build.shader,
+                                           nir_intrinsic_load_var);
+         load->num_components = 4;
+         load->variables[0] =
+               ttn_array_deref(c, load, c->temp_regs[index].var,
+                               c->temp_regs[index].offset,
+                               indirect);
+
+         load->dest = nir_dest_for_reg(reg);
+
+         nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+      } else {
+         assert(!tgsi_dst->Indirect);
+         dest.dest.reg.reg = c->temp_regs[index].reg;
+         dest.dest.reg.base_offset = c->temp_regs[index].offset;
+      }
+   } else if (tgsi_dst->File == TGSI_FILE_OUTPUT) {
+      dest.dest.reg.reg = c->output_regs[index].reg;
+      dest.dest.reg.base_offset = c->output_regs[index].offset;
+   } else if (tgsi_dst->File == TGSI_FILE_ADDRESS) {
+      assert(index == 0);
+      dest.dest.reg.reg = c->addr_reg;
+   }
+
+   dest.write_mask = tgsi_dst->WriteMask;
+   dest.saturate = false;
+
+   if (tgsi_dst->Indirect && (tgsi_dst->File != TGSI_FILE_TEMPORARY)) {
+      nir_src *indirect = ralloc(c->build.shader, nir_src);
+      *indirect = ttn_src_for_indirect(c, &tgsi_fdst->Indirect);
+      dest.dest.reg.indirect = indirect;
+   }
+
+   return dest;
+}
+
+static nir_variable *
+ttn_get_var(struct ttn_compile *c, struct tgsi_full_dst_register *tgsi_fdst)
+{
+   struct tgsi_dst_register *tgsi_dst = &tgsi_fdst->Register;
+   unsigned index = tgsi_dst->Index;
+
+   if (tgsi_dst->File == TGSI_FILE_TEMPORARY) {
+      /* we should not have an indirect when there is no var! */
+      if (!c->temp_regs[index].var)
+         assert(!tgsi_dst->Indirect);
+      return c->temp_regs[index].var;
+   }
+
+   return NULL;
+}
+
+static nir_ssa_def *
+ttn_get_src(struct ttn_compile *c, struct tgsi_full_src_register *tgsi_fsrc)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_src_register *tgsi_src = &tgsi_fsrc->Register;
+   unsigned tgsi_opcode = c->token->FullInstruction.Instruction.Opcode;
+   unsigned tgsi_src_type = tgsi_opcode_infer_src_type(tgsi_opcode);
+   bool src_is_float = !(tgsi_src_type == TGSI_TYPE_SIGNED ||
+                         tgsi_src_type == TGSI_TYPE_UNSIGNED);
+   nir_alu_src src;
+
+   memset(&src, 0, sizeof(src));
+
+   if (tgsi_src->File == TGSI_FILE_NULL) {
+      return nir_imm_float(b, 0.0);
+   } else if (tgsi_src->File == TGSI_FILE_SAMPLER) {
+      /* Only the index of the sampler gets used in texturing, and it will
+       * handle looking that up on its own instead of using the nir_alu_src.
+       */
+      assert(!tgsi_src->Indirect);
+      return NULL;
+   } else {
+      struct tgsi_ind_register *ind = NULL;
+      struct tgsi_dimension *dim = NULL;
+      struct tgsi_ind_register *dimind = NULL;
+      if (tgsi_src->Indirect)
+         ind = &tgsi_fsrc->Indirect;
+      if (tgsi_src->Dimension) {
+         dim = &tgsi_fsrc->Dimension;
+         if (dim->Indirect)
+            dimind = &tgsi_fsrc->DimIndirect;
+      }
+      src.src = ttn_src_for_file_and_index(c,
+                                           tgsi_src->File,
+                                           tgsi_src->Index,
+                                           ind, dim, dimind);
+   }
+
+   src.swizzle[0] = tgsi_src->SwizzleX;
+   src.swizzle[1] = tgsi_src->SwizzleY;
+   src.swizzle[2] = tgsi_src->SwizzleZ;
+   src.swizzle[3] = tgsi_src->SwizzleW;
+
+   nir_ssa_def *def = nir_fmov_alu(b, src, 4);
+
+   if (tgsi_src->Absolute) {
+      if (src_is_float)
+         def = nir_fabs(b, def);
+      else
+         def = nir_iabs(b, def);
+   }
+
+   if (tgsi_src->Negate) {
+      if (src_is_float)
+         def = nir_fneg(b, def);
+      else
+         def = nir_ineg(b, def);
+   }
+
+   return def;
+}
+
+static void
+ttn_alu(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   unsigned num_srcs = nir_op_infos[op].num_inputs;
+   nir_alu_instr *instr = nir_alu_instr_create(b->shader, op);
+   unsigned i;
+
+   for (i = 0; i < num_srcs; i++)
+      instr->src[i].src = nir_src_for_ssa(src[i]);
+
+   instr->dest = dest;
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+}
+
+static void
+ttn_move_dest_masked(nir_builder *b, nir_alu_dest dest,
+                     nir_ssa_def *def, unsigned write_mask)
+{
+   if (!(dest.write_mask & write_mask))
+      return;
+
+   nir_alu_instr *mov = nir_alu_instr_create(b->shader, nir_op_imov);
+   mov->dest = dest;
+   mov->dest.write_mask &= write_mask;
+   mov->src[0].src = nir_src_for_ssa(def);
+   for (unsigned i = def->num_components; i < 4; i++)
+      mov->src[0].swizzle[i] = def->num_components - 1;
+   nir_instr_insert_after_cf_list(b->cf_node_list, &mov->instr);
+}
+
+static void
+ttn_move_dest(nir_builder *b, nir_alu_dest dest, nir_ssa_def *def)
+{
+   ttn_move_dest_masked(b, dest, def, TGSI_WRITEMASK_XYZW);
+}
+
+static void
+ttn_arl(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_f2i(b, nir_ffloor(b, src[0])));
+}
+
+/* EXP - Approximate Exponential Base 2
+ *  dst.x = 2^{\lfloor src.x\rfloor}
+ *  dst.y = src.x - \lfloor src.x\rfloor
+ *  dst.z = 2^{src.x}
+ *  dst.w = 1.0
+ */
+static void
+ttn_exp(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *srcx = ttn_channel(b, src[0], X);
+
+   ttn_move_dest_masked(b, dest, nir_fexp2(b, nir_ffloor(b, srcx)),
+                        TGSI_WRITEMASK_X);
+   ttn_move_dest_masked(b, dest, nir_fsub(b, srcx, nir_ffloor(b, srcx)),
+                        TGSI_WRITEMASK_Y);
+   ttn_move_dest_masked(b, dest, nir_fexp2(b, srcx), TGSI_WRITEMASK_Z);
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_W);
+}
+
+/* LOG - Approximate Logarithm Base 2
+ *  dst.x = \lfloor\log_2{|src.x|}\rfloor
+ *  dst.y = \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}}
+ *  dst.z = \log_2{|src.x|}
+ *  dst.w = 1.0
+ */
+static void
+ttn_log(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *abs_srcx = nir_fabs(b, ttn_channel(b, src[0], X));
+   nir_ssa_def *log2 = nir_flog2(b, abs_srcx);
+
+   ttn_move_dest_masked(b, dest, nir_ffloor(b, log2), TGSI_WRITEMASK_X);
+   ttn_move_dest_masked(b, dest,
+                        nir_fdiv(b, abs_srcx, nir_fexp2(b, nir_ffloor(b, log2))),
+                        TGSI_WRITEMASK_Y);
+   ttn_move_dest_masked(b, dest, nir_flog2(b, abs_srcx), TGSI_WRITEMASK_Z);
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_W);
+}
+
+/* DST - Distance Vector
+ *   dst.x = 1.0
+ *   dst.y = src0.y \times src1.y
+ *   dst.z = src0.z
+ *   dst.w = src1.w
+ */
+static void
+ttn_dst(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_X);
+   ttn_move_dest_masked(b, dest, nir_fmul(b, src[0], src[1]), TGSI_WRITEMASK_Y);
+   ttn_move_dest_masked(b, dest, nir_fmov(b, src[0]), TGSI_WRITEMASK_Z);
+   ttn_move_dest_masked(b, dest, nir_fmov(b, src[1]), TGSI_WRITEMASK_W);
+}
+
+/* LIT - Light Coefficients
+ *  dst.x = 1.0
+ *  dst.y = max(src.x, 0.0)
+ *  dst.z = (src.x > 0.0) ? max(src.y, 0.0)^{clamp(src.w, -128.0, 128.0))} : 0
+ *  dst.w = 1.0
+ */
+static void
+ttn_lit(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_XW);
+
+   ttn_move_dest_masked(b, dest, nir_fmax(b, ttn_channel(b, src[0], X),
+                                          nir_imm_float(b, 0.0)), TGSI_WRITEMASK_Y);
+
+   if (dest.write_mask & TGSI_WRITEMASK_Z) {
+      nir_ssa_def *src0_y = ttn_channel(b, src[0], Y);
+      nir_ssa_def *wclamp = nir_fmax(b, nir_fmin(b, ttn_channel(b, src[0], W),
+                                                 nir_imm_float(b, 128.0)),
+                                     nir_imm_float(b, -128.0));
+      nir_ssa_def *pow = nir_fpow(b, nir_fmax(b, src0_y, nir_imm_float(b, 0.0)),
+                                  wclamp);
+
+      ttn_move_dest_masked(b, dest,
+                           nir_bcsel(b,
+                                     nir_fge(b,
+                                             nir_imm_float(b, 0.0),
+                                             ttn_channel(b, src[0], X)),
+                                     nir_imm_float(b, 0.0),
+                                     pow),
+                           TGSI_WRITEMASK_Z);
+   }
+}
+
+/* SCS - Sine Cosine
+ *   dst.x = \cos{src.x}
+ *   dst.y = \sin{src.x}
+ *   dst.z = 0.0
+ *   dst.w = 1.0
+ */
+static void
+ttn_scs(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest_masked(b, dest, nir_fcos(b, ttn_channel(b, src[0], X)),
+                        TGSI_WRITEMASK_X);
+   ttn_move_dest_masked(b, dest, nir_fsin(b, ttn_channel(b, src[0], X)),
+                        TGSI_WRITEMASK_Y);
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 0.0), TGSI_WRITEMASK_Z);
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_W);
+}
+
+static void
+ttn_sle(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_sge(b, src[1], src[0]));
+}
+
+static void
+ttn_sgt(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_slt(b, src[1], src[0]));
+}
+
+static void
+ttn_clamp(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_fmin(b, nir_fmax(b, src[0], src[1]), src[2]));
+}
+
+static void
+ttn_xpd(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest_masked(b, dest,
+                        nir_fsub(b,
+                                 nir_fmul(b,
+                                          ttn_swizzle(b, src[0], Y, Z, X, X),
+                                          ttn_swizzle(b, src[1], Z, X, Y, X)),
+                                 nir_fmul(b,
+                                          ttn_swizzle(b, src[1], Y, Z, X, X),
+                                          ttn_swizzle(b, src[0], Z, X, Y, X))),
+                        TGSI_WRITEMASK_XYZ);
+   ttn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), TGSI_WRITEMASK_W);
+}
+
+static void
+ttn_dp2a(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest,
+                 ttn_channel(b, nir_fadd(b, nir_fdot2(b, src[0], src[1]),
+                                         src[2]),
+                             X));
+}
+
+static void
+ttn_dp2(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_fdot2(b, src[0], src[1]));
+}
+
+static void
+ttn_dp3(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_fdot3(b, src[0], src[1]));
+}
+
+static void
+ttn_dp4(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_fdot4(b, src[0], src[1]));
+}
+
+static void
+ttn_dph(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_fadd(b, nir_fdot3(b, src[0], src[1]),
+                                   ttn_channel(b, src[1], W)));
+}
+
+static void
+ttn_umad(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_iadd(b, nir_imul(b, src[0], src[1]), src[2]));
+}
+
+static void
+ttn_arr(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_ffloor(b, nir_fadd(b, src[0], nir_imm_float(b, 0.5))));
+}
+
+static void
+ttn_cmp(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_bcsel(b,
+                                    nir_flt(b, src[0], nir_imm_float(b, 0.0)),
+                                    src[1], src[2]));
+}
+
+static void
+ttn_ucmp(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ttn_move_dest(b, dest, nir_bcsel(b,
+                                    nir_ine(b, src[0], nir_imm_int(b, 0)),
+                                    src[1], src[2]));
+}
+
+static void
+ttn_kill(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_intrinsic_instr *discard =
+      nir_intrinsic_instr_create(b->shader, nir_intrinsic_discard);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &discard->instr);
+}
+
+static void
+ttn_kill_if(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *cmp = nir_bany4(b, nir_flt(b, src[0], nir_imm_float(b, 0.0)));
+   nir_intrinsic_instr *discard =
+      nir_intrinsic_instr_create(b->shader, nir_intrinsic_discard_if);
+   discard->src[0] = nir_src_for_ssa(cmp);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &discard->instr);
+}
+
+static void
+ttn_if(struct ttn_compile *c, nir_ssa_def *src, bool is_uint)
+{
+   nir_builder *b = &c->build;
+
+   /* Save the outside-of-the-if-statement node list. */
+   c->if_stack[c->if_stack_pos] = b->cf_node_list;
+   c->if_stack_pos++;
+
+   src = ttn_channel(b, src, X);
+
+   nir_if *if_stmt = nir_if_create(b->shader);
+   if (is_uint) {
+      if_stmt->condition = nir_src_for_ssa(nir_ine(b, src, nir_imm_int(b, 0)));
+   } else {
+      if_stmt->condition = nir_src_for_ssa(nir_fne(b, src, nir_imm_int(b, 0)));
+   }
+   nir_cf_node_insert_end(b->cf_node_list, &if_stmt->cf_node);
+
+   nir_builder_insert_after_cf_list(b, &if_stmt->then_list);
+
+   c->if_stack[c->if_stack_pos] = &if_stmt->else_list;
+   c->if_stack_pos++;
+}
+
+static void
+ttn_else(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   nir_builder_insert_after_cf_list(b, c->if_stack[c->if_stack_pos - 1]);
+}
+
+static void
+ttn_endif(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   c->if_stack_pos -= 2;
+   nir_builder_insert_after_cf_list(b, c->if_stack[c->if_stack_pos]);
+}
+
+static void
+ttn_bgnloop(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   /* Save the outside-of-the-loop node list. */
+   c->loop_stack[c->loop_stack_pos] = b->cf_node_list;
+   c->loop_stack_pos++;
+
+   nir_loop *loop = nir_loop_create(b->shader);
+   nir_cf_node_insert_end(b->cf_node_list, &loop->cf_node);
+
+   nir_builder_insert_after_cf_list(b, &loop->body);
+}
+
+static void
+ttn_cont(nir_builder *b)
+{
+   nir_jump_instr *instr = nir_jump_instr_create(b->shader, nir_jump_continue);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+}
+
+static void
+ttn_brk(nir_builder *b)
+{
+   nir_jump_instr *instr = nir_jump_instr_create(b->shader, nir_jump_break);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+}
+
+static void
+ttn_endloop(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   c->loop_stack_pos--;
+   nir_builder_insert_after_cf_list(b, c->loop_stack[c->loop_stack_pos]);
+}
+
+static void
+setup_texture_info(nir_tex_instr *instr, unsigned texture)
+{
+   switch (texture) {
+   case TGSI_TEXTURE_1D:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_1D;
+      break;
+   case TGSI_TEXTURE_1D_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_1D;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_SHADOW1D:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_1D;
+      instr->is_shadow = true;
+      break;
+   case TGSI_TEXTURE_SHADOW1D_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_1D;
+      instr->is_shadow = true;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_2D:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_2D;
+      break;
+   case TGSI_TEXTURE_2D_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_2D;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_2D_MSAA:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_MS;
+      break;
+   case TGSI_TEXTURE_2D_ARRAY_MSAA:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_MS;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_SHADOW2D:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_2D;
+      instr->is_shadow = true;
+      break;
+   case TGSI_TEXTURE_SHADOW2D_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_2D;
+      instr->is_shadow = true;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_3D:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_3D;
+      break;
+   case TGSI_TEXTURE_CUBE:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_CUBE;
+      break;
+   case TGSI_TEXTURE_CUBE_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_CUBE;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_SHADOWCUBE:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_CUBE;
+      instr->is_shadow = true;
+      break;
+   case TGSI_TEXTURE_SHADOWCUBE_ARRAY:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_CUBE;
+      instr->is_shadow = true;
+      instr->is_array = true;
+      break;
+   case TGSI_TEXTURE_RECT:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_RECT;
+      break;
+   case TGSI_TEXTURE_SHADOWRECT:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_RECT;
+      instr->is_shadow = true;
+      break;
+   default:
+      fprintf(stderr, "Unknown TGSI texture target %d\n", texture);
+      abort();
+   }
+}
+
+static void
+ttn_tex(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_full_instruction *tgsi_inst = &c->token->FullInstruction;
+   nir_tex_instr *instr;
+   nir_texop op;
+   unsigned num_srcs, samp = 1, i;
+
+   switch (tgsi_inst->Instruction.Opcode) {
+   case TGSI_OPCODE_TEX:
+      op = nir_texop_tex;
+      num_srcs = 1;
+      break;
+   case TGSI_OPCODE_TXP:
+      op = nir_texop_tex;
+      num_srcs = 2;
+      break;
+   case TGSI_OPCODE_TXB:
+      op = nir_texop_txb;
+      num_srcs = 2;
+      break;
+   case TGSI_OPCODE_TXL:
+      op = nir_texop_txl;
+      num_srcs = 2;
+      break;
+   case TGSI_OPCODE_TXL2:
+      op = nir_texop_txl;
+      num_srcs = 2;
+      samp = 2;
+      break;
+   case TGSI_OPCODE_TXF:
+      op = nir_texop_txf;
+      num_srcs = 2;
+      break;
+   case TGSI_OPCODE_TXD:
+      op = nir_texop_txd;
+      num_srcs = 3;
+      samp = 3;
+      break;
+
+   default:
+      fprintf(stderr, "unknown TGSI tex op %d\n", tgsi_inst->Instruction.Opcode);
+      abort();
+   }
+
+   if (tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOW1D ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOW1D_ARRAY ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOW2D ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOW2D_ARRAY ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOWRECT ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOWCUBE ||
+       tgsi_inst->Texture.Texture == TGSI_TEXTURE_SHADOWCUBE_ARRAY) {
+      num_srcs++;
+   }
+
+   num_srcs += tgsi_inst->Texture.NumOffsets;
+
+   instr = nir_tex_instr_create(b->shader, num_srcs);
+   instr->op = op;
+
+   setup_texture_info(instr, tgsi_inst->Texture.Texture);
+
+   switch (instr->sampler_dim) {
+   case GLSL_SAMPLER_DIM_1D:
+   case GLSL_SAMPLER_DIM_BUF:
+      instr->coord_components = 1;
+      break;
+   case GLSL_SAMPLER_DIM_2D:
+   case GLSL_SAMPLER_DIM_RECT:
+   case GLSL_SAMPLER_DIM_EXTERNAL:
+   case GLSL_SAMPLER_DIM_MS:
+      instr->coord_components = 2;
+      break;
+   case GLSL_SAMPLER_DIM_3D:
+   case GLSL_SAMPLER_DIM_CUBE:
+      instr->coord_components = 3;
+      break;
+   }
+
+   if (instr->is_array)
+      instr->coord_components++;
+
+   assert(tgsi_inst->Src[samp].Register.File == TGSI_FILE_SAMPLER);
+   instr->sampler_index = tgsi_inst->Src[samp].Register.Index;
+
+   unsigned src_number = 0;
+
+   instr->src[src_number].src =
+      nir_src_for_ssa(nir_swizzle(b, src[0], SWIZ(X, Y, Z, W),
+                                  instr->coord_components, false));
+   instr->src[src_number].src_type = nir_tex_src_coord;
+   src_number++;
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXP) {
+      instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_projector;
+      src_number++;
+   }
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXB) {
+      instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_bias;
+      src_number++;
+   }
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXL) {
+      instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_lod;
+      src_number++;
+   }
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXL2) {
+      instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[1], X));
+      instr->src[src_number].src_type = nir_tex_src_lod;
+      src_number++;
+   }
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXF) {
+      instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_lod;
+      src_number++;
+   }
+
+   if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXD) {
+      instr->src[src_number].src =
+         nir_src_for_ssa(nir_swizzle(b, src[1], SWIZ(X, Y, Z, W),
+              instr->coord_components, false));
+      instr->src[src_number].src_type = nir_tex_src_ddx;
+      src_number++;
+      instr->src[src_number].src =
+         nir_src_for_ssa(nir_swizzle(b, src[2], SWIZ(X, Y, Z, W),
+              instr->coord_components, false));
+      instr->src[src_number].src_type = nir_tex_src_ddy;
+      src_number++;
+   }
+
+   if (instr->is_shadow) {
+      if (instr->coord_components < 3)
+         instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], Z));
+      else
+         instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W));
+
+      instr->src[src_number].src_type = nir_tex_src_comparitor;
+      src_number++;
+   }
+
+   for (i = 0; i < tgsi_inst->Texture.NumOffsets; i++) {
+      struct tgsi_texture_offset *tex_offset = &tgsi_inst->TexOffsets[i];
+      /* since TexOffset ins't using tgsi_full_src_register we get to
+       * do some extra gymnastics:
+       */
+      nir_alu_src src;
+
+      memset(&src, 0, sizeof(src));
+
+      src.src = ttn_src_for_file_and_index(c,
+                                           tex_offset->File,
+                                           tex_offset->Index,
+                                           NULL, NULL, NULL);
+
+      src.swizzle[0] = tex_offset->SwizzleX;
+      src.swizzle[1] = tex_offset->SwizzleY;
+      src.swizzle[2] = tex_offset->SwizzleZ;
+      src.swizzle[3] = TGSI_SWIZZLE_W;
+
+      instr->src[src_number].src_type = nir_tex_src_offset;
+      instr->src[src_number].src = nir_src_for_ssa(
+         nir_fmov_alu(b, src, nir_tex_instr_src_size(instr, src_number)));
+      src_number++;
+   }
+
+   assert(src_number == num_srcs);
+
+   nir_ssa_dest_init(&instr->instr, &instr->dest, 4, NULL);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+
+   /* Resolve the writemask on the texture op. */
+   ttn_move_dest(b, dest, &instr->dest.ssa);
+}
+
+/* TGSI_OPCODE_TXQ is actually two distinct operations:
+ *
+ *     dst.x = texture\_width(unit, lod)
+ *     dst.y = texture\_height(unit, lod)
+ *     dst.z = texture\_depth(unit, lod)
+ *     dst.w = texture\_levels(unit)
+ *
+ * dst.xyz map to NIR txs opcode, and dst.w maps to query_levels
+ */
+static void
+ttn_txq(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_full_instruction *tgsi_inst = &c->token->FullInstruction;
+   nir_tex_instr *txs, *qlv;
+
+   txs = nir_tex_instr_create(b->shader, 1);
+   txs->op = nir_texop_txs;
+   setup_texture_info(txs, tgsi_inst->Texture.Texture);
+
+   qlv = nir_tex_instr_create(b->shader, 0);
+   qlv->op = nir_texop_query_levels;
+   setup_texture_info(qlv, tgsi_inst->Texture.Texture);
+
+   assert(tgsi_inst->Src[1].Register.File == TGSI_FILE_SAMPLER);
+   txs->sampler_index = tgsi_inst->Src[1].Register.Index;
+   qlv->sampler_index = tgsi_inst->Src[1].Register.Index;
+
+   /* only single src, the lod: */
+   txs->src[0].src = nir_src_for_ssa(ttn_channel(b, src[0], X));
+   txs->src[0].src_type = nir_tex_src_lod;
+
+   nir_ssa_dest_init(&txs->instr, &txs->dest, 3, NULL);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &txs->instr);
+
+   nir_ssa_dest_init(&qlv->instr, &qlv->dest, 1, NULL);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &qlv->instr);
+
+   ttn_move_dest_masked(b, dest, &txs->dest.ssa, TGSI_WRITEMASK_XYZ);
+   ttn_move_dest_masked(b, dest, &qlv->dest.ssa, TGSI_WRITEMASK_W);
+}
+
+static const nir_op op_trans[TGSI_OPCODE_LAST] = {
+   [TGSI_OPCODE_ARL] = 0,
+   [TGSI_OPCODE_MOV] = nir_op_fmov,
+   [TGSI_OPCODE_LIT] = 0,
+   [TGSI_OPCODE_RCP] = nir_op_frcp,
+   [TGSI_OPCODE_RSQ] = nir_op_frsq,
+   [TGSI_OPCODE_EXP] = 0,
+   [TGSI_OPCODE_LOG] = 0,
+   [TGSI_OPCODE_MUL] = nir_op_fmul,
+   [TGSI_OPCODE_ADD] = nir_op_fadd,
+   [TGSI_OPCODE_DP3] = 0,
+   [TGSI_OPCODE_DP4] = 0,
+   [TGSI_OPCODE_DST] = 0,
+   [TGSI_OPCODE_MIN] = nir_op_fmin,
+   [TGSI_OPCODE_MAX] = nir_op_fmax,
+   [TGSI_OPCODE_SLT] = nir_op_slt,
+   [TGSI_OPCODE_SGE] = nir_op_sge,
+   [TGSI_OPCODE_MAD] = nir_op_ffma,
+   [TGSI_OPCODE_SUB] = nir_op_fsub,
+   [TGSI_OPCODE_LRP] = 0,
+   [TGSI_OPCODE_SQRT] = nir_op_fsqrt,
+   [TGSI_OPCODE_DP2A] = 0,
+   [TGSI_OPCODE_FRC] = nir_op_ffract,
+   [TGSI_OPCODE_CLAMP] = 0,
+   [TGSI_OPCODE_FLR] = nir_op_ffloor,
+   [TGSI_OPCODE_ROUND] = nir_op_fround_even,
+   [TGSI_OPCODE_EX2] = nir_op_fexp2,
+   [TGSI_OPCODE_LG2] = nir_op_flog2,
+   [TGSI_OPCODE_POW] = nir_op_fpow,
+   [TGSI_OPCODE_XPD] = 0,
+   [TGSI_OPCODE_ABS] = nir_op_fabs,
+   [TGSI_OPCODE_DPH] = 0,
+   [TGSI_OPCODE_COS] = nir_op_fcos,
+   [TGSI_OPCODE_DDX] = nir_op_fddx,
+   [TGSI_OPCODE_DDY] = nir_op_fddy,
+   [TGSI_OPCODE_KILL] = 0,
+   [TGSI_OPCODE_PK2H] = 0, /* XXX */
+   [TGSI_OPCODE_PK2US] = 0, /* XXX */
+   [TGSI_OPCODE_PK4B] = 0, /* XXX */
+   [TGSI_OPCODE_PK4UB] = 0, /* XXX */
+   [TGSI_OPCODE_SEQ] = nir_op_seq,
+   [TGSI_OPCODE_SGT] = 0,
+   [TGSI_OPCODE_SIN] = nir_op_fsin,
+   [TGSI_OPCODE_SLE] = 0,
+   [TGSI_OPCODE_TEX] = 0,
+   [TGSI_OPCODE_TXD] = 0,
+   [TGSI_OPCODE_TXP] = 0,
+   [TGSI_OPCODE_UP2H] = 0, /* XXX */
+   [TGSI_OPCODE_UP2US] = 0, /* XXX */
+   [TGSI_OPCODE_UP4B] = 0, /* XXX */
+   [TGSI_OPCODE_UP4UB] = 0, /* XXX */
+   [TGSI_OPCODE_ARR] = 0,
+
+   /* No function calls, yet. */
+   [TGSI_OPCODE_CAL] = 0, /* XXX */
+   [TGSI_OPCODE_RET] = 0, /* XXX */
+
+   [TGSI_OPCODE_SSG] = nir_op_fsign,
+   [TGSI_OPCODE_CMP] = 0,
+   [TGSI_OPCODE_SCS] = 0,
+   [TGSI_OPCODE_TXB] = 0,
+   [TGSI_OPCODE_DIV] = nir_op_fdiv,
+   [TGSI_OPCODE_DP2] = 0,
+   [TGSI_OPCODE_DP2A] = 0,
+   [TGSI_OPCODE_TXL] = 0,
+
+   [TGSI_OPCODE_BRK] = 0,
+   [TGSI_OPCODE_IF] = 0,
+   [TGSI_OPCODE_UIF] = 0,
+   [TGSI_OPCODE_ELSE] = 0,
+   [TGSI_OPCODE_ENDIF] = 0,
+
+   [TGSI_OPCODE_DDX_FINE] = nir_op_fddx_fine,
+   [TGSI_OPCODE_DDY_FINE] = nir_op_fddy_fine,
+
+   [TGSI_OPCODE_PUSHA] = 0, /* XXX */
+   [TGSI_OPCODE_POPA] = 0, /* XXX */
+
+   [TGSI_OPCODE_CEIL] = nir_op_fceil,
+   [TGSI_OPCODE_I2F] = nir_op_i2f,
+   [TGSI_OPCODE_NOT] = nir_op_inot,
+   [TGSI_OPCODE_TRUNC] = nir_op_ftrunc,
+   [TGSI_OPCODE_SHL] = nir_op_ishl,
+   [TGSI_OPCODE_AND] = nir_op_iand,
+   [TGSI_OPCODE_OR] = nir_op_ior,
+   [TGSI_OPCODE_MOD] = nir_op_umod,
+   [TGSI_OPCODE_XOR] = nir_op_ixor,
+   [TGSI_OPCODE_SAD] = 0, /* XXX */
+   [TGSI_OPCODE_TXF] = 0,
+   [TGSI_OPCODE_TXQ] = 0,
+
+   [TGSI_OPCODE_CONT] = 0,
+
+   [TGSI_OPCODE_EMIT] = 0, /* XXX */
+   [TGSI_OPCODE_ENDPRIM] = 0, /* XXX */
+
+   [TGSI_OPCODE_BGNLOOP] = 0,
+   [TGSI_OPCODE_BGNSUB] = 0, /* XXX: no function calls */
+   [TGSI_OPCODE_ENDLOOP] = 0,
+   [TGSI_OPCODE_ENDSUB] = 0, /* XXX: no function calls */
+
+   [TGSI_OPCODE_TXQ_LZ] = 0,
+   [TGSI_OPCODE_NOP] = 0,
+   [TGSI_OPCODE_FSEQ] = nir_op_feq,
+   [TGSI_OPCODE_FSGE] = nir_op_fge,
+   [TGSI_OPCODE_FSLT] = nir_op_flt,
+   [TGSI_OPCODE_FSNE] = nir_op_fne,
+
+   /* No control flow yet */
+   [TGSI_OPCODE_CALLNZ] = 0, /* XXX */
+   [TGSI_OPCODE_BREAKC] = 0, /* not emitted by glsl_to_tgsi.cpp */
+
+   [TGSI_OPCODE_KILL_IF] = 0,
+
+   [TGSI_OPCODE_END] = 0,
+
+   [TGSI_OPCODE_F2I] = nir_op_f2i,
+   [TGSI_OPCODE_IDIV] = nir_op_idiv,
+   [TGSI_OPCODE_IMAX] = nir_op_imax,
+   [TGSI_OPCODE_IMIN] = nir_op_imin,
+   [TGSI_OPCODE_INEG] = nir_op_ineg,
+   [TGSI_OPCODE_ISGE] = nir_op_ige,
+   [TGSI_OPCODE_ISHR] = nir_op_ishr,
+   [TGSI_OPCODE_ISLT] = nir_op_ilt,
+   [TGSI_OPCODE_F2U] = nir_op_f2u,
+   [TGSI_OPCODE_U2F] = nir_op_u2f,
+   [TGSI_OPCODE_UADD] = nir_op_iadd,
+   [TGSI_OPCODE_UDIV] = nir_op_udiv,
+   [TGSI_OPCODE_UMAD] = 0,
+   [TGSI_OPCODE_UMAX] = nir_op_umax,
+   [TGSI_OPCODE_UMIN] = nir_op_umin,
+   [TGSI_OPCODE_UMOD] = nir_op_umod,
+   [TGSI_OPCODE_UMUL] = nir_op_imul,
+   [TGSI_OPCODE_USEQ] = nir_op_ieq,
+   [TGSI_OPCODE_USGE] = nir_op_uge,
+   [TGSI_OPCODE_USHR] = nir_op_ushr,
+   [TGSI_OPCODE_USLT] = nir_op_ult,
+   [TGSI_OPCODE_USNE] = nir_op_ine,
+
+   [TGSI_OPCODE_SWITCH] = 0, /* not emitted by glsl_to_tgsi.cpp */
+   [TGSI_OPCODE_CASE] = 0, /* not emitted by glsl_to_tgsi.cpp */
+   [TGSI_OPCODE_DEFAULT] = 0, /* not emitted by glsl_to_tgsi.cpp */
+   [TGSI_OPCODE_ENDSWITCH] = 0, /* not emitted by glsl_to_tgsi.cpp */
+
+   /* XXX: SAMPLE opcodes */
+
+   [TGSI_OPCODE_UARL] = nir_op_imov,
+   [TGSI_OPCODE_UCMP] = 0,
+   [TGSI_OPCODE_IABS] = nir_op_iabs,
+   [TGSI_OPCODE_ISSG] = nir_op_isign,
+
+   /* XXX: atomics */
+
+   [TGSI_OPCODE_TEX2] = 0,
+   [TGSI_OPCODE_TXB2] = 0,
+   [TGSI_OPCODE_TXL2] = 0,
+
+   [TGSI_OPCODE_IMUL_HI] = nir_op_imul_high,
+   [TGSI_OPCODE_UMUL_HI] = nir_op_umul_high,
+
+   [TGSI_OPCODE_TG4] = 0,
+   [TGSI_OPCODE_LODQ] = 0, /* XXX */
+
+   [TGSI_OPCODE_IBFE] = nir_op_ibitfield_extract,
+   [TGSI_OPCODE_UBFE] = nir_op_ubitfield_extract,
+   [TGSI_OPCODE_BFI] = nir_op_bitfield_insert,
+   [TGSI_OPCODE_BREV] = nir_op_bitfield_reverse,
+   [TGSI_OPCODE_POPC] = nir_op_bit_count,
+   [TGSI_OPCODE_LSB] = nir_op_find_lsb,
+   [TGSI_OPCODE_IMSB] = nir_op_ifind_msb,
+   [TGSI_OPCODE_UMSB] = nir_op_ifind_msb, /* XXX: signed vs unsigned */
+
+   [TGSI_OPCODE_INTERP_CENTROID] = 0, /* XXX */
+   [TGSI_OPCODE_INTERP_SAMPLE] = 0, /* XXX */
+   [TGSI_OPCODE_INTERP_OFFSET] = 0, /* XXX */
+};
+
+static void
+ttn_emit_instruction(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+   struct tgsi_full_instruction *tgsi_inst = &c->token->FullInstruction;
+   unsigned i;
+   unsigned tgsi_op = tgsi_inst->Instruction.Opcode;
+   struct tgsi_full_dst_register *tgsi_dst = &tgsi_inst->Dst[0];
+
+   if (tgsi_op == TGSI_OPCODE_END)
+      return;
+
+   nir_ssa_def *src[TGSI_FULL_MAX_SRC_REGISTERS];
+   for (i = 0; i < TGSI_FULL_MAX_SRC_REGISTERS; i++) {
+      src[i] = ttn_get_src(c, &tgsi_inst->Src[i]);
+   }
+   nir_alu_dest dest = ttn_get_dest(c, tgsi_dst);
+
+   switch (tgsi_op) {
+   case TGSI_OPCODE_RSQ:
+      ttn_move_dest(b, dest, nir_frsq(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_SQRT:
+      ttn_move_dest(b, dest, nir_fsqrt(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_RCP:
+      ttn_move_dest(b, dest, nir_frcp(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_EX2:
+      ttn_move_dest(b, dest, nir_fexp2(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_LG2:
+      ttn_move_dest(b, dest, nir_flog2(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_POW:
+      ttn_move_dest(b, dest, nir_fpow(b,
+                                      ttn_channel(b, src[0], X),
+                                      ttn_channel(b, src[1], X)));
+      break;
+
+   case TGSI_OPCODE_COS:
+      ttn_move_dest(b, dest, nir_fcos(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_SIN:
+      ttn_move_dest(b, dest, nir_fsin(b, ttn_channel(b, src[0], X)));
+      break;
+
+   case TGSI_OPCODE_ARL:
+      ttn_arl(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_EXP:
+      ttn_exp(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_LOG:
+      ttn_log(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DST:
+      ttn_dst(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_LIT:
+      ttn_lit(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_CLAMP:
+      ttn_clamp(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_XPD:
+      ttn_xpd(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DP2:
+      ttn_dp2(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DP3:
+      ttn_dp3(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DP4:
+      ttn_dp4(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DP2A:
+      ttn_dp2a(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_DPH:
+      ttn_dph(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_UMAD:
+      ttn_umad(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_LRP:
+      ttn_move_dest(b, dest, nir_flrp(b, src[2], src[1], src[0]));
+      break;
+
+   case TGSI_OPCODE_KILL:
+      ttn_kill(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_ARR:
+      ttn_arr(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_CMP:
+      ttn_cmp(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_UCMP:
+      ttn_ucmp(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_SCS:
+      ttn_scs(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_SGT:
+      ttn_sgt(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_SLE:
+      ttn_sle(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_KILL_IF:
+      ttn_kill_if(b, op_trans[tgsi_op], dest, src);
+      break;
+
+   case TGSI_OPCODE_TEX:
+   case TGSI_OPCODE_TXP:
+   case TGSI_OPCODE_TXL:
+   case TGSI_OPCODE_TXB:
+   case TGSI_OPCODE_TXD:
+   case TGSI_OPCODE_TXL2:
+   case TGSI_OPCODE_TXB2:
+   case TGSI_OPCODE_TXQ_LZ:
+   case TGSI_OPCODE_TXF:
+   case TGSI_OPCODE_TG4:
+      ttn_tex(c, dest, src);
+      break;
+
+   case TGSI_OPCODE_TXQ:
+      ttn_txq(c, dest, src);
+      break;
+
+   case TGSI_OPCODE_NOP:
+      break;
+
+   case TGSI_OPCODE_IF:
+      ttn_if(c, src[0], false);
+      break;
+
+   case TGSI_OPCODE_UIF:
+      ttn_if(c, src[0], true);
+      break;
+
+   case TGSI_OPCODE_ELSE:
+      ttn_else(c);
+      break;
+
+   case TGSI_OPCODE_ENDIF:
+      ttn_endif(c);
+      break;
+
+   case TGSI_OPCODE_BGNLOOP:
+      ttn_bgnloop(c);
+      break;
+
+   case TGSI_OPCODE_BRK:
+      ttn_brk(b);
+      break;
+
+   case TGSI_OPCODE_CONT:
+      ttn_cont(b);
+      break;
+
+   case TGSI_OPCODE_ENDLOOP:
+      ttn_endloop(c);
+      break;
+
+   default:
+      if (op_trans[tgsi_op] != 0 || tgsi_op == TGSI_OPCODE_MOV) {
+         ttn_alu(b, op_trans[tgsi_op], dest, src);
+      } else {
+         fprintf(stderr, "unknown TGSI opcode: %s\n",
+                 tgsi_get_opcode_name(tgsi_op));
+         abort();
+      }
+      break;
+   }
+
+   if (tgsi_inst->Instruction.Saturate) {
+      assert(tgsi_inst->Instruction.Saturate == TGSI_SAT_ZERO_ONE);
+      assert(!dest.dest.is_ssa);
+      ttn_move_dest(b, dest, nir_fsat(b, ttn_src_for_dest(b, &dest)));
+   }
+
+   /* if the dst has a matching var, append store_global to move
+    * output from reg to var
+    */
+   nir_variable *var = ttn_get_var(c, tgsi_dst);
+   if (var) {
+      unsigned index = tgsi_dst->Register.Index;
+      unsigned offset = c->temp_regs[index].offset;
+      nir_intrinsic_instr *store =
+         nir_intrinsic_instr_create(b->shader, nir_intrinsic_store_var);
+      struct tgsi_ind_register *indirect = tgsi_dst->Register.Indirect ?
+                                           &tgsi_dst->Indirect : NULL;
+
+      store->num_components = 4;
+      store->variables[0] = ttn_array_deref(c, store, var, offset, indirect);
+      store->src[0] = nir_src_for_reg(dest.dest.reg.reg);
+
+      nir_instr_insert_after_cf_list(b->cf_node_list, &store->instr);
+   }
+}
+
+/**
+ * Puts a NIR intrinsic to store of each TGSI_FILE_OUTPUT value to the output
+ * variables at the end of the shader.
+ *
+ * We don't generate these incrementally as the TGSI_FILE_OUTPUT values are
+ * written, because there's no output load intrinsic, which means we couldn't
+ * handle writemasks.
+ */
+static void
+ttn_add_output_stores(struct ttn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   foreach_list_typed(nir_variable, var, node, &b->shader->outputs) {
+      unsigned array_len = MAX2(glsl_get_length(var->type), 1);
+      unsigned i;
+
+      for (i = 0; i < array_len; i++) {
+         nir_intrinsic_instr *store =
+            nir_intrinsic_instr_create(b->shader, nir_intrinsic_store_output);
+         store->num_components = 4;
+         store->const_index[0] = var->data.driver_location + i;
+         store->const_index[1] = 1;
+         store->src[0].reg.reg = c->output_regs[var->data.driver_location].reg;
+         nir_instr_insert_after_cf_list(b->cf_node_list, &store->instr);
+      }
+   }
+}
+
+struct nir_shader *
+tgsi_to_nir(const void *tgsi_tokens,
+            const nir_shader_compiler_options *options)
+{
+   struct tgsi_parse_context parser;
+   struct tgsi_shader_info scan;
+   struct ttn_compile *c;
+   struct nir_shader *s;
+   int ret;
+
+   c = rzalloc(NULL, struct ttn_compile);
+   s = nir_shader_create(NULL, options);
+
+   nir_function *func = nir_function_create(s, "main");
+   nir_function_overload *overload = nir_function_overload_create(func);
+   nir_function_impl *impl = nir_function_impl_create(overload);
+
+   nir_builder_init(&c->build, impl);
+   nir_builder_insert_after_cf_list(&c->build, &impl->body);
+
+   tgsi_scan_shader(tgsi_tokens, &scan);
+   c->scan = &scan;
+
+   s->num_inputs = scan.file_max[TGSI_FILE_INPUT] + 1;
+   s->num_uniforms = scan.const_file_max[0] + 1;
+   s->num_outputs = scan.file_max[TGSI_FILE_OUTPUT] + 1;
+
+   c->output_regs = rzalloc_array(c, struct ttn_reg_info,
+                                  scan.file_max[TGSI_FILE_OUTPUT] + 1);
+   c->temp_regs = rzalloc_array(c, struct ttn_reg_info,
+                                scan.file_max[TGSI_FILE_TEMPORARY] + 1);
+   c->imm_defs = rzalloc_array(c, nir_ssa_def *,
+                               scan.file_max[TGSI_FILE_IMMEDIATE] + 1);
+
+   c->if_stack = rzalloc_array(c, struct exec_list *,
+                               (scan.opcode_count[TGSI_OPCODE_IF] +
+                                scan.opcode_count[TGSI_OPCODE_UIF]) * 2);
+   c->loop_stack = rzalloc_array(c, struct exec_list *,
+                                 scan.opcode_count[TGSI_OPCODE_BGNLOOP]);
+
+   ret = tgsi_parse_init(&parser, tgsi_tokens);
+   assert(ret == TGSI_PARSE_OK);
+
+   while (!tgsi_parse_end_of_tokens(&parser)) {
+      tgsi_parse_token(&parser);
+      c->token = &parser.FullToken;
+
+      switch (parser.FullToken.Token.Type) {
+      case TGSI_TOKEN_TYPE_DECLARATION:
+         ttn_emit_declaration(c);
+         break;
+
+      case TGSI_TOKEN_TYPE_INSTRUCTION:
+         ttn_emit_instruction(c);
+         break;
+
+      case TGSI_TOKEN_TYPE_IMMEDIATE:
+         ttn_emit_immediate(c);
+         break;
+      }
+   }
+
+   tgsi_parse_free(&parser);
+
+   ttn_add_output_stores(c);
+
+   ralloc_free(c);
+   return s;
+}
diff --git a/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.h b/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.h
new file mode 100644
index 000000000..687348a80
--- /dev/null
+++ b/mesalib/src/gallium/auxiliary/nir/tgsi_to_nir.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright © 2014 Broadcom
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "glsl/nir/nir.h"
+
+struct nir_shader_compiler_options *options;
+
+struct nir_shader *
+tgsi_to_nir(const void *tgsi_tokens,
+            const struct nir_shader_compiler_options *options);
diff --git a/mesalib/src/gallium/auxiliary/util/u_debug_symbol.c b/mesalib/src/gallium/auxiliary/util/u_debug_symbol.c
index bba9122e7..542493252 100644
--- a/mesalib/src/gallium/auxiliary/util/u_debug_symbol.c
+++ b/mesalib/src/gallium/auxiliary/util/u_debug_symbol.c
@@ -82,18 +82,10 @@ getDbgHelpProcAddress(LPCSTR lpProcName)
          hModule = LoadLibraryA("mgwhelp.dll");
          if (!hModule) {
             _debug_printf("warning: mgwhelp.dll not found: symbol names will not be resolved\n"
-                          "warning: download it from http://code.google.com/p/jrfonseca/wiki/DrMingw#MgwHelp\n");
+                          "warning: download it from https://github.com/jrfonseca/drmingw/#mgwhelp\n");
          }
       }
-
-      /*
-       * bfdhelp.dll was the predecessor of mgwhelp.dll.  It is available from
-       * http://people.freedesktop.org/~jrfonseca/bfdhelp/ for now.
-       */
-      if (!hModule) {
-         hModule = LoadLibraryA("bfdhelp.dll");
-      }
-   #endif
+#endif
 
       /*
        * Fallback to the real DbgHelp.
diff --git a/mesalib/src/gallium/auxiliary/util/u_linkage.c b/mesalib/src/gallium/auxiliary/util/u_linkage.c
deleted file mode 100644
index 245d941dc..000000000
--- a/mesalib/src/gallium/auxiliary/util/u_linkage.c
+++ /dev/null
@@ -1,149 +0,0 @@
-/**************************************************************************
- *
- * Copyright 2010 Luca Barbieri
- *
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial
- * portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
- * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
- * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
- * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-#include "util/u_debug.h"
-#include "pipe/p_shader_tokens.h"
-#include "tgsi/tgsi_parse.h"
-#include "tgsi/tgsi_scan.h"
-#include "util/u_linkage.h"
-
-/* we must only record the registers that are actually used, not just declared */
-static INLINE boolean
-util_semantic_set_test_and_set(struct util_semantic_set *set, unsigned value)
-{
-   unsigned mask = 1 << (value % (sizeof(long) * 8));
-   unsigned long *p = &set->masks[value / (sizeof(long) * 8)];
-   unsigned long v = *p & mask;
-   *p |= mask;
-   return !!v;
-}
-
-unsigned
-util_semantic_set_from_program_file(struct util_semantic_set *set, const struct tgsi_token *tokens, enum tgsi_file_type file)
-{
-   struct tgsi_shader_info info;
-   struct tgsi_parse_context parse;
-   unsigned count = 0;
-   ubyte *semantic_name;
-   ubyte *semantic_index;
-
-   tgsi_scan_shader(tokens, &info);
-
-   if(file == TGSI_FILE_INPUT)
-   {
-      semantic_name = info.input_semantic_name;
-      semantic_index = info.input_semantic_index;
-   }
-   else if(file == TGSI_FILE_OUTPUT)
-   {
-      semantic_name = info.output_semantic_name;
-      semantic_index = info.output_semantic_index;
-   }
-   else
-   {
-      assert(0);
-      semantic_name = NULL;
-      semantic_index = NULL;
-   }
-
-   tgsi_parse_init(&parse, tokens);
-
-   memset(set->masks, 0, sizeof(set->masks));
-   while(!tgsi_parse_end_of_tokens(&parse))
-   {
-      tgsi_parse_token(&parse);
-
-      if(parse.FullToken.Token.Type == TGSI_TOKEN_TYPE_INSTRUCTION)
-      {
-	 const struct tgsi_full_instruction *finst = &parse.FullToken.FullInstruction;
-	 unsigned i;
-	 for(i = 0; i < finst->Instruction.NumDstRegs; ++i)
-	 {
-	    if(finst->Dst[i].Register.File == file)
-	    {
-	       unsigned idx = finst->Dst[i].Register.Index;
-	       if(semantic_name[idx] == TGSI_SEMANTIC_GENERIC)
-	       {
-		  if(!util_semantic_set_test_and_set(set, semantic_index[idx]))
-		     ++count;
-	       }
-	    }
-	 }
-
-	 for(i = 0; i < finst->Instruction.NumSrcRegs; ++i)
-	 {
-	    if(finst->Src[i].Register.File == file)
-	    {
-	       unsigned idx = finst->Src[i].Register.Index;
-	       if(semantic_name[idx] == TGSI_SEMANTIC_GENERIC)
-	       {
-		  if(!util_semantic_set_test_and_set(set, semantic_index[idx]))
-		     ++count;
-	       }
-	    }
-	 }
-      }
-   }
-   tgsi_parse_free(&parse);
-
-   return count;
-}
-
-#define UTIL_SEMANTIC_SET_FOR_EACH(i, set) for(i = 0; i < 256; ++i) if(set->masks[i / (sizeof(long) * 8)] & (1 << (i % (sizeof(long) * 8))))
-
-void
-util_semantic_layout_from_set(unsigned char *layout, const struct util_semantic_set *set, unsigned efficient_slots, unsigned num_slots)
-{
-   int first = -1;
-   int last = -1;
-   unsigned i;
-
-   memset(layout, 0xff, num_slots);
-
-   UTIL_SEMANTIC_SET_FOR_EACH(i, set)
-   {
-      if(first < 0)
-	 first = i;
-      last = i;
-   }
-
-   if (last < (int) efficient_slots)
-   {
-      UTIL_SEMANTIC_SET_FOR_EACH(i, set)
-         layout[i] = i;
-   }
-   else if ((last - first) < (int) efficient_slots)
-   {
-      UTIL_SEMANTIC_SET_FOR_EACH(i, set)
-         layout[i - first] = i;
-   }
-   else
-   {
-      unsigned idx = 0;
-      UTIL_SEMANTIC_SET_FOR_EACH(i, set)
-         layout[idx++] = i;
-   }
-}
diff --git a/mesalib/src/gallium/auxiliary/util/u_linkage.h b/mesalib/src/gallium/auxiliary/util/u_linkage.h
deleted file mode 100644
index 7b23123d3..000000000
--- a/mesalib/src/gallium/auxiliary/util/u_linkage.h
+++ /dev/null
@@ -1,67 +0,0 @@
-/**************************************************************************
- *
- * Copyright 2010 Luca Barbieri
- *
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial
- * portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
- * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
- * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
- * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-#ifndef U_LINKAGE_H_
-#define U_LINKAGE_H_
-
-#include "pipe/p_compiler.h"
-#include "pipe/p_shader_tokens.h"
-
-struct util_semantic_set
-{
-   unsigned long masks[256 / 8 / sizeof(unsigned long)];
-};
-
-static INLINE boolean
-util_semantic_set_contains(struct util_semantic_set *set, unsigned char value)
-{
-   return !!(set->masks[value / (sizeof(long) * 8)] & (1 << (value / (sizeof(long) * 8))));
-}
-
-unsigned util_semantic_set_from_program_file(struct util_semantic_set *set, const struct tgsi_token *tokens, enum tgsi_file_type file);
-
-/* efficient_slots is the number of slots such that hardware performance is
- * the same for using that amount, with holes, or less slots but with less
- * holes.
- *
- * num_slots is the size of the layout array and hardware limit instead.
- *
- * efficient_slots == 0 or efficient_slots == num_slots are typical settings.
- */
-void util_semantic_layout_from_set(unsigned char *layout, const struct util_semantic_set *set, unsigned efficient_slots, unsigned num_slots);
-
-static INLINE void
-util_semantic_table_from_layout(unsigned char *table, size_t table_size, unsigned char *layout,
-                                unsigned char first_slot_value, unsigned char num_slots)
-{
-   unsigned char i;
-   memset(table, 0xff, table_size);
-
-   for(i = 0; i < num_slots; ++i)
-      table[layout[i]] = first_slot_value + i;
-}
-
-#endif /* U_LINKAGE_H_ */
diff --git a/mesalib/src/gallium/auxiliary/util/u_math.h b/mesalib/src/gallium/auxiliary/util/u_math.h
index 8f62cac66..3d27a59e8 100644
--- a/mesalib/src/gallium/auxiliary/util/u_math.h
+++ b/mesalib/src/gallium/auxiliary/util/u_math.h
@@ -383,14 +383,28 @@ unsigned ffs( unsigned u )
 
    return i;
 }
-#elif defined(__MINGW32__) || defined(PIPE_OS_ANDROID)
+#elif defined(__MINGW32__) || defined(PIPE_OS_ANDROID) || \
+    defined(HAVE___BUILTIN_FFS)
 #define ffs __builtin_ffs
-#define ffsll __builtin_ffsll
 #endif
 
 #endif /* FFS_DEFINED */
 
 /**
+ * Find first bit set in long long.  Least significant bit is 1.
+ * Return 0 if no bits set.
+ */
+#ifndef FFSLL_DEFINED
+#define FFSLL_DEFINED 1
+
+#if defined(__MINGW32__) || defined(PIPE_OS_ANDROID) || \
+    defined(HAVE___BUILTIN_FFSLL)
+#define ffsll __builtin_ffsll
+#endif
+
+#endif /* FFSLL_DEFINED */
+
+/**
  * Find last bit set in a word.  The least significant bit is 1.
  * Return 0 if no bits are set.
  */
diff --git a/mesalib/src/gallium/auxiliary/util/u_tile.c b/mesalib/src/gallium/auxiliary/util/u_tile.c
index 6252e5dcf..8e199200f 100644
--- a/mesalib/src/gallium/auxiliary/util/u_tile.c
+++ b/mesalib/src/gallium/auxiliary/util/u_tile.c
@@ -214,13 +214,13 @@ s8x24_get_tile_rgba(const unsigned *src,
    unsigned i, j;
 
    for (i = 0; i < h; i++) {
-      float *pRow = p;
+      uint32_t *pRow = (uint32_t *)p;
 
       for (j = 0; j < w; j++, pRow += 4) {
          pRow[0] =
          pRow[1] =
          pRow[2] =
-         pRow[3] = (float)((*src++ >> 24) & 0xff);
+         pRow[3] = ((*src++ >> 24) & 0xff);
       }
 
       p += dst_stride;
@@ -241,12 +241,12 @@ x24s8_get_tile_rgba(const unsigned *src,
    unsigned i, j;
 
    for (i = 0; i < h; i++) {
-      float *pRow = p;
+      uint32_t *pRow = (uint32_t *)p;
       for (j = 0; j < w; j++, pRow += 4) {
          pRow[0] =
          pRow[1] =
          pRow[2] =
-         pRow[3] = (float)(*src++ & 0xff);
+         pRow[3] = (*src++ & 0xff);
       }
       p += dst_stride;
    }
@@ -265,12 +265,12 @@ s8_get_tile_rgba(const unsigned char *src,
    unsigned i, j;
 
    for (i = 0; i < h; i++) {
-      float *pRow = p;
+      uint32_t *pRow = (uint32_t *)p;
       for (j = 0; j < w; j++, pRow += 4) {
          pRow[0] =
          pRow[1] =
          pRow[2] =
-         pRow[3] = (float)(*src++ & 0xff);
+         pRow[3] = (*src++ & 0xff);
       }
       p += dst_stride;
    }
diff --git a/mesalib/src/glsl/Makefile.am b/mesalib/src/glsl/Makefile.am
index b466a3b5c..23c6fe8bb 100644
--- a/mesalib/src/glsl/Makefile.am
+++ b/mesalib/src/glsl/Makefile.am
@@ -46,6 +46,7 @@ EXTRA_DIST = tests glcpp/tests README TODO glcpp/README	\
 	glcpp/glcpp-lex.l				\
 	glcpp/glcpp-parse.y				\
 	nir/nir_algebraic.py				\
+	nir/nir_builder_opcodes_h.py			\
 	nir/nir_constant_expressions.py			\
 	nir/nir_opcodes.py				\
 	nir/nir_opcodes_c.py				\
@@ -67,7 +68,7 @@ TESTS_ENVIRONMENT= \
 	export PYTHON2=$(PYTHON2); \
 	export PYTHON_FLAGS=$(PYTHON_FLAGS);
 
-noinst_LTLIBRARIES = libglsl.la libglcpp.la
+noinst_LTLIBRARIES = libnir.la libglsl.la libglcpp.la
 check_PROGRAMS =					\
 	glcpp/glcpp					\
 	glsl_test					\
@@ -147,6 +148,12 @@ libglsl_la_SOURCES =					\
 	$(LIBGLSL_FILES)				\
 	$(NIR_FILES)
 
+libnir_la_SOURCES =					\
+	glsl_types.cpp					\
+	builtin_types.cpp				\
+	glsl_symbol_table.cpp				\
+	$(NIR_FILES)
+
 glsl_compiler_SOURCES = \
 	$(GLSL_COMPILER_CXX_FILES)
 
@@ -251,8 +258,6 @@ nir/nir_opcodes.h: nir/nir_opcodes.py nir/nir_opcodes_h.py
 	$(MKDIR_P) nir;							\
 	$(PYTHON2) $(PYTHON_FLAGS) $(srcdir)/nir/nir_opcodes_h.py > $@
 
-nir/nir.h: nir/nir_opcodes.h
-
 nir/nir_opcodes.c: nir/nir_opcodes.py nir/nir_opcodes_c.py
 	$(MKDIR_P) nir;							\
 	$(PYTHON2) $(PYTHON_FLAGS) $(srcdir)/nir/nir_opcodes_c.py > $@
diff --git a/mesalib/src/glsl/Makefile.sources b/mesalib/src/glsl/Makefile.sources
index b876642e8..c471eca23 100644
--- a/mesalib/src/glsl/Makefile.sources
+++ b/mesalib/src/glsl/Makefile.sources
@@ -22,6 +22,7 @@ NIR_FILES = \
 	nir/glsl_to_nir.h \
 	nir/nir.c \
 	nir/nir.h \
+	nir/nir_builder.h \
 	nir/nir_constant_expressions.h \
 	nir/nir_dominance.c \
 	nir/nir_from_ssa.c \
@@ -32,21 +33,25 @@ NIR_FILES = \
 	nir/nir_lower_atomics.c \
 	nir/nir_lower_global_vars_to_local.c \
 	nir/nir_lower_locals_to_regs.c \
+	nir/nir_lower_idiv.c \
 	nir/nir_lower_io.c \
 	nir/nir_lower_phis_to_scalar.c \
 	nir/nir_lower_samplers.cpp \
 	nir/nir_lower_system_values.c \
+	nir/nir_lower_tex_projector.c \
 	nir/nir_lower_to_source_mods.c \
 	nir/nir_lower_vars_to_ssa.c \
 	nir/nir_lower_var_copies.c \
 	nir/nir_lower_vec_to_movs.c \
 	nir/nir_metadata.c \
+	nir/nir_normalize_cubemap_coords.c \
 	nir/nir_opt_constant_folding.c \
 	nir/nir_opt_copy_propagate.c \
 	nir/nir_opt_cse.c \
 	nir/nir_opt_dce.c \
 	nir/nir_opt_gcm.c \
 	nir/nir_opt_global_to_local.c \
+	nir/nir_opt_peephole_ffma.c \
 	nir/nir_opt_peephole_select.c \
 	nir/nir_opt_remove_phis.c \
 	nir/nir_print.c \
@@ -54,9 +59,11 @@ NIR_FILES = \
 	nir/nir_search.c \
 	nir/nir_search.h \
 	nir/nir_split_var_copies.c \
+	nir/nir_sweep.c \
 	nir/nir_to_ssa.c \
 	nir/nir_types.h \
 	nir/nir_validate.c \
+	nir/nir_vla.h \
 	nir/nir_worklist.c \
 	nir/nir_worklist.h \
 	nir/nir_types.cpp \
@@ -183,7 +190,8 @@ LIBGLSL_FILES = \
 	opt_vectorize.cpp \
 	program.h \
 	s_expression.cpp \
-	s_expression.h
+	s_expression.h \
+	shader_enums.h
 
 # glsl_compiler
 
diff --git a/mesalib/src/glsl/ast_function.cpp b/mesalib/src/glsl/ast_function.cpp
index 918be6966..87df93e68 100644
--- a/mesalib/src/glsl/ast_function.cpp
+++ b/mesalib/src/glsl/ast_function.cpp
@@ -1370,71 +1370,59 @@ emit_inline_matrix_constructor(const glsl_type *type,
    } else {
       const unsigned cols = type->matrix_columns;
       const unsigned rows = type->vector_elements;
+      unsigned remaining_slots = rows * cols;
       unsigned col_idx = 0;
       unsigned row_idx = 0;
 
       foreach_in_list(ir_rvalue, rhs, parameters) {
-	 const unsigned components_remaining_this_column = rows - row_idx;
-	 unsigned rhs_components = rhs->type->components();
-	 unsigned rhs_base = 0;
-
-	 /* Since the parameter might be used in the RHS of two assignments,
-	  * generate a temporary and copy the paramter there.
-	  */
-	 ir_variable *rhs_var =
-	    new(ctx) ir_variable(rhs->type, "mat_ctor_vec", ir_var_temporary);
-	 instructions->push_tail(rhs_var);
-
-	 ir_dereference *rhs_var_ref =
-	    new(ctx) ir_dereference_variable(rhs_var);
-	 ir_instruction *inst = new(ctx) ir_assignment(rhs_var_ref, rhs, NULL);
-	 instructions->push_tail(inst);
-
-	 /* Assign the current parameter to as many components of the matrix
-	  * as it will fill.
-	  *
-	  * NOTE: A single vector parameter can span two matrix columns.  A
-	  * single vec4, for example, can completely fill a mat2.
-	  */
-	 if (rhs_components >= components_remaining_this_column) {
-	    const unsigned count = MIN2(rhs_components,
-					components_remaining_this_column);
-
-	    rhs_var_ref = new(ctx) ir_dereference_variable(rhs_var);
-
-	    ir_instruction *inst = assign_to_matrix_column(var, col_idx,
-							   row_idx,
-							   rhs_var_ref, 0,
-							   count, ctx);
-	    instructions->push_tail(inst);
-
-	    rhs_base = count;
-
-	    col_idx++;
-	    row_idx = 0;
-	 }
-
-	 /* If there is data left in the parameter and components left to be
-	  * set in the destination, emit another assignment.  It is possible
-	  * that the assignment could be of a vec4 to the last element of the
-	  * matrix.  In this case col_idx==cols, but there is still data
-	  * left in the source parameter.  Obviously, don't emit an assignment
-	  * to data outside the destination matrix.
-	  */
-	 if ((col_idx < cols) && (rhs_base < rhs_components)) {
-	    const unsigned count = rhs_components - rhs_base;
-
-	    rhs_var_ref = new(ctx) ir_dereference_variable(rhs_var);
-
-	    ir_instruction *inst = assign_to_matrix_column(var, col_idx,
-							   row_idx,
-							   rhs_var_ref,
-							   rhs_base,
-							   count, ctx);
-	    instructions->push_tail(inst);
-
-	    row_idx += count;
-	 }
+         unsigned rhs_components = rhs->type->components();
+         unsigned rhs_base = 0;
+
+         if (remaining_slots == 0)
+            break;
+
+         /* Since the parameter might be used in the RHS of two assignments,
+          * generate a temporary and copy the paramter there.
+          */
+         ir_variable *rhs_var =
+            new(ctx) ir_variable(rhs->type, "mat_ctor_vec", ir_var_temporary);
+         instructions->push_tail(rhs_var);
+
+         ir_dereference *rhs_var_ref =
+            new(ctx) ir_dereference_variable(rhs_var);
+         ir_instruction *inst = new(ctx) ir_assignment(rhs_var_ref, rhs, NULL);
+         instructions->push_tail(inst);
+
+         do {
+            /* Assign the current parameter to as many components of the matrix
+             * as it will fill.
+             *
+             * NOTE: A single vector parameter can span two matrix columns.  A
+             * single vec4, for example, can completely fill a mat2.
+             */
+            unsigned count = MIN2(rows - row_idx,
+                                  rhs_components - rhs_base);
+
+            rhs_var_ref = new(ctx) ir_dereference_variable(rhs_var);
+            ir_instruction *inst = assign_to_matrix_column(var, col_idx,
+                                                         row_idx,
+                                                         rhs_var_ref,
+                                                         rhs_base,
+                                                         count, ctx);
+            instructions->push_tail(inst);
+            rhs_base += count;
+            row_idx += count;
+            remaining_slots -= count;
+
+            /* Sometimes, there is still data left in the parameters and
+             * components left to be set in the destination but in other
+             * column.
+             */
+            if (row_idx >= rows) {
+               row_idx = 0;
+               col_idx++;
+            }
+         } while(remaining_slots > 0 && rhs_base < rhs_components);
       }
    }
 
@@ -1791,7 +1779,7 @@ ast_function_expression::hir(exec_list *instructions,
       return value;
    }
 
-   return ir_rvalue::error_value(ctx);
+   unreachable("not reached");
 }
 
 ir_rvalue *
diff --git a/mesalib/src/glsl/ast_to_hir.cpp b/mesalib/src/glsl/ast_to_hir.cpp
index d387b2e35..78369360f 100644
--- a/mesalib/src/glsl/ast_to_hir.cpp
+++ b/mesalib/src/glsl/ast_to_hir.cpp
@@ -375,66 +375,14 @@ arithmetic_result_type(ir_rvalue * &value_a, ir_rvalue * &value_b,
       if (type_a == type_b)
          return type_a;
    } else {
-      if (type_a->is_matrix() && type_b->is_matrix()) {
-         /* Matrix multiply.  The columns of A must match the rows of B.  Given
-          * the other previously tested constraints, this means the vector type
-          * of a row from A must be the same as the vector type of a column from
-          * B.
-          */
-         if (type_a->row_type() == type_b->column_type()) {
-            /* The resulting matrix has the number of columns of matrix B and
-             * the number of rows of matrix A.  We get the row count of A by
-             * looking at the size of a vector that makes up a column.  The
-             * transpose (size of a row) is done for B.
-             */
-            const glsl_type *const type =
-               glsl_type::get_instance(type_a->base_type,
-                                       type_a->column_type()->vector_elements,
-                                       type_b->row_type()->vector_elements);
-            assert(type != glsl_type::error_type);
-
-            return type;
-         }
-      } else if (type_a->is_matrix()) {
-         /* A is a matrix and B is a column vector.  Columns of A must match
-          * rows of B.  Given the other previously tested constraints, this
-          * means the vector type of a row from A must be the same as the
-          * vector the type of B.
-          */
-         if (type_a->row_type() == type_b) {
-            /* The resulting vector has a number of elements equal to
-             * the number of rows of matrix A. */
-            const glsl_type *const type =
-               glsl_type::get_instance(type_a->base_type,
-                                       type_a->column_type()->vector_elements,
-                                       1);
-            assert(type != glsl_type::error_type);
-
-            return type;
-         }
-      } else {
-         assert(type_b->is_matrix());
+      const glsl_type *type = glsl_type::get_mul_type(type_a, type_b);
 
-         /* A is a row vector and B is a matrix.  Columns of A must match rows
-          * of B.  Given the other previously tested constraints, this means
-          * the type of A must be the same as the vector type of a column from
-          * B.
-          */
-         if (type_a == type_b->column_type()) {
-            /* The resulting vector has a number of elements equal to
-             * the number of columns of matrix B. */
-            const glsl_type *const type =
-               glsl_type::get_instance(type_a->base_type,
-                                       type_b->row_type()->vector_elements,
-                                       1);
-            assert(type != glsl_type::error_type);
-
-            return type;
-         }
+      if (type == glsl_type::error_type) {
+         _mesa_glsl_error(loc, state,
+                          "size mismatch for matrix multiplication");
       }
 
-      _mesa_glsl_error(loc, state, "size mismatch for matrix multiplication");
-      return glsl_type::error_type;
+      return type;
    }
 
 
@@ -5776,6 +5724,9 @@ ast_interface_block::hir(exec_list *instructions,
       var->data.matrix_layout = matrix_layout == GLSL_MATRIX_LAYOUT_INHERITED
          ? GLSL_MATRIX_LAYOUT_COLUMN_MAJOR : matrix_layout;
 
+      if (var_mode == ir_var_shader_in || var_mode == ir_var_uniform)
+         var->data.read_only = true;
+
       if (state->stage == MESA_SHADER_GEOMETRY && var_mode == ir_var_shader_in)
          handle_geometry_shader_input_decl(state, loc, var);
 
@@ -5816,6 +5767,9 @@ ast_interface_block::hir(exec_list *instructions,
          var->data.sample = fields[i].sample;
          var->init_interface_type(block_type);
 
+         if (var_mode == ir_var_shader_in || var_mode == ir_var_uniform)
+            var->data.read_only = true;
+
          if (fields[i].matrix_layout == GLSL_MATRIX_LAYOUT_INHERITED) {
             var->data.matrix_layout = matrix_layout == GLSL_MATRIX_LAYOUT_INHERITED
                ? GLSL_MATRIX_LAYOUT_COLUMN_MAJOR : matrix_layout;
diff --git a/mesalib/src/glsl/builtin_functions.cpp b/mesalib/src/glsl/builtin_functions.cpp
index e46db8552..98d0b1887 100755
--- a/mesalib/src/glsl/builtin_functions.cpp
+++ b/mesalib/src/glsl/builtin_functions.cpp
@@ -60,7 +60,7 @@
 #include "ir_builder.h"
 #include "glsl_parser_extras.h"
 #include "program/prog_instruction.h"
-#include <limits>
+#include <math.h>
 
 #define M_PIf   ((float) M_PI)
 #define M_PI_2f ((float) M_PI/2.0f)
@@ -3215,7 +3215,7 @@ builtin_builder::_isinf(builtin_available_predicate avail, const glsl_type *type
 
    ir_constant_data infinities;
    for (int i = 0; i < type->vector_elements; i++) {
-      infinities.f[i] = std::numeric_limits<float>::infinity();
+      infinities.f[i] = INFINITY;
    }
 
    body.emit(ret(equal(abs(x), imm(type, infinities))));
diff --git a/mesalib/src/glsl/glsl_lexer.ll b/mesalib/src/glsl/glsl_lexer.ll
index 8dc3d106b..2785ed168 100644
--- a/mesalib/src/glsl/glsl_lexer.ll
+++ b/mesalib/src/glsl/glsl_lexer.ll
@@ -36,14 +36,13 @@ static int classify_identifier(struct _mesa_glsl_parse_state *, const char *);
 
 #define YY_USER_ACTION						\
    do {								\
-      yylloc->source = 0;					\
       yylloc->first_column = yycolumn + 1;			\
       yylloc->first_line = yylloc->last_line = yylineno + 1;	\
       yycolumn += yyleng;					\
       yylloc->last_column = yycolumn + 1;			\
    } while(0);
 
-#define YY_USER_INIT yylineno = 0; yycolumn = 0;
+#define YY_USER_INIT yylineno = 0; yycolumn = 0; yylloc->source = 0;
 
 /* A macro for handling reserved words and keywords across language versions.
  *
@@ -188,6 +187,15 @@ HASH		^{SPC}#{SPC}
 				    * one-based.
 				    */
 				   yylineno = strtol(ptr, &ptr, 0) - 1;
+
+                                   /* From GLSL 3.30 and GLSL ES on, after processing the
+                                    * line directive (including its new-line), the implementation
+                                    * will behave as if it is compiling at the line number passed
+                                    * as argument. It was line number + 1 in older specifications.
+                                    */
+                                   if (yyextra->is_version(330, 100))
+                                      yylineno--;
+
 				   yylloc->source = strtol(ptr, NULL, 0);
 				}
 {HASH}line{SPCP}{INT}{SPC}$	{
@@ -203,6 +211,14 @@ HASH		^{SPC}#{SPC}
 				    * one-based.
 				    */
 				   yylineno = strtol(ptr, &ptr, 0) - 1;
+
+                                   /* From GLSL 3.30 and GLSL ES on, after processing the
+                                    * line directive (including its new-line), the implementation
+                                    * will behave as if it is compiling at the line number passed
+                                    * as argument. It was line number + 1 in older specifications.
+                                    */
+                                   if (yyextra->is_version(330, 100))
+                                      yylineno--;
 				}
 ^{SPC}#{SPC}pragma{SPCP}debug{SPC}\({SPC}on{SPC}\) {
 				  BEGIN PP;
diff --git a/mesalib/src/glsl/glsl_parser_extras.cpp b/mesalib/src/glsl/glsl_parser_extras.cpp
index 79624bc26..0aa3c54fc 100644
--- a/mesalib/src/glsl/glsl_parser_extras.cpp
+++ b/mesalib/src/glsl/glsl_parser_extras.cpp
@@ -73,8 +73,8 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    this->uses_builtin_functions = false;
 
    /* Set default language version and extensions */
-   this->language_version = ctx->Const.ForceGLSLVersion ?
-                            ctx->Const.ForceGLSLVersion : 110;
+   this->language_version = 110;
+   this->forced_language_version = ctx->Const.ForceGLSLVersion;
    this->es_shader = false;
    this->ARB_texture_rectangle_enable = true;
 
@@ -320,11 +320,14 @@ _mesa_glsl_parse_state::process_version_directive(YYLTYPE *locp, int version,
       this->ARB_texture_rectangle_enable = false;
    }
 
-   this->language_version = version;
+   if (this->forced_language_version)
+      this->language_version = this->forced_language_version;
+   else
+      this->language_version = version;
 
    bool supported = false;
    for (unsigned i = 0; i < this->num_supported_versions; i++) {
-      if (this->supported_versions[i].ver == (unsigned) version
+      if (this->supported_versions[i].ver == this->language_version
           && this->supported_versions[i].es == this->es_shader) {
          supported = true;
          break;
diff --git a/mesalib/src/glsl/glsl_parser_extras.h b/mesalib/src/glsl/glsl_parser_extras.h
index c5670fdb1..5047049fb 100644
--- a/mesalib/src/glsl/glsl_parser_extras.h
+++ b/mesalib/src/glsl/glsl_parser_extras.h
@@ -109,8 +109,10 @@ struct _mesa_glsl_parse_state {
    {
       unsigned required_version = this->es_shader ?
          required_glsl_es_version : required_glsl_version;
+      unsigned this_version = this->forced_language_version
+         ? this->forced_language_version : this->language_version;
       return required_version != 0
-         && this->language_version >= required_version;
+         && this_version >= required_version;
    }
 
    bool check_version(unsigned required_glsl_version,
@@ -230,6 +232,7 @@ struct _mesa_glsl_parse_state {
 
    bool es_shader;
    unsigned language_version;
+   unsigned forced_language_version;
    gl_shader_stage stage;
 
    /**
diff --git a/mesalib/src/glsl/glsl_types.cpp b/mesalib/src/glsl/glsl_types.cpp
index 38b37a6a9..4aa36a794 100644
--- a/mesalib/src/glsl/glsl_types.cpp
+++ b/mesalib/src/glsl/glsl_types.cpp
@@ -825,6 +825,73 @@ glsl_type::get_interface_instance(const glsl_struct_field *fields,
 
 
 const glsl_type *
+glsl_type::get_mul_type(const glsl_type *type_a, const glsl_type *type_b)
+{
+   if (type_a == type_b) {
+      return type_a;
+   } else if (type_a->is_matrix() && type_b->is_matrix()) {
+      /* Matrix multiply.  The columns of A must match the rows of B.  Given
+       * the other previously tested constraints, this means the vector type
+       * of a row from A must be the same as the vector type of a column from
+       * B.
+       */
+      if (type_a->row_type() == type_b->column_type()) {
+         /* The resulting matrix has the number of columns of matrix B and
+          * the number of rows of matrix A.  We get the row count of A by
+          * looking at the size of a vector that makes up a column.  The
+          * transpose (size of a row) is done for B.
+          */
+         const glsl_type *const type =
+            get_instance(type_a->base_type,
+                         type_a->column_type()->vector_elements,
+                         type_b->row_type()->vector_elements);
+         assert(type != error_type);
+
+         return type;
+      }
+   } else if (type_a->is_matrix()) {
+      /* A is a matrix and B is a column vector.  Columns of A must match
+       * rows of B.  Given the other previously tested constraints, this
+       * means the vector type of a row from A must be the same as the
+       * vector the type of B.
+       */
+      if (type_a->row_type() == type_b) {
+         /* The resulting vector has a number of elements equal to
+          * the number of rows of matrix A. */
+         const glsl_type *const type =
+            get_instance(type_a->base_type,
+                         type_a->column_type()->vector_elements,
+                         1);
+         assert(type != error_type);
+
+         return type;
+      }
+   } else {
+      assert(type_b->is_matrix());
+
+      /* A is a row vector and B is a matrix.  Columns of A must match rows
+       * of B.  Given the other previously tested constraints, this means
+       * the type of A must be the same as the vector type of a column from
+       * B.
+       */
+      if (type_a == type_b->column_type()) {
+         /* The resulting vector has a number of elements equal to
+          * the number of columns of matrix B. */
+         const glsl_type *const type =
+            get_instance(type_a->base_type,
+                         type_b->row_type()->vector_elements,
+                         1);
+         assert(type != error_type);
+
+         return type;
+      }
+   }
+
+   return error_type;
+}
+
+
+const glsl_type *
 glsl_type::field_type(const char *name) const
 {
    if (this->base_type != GLSL_TYPE_STRUCT
@@ -1077,15 +1144,6 @@ glsl_type::std140_base_alignment(bool row_major) const
       return base_alignment;
    }
 
-   /* A sampler may never occur in a UBO (without bindless of some sort),
-    * however it is convenient to use this alignment function even with
-    * regular uniforms. This allows use of this function on uniform structs
-    * that contain samplers.
-    */
-   if (this->is_sampler()) {
-      return 0;
-   }
-
    assert(!"not reached");
    return -1;
 }
diff --git a/mesalib/src/glsl/glsl_types.h b/mesalib/src/glsl/glsl_types.h
index 7359e9476..d383dd5be 100644
--- a/mesalib/src/glsl/glsl_types.h
+++ b/mesalib/src/glsl/glsl_types.h
@@ -276,6 +276,12 @@ struct glsl_type {
 						  const char *block_name);
 
    /**
+    * Get the type resulting from a multiplication of \p type_a * \p type_b
+    */
+   static const glsl_type *get_mul_type(const glsl_type *type_a,
+                                        const glsl_type *type_b);
+
+   /**
     * Query the total number of scalars that make up a scalar, vector or matrix
     */
    unsigned components() const
diff --git a/mesalib/src/glsl/ir.cpp b/mesalib/src/glsl/ir.cpp
index 516a53499..68c37275c 100755
--- a/mesalib/src/glsl/ir.cpp
+++ b/mesalib/src/glsl/ir.cpp
@@ -240,8 +240,6 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_round_even:
    case ir_unop_sin:
    case ir_unop_cos:
-   case ir_unop_sin_reduced:
-   case ir_unop_cos_reduced:
    case ir_unop_dFdx:
    case ir_unop_dFdx_coarse:
    case ir_unop_dFdx_fine:
@@ -380,10 +378,12 @@ ir_expression::ir_expression(int op, ir_rvalue *op0, ir_rvalue *op1)
       } else if (op1->type->is_scalar()) {
 	 this->type = op0->type;
       } else {
-	 /* FINISHME: matrix types */
-	 assert(!op0->type->is_matrix() && !op1->type->is_matrix());
-	 assert(op0->type == op1->type);
-	 this->type = op0->type;
+         if (this->operation == ir_binop_mul) {
+            this->type = glsl_type::get_mul_type(op0->type, op1->type);
+         } else {
+            assert(op0->type == op1->type);
+            this->type = op0->type;
+         }
       }
       break;
 
@@ -540,8 +540,6 @@ static const char *const operator_strs[] = {
    "round_even",
    "sin",
    "cos",
-   "sin_reduced",
-   "cos_reduced",
    "dFdx",
    "dFdxCoarse",
    "dFdxFine",
diff --git a/mesalib/src/glsl/ir.h b/mesalib/src/glsl/ir.h
index fdc22edf1..fab1cd2d2 100644
--- a/mesalib/src/glsl/ir.h
+++ b/mesalib/src/glsl/ir.h
@@ -109,6 +109,31 @@ public:
    virtual ir_instruction *clone(void *mem_ctx,
 				 struct hash_table *ht) const = 0;
 
+   bool is_rvalue() const
+   {
+      return ir_type == ir_type_dereference_array ||
+             ir_type == ir_type_dereference_record ||
+             ir_type == ir_type_dereference_variable ||
+             ir_type == ir_type_constant ||
+             ir_type == ir_type_expression ||
+             ir_type == ir_type_swizzle ||
+             ir_type == ir_type_texture;
+   }
+
+   bool is_dereference() const
+   {
+      return ir_type == ir_type_dereference_array ||
+             ir_type == ir_type_dereference_record ||
+             ir_type == ir_type_dereference_variable;
+   }
+
+   bool is_jump() const
+   {
+      return ir_type == ir_type_loop_jump ||
+             ir_type == ir_type_return ||
+             ir_type == ir_type_discard;
+   }
+
    /**
     * \name IR instruction downcast functions
     *
@@ -117,45 +142,33 @@ public:
     * Additional downcast functions will be added as needed.
     */
    /*@{*/
-   class ir_rvalue *as_rvalue()
-   {
-      assume(this != NULL);
-      if (ir_type == ir_type_dereference_array ||
-          ir_type == ir_type_dereference_record ||
-          ir_type == ir_type_dereference_variable ||
-          ir_type == ir_type_constant ||
-          ir_type == ir_type_expression ||
-          ir_type == ir_type_swizzle ||
-          ir_type == ir_type_texture)
-         return (class ir_rvalue *) this;
-      return NULL;
-   }
-
-   class ir_dereference *as_dereference()
-   {
-      assume(this != NULL);
-      if (ir_type == ir_type_dereference_array ||
-          ir_type == ir_type_dereference_record ||
-          ir_type == ir_type_dereference_variable)
-         return (class ir_dereference *) this;
-      return NULL;
-   }
-
-   class ir_jump *as_jump()
-   {
-      assume(this != NULL);
-      if (ir_type == ir_type_loop_jump ||
-          ir_type == ir_type_return ||
-          ir_type == ir_type_discard)
-         return (class ir_jump *) this;
-      return NULL;
-   }
+   #define AS_BASE(TYPE)                                \
+   class ir_##TYPE *as_##TYPE()                         \
+   {                                                    \
+      assume(this != NULL);                             \
+      return is_##TYPE() ? (ir_##TYPE *) this : NULL;   \
+   }                                                    \
+   const class ir_##TYPE *as_##TYPE() const             \
+   {                                                    \
+      assume(this != NULL);                             \
+      return is_##TYPE() ? (ir_##TYPE *) this : NULL;   \
+   }
+
+   AS_BASE(rvalue)
+   AS_BASE(dereference)
+   AS_BASE(jump)
+   #undef AS_BASE
 
    #define AS_CHILD(TYPE) \
    class ir_##TYPE * as_##TYPE() \
    { \
       assume(this != NULL);                                         \
       return ir_type == ir_type_##TYPE ? (ir_##TYPE *) this : NULL; \
+   }                                                                      \
+   const class ir_##TYPE * as_##TYPE() const                              \
+   {                                                                      \
+      assume(this != NULL);                                               \
+      return ir_type == ir_type_##TYPE ? (const ir_##TYPE *) this : NULL; \
    }
    AS_CHILD(variable)
    AS_CHILD(function)
@@ -183,7 +196,8 @@ public:
     * in particular.  No support for other instruction types (assignments,
     * jumps, calls, etc.) is planned.
     */
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
 protected:
    ir_instruction(enum ir_node_type t)
@@ -1300,8 +1314,6 @@ enum ir_expression_operation {
    /*@{*/
    ir_unop_sin,
    ir_unop_cos,
-   ir_unop_sin_reduced,    /**< Reduced range sin. [-pi, pi] */
-   ir_unop_cos_reduced,    /**< Reduced range cos. [-pi, pi] */
    /*@}*/
 
    /**
@@ -1598,7 +1610,8 @@ public:
     */
    ir_expression(int op, ir_rvalue *op0, ir_rvalue *op1, ir_rvalue *op2);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    virtual ir_expression *clone(void *mem_ctx, struct hash_table *ht) const;
 
@@ -1909,7 +1922,8 @@ public:
 
    virtual ir_visitor_status accept(ir_hierarchical_visitor *);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    /**
     * Return a string representing the ir_texture_opcode.
@@ -2010,7 +2024,8 @@ public:
 
    virtual ir_visitor_status accept(ir_hierarchical_visitor *);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    bool is_lvalue() const
    {
@@ -2063,7 +2078,8 @@ public:
 
    virtual ir_constant *constant_expression_value(struct hash_table *variable_context = NULL);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    /**
     * Get the variable that is ultimately referenced by an r-value
@@ -2109,7 +2125,8 @@ public:
 
    virtual ir_constant *constant_expression_value(struct hash_table *variable_context = NULL);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    /**
     * Get the variable that is ultimately referenced by an r-value
@@ -2219,7 +2236,8 @@ public:
 
    virtual ir_visitor_status accept(ir_hierarchical_visitor *);
 
-   virtual bool equals(ir_instruction *ir, enum ir_node_type ignore = ir_type_unset);
+   virtual bool equals(const ir_instruction *ir,
+                       enum ir_node_type ignore = ir_type_unset) const;
 
    /**
     * Get a particular component of a constant as a specific type
diff --git a/mesalib/src/glsl/ir_constant_expression.cpp b/mesalib/src/glsl/ir_constant_expression.cpp
index d198276c6..bbf2b0ca8 100755
--- a/mesalib/src/glsl/ir_constant_expression.cpp
+++ b/mesalib/src/glsl/ir_constant_expression.cpp
@@ -774,7 +774,6 @@ ir_expression::constant_expression_value(struct hash_table *variable_context)
       break;
 
    case ir_unop_sin:
-   case ir_unop_sin_reduced:
       assert(op[0]->type->base_type == GLSL_TYPE_FLOAT);
       for (unsigned c = 0; c < op[0]->type->components(); c++) {
 	 data.f[c] = sinf(op[0]->value.f[c]);
@@ -782,7 +781,6 @@ ir_expression::constant_expression_value(struct hash_table *variable_context)
       break;
 
    case ir_unop_cos:
-   case ir_unop_cos_reduced:
       assert(op[0]->type->base_type == GLSL_TYPE_FLOAT);
       for (unsigned c = 0; c < op[0]->type->components(); c++) {
 	 data.f[c] = cosf(op[0]->value.f[c]);
diff --git a/mesalib/src/glsl/ir_equals.cpp b/mesalib/src/glsl/ir_equals.cpp
index 65376cd94..cc1964eef 100644
--- a/mesalib/src/glsl/ir_equals.cpp
+++ b/mesalib/src/glsl/ir_equals.cpp
@@ -28,7 +28,8 @@
  * can't access a's vtable in that case.
  */
 static bool
-possibly_null_equals(ir_instruction *a, ir_instruction *b, enum ir_node_type ignore)
+possibly_null_equals(const ir_instruction *a, const ir_instruction *b,
+                     enum ir_node_type ignore)
 {
    if (!a || !b)
       return !a && !b;
@@ -41,13 +42,13 @@ possibly_null_equals(ir_instruction *a, ir_instruction *b, enum ir_node_type ign
  * about.
  */
 bool
-ir_instruction::equals(ir_instruction *, enum ir_node_type)
+ir_instruction::equals(const ir_instruction *, enum ir_node_type) const
 {
    return false;
 }
 
 bool
-ir_constant::equals(ir_instruction *ir, enum ir_node_type)
+ir_constant::equals(const ir_instruction *ir, enum ir_node_type) const
 {
    const ir_constant *other = ir->as_constant();
    if (!other)
@@ -65,7 +66,8 @@ ir_constant::equals(ir_instruction *ir, enum ir_node_type)
 }
 
 bool
-ir_dereference_variable::equals(ir_instruction *ir, enum ir_node_type)
+ir_dereference_variable::equals(const ir_instruction *ir,
+                                enum ir_node_type) const
 {
    const ir_dereference_variable *other = ir->as_dereference_variable();
    if (!other)
@@ -75,7 +77,8 @@ ir_dereference_variable::equals(ir_instruction *ir, enum ir_node_type)
 }
 
 bool
-ir_dereference_array::equals(ir_instruction *ir, enum ir_node_type ignore)
+ir_dereference_array::equals(const ir_instruction *ir,
+                             enum ir_node_type ignore) const
 {
    const ir_dereference_array *other = ir->as_dereference_array();
    if (!other)
@@ -94,7 +97,8 @@ ir_dereference_array::equals(ir_instruction *ir, enum ir_node_type ignore)
 }
 
 bool
-ir_swizzle::equals(ir_instruction *ir, enum ir_node_type ignore)
+ir_swizzle::equals(const ir_instruction *ir,
+                   enum ir_node_type ignore) const
 {
    const ir_swizzle *other = ir->as_swizzle();
    if (!other)
@@ -116,7 +120,7 @@ ir_swizzle::equals(ir_instruction *ir, enum ir_node_type ignore)
 }
 
 bool
-ir_texture::equals(ir_instruction *ir, enum ir_node_type ignore)
+ir_texture::equals(const ir_instruction *ir, enum ir_node_type ignore) const
 {
    const ir_texture *other = ir->as_texture();
    if (!other)
@@ -179,7 +183,7 @@ ir_texture::equals(ir_instruction *ir, enum ir_node_type ignore)
 }
 
 bool
-ir_expression::equals(ir_instruction *ir, enum ir_node_type ignore)
+ir_expression::equals(const ir_instruction *ir, enum ir_node_type ignore) const
 {
    const ir_expression *other = ir->as_expression();
    if (!other)
diff --git a/mesalib/src/glsl/ir_validate.cpp b/mesalib/src/glsl/ir_validate.cpp
index 7a7688cb2..cfe0df3dc 100644
--- a/mesalib/src/glsl/ir_validate.cpp
+++ b/mesalib/src/glsl/ir_validate.cpp
@@ -334,8 +334,6 @@ ir_validate::visit_leave(ir_expression *ir)
       break;
    case ir_unop_sin:
    case ir_unop_cos:
-   case ir_unop_sin_reduced:
-   case ir_unop_cos_reduced:
    case ir_unop_dFdx:
    case ir_unop_dFdx_coarse:
    case ir_unop_dFdx_fine:
@@ -543,9 +541,9 @@ ir_validate::visit_leave(ir_expression *ir)
    case ir_binop_logic_and:
    case ir_binop_logic_xor:
    case ir_binop_logic_or:
-      assert(ir->type == glsl_type::bool_type);
-      assert(ir->operands[0]->type == glsl_type::bool_type);
-      assert(ir->operands[1]->type == glsl_type::bool_type);
+      assert(ir->type->base_type == GLSL_TYPE_BOOL);
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_BOOL);
+      assert(ir->operands[1]->type->base_type == GLSL_TYPE_BOOL);
       break;
 
    case ir_binop_dot:
diff --git a/mesalib/src/glsl/link_uniforms.cpp b/mesalib/src/glsl/link_uniforms.cpp
index 972ba8100..51b939098 100644
--- a/mesalib/src/glsl/link_uniforms.cpp
+++ b/mesalib/src/glsl/link_uniforms.cpp
@@ -547,6 +547,8 @@ private:
    virtual void enter_record(const glsl_type *type, const char *name,
                              bool row_major) {
       assert(type->is_record());
+      if (this->ubo_block_index == -1)
+         return;
       this->ubo_byte_offset = glsl_align(
             this->ubo_byte_offset, type->std140_base_alignment(row_major));
    }
@@ -554,6 +556,8 @@ private:
    virtual void leave_record(const glsl_type *type, const char *name,
                              bool row_major) {
       assert(type->is_record());
+      if (this->ubo_block_index == -1)
+         return;
       this->ubo_byte_offset = glsl_align(
             this->ubo_byte_offset, type->std140_base_alignment(row_major));
    }
diff --git a/mesalib/src/glsl/link_varyings.cpp b/mesalib/src/glsl/link_varyings.cpp
index 22617990f..605748a9c 100644
--- a/mesalib/src/glsl/link_varyings.cpp
+++ b/mesalib/src/glsl/link_varyings.cpp
@@ -263,6 +263,19 @@ cross_validate_outputs_to_inputs(struct gl_shader_program *prog,
          if (output != NULL) {
             cross_validate_types_and_qualifiers(prog, input, output,
                                                 consumer->Stage, producer->Stage);
+         } else {
+            /* Check for input vars with unmatched output vars in prev stage
+             * taking into account that interface blocks could have a matching
+             * output but with different name, so we ignore them.
+             */
+            assert(!input->data.assigned);
+            if (input->data.used && !input->get_interface_type() &&
+                !input->data.explicit_location && !prog->SeparateShader)
+               linker_error(prog,
+                            "%s shader input `%s' "
+                            "has no matching output in the previous stage\n",
+                            _mesa_shader_stage_to_string(consumer->Stage),
+                            input->name);
          }
       }
    }
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index 4349f0973..b6baa5d36 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -1377,24 +1377,13 @@ link_fs_input_layout_qualifiers(struct gl_shader_program *prog,
        *   "If gl_FragCoord is redeclared in any fragment shader in a program,
        *    it must be redeclared in all the fragment shaders in that program
        *    that have a static use gl_FragCoord."
-       *
-       * Exclude the case when one of the 'linked_shader' or 'shader' redeclares
-       * gl_FragCoord with no layout qualifiers but the other one doesn't
-       * redeclare it. If we strictly follow GLSL 1.50 spec's language, it
-       * should be a link error. But, generating link error for this case will
-       * be a wrong behaviour which spec didn't intend to do and it could also
-       * break some applications.
        */
       if ((linked_shader->redeclares_gl_fragcoord
            && !shader->redeclares_gl_fragcoord
-           && shader->uses_gl_fragcoord
-           && (linked_shader->origin_upper_left
-               || linked_shader->pixel_center_integer))
+           && shader->uses_gl_fragcoord)
           || (shader->redeclares_gl_fragcoord
               && !linked_shader->redeclares_gl_fragcoord
-              && linked_shader->uses_gl_fragcoord
-              && (shader->origin_upper_left
-                  || shader->pixel_center_integer))) {
+              && linked_shader->uses_gl_fragcoord)) {
              linker_error(prog, "fragment shader defined with conflicting "
                          "layout qualifiers for gl_FragCoord\n");
       }
@@ -2503,6 +2492,194 @@ check_explicit_uniform_locations(struct gl_context *ctx,
    delete uniform_map;
 }
 
+static bool
+add_program_resource(struct gl_shader_program *prog, GLenum type,
+                     const void *data, uint8_t stages)
+{
+   assert(data);
+
+   /* If resource already exists, do not add it again. */
+   for (unsigned i = 0; i < prog->NumProgramResourceList; i++)
+      if (prog->ProgramResourceList[i].Data == data)
+         return true;
+
+   prog->ProgramResourceList =
+      reralloc(prog,
+               prog->ProgramResourceList,
+               gl_program_resource,
+               prog->NumProgramResourceList + 1);
+
+   if (!prog->ProgramResourceList) {
+      linker_error(prog, "Out of memory during linking.\n");
+      return false;
+   }
+
+   struct gl_program_resource *res =
+      &prog->ProgramResourceList[prog->NumProgramResourceList];
+
+   res->Type = type;
+   res->Data = data;
+   res->StageReferences = stages;
+
+   prog->NumProgramResourceList++;
+
+   return true;
+}
+
+/**
+ * Function builds a stage reference bitmask from variable name.
+ */
+static uint8_t
+build_stageref(struct gl_shader_program *shProg, const char *name)
+{
+   uint8_t stages = 0;
+
+   /* Note, that we assume MAX 8 stages, if there will be more stages, type
+    * used for reference mask in gl_program_resource will need to be changed.
+    */
+   assert(MESA_SHADER_STAGES < 8);
+
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      struct gl_shader *sh = shProg->_LinkedShaders[i];
+      if (!sh)
+         continue;
+      ir_variable *var = sh->symbols->get_variable(name);
+      if (var)
+         stages |= (1 << i);
+   }
+   return stages;
+}
+
+static bool
+add_interface_variables(struct gl_shader_program *shProg,
+                        struct gl_shader *sh, GLenum programInterface)
+{
+   foreach_in_list(ir_instruction, node, sh->ir) {
+      ir_variable *var = node->as_variable();
+
+      if (!var)
+         continue;
+
+      switch (var->data.mode) {
+      /* From GL 4.3 core spec, section 11.1.1 (Vertex Attributes):
+       * "For GetActiveAttrib, all active vertex shader input variables
+       * are enumerated, including the special built-in inputs gl_VertexID
+       * and gl_InstanceID."
+       */
+      case ir_var_system_value:
+         if (var->data.location != SYSTEM_VALUE_VERTEX_ID &&
+             var->data.location != SYSTEM_VALUE_VERTEX_ID_ZERO_BASE &&
+             var->data.location != SYSTEM_VALUE_INSTANCE_ID)
+         continue;
+      case ir_var_shader_in:
+         if (programInterface != GL_PROGRAM_INPUT)
+            continue;
+         break;
+      case ir_var_shader_out:
+         if (programInterface != GL_PROGRAM_OUTPUT)
+            continue;
+         break;
+      default:
+         continue;
+      };
+
+      if (!add_program_resource(shProg, programInterface, var,
+                                build_stageref(shProg, var->name)))
+         return false;
+   }
+   return true;
+}
+
+/**
+ * Builds up a list of program resources that point to existing
+ * resource data.
+ */
+static void
+build_program_resource_list(struct gl_context *ctx,
+                            struct gl_shader_program *shProg)
+{
+   /* Rebuild resource list. */
+   if (shProg->ProgramResourceList) {
+      ralloc_free(shProg->ProgramResourceList);
+      shProg->ProgramResourceList = NULL;
+      shProg->NumProgramResourceList = 0;
+   }
+
+   int input_stage = MESA_SHADER_STAGES, output_stage = 0;
+
+   /* Determine first input and final output stage. These are used to
+    * detect which variables should be enumerated in the resource list
+    * for GL_PROGRAM_INPUT and GL_PROGRAM_OUTPUT.
+    */
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      if (!shProg->_LinkedShaders[i])
+         continue;
+      if (input_stage == MESA_SHADER_STAGES)
+         input_stage = i;
+      output_stage = i;
+   }
+
+   /* Empty shader, no resources. */
+   if (input_stage == MESA_SHADER_STAGES && output_stage == 0)
+      return;
+
+   /* Add inputs and outputs to the resource list. */
+   if (!add_interface_variables(shProg, shProg->_LinkedShaders[input_stage],
+                                GL_PROGRAM_INPUT))
+      return;
+
+   if (!add_interface_variables(shProg, shProg->_LinkedShaders[output_stage],
+                                GL_PROGRAM_OUTPUT))
+      return;
+
+   /* Add transform feedback varyings. */
+   if (shProg->LinkedTransformFeedback.NumVarying > 0) {
+      for (int i = 0; i < shProg->LinkedTransformFeedback.NumVarying; i++) {
+         uint8_t stageref =
+            build_stageref(shProg,
+                           shProg->LinkedTransformFeedback.Varyings[i].Name);
+         if (!add_program_resource(shProg, GL_TRANSFORM_FEEDBACK_VARYING,
+                                   &shProg->LinkedTransformFeedback.Varyings[i],
+                                   stageref))
+         return;
+      }
+   }
+
+   /* Add uniforms from uniform storage. */
+   for (unsigned i = 0; i < shProg->NumUserUniformStorage; i++) {
+      /* Do not add uniforms internally used by Mesa. */
+      if (shProg->UniformStorage[i].hidden)
+         continue;
+
+      uint8_t stageref =
+         build_stageref(shProg, shProg->UniformStorage[i].name);
+      if (!add_program_resource(shProg, GL_UNIFORM,
+                                &shProg->UniformStorage[i], stageref))
+         return;
+   }
+
+   /* Add program uniform blocks. */
+   for (unsigned i = 0; i < shProg->NumUniformBlocks; i++) {
+      if (!add_program_resource(shProg, GL_UNIFORM_BLOCK,
+          &shProg->UniformBlocks[i], 0))
+         return;
+   }
+
+   /* Add atomic counter buffers. */
+   for (unsigned i = 0; i < shProg->NumAtomicBuffers; i++) {
+      if (!add_program_resource(shProg, GL_ATOMIC_COUNTER_BUFFER,
+                                &shProg->AtomicBuffers[i], 0))
+         return;
+   }
+
+   /* TODO - following extensions will require more resource types:
+    *
+    *    GL_ARB_shader_storage_buffer_object
+    *    GL_ARB_shader_subroutine
+    */
+}
+
+
 void
 link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
 {
@@ -2737,10 +2914,18 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
       goto done;
    }
 
-   unsigned first;
-   for (first = 0; first <= MESA_SHADER_FRAGMENT; first++) {
-      if (prog->_LinkedShaders[first] != NULL)
-	 break;
+   unsigned first, last;
+
+   first = MESA_SHADER_STAGES;
+   last = 0;
+
+   /* Determine first and last stage. */
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      if (!prog->_LinkedShaders[i])
+         continue;
+      if (first == MESA_SHADER_STAGES)
+         first = i;
+      last = i;
    }
 
    if (num_tfeedback_decls != 0) {
@@ -2769,13 +2954,9 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
     * ensures that inter-shader outputs written to in an earlier stage are
     * eliminated if they are (transitively) not used in a later stage.
     */
-   int last, next;
-   for (last = MESA_SHADER_FRAGMENT; last >= 0; last--) {
-      if (prog->_LinkedShaders[last] != NULL)
-         break;
-   }
+   int next;
 
-   if (last >= 0 && last < MESA_SHADER_FRAGMENT) {
+   if (first < MESA_SHADER_FRAGMENT) {
       gl_shader *const sh = prog->_LinkedShaders[last];
 
       if (first == MESA_SHADER_GEOMETRY) {
@@ -2787,13 +2968,14 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
           * MESA_SHADER_GEOMETRY.
           */
          if (!assign_varying_locations(ctx, mem_ctx, prog,
-                                       NULL, sh,
+                                       NULL, prog->_LinkedShaders[first],
                                        num_tfeedback_decls, tfeedback_decls,
                                        prog->Geom.VerticesIn))
             goto done;
       }
 
-      if (num_tfeedback_decls != 0 || prog->SeparateShader) {
+      if (last != MESA_SHADER_FRAGMENT &&
+         (num_tfeedback_decls != 0 || prog->SeparateShader)) {
          /* There was no fragment shader, but we still have to assign varying
           * locations for use by transform feedback.
           */
@@ -2905,6 +3087,10 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
       }
    }
 
+   build_program_resource_list(ctx, prog);
+   if (!prog->LinkStatus)
+      goto done;
+
    /* FINISHME: Assign fragment shader output locations. */
 
 done:
diff --git a/mesalib/src/glsl/loop_controls.cpp b/mesalib/src/glsl/loop_controls.cpp
index d7f0b2809..51804bb5f 100644
--- a/mesalib/src/glsl/loop_controls.cpp
+++ b/mesalib/src/glsl/loop_controls.cpp
@@ -139,7 +139,7 @@ calculate_iterations(ir_rvalue *from, ir_rvalue *to, ir_rvalue *increment,
          iter = new(mem_ctx) ir_constant(double(iter_value + bias[i]));
          break;
       default:
-          unreachable(!"Unsupported type for loop iterator.");
+          unreachable("Unsupported type for loop iterator.");
       }
 
       ir_expression *const mul =
diff --git a/mesalib/src/glsl/lower_const_arrays_to_uniforms.cpp b/mesalib/src/glsl/lower_const_arrays_to_uniforms.cpp
index 2243f479a..44967dcdb 100644
--- a/mesalib/src/glsl/lower_const_arrays_to_uniforms.cpp
+++ b/mesalib/src/glsl/lower_const_arrays_to_uniforms.cpp
@@ -49,7 +49,6 @@ public:
    {
       instructions = insts;
       progress = false;
-      index = 0;
    }
 
    bool run()
@@ -63,7 +62,6 @@ public:
 private:
    exec_list *instructions;
    bool progress;
-   unsigned index;
 };
 
 void
@@ -82,7 +80,7 @@ lower_const_array_visitor::handle_rvalue(ir_rvalue **rvalue)
 
    void *mem_ctx = ralloc_parent(con);
 
-   char *uniform_name = ralloc_asprintf(mem_ctx, "constarray__%d", index++);
+   char *uniform_name = ralloc_asprintf(mem_ctx, "constarray__%p", dra);
 
    ir_variable *uni =
       new(mem_ctx) ir_variable(con->type, uniform_name, ir_var_uniform);
diff --git a/mesalib/src/glsl/nir/glsl_to_nir.cpp b/mesalib/src/glsl/nir/glsl_to_nir.cpp
index 357944da6..f6b8331d4 100644
--- a/mesalib/src/glsl/nir/glsl_to_nir.cpp
+++ b/mesalib/src/glsl/nir/glsl_to_nir.cpp
@@ -88,6 +88,8 @@ private:
    exec_list *cf_node_list;
    nir_instr *result; /* result of the expression tree last visited */
 
+   nir_deref_var *evaluate_deref(nir_instr *mem_ctx, ir_instruction *ir);
+
    /* the head of the dereference chain we're creating */
    nir_deref_var *deref_head;
    /* the tail of the dereference chain we're creating */
@@ -156,6 +158,14 @@ nir_visitor::~nir_visitor()
    _mesa_hash_table_destroy(this->overload_table, NULL);
 }
 
+nir_deref_var *
+nir_visitor::evaluate_deref(nir_instr *mem_ctx, ir_instruction *ir)
+{
+   ir->accept(this);
+   ralloc_steal(mem_ctx, this->deref_head);
+   return this->deref_head;
+}
+
 static nir_constant *
 constant_copy(ir_constant *ir, void *mem_ctx)
 {
@@ -582,13 +592,11 @@ void
 nir_visitor::visit(ir_return *ir)
 {
    if (ir->value != NULL) {
-      ir->value->accept(this);
       nir_intrinsic_instr *copy =
          nir_intrinsic_instr_create(this->shader, nir_intrinsic_copy_var);
 
-      copy->variables[0] = nir_deref_var_create(this->shader,
-                                                this->impl->return_var);
-      copy->variables[1] = this->deref_head;
+      copy->variables[0] = nir_deref_var_create(copy, this->impl->return_var);
+      copy->variables[1] = evaluate_deref(&copy->instr, ir->value);
    }
 
    nir_jump_instr *instr = nir_jump_instr_create(this->shader, nir_jump_return);
@@ -613,8 +621,7 @@ nir_visitor::visit(ir_call *ir)
       nir_intrinsic_instr *instr = nir_intrinsic_instr_create(shader, op);
       ir_dereference *param =
          (ir_dereference *) ir->actual_parameters.get_head();
-      param->accept(this);
-      instr->variables[0] = this->deref_head;
+      instr->variables[0] = evaluate_deref(&instr->instr, param);
       nir_ssa_dest_init(&instr->instr, &instr->dest, 1, NULL);
 
       nir_instr_insert_after_cf_list(this->cf_node_list, &instr->instr);
@@ -623,8 +630,7 @@ nir_visitor::visit(ir_call *ir)
          nir_intrinsic_instr_create(shader, nir_intrinsic_store_var);
       store_instr->num_components = 1;
 
-      ir->return_deref->accept(this);
-      store_instr->variables[0] = this->deref_head;
+      store_instr->variables[0] = evaluate_deref(&store_instr->instr, ir->return_deref);
       store_instr->src[0].is_ssa = true;
       store_instr->src[0].ssa = &instr->dest.ssa;
 
@@ -642,13 +648,11 @@ nir_visitor::visit(ir_call *ir)
 
    unsigned i = 0;
    foreach_in_list(ir_dereference, param, &ir->actual_parameters) {
-      param->accept(this);
-      instr->params[i] = this->deref_head;
+      instr->params[i] = evaluate_deref(&instr->instr, param);
       i++;
    }
 
-   ir->return_deref->accept(this);
-   instr->return_deref = this->deref_head;
+   instr->return_deref = evaluate_deref(&instr->instr, ir->return_deref);
    nir_instr_insert_after_cf_list(this->cf_node_list, &instr->instr);
 }
 
@@ -663,12 +667,8 @@ nir_visitor::visit(ir_assignment *ir)
       nir_intrinsic_instr *copy =
          nir_intrinsic_instr_create(this->shader, nir_intrinsic_copy_var);
 
-      ir->lhs->accept(this);
-      copy->variables[0] = this->deref_head;
-
-      ir->rhs->accept(this);
-      copy->variables[1] = this->deref_head;
-
+      copy->variables[0] = evaluate_deref(&copy->instr, ir->lhs);
+      copy->variables[1] = evaluate_deref(&copy->instr, ir->rhs);
 
       if (ir->condition) {
          nir_if *if_stmt = nir_if_create(this->shader);
@@ -700,6 +700,7 @@ nir_visitor::visit(ir_assignment *ir)
       load->num_components = ir->lhs->type->vector_elements;
       nir_ssa_dest_init(&load->instr, &load->dest, num_components, NULL);
       load->variables[0] = lhs_deref;
+      ralloc_steal(load, load->variables[0]);
       nir_instr_insert_after_cf_list(this->cf_node_list, &load->instr);
 
       nir_op vec_op;
@@ -741,7 +742,7 @@ nir_visitor::visit(ir_assignment *ir)
    nir_intrinsic_instr *store =
       nir_intrinsic_instr_create(this->shader, nir_intrinsic_store_var);
    store->num_components = ir->lhs->type->vector_elements;
-   nir_deref *store_deref = nir_copy_deref(this->shader, &lhs_deref->deref);
+   nir_deref *store_deref = nir_copy_deref(store, &lhs_deref->deref);
    store->variables[0] = nir_deref_as_var(store_deref);
    store->src[0] = src;
 
@@ -816,6 +817,7 @@ nir_visitor::evaluate_rvalue(ir_rvalue* ir)
          nir_intrinsic_instr_create(this->shader, nir_intrinsic_load_var);
       load_instr->num_components = ir->type->vector_elements;
       load_instr->variables[0] = this->deref_head;
+      ralloc_steal(load_instr, load_instr->variables[0]);
       add_instr(&load_instr->instr, ir->type->vector_elements);
    }
 
@@ -959,6 +961,7 @@ nir_visitor::visit(ir_expression *ir)
       nir_intrinsic_instr *intrin = nir_intrinsic_instr_create(shader, op);
       intrin->num_components = deref->type->vector_elements;
       intrin->variables[0] = this->deref_head;
+      ralloc_steal(intrin, intrin->variables[0]);
 
       if (intrin->intrinsic == nir_intrinsic_interp_var_at_offset ||
           intrin->intrinsic == nir_intrinsic_interp_var_at_sample)
@@ -1087,12 +1090,6 @@ nir_visitor::visit(ir_expression *ir)
    case ir_unop_round_even: emit(nir_op_fround_even, dest_size, srcs); break;
    case ir_unop_sin:   emit(nir_op_fsin,   dest_size, srcs); break;
    case ir_unop_cos:   emit(nir_op_fcos,   dest_size, srcs); break;
-   case ir_unop_sin_reduced:
-      emit(nir_op_fsin_reduced, dest_size, srcs);
-      break;
-   case ir_unop_cos_reduced:
-      emit(nir_op_fcos_reduced, dest_size, srcs);
-      break;
    case ir_unop_dFdx:        emit(nir_op_fddx,        dest_size, srcs); break;
    case ir_unop_dFdy:        emit(nir_op_fddy,        dest_size, srcs); break;
    case ir_unop_dFdx_fine:   emit(nir_op_fddx_fine,   dest_size, srcs); break;
@@ -1210,6 +1207,9 @@ nir_visitor::visit(ir_expression *ir)
    case ir_binop_bit_and:
    case ir_binop_bit_or:
    case ir_binop_bit_xor:
+   case ir_binop_logic_and:
+   case ir_binop_logic_or:
+   case ir_binop_logic_xor:
    case ir_binop_lshift:
    case ir_binop_rshift:
       switch (ir->operation) {
@@ -1270,6 +1270,24 @@ nir_visitor::visit(ir_expression *ir)
       case ir_binop_bit_xor:
          op = nir_op_ixor;
          break;
+      case ir_binop_logic_and:
+         if (supports_ints)
+            op = nir_op_iand;
+         else
+            op = nir_op_fand;
+         break;
+      case ir_binop_logic_or:
+         if (supports_ints)
+            op = nir_op_ior;
+         else
+            op = nir_op_for;
+         break;
+      case ir_binop_logic_xor:
+         if (supports_ints)
+            op = nir_op_ixor;
+         else
+            op = nir_op_fxor;
+         break;
       case ir_binop_lshift:
          op = nir_op_ishl;
          break;
@@ -1444,24 +1462,6 @@ nir_visitor::visit(ir_expression *ir)
          }
       }
       break;
-   case ir_binop_logic_and:
-      if (supports_ints)
-         emit(nir_op_iand, dest_size, srcs);
-      else
-         emit(nir_op_fand, dest_size, srcs);
-      break;
-   case ir_binop_logic_or:
-      if (supports_ints)
-         emit(nir_op_ior, dest_size, srcs);
-      else
-         emit(nir_op_for, dest_size, srcs);
-      break;
-   case ir_binop_logic_xor:
-      if (supports_ints)
-         emit(nir_op_ixor, dest_size, srcs);
-      else
-         emit(nir_op_fxor, dest_size, srcs);
-      break;
    case ir_binop_dot:
       switch (ir->operands[0]->type->vector_elements) {
          case 2: emit(nir_op_fdot2, dest_size, srcs); break;
@@ -1633,8 +1633,7 @@ nir_visitor::visit(ir_texture *ir)
       unreachable("not reached");
    }
 
-   ir->sampler->accept(this);
-   instr->sampler = this->deref_head;
+   instr->sampler = evaluate_deref(&instr->instr, ir->sampler);
 
    unsigned src_number = 0;
 
@@ -1759,7 +1758,7 @@ nir_visitor::visit(ir_dereference_record *ir)
    int field_index = this->deref_tail->type->field_index(ir->field);
    assert(field_index >= 0);
 
-   nir_deref_struct *deref = nir_deref_struct_create(this->shader, field_index);
+   nir_deref_struct *deref = nir_deref_struct_create(this->deref_tail, field_index);
    deref->deref.type = ir->type;
    this->deref_tail->child = &deref->deref;
    this->deref_tail = &deref->deref;
@@ -1783,5 +1782,6 @@ nir_visitor::visit(ir_dereference_array *ir)
    ir->array->accept(this);
 
    this->deref_tail->child = &deref->deref;
+   ralloc_steal(this->deref_tail, deref);
    this->deref_tail = &deref->deref;
 }
diff --git a/mesalib/src/glsl/nir/nir.c b/mesalib/src/glsl/nir/nir.c
index 6459d5108..c6e53612b 100644
--- a/mesalib/src/glsl/nir/nir.c
+++ b/mesalib/src/glsl/nir/nir.c
@@ -58,11 +58,11 @@ reg_create(void *mem_ctx, struct exec_list *list)
    nir_register *reg = ralloc(mem_ctx, nir_register);
 
    reg->parent_instr = NULL;
-   reg->uses = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   reg->uses = _mesa_set_create(reg, _mesa_hash_pointer,
                                 _mesa_key_pointer_equal);
-   reg->defs = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   reg->defs = _mesa_set_create(reg, _mesa_hash_pointer,
                                 _mesa_key_pointer_equal);
-   reg->if_uses = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   reg->if_uses = _mesa_set_create(reg, _mesa_hash_pointer,
                                    _mesa_key_pointer_equal);
 
    reg->num_components = 0;
@@ -108,7 +108,7 @@ nir_function_create(nir_shader *shader, const char *name)
 
    exec_list_push_tail(&shader->functions, &func->node);
    exec_list_make_empty(&func->overload_list);
-   func->name = name;
+   func->name = ralloc_strdup(func, name);
    func->shader = shader;
 
    return func;
@@ -285,10 +285,10 @@ nir_block_create(void *mem_ctx)
    cf_init(&block->cf_node, nir_cf_node_block);
 
    block->successors[0] = block->successors[1] = NULL;
-   block->predecessors = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   block->predecessors = _mesa_set_create(block, _mesa_hash_pointer,
                                           _mesa_key_pointer_equal);
    block->imm_dom = NULL;
-   block->dom_frontier = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   block->dom_frontier = _mesa_set_create(block, _mesa_hash_pointer,
                                           _mesa_key_pointer_equal);
 
    exec_list_make_empty(&block->instr_list);
@@ -381,11 +381,11 @@ alu_src_init(nir_alu_src *src)
 }
 
 nir_alu_instr *
-nir_alu_instr_create(void *mem_ctx, nir_op op)
+nir_alu_instr_create(nir_shader *shader, nir_op op)
 {
    unsigned num_srcs = nir_op_infos[op].num_inputs;
    nir_alu_instr *instr =
-      ralloc_size(mem_ctx,
+      ralloc_size(shader,
                   sizeof(nir_alu_instr) + num_srcs * sizeof(nir_alu_src));
 
    instr_init(&instr->instr, nir_instr_type_alu);
@@ -398,18 +398,18 @@ nir_alu_instr_create(void *mem_ctx, nir_op op)
 }
 
 nir_jump_instr *
-nir_jump_instr_create(void *mem_ctx, nir_jump_type type)
+nir_jump_instr_create(nir_shader *shader, nir_jump_type type)
 {
-   nir_jump_instr *instr = ralloc(mem_ctx, nir_jump_instr);
+   nir_jump_instr *instr = ralloc(shader, nir_jump_instr);
    instr_init(&instr->instr, nir_instr_type_jump);
    instr->type = type;
    return instr;
 }
 
 nir_load_const_instr *
-nir_load_const_instr_create(void *mem_ctx, unsigned num_components)
+nir_load_const_instr_create(nir_shader *shader, unsigned num_components)
 {
-   nir_load_const_instr *instr = ralloc(mem_ctx, nir_load_const_instr);
+   nir_load_const_instr *instr = ralloc(shader, nir_load_const_instr);
    instr_init(&instr->instr, nir_instr_type_load_const);
 
    nir_ssa_def_init(&instr->instr, &instr->def, num_components, NULL);
@@ -418,11 +418,11 @@ nir_load_const_instr_create(void *mem_ctx, unsigned num_components)
 }
 
 nir_intrinsic_instr *
-nir_intrinsic_instr_create(void *mem_ctx, nir_intrinsic_op op)
+nir_intrinsic_instr_create(nir_shader *shader, nir_intrinsic_op op)
 {
    unsigned num_srcs = nir_intrinsic_infos[op].num_srcs;
    nir_intrinsic_instr *instr =
-      ralloc_size(mem_ctx,
+      ralloc_size(shader,
                   sizeof(nir_intrinsic_instr) + num_srcs * sizeof(nir_src));
 
    instr_init(&instr->instr, nir_instr_type_intrinsic);
@@ -438,29 +438,29 @@ nir_intrinsic_instr_create(void *mem_ctx, nir_intrinsic_op op)
 }
 
 nir_call_instr *
-nir_call_instr_create(void *mem_ctx, nir_function_overload *callee)
+nir_call_instr_create(nir_shader *shader, nir_function_overload *callee)
 {
-   nir_call_instr *instr = ralloc(mem_ctx, nir_call_instr);
+   nir_call_instr *instr = ralloc(shader, nir_call_instr);
    instr_init(&instr->instr, nir_instr_type_call);
 
    instr->callee = callee;
    instr->num_params = callee->num_params;
-   instr->params = ralloc_array(mem_ctx, nir_deref_var *, instr->num_params);
+   instr->params = ralloc_array(instr, nir_deref_var *, instr->num_params);
    instr->return_deref = NULL;
 
    return instr;
 }
 
 nir_tex_instr *
-nir_tex_instr_create(void *mem_ctx, unsigned num_srcs)
+nir_tex_instr_create(nir_shader *shader, unsigned num_srcs)
 {
-   nir_tex_instr *instr = ralloc(mem_ctx, nir_tex_instr);
+   nir_tex_instr *instr = ralloc(shader, nir_tex_instr);
    instr_init(&instr->instr, nir_instr_type_tex);
 
    dest_init(&instr->dest);
 
    instr->num_srcs = num_srcs;
-   instr->src = ralloc_array(mem_ctx, nir_tex_src, num_srcs);
+   instr->src = ralloc_array(instr, nir_tex_src, num_srcs);
    for (unsigned i = 0; i < num_srcs; i++)
       src_init(&instr->src[i].src);
 
@@ -472,9 +472,9 @@ nir_tex_instr_create(void *mem_ctx, unsigned num_srcs)
 }
 
 nir_phi_instr *
-nir_phi_instr_create(void *mem_ctx)
+nir_phi_instr_create(nir_shader *shader)
 {
-   nir_phi_instr *instr = ralloc(mem_ctx, nir_phi_instr);
+   nir_phi_instr *instr = ralloc(shader, nir_phi_instr);
    instr_init(&instr->instr, nir_instr_type_phi);
 
    dest_init(&instr->dest);
@@ -483,9 +483,9 @@ nir_phi_instr_create(void *mem_ctx)
 }
 
 nir_parallel_copy_instr *
-nir_parallel_copy_instr_create(void *mem_ctx)
+nir_parallel_copy_instr_create(nir_shader *shader)
 {
-   nir_parallel_copy_instr *instr = ralloc(mem_ctx, nir_parallel_copy_instr);
+   nir_parallel_copy_instr *instr = ralloc(shader, nir_parallel_copy_instr);
    instr_init(&instr->instr, nir_instr_type_parallel_copy);
 
    exec_list_make_empty(&instr->entries);
@@ -494,9 +494,9 @@ nir_parallel_copy_instr_create(void *mem_ctx)
 }
 
 nir_ssa_undef_instr *
-nir_ssa_undef_instr_create(void *mem_ctx, unsigned num_components)
+nir_ssa_undef_instr_create(nir_shader *shader, unsigned num_components)
 {
-   nir_ssa_undef_instr *instr = ralloc(mem_ctx, nir_ssa_undef_instr);
+   nir_ssa_undef_instr *instr = ralloc(shader, nir_ssa_undef_instr);
    instr_init(&instr->instr, nir_instr_type_ssa_undef);
 
    nir_ssa_def_init(&instr->instr, &instr->def, num_components, NULL);
@@ -543,7 +543,7 @@ copy_deref_var(void *mem_ctx, nir_deref_var *deref)
    nir_deref_var *ret = nir_deref_var_create(mem_ctx, deref->var);
    ret->deref.type = deref->deref.type;
    if (deref->deref.child)
-      ret->deref.child = nir_copy_deref(mem_ctx, deref->deref.child);
+      ret->deref.child = nir_copy_deref(ret, deref->deref.child);
    return ret;
 }
 
@@ -558,7 +558,7 @@ copy_deref_array(void *mem_ctx, nir_deref_array *deref)
    }
    ret->deref.type = deref->deref.type;
    if (deref->deref.child)
-      ret->deref.child = nir_copy_deref(mem_ctx, deref->deref.child);
+      ret->deref.child = nir_copy_deref(ret, deref->deref.child);
    return ret;
 }
 
@@ -568,7 +568,7 @@ copy_deref_struct(void *mem_ctx, nir_deref_struct *deref)
    nir_deref_struct *ret = nir_deref_struct_create(mem_ctx, deref->index);
    ret->deref.type = deref->deref.type;
    if (deref->deref.child)
-      ret->deref.child = nir_copy_deref(mem_ctx, deref->deref.child);
+      ret->deref.child = nir_copy_deref(ret, deref->deref.child);
    return ret;
 }
 
@@ -1834,13 +1834,11 @@ void
 nir_ssa_def_init(nir_instr *instr, nir_ssa_def *def,
                  unsigned num_components, const char *name)
 {
-   void *mem_ctx = ralloc_parent(instr);
-
    def->name = name;
    def->parent_instr = instr;
-   def->uses = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   def->uses = _mesa_set_create(instr, _mesa_hash_pointer,
                                 _mesa_key_pointer_equal);
-   def->if_uses = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
+   def->if_uses = _mesa_set_create(instr, _mesa_hash_pointer,
                                    _mesa_key_pointer_equal);
    def->num_components = num_components;
 
diff --git a/mesalib/src/glsl/nir/nir.h b/mesalib/src/glsl/nir/nir.h
index 29fe94243..74772c798 100644
--- a/mesalib/src/glsl/nir/nir.h
+++ b/mesalib/src/glsl/nir/nir.h
@@ -34,6 +34,7 @@
 #include "util/set.h"
 #include "util/bitset.h"
 #include "nir_types.h"
+#include "glsl/shader_enums.h"
 #include <stdio.h>
 
 #include "nir_opcodes.h"
@@ -529,6 +530,16 @@ nir_src_for_reg(nir_register *reg)
    return src;
 }
 
+static inline nir_instr *
+nir_src_get_parent_instr(const nir_src *src)
+{
+   if (src->is_ssa) {
+      return src->ssa->parent_instr;
+   } else {
+      return src->reg.reg->parent_instr;
+   }
+}
+
 static inline nir_dest
 nir_dest_for_reg(nir_register *reg)
 {
@@ -1365,11 +1376,17 @@ typedef struct nir_function {
 
 typedef struct nir_shader_compiler_options {
    bool lower_ffma;
+   bool lower_flrp;
    bool lower_fpow;
    bool lower_fsat;
    bool lower_fsqrt;
    /** lowers fneg and ineg to fsub and isub. */
    bool lower_negate;
+   /** lowers fsub and isub to fadd+fneg and iadd+ineg. */
+   bool lower_sub;
+
+   /* lower {slt,sge,seq,sne} to {flt,fge,feq,fne} + b2f: */
+   bool lower_scmp;
 
    /**
     * Does the driver support real 32-bit integers?  (Otherwise, integers
@@ -1414,6 +1431,9 @@ typedef struct nir_shader {
     * access plus one
     */
    unsigned num_inputs, num_uniforms, num_outputs;
+
+   /** the number of uniforms that are only accessed directly */
+   unsigned num_direct_uniforms;
 } nir_shader;
 
 #define nir_foreach_overload(shader, overload)                        \
@@ -1466,26 +1486,26 @@ void nir_metadata_require(nir_function_impl *impl, nir_metadata required);
 void nir_metadata_preserve(nir_function_impl *impl, nir_metadata preserved);
 
 /** creates an instruction with default swizzle/writemask/etc. with NULL registers */
-nir_alu_instr *nir_alu_instr_create(void *mem_ctx, nir_op op);
+nir_alu_instr *nir_alu_instr_create(nir_shader *shader, nir_op op);
 
-nir_jump_instr *nir_jump_instr_create(void *mem_ctx, nir_jump_type type);
+nir_jump_instr *nir_jump_instr_create(nir_shader *shader, nir_jump_type type);
 
-nir_load_const_instr *nir_load_const_instr_create(void *mem_ctx,
+nir_load_const_instr *nir_load_const_instr_create(nir_shader *shader,
                                                   unsigned num_components);
 
-nir_intrinsic_instr *nir_intrinsic_instr_create(void *mem_ctx,
+nir_intrinsic_instr *nir_intrinsic_instr_create(nir_shader *shader,
                                                 nir_intrinsic_op op);
 
-nir_call_instr *nir_call_instr_create(void *mem_ctx,
+nir_call_instr *nir_call_instr_create(nir_shader *shader,
                                       nir_function_overload *callee);
 
-nir_tex_instr *nir_tex_instr_create(void *mem_ctx, unsigned num_srcs);
+nir_tex_instr *nir_tex_instr_create(nir_shader *shader, unsigned num_srcs);
 
-nir_phi_instr *nir_phi_instr_create(void *mem_ctx);
+nir_phi_instr *nir_phi_instr_create(nir_shader *shader);
 
-nir_parallel_copy_instr *nir_parallel_copy_instr_create(void *mem_ctx);
+nir_parallel_copy_instr *nir_parallel_copy_instr_create(nir_shader *shader);
 
-nir_ssa_undef_instr *nir_ssa_undef_instr_create(void *mem_ctx,
+nir_ssa_undef_instr *nir_ssa_undef_instr_create(nir_shader *shader,
                                                 unsigned num_components);
 
 nir_deref_var *nir_deref_var_create(void *mem_ctx, nir_variable *var);
@@ -1550,7 +1570,7 @@ void nir_print_instr(const nir_instr *instr, FILE *fp);
 #ifdef DEBUG
 void nir_validate_shader(nir_shader *shader);
 #else
-static inline void nir_validate_shader(nir_shader *shader) { }
+static inline void nir_validate_shader(nir_shader *shader) { (void) shader; }
 #endif /* DEBUG */
 
 void nir_calc_dominance_impl(nir_function_impl *impl);
@@ -1596,14 +1616,18 @@ void nir_lower_alu_to_scalar(nir_shader *shader);
 void nir_lower_phis_to_scalar(nir_shader *shader);
 
 void nir_lower_samplers(nir_shader *shader,
-                        struct gl_shader_program *shader_program,
-                        struct gl_program *prog);
+                        const struct gl_shader_program *shader_program,
+                        gl_shader_stage stage);
 
 void nir_lower_system_values(nir_shader *shader);
+void nir_lower_tex_projector(nir_shader *shader);
+void nir_lower_idiv(nir_shader *shader);
 
 void nir_lower_atomics(nir_shader *shader);
 void nir_lower_to_source_mods(nir_shader *shader);
 
+void nir_normalize_cubemap_coords(nir_shader *shader);
+
 void nir_live_variables_impl(nir_function_impl *impl);
 bool nir_ssa_defs_interfere(nir_ssa_def *a, nir_ssa_def *b);
 
@@ -1612,6 +1636,7 @@ void nir_convert_to_ssa(nir_shader *shader);
 void nir_convert_from_ssa(nir_shader *shader);
 
 bool nir_opt_algebraic(nir_shader *shader);
+bool nir_opt_algebraic_late(nir_shader *shader);
 bool nir_opt_constant_folding(nir_shader *shader);
 
 bool nir_opt_global_to_local(nir_shader *shader);
@@ -1631,6 +1656,8 @@ bool nir_opt_peephole_ffma(nir_shader *shader);
 
 bool nir_opt_remove_phis(nir_shader *shader);
 
+void nir_sweep(nir_shader *shader);
+
 #ifdef __cplusplus
 } /* extern "C" */
 #endif
diff --git a/mesalib/src/glsl/nir/nir_algebraic.py b/mesalib/src/glsl/nir/nir_algebraic.py
index afab1a008..bbf4f08ef 100644
--- a/mesalib/src/glsl/nir/nir_algebraic.py
+++ b/mesalib/src/glsl/nir/nir_algebraic.py
@@ -181,12 +181,23 @@ _algebraic_pass_template = mako.template.Template("""
 #include "nir.h"
 #include "nir_search.h"
 
+#ifndef NIR_OPT_ALGEBRAIC_STRUCT_DEFS
+#define NIR_OPT_ALGEBRAIC_STRUCT_DEFS
+
 struct transform {
    const nir_search_expression *search;
    const nir_search_value *replace;
    unsigned condition_offset;
 };
 
+struct opt_state {
+   void *mem_ctx;
+   bool progress;
+   const bool *condition_flags;
+};
+
+#endif
+
 % for (opcode, xform_list) in xform_dict.iteritems():
 % for xform in xform_list:
    ${xform.search.render()}
@@ -200,12 +211,6 @@ static const struct transform ${pass_name}_${opcode}_xforms[] = {
 };
 % endfor
 
-struct opt_state {
-   void *mem_ctx;
-   bool progress;
-   const bool *condition_flags;
-};
-
 static bool
 ${pass_name}_block(nir_block *block, void *void_state)
 {
diff --git a/mesalib/src/glsl/nir/nir_builder.h b/mesalib/src/glsl/nir/nir_builder.h
index 7c4f7fd96..d1419ee21 100644
--- a/mesalib/src/glsl/nir/nir_builder.h
+++ b/mesalib/src/glsl/nir/nir_builder.h
@@ -28,6 +28,9 @@ struct exec_list;
 
 typedef struct nir_builder {
    struct exec_list *cf_node_list;
+   nir_instr *before_instr;
+   nir_instr *after_instr;
+
    nir_shader *shader;
    nir_function_impl *impl;
 } nir_builder;
@@ -45,8 +48,75 @@ nir_builder_insert_after_cf_list(nir_builder *build,
                                  struct exec_list *cf_node_list)
 {
    build->cf_node_list = cf_node_list;
+   build->before_instr = NULL;
+   build->after_instr = NULL;
+}
+
+static inline void
+nir_builder_insert_before_instr(nir_builder *build, nir_instr *before_instr)
+{
+   build->cf_node_list = NULL;
+   build->before_instr = before_instr;
+   build->after_instr = NULL;
 }
 
+static inline void
+nir_builder_insert_after_instr(nir_builder *build, nir_instr *after_instr)
+{
+   build->cf_node_list = NULL;
+   build->before_instr = NULL;
+   build->after_instr = after_instr;
+}
+
+static inline void
+nir_builder_instr_insert(nir_builder *build, nir_instr *instr)
+{
+   if (build->cf_node_list) {
+      nir_instr_insert_after_cf_list(build->cf_node_list, instr);
+   } else if (build->before_instr) {
+      nir_instr_insert_before(build->before_instr, instr);
+   } else {
+      assert(build->after_instr);
+      nir_instr_insert_after(build->after_instr, instr);
+      build->after_instr = instr;
+   }
+}
+
+static inline nir_ssa_def *
+nir_build_imm(nir_builder *build, unsigned num_components, nir_const_value value)
+{
+   nir_load_const_instr *load_const =
+      nir_load_const_instr_create(build->shader, num_components);
+   if (!load_const)
+      return NULL;
+
+   load_const->value = value;
+
+   nir_builder_instr_insert(build, &load_const->instr);
+
+   return &load_const->def;
+}
+
+static inline nir_ssa_def *
+nir_imm_float(nir_builder *build, float x)
+{
+   nir_const_value v = { { .f = {x, 0, 0, 0} } };
+   return nir_build_imm(build, 1, v);
+}
+
+static inline nir_ssa_def *
+nir_imm_vec4(nir_builder *build, float x, float y, float z, float w)
+{
+   nir_const_value v = { { .f = {x, y, z, w} } };
+   return nir_build_imm(build, 4, v);
+}
+
+static inline nir_ssa_def *
+nir_imm_int(nir_builder *build, int x)
+{
+   nir_const_value v = { { .i = {x, 0, 0, 0} } };
+   return nir_build_imm(build, 1, v);
+}
 
 static inline nir_ssa_def *
 nir_build_alu(nir_builder *build, nir_op op, nir_ssa_def *src0,
@@ -90,7 +160,7 @@ nir_build_alu(nir_builder *build, nir_op op, nir_ssa_def *src0,
    nir_ssa_dest_init(&instr->instr, &instr->dest.dest, num_components, NULL);
    instr->dest.write_mask = (1 << num_components) - 1;
 
-   nir_instr_insert_after_cf_list(build->cf_node_list, &instr->instr);
+   nir_builder_instr_insert(build, &instr->instr);
 
    return &instr->dest.dest.ssa;
 }
@@ -127,4 +197,67 @@ nir_##op(nir_builder *build, nir_ssa_def *src0,                           \
 
 #include "nir_builder_opcodes.h"
 
+/**
+ * Similar to nir_fmov, but takes a nir_alu_src instead of a nir_ssa_def.
+ */
+static inline nir_ssa_def *
+nir_fmov_alu(nir_builder *build, nir_alu_src src, unsigned num_components)
+{
+   nir_alu_instr *mov = nir_alu_instr_create(build->shader, nir_op_fmov);
+   nir_ssa_dest_init(&mov->instr, &mov->dest.dest, num_components, NULL);
+   mov->dest.write_mask = (1 << num_components) - 1;
+   mov->src[0] = src;
+   nir_builder_instr_insert(build, &mov->instr);
+
+   return &mov->dest.dest.ssa;
+}
+
+static inline nir_ssa_def *
+nir_imov_alu(nir_builder *build, nir_alu_src src, unsigned num_components)
+{
+   nir_alu_instr *mov = nir_alu_instr_create(build->shader, nir_op_imov);
+   nir_ssa_dest_init(&mov->instr, &mov->dest.dest, num_components, NULL);
+   mov->dest.write_mask = (1 << num_components) - 1;
+   mov->src[0] = src;
+   nir_builder_instr_insert(build, &mov->instr);
+
+   return &mov->dest.dest.ssa;
+}
+
+/**
+ * Construct an fmov or imov that reswizzles the source's components.
+ */
+static inline nir_ssa_def *
+nir_swizzle(nir_builder *build, nir_ssa_def *src, unsigned swiz[4],
+            unsigned num_components, bool use_fmov)
+{
+   nir_alu_src alu_src;
+   memset(&alu_src, 0, sizeof(alu_src));
+   alu_src.src = nir_src_for_ssa(src);
+   for (int i = 0; i < 4; i++)
+      alu_src.swizzle[i] = swiz[i];
+
+   return use_fmov ? nir_fmov_alu(build, alu_src, num_components) :
+                     nir_imov_alu(build, alu_src, num_components);
+}
+
+/**
+ * Turns a nir_src into a nir_ssa_def * so it can be passed to
+ * nir_build_alu()-based builder calls.
+ */
+static inline nir_ssa_def *
+nir_ssa_for_src(nir_builder *build, nir_src src, int num_components)
+{
+   if (src.is_ssa && src.ssa->num_components == num_components)
+      return src.ssa;
+
+   nir_alu_src alu;
+   memset(&alu, 0, sizeof(alu));
+   alu.src = src;
+   for (int j = 0; j < 4; j++)
+      alu.swizzle[j] = j;
+
+   return nir_imov_alu(build, alu, num_components);
+}
+
 #endif /* NIR_BUILDER_H */
diff --git a/mesalib/src/glsl/nir/nir_from_ssa.c b/mesalib/src/glsl/nir/nir_from_ssa.c
index c3090fb06..184698abd 100644
--- a/mesalib/src/glsl/nir/nir_from_ssa.c
+++ b/mesalib/src/glsl/nir/nir_from_ssa.c
@@ -509,12 +509,13 @@ get_register_for_ssa_def(nir_ssa_def *def, struct from_ssa_state *state)
       reg->num_components = def->num_components;
       reg->num_array_elems = 0;
 
-      /* This register comes from an SSA definition that was not part of a
-       * phi-web.  Therefore, we know it has a single unique definition
-       * that dominates all of its uses.  Therefore, we can copy the
+      /* This register comes from an SSA definition that is defined and not
+       * part of a phi-web.  Therefore, we know it has a single unique
+       * definition that dominates all of its uses; we can copy the
        * parent_instr from the SSA def safely.
        */
-      reg->parent_instr = def->parent_instr;
+      if (def->parent_instr->type != nir_instr_type_ssa_undef)
+         reg->parent_instr = def->parent_instr;
 
       _mesa_hash_table_insert(state->ssa_table, def, reg);
       return reg;
diff --git a/mesalib/src/glsl/nir/nir_lower_idiv.c b/mesalib/src/glsl/nir/nir_lower_idiv.c
new file mode 100644
index 000000000..7b6803207
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_lower_idiv.c
@@ -0,0 +1,155 @@
+/*
+ * Copyright © 2015 Red Hat
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Rob Clark <robclark@freedesktop.org>
+ */
+
+#include "nir.h"
+#include "nir_builder.h"
+
+/* Lowers idiv/udiv/umod
+ * Based on NV50LegalizeSSA::handleDIV()
+ *
+ * Note that this is probably not enough precision for compute shaders.
+ * Perhaps we want a second higher precision (looping) version of this?
+ * Or perhaps we assume if you can do compute shaders you can also
+ * branch out to a pre-optimized shader library routine..
+ */
+
+static void
+convert_instr(nir_builder *bld, nir_alu_instr *alu)
+{
+   nir_ssa_def *numer, *denom, *af, *bf, *a, *b, *q, *r;
+   nir_op op = alu->op;
+   bool is_signed;
+
+   if ((op != nir_op_idiv) &&
+       (op != nir_op_udiv) &&
+       (op != nir_op_umod))
+      return;
+
+   is_signed = (op == nir_op_idiv);
+
+   nir_builder_insert_before_instr(bld, &alu->instr);
+
+   numer = nir_ssa_for_src(bld, alu->src[0].src,
+                           nir_ssa_alu_instr_src_components(alu, 0));
+   denom = nir_ssa_for_src(bld, alu->src[1].src,
+                           nir_ssa_alu_instr_src_components(alu, 1));
+
+   if (is_signed) {
+      af = nir_i2f(bld, numer);
+      bf = nir_i2f(bld, denom);
+      af = nir_fabs(bld, af);
+      bf = nir_fabs(bld, bf);
+      a  = nir_iabs(bld, numer);
+      b  = nir_iabs(bld, denom);
+   } else {
+      af = nir_u2f(bld, numer);
+      bf = nir_u2f(bld, denom);
+      a  = numer;
+      b  = denom;
+   }
+
+   /* get first result: */
+   bf = nir_frcp(bld, bf);
+   bf = nir_isub(bld, bf, nir_imm_int(bld, 2));  /* yes, really */
+   q  = nir_fmul(bld, af, bf);
+
+   if (is_signed) {
+      q = nir_f2i(bld, q);
+   } else {
+      q = nir_f2u(bld, q);
+   }
+
+   /* get error of first result: */
+   r = nir_imul(bld, q, b);
+   r = nir_isub(bld, a, r);
+   r = nir_u2f(bld, r);
+   r = nir_fmul(bld, r, bf);
+   r = nir_f2u(bld, r);
+
+   /* add quotients: */
+   q = nir_iadd(bld, q, r);
+
+   /* correction: if modulus >= divisor, add 1 */
+   r = nir_imul(bld, q, b);
+   r = nir_isub(bld, a, r);
+
+   r = nir_ige(bld, r, b);
+   r = nir_b2i(bld, r);
+
+   q = nir_iadd(bld, q, r);
+   if (is_signed)  {
+      /* fix the sign: */
+      r = nir_ixor(bld, numer, denom);
+      r = nir_ushr(bld, r, nir_imm_int(bld, 31));
+      r = nir_i2b(bld, r);
+      b = nir_ineg(bld, q);
+      q = nir_bcsel(bld, r, b, q);
+   }
+
+   if (op == nir_op_umod) {
+      /* division result in q */
+      r = nir_imul(bld, q, b);
+      q = nir_isub(bld, a, r);
+   }
+
+   assert(alu->dest.dest.is_ssa);
+   nir_ssa_def_rewrite_uses(&alu->dest.dest.ssa,
+                            nir_src_for_ssa(q),
+                            ralloc_parent(alu));
+}
+
+static bool
+convert_block(nir_block *block, void *state)
+{
+   nir_builder *b = state;
+
+   nir_foreach_instr_safe(block, instr) {
+      if (instr->type == nir_instr_type_alu)
+         convert_instr(b, nir_instr_as_alu(instr));
+   }
+
+   return true;
+}
+
+static void
+convert_impl(nir_function_impl *impl)
+{
+   nir_builder b;
+   nir_builder_init(&b, impl);
+
+   nir_foreach_block(impl, convert_block, &b);
+   nir_metadata_preserve(impl, nir_metadata_block_index |
+                               nir_metadata_dominance);
+}
+
+void
+nir_lower_idiv(nir_shader *shader)
+{
+   nir_foreach_overload(shader, overload) {
+      if (overload->impl)
+         convert_impl(overload->impl);
+   }
+}
diff --git a/mesalib/src/glsl/nir/nir_lower_phis_to_scalar.c b/mesalib/src/glsl/nir/nir_lower_phis_to_scalar.c
index 7cd93ea0a..4bdb80072 100644
--- a/mesalib/src/glsl/nir/nir_lower_phis_to_scalar.c
+++ b/mesalib/src/glsl/nir/nir_lower_phis_to_scalar.c
@@ -223,7 +223,7 @@ lower_phis_to_scalar_block(nir_block *block, void *void_state)
             else
                nir_instr_insert_after_block(src->pred, &mov->instr);
 
-            nir_phi_src *new_src = ralloc(state->mem_ctx, nir_phi_src);
+            nir_phi_src *new_src = ralloc(new_phi, nir_phi_src);
             new_src->pred = src->pred;
             new_src->src = nir_src_for_ssa(&mov->dest.dest.ssa);
 
diff --git a/mesalib/src/glsl/nir/nir_lower_samplers.cpp b/mesalib/src/glsl/nir/nir_lower_samplers.cpp
index 3015dbd09..cf8ab8325 100644
--- a/mesalib/src/glsl/nir/nir_lower_samplers.cpp
+++ b/mesalib/src/glsl/nir/nir_lower_samplers.cpp
@@ -36,33 +36,26 @@ extern "C" {
 }
 
 static unsigned
-get_sampler_index(struct gl_shader_program *shader_program, const char *name,
-                  const struct gl_program *prog)
+get_sampler_index(const struct gl_shader_program *shader_program,
+                  gl_shader_stage stage, const char *name)
 {
-   GLuint shader = _mesa_program_enum_to_shader_stage(prog->Target);
-
    unsigned location;
    if (!shader_program->UniformHash->get(location, name)) {
-      linker_error(shader_program,
-                   "failed to find sampler named %s.\n", name);
+      assert(!"failed to find sampler");
       return 0;
    }
 
-   if (!shader_program->UniformStorage[location].sampler[shader].active) {
-      assert(0 && "cannot return a sampler");
-      linker_error(shader_program,
-                   "cannot return a sampler named %s, because it is not "
-                   "used in this shader stage. This is a driver bug.\n",
-                   name);
+   if (!shader_program->UniformStorage[location].sampler[stage].active) {
+      assert(!"cannot return a sampler");
       return 0;
    }
 
-   return shader_program->UniformStorage[location].sampler[shader].index;
+   return shader_program->UniformStorage[location].sampler[stage].index;
 }
 
 static void
-lower_sampler(nir_tex_instr *instr, struct gl_shader_program *shader_program,
-              const struct gl_program *prog, void *mem_ctx)
+lower_sampler(nir_tex_instr *instr, const struct gl_shader_program *shader_program,
+              gl_shader_stage stage, void *mem_ctx)
 {
    if (instr->sampler == NULL)
       return;
@@ -90,7 +83,7 @@ lower_sampler(nir_tex_instr *instr, struct gl_shader_program *shader_program,
                ralloc_asprintf_append(&name, "[%u]", deref_array->base_offset);
             break;
          case nir_deref_array_type_indirect: {
-            instr->src = reralloc(mem_ctx, instr->src, nir_tex_src,
+            instr->src = reralloc(instr, instr->src, nir_tex_src,
                                   instr->num_srcs + 1);
             memset(&instr->src[instr->num_srcs], 0, sizeof *instr->src);
             instr->src[instr->num_srcs].src_type = nir_tex_src_sampler_offset;
@@ -133,15 +126,15 @@ lower_sampler(nir_tex_instr *instr, struct gl_shader_program *shader_program,
       }
    }
 
-   instr->sampler_index += get_sampler_index(shader_program, name, prog);
+   instr->sampler_index += get_sampler_index(shader_program, stage, name);
 
    instr->sampler = NULL;
 }
 
 typedef struct {
    void *mem_ctx;
-   struct gl_shader_program *shader_program;
-   struct gl_program *prog;
+   const struct gl_shader_program *shader_program;
+   gl_shader_stage stage;
 } lower_state;
 
 static bool
@@ -152,7 +145,7 @@ lower_block_cb(nir_block *block, void *_state)
    nir_foreach_instr(block, instr) {
       if (instr->type == nir_instr_type_tex) {
          nir_tex_instr *tex_instr = nir_instr_as_tex(instr);
-         lower_sampler(tex_instr, state->shader_program, state->prog,
+         lower_sampler(tex_instr, state->shader_program, state->stage,
                        state->mem_ctx);
       }
    }
@@ -161,24 +154,24 @@ lower_block_cb(nir_block *block, void *_state)
 }
 
 static void
-lower_impl(nir_function_impl *impl, struct gl_shader_program *shader_program,
-           struct gl_program *prog)
+lower_impl(nir_function_impl *impl, const struct gl_shader_program *shader_program,
+           gl_shader_stage stage)
 {
    lower_state state;
 
    state.mem_ctx = ralloc_parent(impl);
    state.shader_program = shader_program;
-   state.prog = prog;
+   state.stage = stage;
 
    nir_foreach_block(impl, lower_block_cb, &state);
 }
 
 extern "C" void
-nir_lower_samplers(nir_shader *shader, struct gl_shader_program *shader_program,
-                   struct gl_program *prog)
+nir_lower_samplers(nir_shader *shader, const struct gl_shader_program *shader_program,
+                   gl_shader_stage stage)
 {
    nir_foreach_overload(shader, overload) {
       if (overload->impl)
-         lower_impl(overload->impl, shader_program, prog);
+         lower_impl(overload->impl, shader_program, stage);
    }
 }
diff --git a/mesalib/src/glsl/nir/nir_lower_tex_projector.c b/mesalib/src/glsl/nir/nir_lower_tex_projector.c
new file mode 100644
index 000000000..6b0e9c340
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_lower_tex_projector.c
@@ -0,0 +1,143 @@
+/*
+ * Copyright © 2015 Broadcom
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+/*
+ * This lowering pass converts the coordinate division for texture projection
+ * to be done in ALU instructions instead of asking the texture operation to
+ * do so.
+ */
+
+#include "nir.h"
+#include "nir_builder.h"
+
+static nir_ssa_def *
+channel(nir_builder *b, nir_ssa_def *def, int c)
+{
+   return nir_swizzle(b, def, (unsigned[4]){c, c, c, c}, 1, false);
+}
+
+static bool
+nir_lower_tex_projector_block(nir_block *block, void *void_state)
+{
+   nir_builder *b = void_state;
+
+   nir_foreach_instr_safe(block, instr) {
+      if (instr->type != nir_instr_type_tex)
+         continue;
+
+      nir_tex_instr *tex = nir_instr_as_tex(instr);
+      nir_builder_insert_before_instr(b, &tex->instr);
+
+      /* Find the projector in the srcs list, if present. */
+      int proj_index;
+      for (proj_index = 0; proj_index < tex->num_srcs; proj_index++) {
+         if (tex->src[proj_index].src_type == nir_tex_src_projector)
+            break;
+      }
+      if (proj_index == tex->num_srcs)
+         continue;
+      nir_ssa_def *inv_proj =
+         nir_frcp(b, nir_ssa_for_src(b, tex->src[proj_index].src, 1));
+
+      /* Walk through the sources projecting the arguments. */
+      for (int i = 0; i < tex->num_srcs; i++) {
+         switch (tex->src[i].src_type) {
+         case nir_tex_src_coord:
+         case nir_tex_src_comparitor:
+            break;
+         default:
+            continue;
+         }
+         nir_ssa_def *unprojected =
+            nir_ssa_for_src(b, tex->src[i].src, nir_tex_instr_src_size(tex, i));
+         nir_ssa_def *projected = nir_fmul(b, unprojected, inv_proj);
+
+         /* Array indices don't get projected, so make an new vector with the
+          * coordinate's array index untouched.
+          */
+         if (tex->is_array && tex->src[i].src_type == nir_tex_src_coord) {
+            switch (tex->coord_components) {
+            case 4:
+               projected = nir_vec4(b,
+                                    channel(b, projected, 0),
+                                    channel(b, projected, 1),
+                                    channel(b, projected, 2),
+                                    channel(b, unprojected, 3));
+               break;
+            case 3:
+               projected = nir_vec3(b,
+                                    channel(b, projected, 0),
+                                    channel(b, projected, 1),
+                                    channel(b, unprojected, 2));
+               break;
+            case 2:
+               projected = nir_vec2(b,
+                                    channel(b, projected, 0),
+                                    channel(b, unprojected, 1));
+               break;
+            default:
+               unreachable("bad texture coord count for array");
+               break;
+            }
+         }
+
+         nir_instr_rewrite_src(&tex->instr,
+                               &tex->src[i].src,
+                               nir_src_for_ssa(projected));
+      }
+
+      /* Now move the later tex sources down the array so that the projector
+       * disappears.
+       */
+      nir_src dead;
+      memset(&dead, 0, sizeof dead);
+      nir_instr_rewrite_src(&tex->instr, &tex->src[proj_index].src, dead);
+      memmove(&tex->src[proj_index],
+              &tex->src[proj_index + 1],
+              (tex->num_srcs - proj_index) * sizeof(*tex->src));
+      tex->num_srcs--;
+   }
+
+   return true;
+}
+
+static void
+nir_lower_tex_projector_impl(nir_function_impl *impl)
+{
+   nir_builder b;
+   nir_builder_init(&b, impl);
+
+   nir_foreach_block(impl, nir_lower_tex_projector_block, &b);
+
+   nir_metadata_preserve(impl, nir_metadata_block_index |
+                               nir_metadata_dominance);
+}
+
+void
+nir_lower_tex_projector(nir_shader *shader)
+{
+   nir_foreach_overload(shader, overload) {
+      if (overload->impl)
+         nir_lower_tex_projector_impl(overload->impl);
+   }
+}
diff --git a/mesalib/src/glsl/nir/nir_lower_var_copies.c b/mesalib/src/glsl/nir/nir_lower_var_copies.c
index 85ebb281c..58389a7c7 100644
--- a/mesalib/src/glsl/nir/nir_lower_var_copies.c
+++ b/mesalib/src/glsl/nir/nir_lower_var_copies.c
@@ -148,13 +148,10 @@ emit_copy_load_store(nir_intrinsic_instr *copy_instr,
 
       unsigned num_components = glsl_get_vector_elements(src_tail->type);
 
-      nir_deref *src_deref = nir_copy_deref(mem_ctx, &src_head->deref);
-      nir_deref *dest_deref = nir_copy_deref(mem_ctx, &dest_head->deref);
-
       nir_intrinsic_instr *load =
          nir_intrinsic_instr_create(mem_ctx, nir_intrinsic_load_var);
       load->num_components = num_components;
-      load->variables[0] = nir_deref_as_var(src_deref);
+      load->variables[0] = nir_deref_as_var(nir_copy_deref(load, &src_head->deref));
       nir_ssa_dest_init(&load->instr, &load->dest, num_components, NULL);
 
       nir_instr_insert_before(&copy_instr->instr, &load->instr);
@@ -162,7 +159,8 @@ emit_copy_load_store(nir_intrinsic_instr *copy_instr,
       nir_intrinsic_instr *store =
          nir_intrinsic_instr_create(mem_ctx, nir_intrinsic_store_var);
       store->num_components = num_components;
-      store->variables[0] = nir_deref_as_var(dest_deref);
+      store->variables[0] = nir_deref_as_var(nir_copy_deref(store, &dest_head->deref));
+
       store->src[0].is_ssa = true;
       store->src[0].ssa = &load->dest.ssa;
 
diff --git a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
index 86e6ab416..2ca74d71b 100644
--- a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
+++ b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
@@ -642,7 +642,7 @@ add_phi_sources(nir_block *block, nir_block *pred,
 
       struct deref_node *node = entry->data;
 
-      nir_phi_src *src = ralloc(state->mem_ctx, nir_phi_src);
+      nir_phi_src *src = ralloc(phi, nir_phi_src);
       src->pred = pred;
       src->src.is_ssa = true;
       src->src.ssa = get_ssa_def_for_block(node, pred, state);
diff --git a/mesalib/src/glsl/nir/nir_normalize_cubemap_coords.c b/mesalib/src/glsl/nir/nir_normalize_cubemap_coords.c
new file mode 100644
index 000000000..0da8447ac
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_normalize_cubemap_coords.c
@@ -0,0 +1,110 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Jason Ekstrand <jason@jlekstrand.net>
+ */
+
+#include "nir.h"
+#include "nir_builder.h"
+
+/**
+ * This file implements a NIR lowering pass to perform the normalization of
+ * the cubemap coordinates to have the largest magnitude component be -1.0
+ * or 1.0.  This is based on the old GLSL IR based pass by Eric.
+ */
+
+static nir_ssa_def *
+channel(nir_builder *b, nir_ssa_def *def, int c)
+{
+   return nir_swizzle(b, def, (unsigned[4]){c, c, c, c}, 1, false);
+}
+
+static bool
+normalize_cubemap_coords_block(nir_block *block, void *void_state)
+{
+   nir_builder *b = void_state;
+
+   nir_foreach_instr(block, instr) {
+      if (instr->type != nir_instr_type_tex)
+         continue;
+
+      nir_tex_instr *tex = nir_instr_as_tex(instr);
+      if (tex->sampler_dim != GLSL_SAMPLER_DIM_CUBE)
+         continue;
+
+      nir_builder_insert_before_instr(b, &tex->instr);
+
+      for (unsigned i = 0; i < tex->num_srcs; i++) {
+         if (tex->src[i].src_type != nir_tex_src_coord)
+            continue;
+
+         nir_ssa_def *orig_coord =
+            nir_ssa_for_src(b, tex->src[i].src, nir_tex_instr_src_size(tex, i));
+         assert(orig_coord->num_components >= 3);
+
+         nir_ssa_def *abs = nir_fabs(b, orig_coord);
+         nir_ssa_def *norm = nir_fmax(b, channel(b, abs, 0),
+                                         nir_fmax(b, channel(b, abs, 1),
+                                                     channel(b, abs, 2)));
+
+         nir_ssa_def *normalized = nir_fmul(b, orig_coord, nir_frcp(b, norm));
+
+         /* Array indices don't have to be normalized, so make a new vector
+          * with the coordinate's array index untouched.
+          */
+         if (tex->coord_components == 4) {
+            normalized = nir_vec4(b,
+                                  channel(b, normalized, 0),
+                                  channel(b, normalized, 1),
+                                  channel(b, normalized, 2),
+                                  channel(b, orig_coord, 3));
+         }
+
+         nir_instr_rewrite_src(&tex->instr,
+                               &tex->src[i].src,
+                               nir_src_for_ssa(normalized));
+      }
+   }
+
+   return true;
+}
+
+static void
+normalize_cubemap_coords_impl(nir_function_impl *impl)
+{
+   nir_builder b;
+   nir_builder_init(&b, impl);
+
+   nir_foreach_block(impl, normalize_cubemap_coords_block, &b);
+
+   nir_metadata_preserve(impl, nir_metadata_block_index |
+                               nir_metadata_dominance);
+}
+
+void
+nir_normalize_cubemap_coords(nir_shader *shader)
+{
+   nir_foreach_overload(shader, overload)
+      if (overload->impl)
+         normalize_cubemap_coords_impl(overload->impl);
+}
diff --git a/mesalib/src/glsl/nir/nir_opcodes.py b/mesalib/src/glsl/nir/nir_opcodes.py
index 062cd628b..264806f5d 100644
--- a/mesalib/src/glsl/nir/nir_opcodes.py
+++ b/mesalib/src/glsl/nir/nir_opcodes.py
@@ -161,12 +161,12 @@ unop_convert("f2i", tfloat, tint, "src0") # Float-to-integer conversion.
 unop_convert("f2u", tfloat, tunsigned, "src0") # Float-to-unsigned conversion
 unop_convert("i2f", tint, tfloat, "src0") # Integer-to-float conversion.
 # Float-to-boolean conversion
-unop_convert("f2b", tfloat, tbool, "src0 == 0.0f")
+unop_convert("f2b", tfloat, tbool, "src0 != 0.0f")
 # Boolean-to-float conversion
 unop_convert("b2f", tbool, tfloat, "src0 ? 1.0f : 0.0f")
 # Int-to-boolean conversion
-unop_convert("i2b", tint, tbool, "src0 == 0")
-unop_convert("b2i", tbool, tint, "src0 ? 0 : -1") # Boolean-to-int conversion
+unop_convert("i2b", tint, tbool, "src0 != 0")
+unop_convert("b2i", tbool, tint, "src0 ? 1 : 0") # Boolean-to-int conversion
 unop_convert("u2f", tunsigned, tfloat, "src0") #Unsigned-to-float conversion.
 
 unop_reduce("bany", 1, tbool, tbool, "{src}", "{src0} || {src1}", "{src}")
@@ -191,8 +191,6 @@ unop("fround_even", tfloat, "_mesa_roundevenf(src0)")
 
 unop("fsin", tfloat, "sinf(src0)")
 unop("fcos", tfloat, "cosf(src0)")
-unop("fsin_reduced", tfloat, "sinf(src0)")
-unop("fcos_reduced", tfloat, "cosf(src0)")
 
 
 # Partial derivatives.
diff --git a/mesalib/src/glsl/nir/nir_opt_algebraic.py b/mesalib/src/glsl/nir/nir_opt_algebraic.py
index ef855aa77..cdb19241c 100644
--- a/mesalib/src/glsl/nir/nir_opt_algebraic.py
+++ b/mesalib/src/glsl/nir/nir_opt_algebraic.py
@@ -75,6 +75,9 @@ optimizations = [
    (('flrp', a, b, 1.0), b),
    (('flrp', a, a, b), a),
    (('flrp', 0.0, a, b), ('fmul', a, b)),
+   (('flrp', a, b, c), ('fadd', ('fmul', c, ('fsub', b, a)), a), 'options->lower_flrp'),
+   (('fadd', ('fmul', a, ('fadd', 1.0, ('fneg', c))), ('fmul', b, c)), ('flrp', a, b, c), '!options->lower_flrp'),
+   (('fadd', a, ('fmul', c, ('fadd', b, ('fneg', a)))), ('flrp', a, b, c), '!options->lower_flrp'),
    (('ffma', a, b, c), ('fadd', ('fmul', a, b), c), 'options->lower_ffma'),
    (('fadd', ('fmul', a, b), c), ('ffma', a, b, c), '!options->lower_ffma'),
    # Comparison simplifications
@@ -82,10 +85,6 @@ optimizations = [
    (('inot', ('fge', a, b)), ('flt', a, b)),
    (('inot', ('ilt', a, b)), ('ige', a, b)),
    (('inot', ('ige', a, b)), ('ilt', a, b)),
-   (('flt', ('fadd', a, b), 0.0), ('flt', a, ('fneg', b))),
-   (('fge', ('fadd', a, b), 0.0), ('fge', a, ('fneg', b))),
-   (('feq', ('fadd', a, b), 0.0), ('feq', a, ('fneg', b))),
-   (('fne', ('fadd', a, b), 0.0), ('fne', a, ('fneg', b))),
    (('fge', ('fneg', ('fabs', a)), 0.0), ('feq', a, 0.0)),
    (('bcsel', ('flt', a, b), a, b), ('fmin', a, b)),
    (('bcsel', ('flt', a, b), b, a), ('fmax', a, b)),
@@ -95,6 +94,18 @@ optimizations = [
    (('fsat', a), ('fmin', ('fmax', a, 0.0), 1.0), 'options->lower_fsat'),
    (('fsat', ('fsat', a)), ('fsat', a)),
    (('fmin', ('fmax', ('fmin', ('fmax', a, 0.0), 1.0), 0.0), 1.0), ('fmin', ('fmax', a, 0.0), 1.0)),
+   (('ior', ('flt', a, b), ('flt', a, c)), ('flt', a, ('fmax', b, c))),
+   (('ior', ('fge', a, b), ('fge', a, c)), ('fge', a, ('fmin', b, c))),
+   (('slt', a, b), ('b2f', ('flt', a, b)), 'options->lower_scmp'),
+   (('sge', a, b), ('b2f', ('fge', a, b)), 'options->lower_scmp'),
+   (('seq', a, b), ('b2f', ('feq', a, b)), 'options->lower_scmp'),
+   (('sne', a, b), ('b2f', ('fne', a, b)), 'options->lower_scmp'),
+   # Emulating booleans
+   (('fmul', ('b2f', a), ('b2f', b)), ('b2f', ('iand', a, b))),
+   (('fsat', ('fadd', ('b2f', a), ('b2f', b))), ('b2f', ('ior', a, b))),
+   (('iand', 'a@bool', 1.0), ('b2f', a)),
+   (('flt', ('fneg', ('b2f', a)), 0), a), # Generated by TGSI KILL_IF.
+   (('flt', ('fsub', 0.0, ('b2f', a)), 0), a), # Generated by TGSI KILL_IF.
    # Comparison with the same args.  Note that these are not done for
    # the float versions because NaN always returns false on float
    # inequalities.
@@ -122,7 +133,7 @@ optimizations = [
    (('ishr', 0, a), 0),
    (('ishr', a, 0), a),
    (('ushr', 0, a), 0),
-   (('ushr', a, 0), 0),
+   (('ushr', a, 0), a),
    # Exponential/logarithmic identities
    (('fexp2', ('flog2', a)), a), # 2^lg2(a) = a
    (('fexp',  ('flog',  a)), a), # e^ln(a)  = a
@@ -134,6 +145,26 @@ optimizations = [
    (('fpow', a, 1.0), a),
    (('fpow', a, 2.0), ('fmul', a, a)),
    (('fpow', 2.0, a), ('fexp2', a)),
+   (('fsqrt', ('fexp2', a)), ('fexp2', ('fmul', 0.5, a))),
+   (('fsqrt', ('fexp', a)), ('fexp', ('fmul', 0.5, a))),
+   (('frcp', ('fexp2', a)), ('fexp2', ('fneg', a))),
+   (('frcp', ('fexp', a)), ('fexp', ('fneg', a))),
+   (('frsq', ('fexp2', a)), ('fexp2', ('fmul', -0.5, a))),
+   (('frsq', ('fexp', a)), ('fexp', ('fmul', -0.5, a))),
+   (('flog2', ('fsqrt', a)), ('fmul', 0.5, ('flog2', a))),
+   (('flog', ('fsqrt', a)), ('fmul', 0.5, ('flog', a))),
+   (('flog2', ('frcp', a)), ('fneg', ('flog2', a))),
+   (('flog', ('frcp', a)), ('fneg', ('flog', a))),
+   (('flog2', ('frsq', a)), ('fmul', -0.5, ('flog2', a))),
+   (('flog', ('frsq', a)), ('fmul', -0.5, ('flog', a))),
+   (('flog2', ('fpow', a, b)), ('fmul', b, ('flog2', a))),
+   (('flog', ('fpow', a, b)), ('fmul', b, ('flog', a))),
+   (('fadd', ('flog2', a), ('flog2', b)), ('flog2', ('fmul', a, b))),
+   (('fadd', ('flog', a), ('flog', b)), ('flog', ('fmul', a, b))),
+   (('fadd', ('flog2', a), ('fneg', ('flog2', b))), ('flog2', ('fdiv', a, b))),
+   (('fadd', ('flog', a), ('fneg', ('flog', b))), ('flog', ('fdiv', a, b))),
+   (('fmul', ('fexp2', a), ('fexp2', b)), ('fexp2', ('fadd', a, b))),
+   (('fmul', ('fexp', a), ('fexp', b)), ('fexp', ('fadd', a, b))),
    # Division and reciprocal
    (('fdiv', 1.0, a), ('frcp', a)),
    (('frcp', ('frcp', a)), a),
@@ -154,18 +185,21 @@ optimizations = [
    (('bcsel', a, b, b), b),
    (('fcsel', a, b, b), b),
 
+   # Conversions
+   (('f2i', ('ftrunc', a)), ('f2i', a)),
+   (('f2u', ('ftrunc', a)), ('f2u', a)),
+
    # Subtracts
    (('fsub', a, ('fsub', 0.0, b)), ('fadd', a, b)),
    (('isub', a, ('isub', 0, b)), ('iadd', a, b)),
+   (('fsub', a, b), ('fadd', a, ('fneg', b)), 'options->lower_sub'),
+   (('isub', a, b), ('iadd', a, ('ineg', b)), 'options->lower_sub'),
    (('fneg', a), ('fsub', 0.0, a), 'options->lower_negate'),
    (('ineg', a), ('isub', 0, a), 'options->lower_negate'),
    (('fadd', a, ('fsub', 0.0, b)), ('fsub', a, b)),
    (('iadd', a, ('isub', 0, b)), ('isub', a, b)),
    (('fabs', ('fsub', 0.0, a)), ('fabs', a)),
    (('iabs', ('isub', 0, a)), ('iabs', a)),
-
-# This one may not be exact
-   (('feq', ('fadd', a, b), 0.0), ('feq', a, ('fneg', b))),
 ]
 
 # Add optimizations to handle the case where the result of a ternary is
@@ -189,4 +223,17 @@ for op in ['flt', 'fge', 'feq', 'fne',
        ('bcsel', 'a', (op, 'd', 'b'), (op, 'd', 'c'))),
    ]
 
+# This section contains "late" optimizations that should be run after the
+# regular optimizations have finished.  Optimizations should go here if
+# they help code generation but do not necessarily produce code that is
+# more easily optimizable.
+late_optimizations = [
+   (('flt', ('fadd', a, b), 0.0), ('flt', a, ('fneg', b))),
+   (('fge', ('fadd', a, b), 0.0), ('fge', a, ('fneg', b))),
+   (('feq', ('fadd', a, b), 0.0), ('feq', a, ('fneg', b))),
+   (('fne', ('fadd', a, b), 0.0), ('fne', a, ('fneg', b))),
+]
+
 print nir_algebraic.AlgebraicPass("nir_opt_algebraic", optimizations).render()
+print nir_algebraic.AlgebraicPass("nir_opt_algebraic_late",
+                                  late_optimizations).render()
diff --git a/mesalib/src/glsl/nir/nir_opt_cse.c b/mesalib/src/glsl/nir/nir_opt_cse.c
index 9b383202d..553906e12 100644
--- a/mesalib/src/glsl/nir/nir_opt_cse.c
+++ b/mesalib/src/glsl/nir/nir_opt_cse.c
@@ -37,20 +37,19 @@ struct cse_state {
 };
 
 static bool
-nir_alu_srcs_equal(nir_alu_src src1, nir_alu_src src2, uint8_t read_mask)
+nir_alu_srcs_equal(nir_alu_instr *alu1, nir_alu_instr *alu2, unsigned src1,
+                   unsigned src2)
 {
-   if (src1.abs != src2.abs || src1.negate != src2.negate)
+   if (alu1->src[src1].abs != alu2->src[src2].abs ||
+       alu1->src[src1].negate != alu2->src[src2].negate)
       return false;
 
-   for (int i = 0; i < 4; ++i) {
-      if (!(read_mask & (1 << i)))
-         continue;
-
-      if (src1.swizzle[i] != src2.swizzle[i])
+   for (unsigned i = 0; i < nir_ssa_alu_instr_src_components(alu1, src1); i++) {
+      if (alu1->src[src1].swizzle[i] != alu2->src[src2].swizzle[i])
          return false;
    }
 
-   return nir_srcs_equal(src1.src, src2.src);
+   return nir_srcs_equal(alu1->src[src1].src, alu2->src[src2].src);
 }
 
 static bool
@@ -73,10 +72,17 @@ nir_instrs_equal(nir_instr *instr1, nir_instr *instr2)
       if (alu1->dest.dest.ssa.num_components != alu2->dest.dest.ssa.num_components)
          return false;
 
-      for (unsigned i = 0; i < nir_op_infos[alu1->op].num_inputs; i++) {
-         if (!nir_alu_srcs_equal(alu1->src[i], alu2->src[i],
-                                 (1 << alu1->dest.dest.ssa.num_components) - 1))
-            return false;
+      if (nir_op_infos[alu1->op].algebraic_properties & NIR_OP_IS_COMMUTATIVE) {
+         assert(nir_op_infos[alu1->op].num_inputs == 2);
+         return (nir_alu_srcs_equal(alu1, alu2, 0, 0) &&
+                 nir_alu_srcs_equal(alu1, alu2, 1, 1)) ||
+                (nir_alu_srcs_equal(alu1, alu2, 0, 1) &&
+                 nir_alu_srcs_equal(alu1, alu2, 1, 0));
+      } else {
+         for (unsigned i = 0; i < nir_op_infos[alu1->op].num_inputs; i++) {
+            if (!nir_alu_srcs_equal(alu1, alu2, i, i))
+               return false;
+         }
       }
       return true;
    }
@@ -154,12 +160,14 @@ nir_instrs_equal(nir_instr *instr1, nir_instr *instr2)
 static bool
 src_is_ssa(nir_src *src, void *data)
 {
+   (void) data;
    return src->is_ssa;
 }
 
 static bool
 dest_is_ssa(nir_dest *dest, void *data)
 {
+   (void) data;
    return dest->is_ssa;
 }
 
diff --git a/mesalib/src/glsl/nir/nir_opt_peephole_ffma.c b/mesalib/src/glsl/nir/nir_opt_peephole_ffma.c
new file mode 100644
index 000000000..9d5646fe6
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_opt_peephole_ffma.c
@@ -0,0 +1,261 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Jason Ekstrand (jason@jlekstrand.net)
+ *
+ */
+
+#include "nir.h"
+
+/*
+ * Implements a small peephole optimization that looks for a multiply that
+ * is only ever used in an add and replaces both with an fma.
+ */
+
+struct peephole_ffma_state {
+   void *mem_ctx;
+   nir_function_impl *impl;
+   bool progress;
+};
+
+static inline bool
+are_all_uses_fadd(nir_ssa_def *def)
+{
+   if (def->if_uses->entries > 0)
+      return false;
+
+   struct set_entry *use_iter;
+   set_foreach(def->uses, use_iter) {
+      nir_instr *use_instr = (nir_instr *)use_iter->key;
+
+      if (use_instr->type != nir_instr_type_alu)
+         return false;
+
+      nir_alu_instr *use_alu = nir_instr_as_alu(use_instr);
+      switch (use_alu->op) {
+      case nir_op_fadd:
+         break; /* This one's ok */
+
+      case nir_op_imov:
+      case nir_op_fmov:
+      case nir_op_fneg:
+      case nir_op_fabs:
+         assert(use_alu->dest.dest.is_ssa);
+         if (!are_all_uses_fadd(&use_alu->dest.dest.ssa))
+            return false;
+         break;
+
+      default:
+         return false;
+      }
+   }
+
+   return true;
+}
+
+static nir_alu_instr *
+get_mul_for_src(nir_alu_src *src, uint8_t swizzle[4], bool *negate, bool *abs)
+{
+   assert(src->src.is_ssa && !src->abs && !src->negate);
+
+   nir_instr *instr = src->src.ssa->parent_instr;
+   if (instr->type != nir_instr_type_alu)
+      return NULL;
+
+   nir_alu_instr *alu = nir_instr_as_alu(instr);
+   switch (alu->op) {
+   case nir_op_imov:
+   case nir_op_fmov:
+      alu = get_mul_for_src(&alu->src[0], swizzle, negate, abs);
+      break;
+
+   case nir_op_fneg:
+      alu = get_mul_for_src(&alu->src[0], swizzle, negate, abs);
+      *negate = !*negate;
+      break;
+
+   case nir_op_fabs:
+      alu = get_mul_for_src(&alu->src[0], swizzle, negate, abs);
+      *negate = false;
+      *abs = true;
+      break;
+
+   case nir_op_fmul:
+      /* Only absorb a fmul into a ffma if the fmul is is only used in fadd
+       * operations.  This prevents us from being too aggressive with our
+       * fusing which can actually lead to more instructions.
+       */
+      if (!are_all_uses_fadd(&alu->dest.dest.ssa))
+         return NULL;
+      break;
+
+   default:
+      return NULL;
+   }
+
+   if (!alu)
+      return NULL;
+
+   for (unsigned i = 0; i < 4; i++) {
+      if (!(alu->dest.write_mask & (1 << i)))
+         break;
+
+      swizzle[i] = swizzle[src->swizzle[i]];
+   }
+
+   return alu;
+}
+
+static bool
+nir_opt_peephole_ffma_block(nir_block *block, void *void_state)
+{
+   struct peephole_ffma_state *state = void_state;
+
+   nir_foreach_instr_safe(block, instr) {
+      if (instr->type != nir_instr_type_alu)
+         continue;
+
+      nir_alu_instr *add = nir_instr_as_alu(instr);
+      if (add->op != nir_op_fadd)
+         continue;
+
+      /* TODO: Maybe bail if this expression is considered "precise"? */
+
+      assert(add->src[0].src.is_ssa && add->src[1].src.is_ssa);
+
+      /* This, is the case a + a.  We would rather handle this with an
+       * algebraic reduction than fuse it.  Also, we want to only fuse
+       * things where the multiply is used only once and, in this case,
+       * it would be used twice by the same instruction.
+       */
+      if (add->src[0].src.ssa == add->src[1].src.ssa)
+         continue;
+
+      nir_alu_instr *mul;
+      uint8_t add_mul_src, swizzle[4];
+      bool negate, abs;
+      for (add_mul_src = 0; add_mul_src < 2; add_mul_src++) {
+         for (unsigned i = 0; i < 4; i++)
+            swizzle[i] = i;
+
+         negate = false;
+         abs = false;
+
+         mul = get_mul_for_src(&add->src[add_mul_src], swizzle, &negate, &abs);
+
+         if (mul != NULL)
+            break;
+      }
+
+      if (mul == NULL)
+         continue;
+
+      nir_ssa_def *mul_src[2];
+      mul_src[0] = mul->src[0].src.ssa;
+      mul_src[1] = mul->src[1].src.ssa;
+
+      if (abs) {
+         for (unsigned i = 0; i < 2; i++) {
+            nir_alu_instr *abs = nir_alu_instr_create(state->mem_ctx,
+                                                      nir_op_fabs);
+            abs->src[0].src = nir_src_for_ssa(mul_src[i]);
+            nir_ssa_dest_init(&abs->instr, &abs->dest.dest,
+                              mul_src[i]->num_components, NULL);
+            abs->dest.write_mask = (1 << mul_src[i]->num_components) - 1;
+            nir_instr_insert_before(&add->instr, &abs->instr);
+            mul_src[i] = &abs->dest.dest.ssa;
+         }
+      }
+
+      if (negate) {
+         nir_alu_instr *neg = nir_alu_instr_create(state->mem_ctx,
+                                                   nir_op_fneg);
+         neg->src[0].src = nir_src_for_ssa(mul_src[0]);
+         nir_ssa_dest_init(&neg->instr, &neg->dest.dest,
+                           mul_src[0]->num_components, NULL);
+         neg->dest.write_mask = (1 << mul_src[0]->num_components) - 1;
+         nir_instr_insert_before(&add->instr, &neg->instr);
+         mul_src[0] = &neg->dest.dest.ssa;
+      }
+
+      nir_alu_instr *ffma = nir_alu_instr_create(state->mem_ctx, nir_op_ffma);
+      ffma->dest.saturate = add->dest.saturate;
+      ffma->dest.write_mask = add->dest.write_mask;
+
+      for (unsigned i = 0; i < 2; i++) {
+         ffma->src[i].src = nir_src_for_ssa(mul_src[i]);
+         for (unsigned j = 0; j < add->dest.dest.ssa.num_components; j++)
+            ffma->src[i].swizzle[j] = mul->src[i].swizzle[swizzle[j]];
+      }
+      nir_alu_src_copy(&ffma->src[2], &add->src[1 - add_mul_src],
+                       state->mem_ctx);
+
+      assert(add->dest.dest.is_ssa);
+
+      nir_ssa_dest_init(&ffma->instr, &ffma->dest.dest,
+                        add->dest.dest.ssa.num_components,
+                        add->dest.dest.ssa.name);
+      nir_ssa_def_rewrite_uses(&add->dest.dest.ssa,
+                               nir_src_for_ssa(&ffma->dest.dest.ssa),
+                               state->mem_ctx);
+
+      nir_instr_insert_before(&add->instr, &ffma->instr);
+      assert(add->dest.dest.ssa.uses->entries == 0);
+      nir_instr_remove(&add->instr);
+
+      state->progress = true;
+   }
+
+   return true;
+}
+
+static bool
+nir_opt_peephole_ffma_impl(nir_function_impl *impl)
+{
+   struct peephole_ffma_state state;
+
+   state.mem_ctx = ralloc_parent(impl);
+   state.impl = impl;
+   state.progress = false;
+
+   nir_foreach_block(impl, nir_opt_peephole_ffma_block, &state);
+
+   if (state.progress)
+      nir_metadata_preserve(impl, nir_metadata_block_index |
+                                  nir_metadata_dominance);
+
+   return state.progress;
+}
+
+bool
+nir_opt_peephole_ffma(nir_shader *shader)
+{
+   bool progress = false;
+
+   nir_foreach_overload(shader, overload) {
+      if (overload->impl)
+         progress |= nir_opt_peephole_ffma_impl(overload->impl);
+   }
+
+   return progress;
+}
diff --git a/mesalib/src/glsl/nir/nir_opt_peephole_select.c b/mesalib/src/glsl/nir/nir_opt_peephole_select.c
index b89451b09..f400cfd66 100644
--- a/mesalib/src/glsl/nir/nir_opt_peephole_select.c
+++ b/mesalib/src/glsl/nir/nir_opt_peephole_select.c
@@ -84,7 +84,9 @@ block_check_for_allowed_instrs(nir_block *block)
       case nir_instr_type_alu: {
          /* It must be a move operation */
          nir_alu_instr *mov = nir_instr_as_alu(instr);
-         if (mov->op != nir_op_fmov && mov->op != nir_op_imov)
+         if (mov->op != nir_op_fmov && mov->op != nir_op_imov &&
+             mov->op != nir_op_fneg && mov->op != nir_op_ineg &&
+             mov->op != nir_op_fabs && mov->op != nir_op_iabs)
             return false;
 
          /* Can't handle saturate */
diff --git a/mesalib/src/glsl/nir/nir_print.c b/mesalib/src/glsl/nir/nir_print.c
index fa11a312e..fb8c9344c 100644
--- a/mesalib/src/glsl/nir/nir_print.c
+++ b/mesalib/src/glsl/nir/nir_print.c
@@ -137,25 +137,37 @@ print_dest(nir_dest *dest, FILE *fp)
 }
 
 static void
-print_alu_src(nir_alu_src *src, FILE *fp)
+print_alu_src(nir_alu_instr *instr, unsigned src, FILE *fp)
 {
-   if (src->negate)
+   if (instr->src[src].negate)
       fprintf(fp, "-");
-   if (src->abs)
+   if (instr->src[src].abs)
       fprintf(fp, "abs(");
 
-   print_src(&src->src, fp);
+   print_src(&instr->src[src].src, fp);
 
-   if (src->swizzle[0] != 0 ||
-       src->swizzle[1] != 1 ||
-       src->swizzle[2] != 2 ||
-       src->swizzle[3] != 3) {
+   bool print_swizzle = false;
+   for (unsigned i = 0; i < 4; i++) {
+      if (!nir_alu_instr_channel_used(instr, src, i))
+         continue;
+
+      if (instr->src[src].swizzle[i] != i) {
+         print_swizzle = true;
+         break;
+      }
+   }
+
+   if (print_swizzle) {
       fprintf(fp, ".");
-      for (unsigned i = 0; i < 4; i++)
-         fprintf(fp, "%c", "xyzw"[src->swizzle[i]]);
+      for (unsigned i = 0; i < 4; i++) {
+         if (!nir_alu_instr_channel_used(instr, src, i))
+            continue;
+
+         fprintf(fp, "%c", "xyzw"[instr->src[src].swizzle[i]]);
+      }
    }
 
-   if (src->abs)
+   if (instr->src[src].abs)
       fprintf(fp, ")");
 }
 
@@ -189,7 +201,7 @@ print_alu_instr(nir_alu_instr *instr, FILE *fp)
       if (i != 0)
          fprintf(fp, ", ");
 
-      print_alu_src(&instr->src[i], fp);
+      print_alu_src(instr, i, fp);
    }
 }
 
diff --git a/mesalib/src/glsl/nir/nir_remove_dead_variables.c b/mesalib/src/glsl/nir/nir_remove_dead_variables.c
index e7f8aeacb..4417e2a48 100644
--- a/mesalib/src/glsl/nir/nir_remove_dead_variables.c
+++ b/mesalib/src/glsl/nir/nir_remove_dead_variables.c
@@ -98,22 +98,14 @@ add_var_use_shader(nir_shader *shader, struct set *live)
 }
 
 static void
-remove_dead_local_vars(nir_function_impl *impl, struct set *live)
+remove_dead_vars(struct exec_list *var_list, struct set *live)
 {
-   foreach_list_typed_safe(nir_variable, var, node, &impl->locals) {
+   foreach_list_typed_safe(nir_variable, var, node, var_list) {
       struct set_entry *entry = _mesa_set_search(live, var);
-      if (entry == NULL)
-         exec_node_remove(&var->node);
-   }
-}
-
-static void
-remove_dead_global_vars(nir_shader *shader, struct set *live)
-{
-   foreach_list_typed_safe(nir_variable, var, node, &shader->globals) {
-      struct set_entry *entry = _mesa_set_search(live, var);
-      if (entry == NULL)
+      if (entry == NULL) {
          exec_node_remove(&var->node);
+         ralloc_free(var);
+      }
    }
 }
 
@@ -125,11 +117,11 @@ nir_remove_dead_variables(nir_shader *shader)
 
    add_var_use_shader(shader, live);
 
-   remove_dead_global_vars(shader, live);
+   remove_dead_vars(&shader->globals, live);
 
    nir_foreach_overload(shader, overload) {
       if (overload->impl)
-         remove_dead_local_vars(overload->impl, live);
+         remove_dead_vars(&overload->impl->locals, live);
    }
 
    _mesa_set_destroy(live, NULL);
diff --git a/mesalib/src/glsl/nir/nir_search.c b/mesalib/src/glsl/nir/nir_search.c
index 73a802be7..5ba016085 100644
--- a/mesalib/src/glsl/nir/nir_search.c
+++ b/mesalib/src/glsl/nir/nir_search.c
@@ -218,8 +218,8 @@ match_expression(const nir_search_expression *expr, nir_alu_instr *instr,
    if (matched)
       return true;
 
-   if (nir_op_infos[instr->op].num_inputs == 2 &&
-       (nir_op_infos[instr->op].algebraic_properties & NIR_OP_IS_COMMUTATIVE)) {
+   if (nir_op_infos[instr->op].algebraic_properties & NIR_OP_IS_COMMUTATIVE) {
+      assert(nir_op_infos[instr->op].num_inputs == 2);
       if (!match_value(expr->srcs[0], instr, 1, num_components,
                        swizzle, state))
          return false;
diff --git a/mesalib/src/glsl/nir/nir_split_var_copies.c b/mesalib/src/glsl/nir/nir_split_var_copies.c
index 4d663b51b..fc72c078c 100644
--- a/mesalib/src/glsl/nir/nir_split_var_copies.c
+++ b/mesalib/src/glsl/nir/nir_split_var_copies.c
@@ -188,8 +188,8 @@ split_var_copy_instr(nir_intrinsic_instr *old_copy,
           * belongs to the copy instruction and b) the deref chains may
           * have some of the same links due to the way we constructed them
           */
-         nir_deref *src = nir_copy_deref(state->mem_ctx, src_head);
-         nir_deref *dest = nir_copy_deref(state->mem_ctx, dest_head);
+         nir_deref *src = nir_copy_deref(new_copy, src_head);
+         nir_deref *dest = nir_copy_deref(new_copy, dest_head);
 
          new_copy->variables[0] = nir_deref_as_var(dest);
          new_copy->variables[1] = nir_deref_as_var(src);
diff --git a/mesalib/src/glsl/nir/nir_sweep.c b/mesalib/src/glsl/nir/nir_sweep.c
new file mode 100644
index 000000000..d3549756a
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_sweep.c
@@ -0,0 +1,172 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "nir.h"
+
+/**
+ * \file nir_sweep.c
+ *
+ * The nir_sweep() pass performs a mark and sweep pass over a nir_shader's associated
+ * memory - anything still connected to the program will be kept, and any dead memory
+ * we dropped on the floor will be freed.
+ *
+ * The expectation is that drivers should call this when finished compiling the shader
+ * (after any optimization, lowering, and so on).  However, it's also fine to call it
+ * earlier, and even many times, trading CPU cycles for memory savings.
+ */
+
+#define steal_list(mem_ctx, type, list) \
+   foreach_list_typed(type, obj, node, list) { ralloc_steal(mem_ctx, obj); }
+
+static void sweep_cf_node(nir_shader *nir, nir_cf_node *cf_node);
+
+static bool
+sweep_src_indirect(nir_src *src, void *nir)
+{
+   if (!src->is_ssa && src->reg.indirect)
+      ralloc_steal(nir, src->reg.indirect);
+
+   return true;
+}
+
+static bool
+sweep_dest_indirect(nir_dest *dest, void *nir)
+{
+   if (!dest->is_ssa && dest->reg.indirect)
+      ralloc_steal(nir, dest->reg.indirect);
+
+   return true;
+}
+
+static void
+sweep_block(nir_shader *nir, nir_block *block)
+{
+   ralloc_steal(nir, block);
+
+   nir_foreach_instr(block, instr) {
+      ralloc_steal(nir, instr);
+
+      nir_foreach_src(instr, sweep_src_indirect, nir);
+      nir_foreach_dest(instr, sweep_dest_indirect, nir);
+   }
+}
+
+static void
+sweep_if(nir_shader *nir, nir_if *iff)
+{
+   ralloc_steal(nir, iff);
+
+   foreach_list_typed(nir_cf_node, cf_node, node, &iff->then_list) {
+      sweep_cf_node(nir, cf_node);
+   }
+
+   foreach_list_typed(nir_cf_node, cf_node, node, &iff->else_list) {
+      sweep_cf_node(nir, cf_node);
+   }
+}
+
+static void
+sweep_loop(nir_shader *nir, nir_loop *loop)
+{
+   ralloc_steal(nir, loop);
+
+   foreach_list_typed(nir_cf_node, cf_node, node, &loop->body) {
+      sweep_cf_node(nir, cf_node);
+   }
+}
+
+static void
+sweep_cf_node(nir_shader *nir, nir_cf_node *cf_node)
+{
+   switch (cf_node->type) {
+   case nir_cf_node_block:
+      sweep_block(nir, nir_cf_node_as_block(cf_node));
+      break;
+   case nir_cf_node_if:
+      sweep_if(nir, nir_cf_node_as_if(cf_node));
+      break;
+   case nir_cf_node_loop:
+      sweep_loop(nir, nir_cf_node_as_loop(cf_node));
+      break;
+   default:
+      unreachable("Invalid CF node type");
+   }
+}
+
+static void
+sweep_impl(nir_shader *nir, nir_function_impl *impl)
+{
+   ralloc_steal(nir, impl);
+
+   ralloc_steal(nir, impl->params);
+   ralloc_steal(nir, impl->return_var);
+   steal_list(nir, nir_variable, &impl->locals);
+   steal_list(nir, nir_register, &impl->registers);
+
+   foreach_list_typed(nir_cf_node, cf_node, node, &impl->body) {
+      sweep_cf_node(nir, cf_node);
+   }
+
+   sweep_block(nir, impl->end_block);
+
+   /* Wipe out all the metadata, if any. */
+   nir_metadata_preserve(impl, nir_metadata_none);
+}
+
+static void
+sweep_function(nir_shader *nir, nir_function *f)
+{
+   ralloc_steal(nir, f);
+
+   foreach_list_typed(nir_function_overload, overload, node, &f->overload_list) {
+      ralloc_steal(nir, overload);
+      ralloc_steal(nir, overload->params);
+      if (overload->impl)
+         sweep_impl(nir, overload->impl);
+   }
+}
+
+void
+nir_sweep(nir_shader *nir)
+{
+   void *rubbish = ralloc_context(NULL);
+
+   /* First, move ownership of all the memory to a temporary context; assume dead. */
+   ralloc_adopt(rubbish, nir);
+
+   /* Variables and registers are not dead.  Steal them back. */
+   steal_list(nir, nir_variable, &nir->uniforms);
+   steal_list(nir, nir_variable, &nir->inputs);
+   steal_list(nir, nir_variable, &nir->outputs);
+   steal_list(nir, nir_variable, &nir->globals);
+   steal_list(nir, nir_variable, &nir->system_values);
+   steal_list(nir, nir_register, &nir->registers);
+
+   /* Recurse into functions, stealing their contents back. */
+   foreach_list_typed(nir_function, func, node, &nir->functions) {
+      sweep_function(nir, func);
+   }
+
+   /* Free everything we didn't steal back. */
+   ralloc_free(rubbish);
+}
diff --git a/mesalib/src/glsl/nir/nir_to_ssa.c b/mesalib/src/glsl/nir/nir_to_ssa.c
index 47cf45393..53ff54766 100644
--- a/mesalib/src/glsl/nir/nir_to_ssa.c
+++ b/mesalib/src/glsl/nir/nir_to_ssa.c
@@ -47,7 +47,7 @@ insert_trivial_phi(nir_register *reg, nir_block *block, void *mem_ctx)
    set_foreach(block->predecessors, entry) {
       nir_block *pred = (nir_block *) entry->key;
 
-      nir_phi_src *src = ralloc(mem_ctx, nir_phi_src);
+      nir_phi_src *src = ralloc(instr, nir_phi_src);
       src->pred = pred;
       src->src.is_ssa = false;
       src->src.reg.base_offset = 0;
diff --git a/mesalib/src/glsl/nir/nir_types.cpp b/mesalib/src/glsl/nir/nir_types.cpp
index a13c3e12a..f0d0b46d2 100644
--- a/mesalib/src/glsl/nir/nir_types.cpp
+++ b/mesalib/src/glsl/nir/nir_types.cpp
@@ -143,6 +143,12 @@ glsl_void_type(void)
 }
 
 const glsl_type *
+glsl_float_type(void)
+{
+   return glsl_type::float_type;
+}
+
+const glsl_type *
 glsl_vec4_type(void)
 {
    return glsl_type::vec4_type;
diff --git a/mesalib/src/glsl/nir/nir_types.h b/mesalib/src/glsl/nir/nir_types.h
index 494051a67..276d4ad62 100644
--- a/mesalib/src/glsl/nir/nir_types.h
+++ b/mesalib/src/glsl/nir/nir_types.h
@@ -69,6 +69,7 @@ bool glsl_type_is_scalar(const struct glsl_type *type);
 bool glsl_type_is_matrix(const struct glsl_type *type);
 
 const struct glsl_type *glsl_void_type(void);
+const struct glsl_type *glsl_float_type(void);
 const struct glsl_type *glsl_vec4_type(void);
 const struct glsl_type *glsl_array_type(const struct glsl_type *base,
                                         unsigned elements);
diff --git a/mesalib/src/glsl/nir/nir_validate.c b/mesalib/src/glsl/nir/nir_validate.c
index f247ae069..a7aa79837 100644
--- a/mesalib/src/glsl/nir/nir_validate.c
+++ b/mesalib/src/glsl/nir/nir_validate.c
@@ -295,6 +295,8 @@ validate_alu_instr(nir_alu_instr *instr, validate_state *state)
 static void
 validate_deref_chain(nir_deref *deref, validate_state *state)
 {
+   assert(deref->child == NULL || ralloc_parent(deref->child) == deref);
+
    nir_deref *parent = NULL;
    while (deref != NULL) {
       switch (deref->deref_type) {
@@ -336,9 +338,10 @@ validate_var_use(nir_variable *var, validate_state *state)
 }
 
 static void
-validate_deref_var(nir_deref_var *deref, validate_state *state)
+validate_deref_var(void *parent_mem_ctx, nir_deref_var *deref, validate_state *state)
 {
    assert(deref != NULL);
+   assert(ralloc_parent(deref) == parent_mem_ctx);
    assert(deref->deref.type == deref->var->type);
 
    validate_var_use(deref->var, state);
@@ -386,7 +389,7 @@ validate_intrinsic_instr(nir_intrinsic_instr *instr, validate_state *state)
 
    unsigned num_vars = nir_intrinsic_infos[instr->intrinsic].num_variables;
    for (unsigned i = 0; i < num_vars; i++) {
-      validate_deref_var(instr->variables[i], state);
+      validate_deref_var(instr, instr->variables[i], state);
    }
 
    switch (instr->intrinsic) {
@@ -423,7 +426,7 @@ validate_tex_instr(nir_tex_instr *instr, validate_state *state)
    }
 
    if (instr->sampler != NULL)
-      validate_deref_var(instr->sampler, state);
+      validate_deref_var(instr, instr->sampler, state);
 }
 
 static void
@@ -438,10 +441,10 @@ validate_call_instr(nir_call_instr *instr, validate_state *state)
 
    for (unsigned i = 0; i < instr->num_params; i++) {
       assert(instr->callee->params[i].type == instr->params[i]->deref.type);
-      validate_deref_var(instr->params[i], state);
+      validate_deref_var(instr, instr->params[i], state);
    }
 
-   validate_deref_var(instr->return_deref, state);
+   validate_deref_var(instr, instr->return_deref, state);
 }
 
 static void
@@ -680,8 +683,7 @@ validate_cf_node(nir_cf_node *node, validate_state *state)
       break;
 
    default:
-      assert(!"Invalid ALU instruction type");
-      break;
+      unreachable("Invalid CF node type");
    }
 }
 
diff --git a/mesalib/src/glsl/opt_algebraic.cpp b/mesalib/src/glsl/opt_algebraic.cpp
index 69c03ea8b..3d2f2ca0b 100644
--- a/mesalib/src/glsl/opt_algebraic.cpp
+++ b/mesalib/src/glsl/opt_algebraic.cpp
@@ -290,6 +290,20 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
    ir_expression *op_expr[4] = {NULL, NULL, NULL, NULL};
    unsigned int i;
 
+   if (ir->operation == ir_binop_mul &&
+       ir->operands[0]->type->is_matrix() &&
+       ir->operands[1]->type->is_vector()) {
+      ir_expression *matrix_mul = ir->operands[0]->as_expression();
+
+      if (matrix_mul && matrix_mul->operation == ir_binop_mul &&
+         matrix_mul->operands[0]->type->is_matrix() &&
+         matrix_mul->operands[1]->type->is_matrix()) {
+
+         return mul(matrix_mul->operands[0],
+                    mul(matrix_mul->operands[1], ir->operands[1]));
+      }
+   }
+
    assert(ir->get_num_operands() <= 4);
    for (i = 0; i < ir->get_num_operands(); i++) {
       if (ir->operands[i]->type->is_matrix())
@@ -421,6 +435,18 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
       break;
    }
 
+   case ir_unop_saturate:
+      if (op_expr[0] && op_expr[0]->operation == ir_binop_add) {
+         ir_expression *b2f_0 = op_expr[0]->operands[0]->as_expression();
+         ir_expression *b2f_1 = op_expr[0]->operands[1]->as_expression();
+
+         if (b2f_0 && b2f_0->operation == ir_unop_b2f &&
+             b2f_1 && b2f_1->operation == ir_unop_b2f) {
+            return b2f(logic_or(b2f_0->operands[0], b2f_1->operands[0]));
+         }
+      }
+      break;
+
    case ir_binop_add:
       if (is_vec_zero(op_const[0]))
 	 return ir->operands[1];
@@ -518,6 +544,10 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
       if (is_vec_negative_one(op_const[1]))
          return neg(ir->operands[0]);
 
+      if (op_expr[0] && op_expr[0]->operation == ir_unop_b2f &&
+          op_expr[1] && op_expr[1]->operation == ir_unop_b2f) {
+         return b2f(logic_and(op_expr[0]->operands[0], op_expr[1]->operands[0]));
+      }
 
       /* Reassociate multiplication of constants so that we can do
        * constant folding.
@@ -544,6 +574,8 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
             continue;
 
          ir_expression *add_expr = floor_expr->operands[0]->as_expression();
+         if (!add_expr)
+            continue;
 
          for (int j = 0; j < 2; j++) {
             ir_expression *abs_expr = add_expr->operands[j]->as_expression();
diff --git a/mesalib/src/glsl/opt_cse.cpp b/mesalib/src/glsl/opt_cse.cpp
index b0b67f496..4b8e9a07b 100644
--- a/mesalib/src/glsl/opt_cse.cpp
+++ b/mesalib/src/glsl/opt_cse.cpp
@@ -63,6 +63,17 @@ public:
       var = NULL;
    }
 
+   void init(ir_instruction *base_ir, ir_rvalue **val)
+   {
+      this->val = val;
+      this->base_ir = base_ir;
+      this->var = NULL;
+
+      assert(val);
+      assert(*val);
+      assert(base_ir);
+   }
+
    /**
     * The pointer to the expression that we might be able to reuse
     *
@@ -116,6 +127,18 @@ private:
    ir_rvalue *try_cse(ir_rvalue *rvalue);
    void add_to_ae(ir_rvalue **rvalue);
 
+   /**
+    * Move all nodes from the ae list to the free list
+    */
+   void empty_ae_list();
+
+   /**
+    * Get and initialize a new ae_entry
+    *
+    * This will either come from the free list or be freshly allocated.
+    */
+   ae_entry *get_ae_entry(ir_rvalue **rvalue);
+
    /** List of ae_entry: The available expressions to reuse */
    exec_list *ae;
 
@@ -126,6 +149,11 @@ private:
     * right.
     */
    exec_list *validate_instructions;
+
+   /**
+    * List of available-for-use ae_entry objects.
+    */
+   exec_list free_ae_entries;
 };
 
 /**
@@ -322,6 +350,25 @@ cse_visitor::try_cse(ir_rvalue *rvalue)
    return NULL;
 }
 
+void
+cse_visitor::empty_ae_list()
+{
+   free_ae_entries.append_list(ae);
+}
+
+ae_entry *
+cse_visitor::get_ae_entry(ir_rvalue **rvalue)
+{
+   ae_entry *entry = (ae_entry *) free_ae_entries.pop_head();
+   if (entry) {
+      entry->init(base_ir, rvalue);
+   } else {
+      entry = new(mem_ctx) ae_entry(base_ir, rvalue);
+   }
+
+   return entry;
+}
+
 /** Add the rvalue to the list of available expressions for CSE. */
 void
 cse_visitor::add_to_ae(ir_rvalue **rvalue)
@@ -332,7 +379,7 @@ cse_visitor::add_to_ae(ir_rvalue **rvalue)
       printf("\n");
    }
 
-   ae->push_tail(new(mem_ctx) ae_entry(base_ir, rvalue));
+   ae->push_tail(get_ae_entry(rvalue));
 
    if (debug)
       dump_ae(ae);
@@ -370,33 +417,33 @@ cse_visitor::visit_enter(ir_if *ir)
 {
    handle_rvalue(&ir->condition);
 
-   ae->make_empty();
+   empty_ae_list();
    visit_list_elements(this, &ir->then_instructions);
 
-   ae->make_empty();
+   empty_ae_list();
    visit_list_elements(this, &ir->else_instructions);
 
-   ae->make_empty();
+   empty_ae_list();
    return visit_continue_with_parent;
 }
 
 ir_visitor_status
 cse_visitor::visit_enter(ir_function_signature *ir)
 {
-   ae->make_empty();
+   empty_ae_list();
    visit_list_elements(this, &ir->body);
 
-   ae->make_empty();
+   empty_ae_list();
    return visit_continue_with_parent;
 }
 
 ir_visitor_status
 cse_visitor::visit_enter(ir_loop *ir)
 {
-   ae->make_empty();
+   empty_ae_list();
    visit_list_elements(this, &ir->body_instructions);
 
-   ae->make_empty();
+   empty_ae_list();
    return visit_continue_with_parent;
 }
 
diff --git a/mesalib/src/glsl/s_expression.cpp b/mesalib/src/glsl/s_expression.cpp
index 7eaa491e2..f82e155a6 100644
--- a/mesalib/src/glsl/s_expression.cpp
+++ b/mesalib/src/glsl/s_expression.cpp
@@ -23,8 +23,8 @@
  */
 
 #include <assert.h>
-#include <limits>
 #include <stdio.h>
+#include <math.h>
 #include "s_expression.h"
 
 s_symbol::s_symbol(const char *str, size_t n)
@@ -70,7 +70,7 @@ read_atom(void *ctx, const char *&src, char *&symbol_buffer)
    // requires strtof to parse '+INF' as +Infinity, but we still support some
    // non-C99-compliant compilers (e.g. MSVC).
    if (n == 4 && strncmp(src, "+INF", 4) == 0) {
-      expr = new(ctx) s_float(std::numeric_limits<float>::infinity());
+      expr = new(ctx) s_float(INFINITY);
    } else {
       // Check if the atom is a number.
       char *float_end = NULL;
diff --git a/mesalib/src/glsl/shader_enums.h b/mesalib/src/glsl/shader_enums.h
new file mode 100644
index 000000000..79e0f6b5f
--- /dev/null
+++ b/mesalib/src/glsl/shader_enums.h
@@ -0,0 +1,187 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
+ * Copyright (C) 2009  VMware, Inc.  All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef SHADER_ENUMS_H
+#define SHADER_ENUMS_H
+
+/**
+ * Shader stages. Note that these will become 5 with tessellation.
+ *
+ * The order must match how shaders are ordered in the pipeline.
+ * The GLSL linker assumes that if i<j, then the j-th shader is
+ * executed later than the i-th shader.
+ */
+typedef enum
+{
+   MESA_SHADER_VERTEX = 0,
+   MESA_SHADER_GEOMETRY = 1,
+   MESA_SHADER_FRAGMENT = 2,
+   MESA_SHADER_COMPUTE = 3,
+} gl_shader_stage;
+
+#define MESA_SHADER_STAGES (MESA_SHADER_COMPUTE + 1)
+
+/**
+ * Bitflags for system values.
+ */
+#define SYSTEM_BIT_SAMPLE_ID ((uint64_t)1 << SYSTEM_VALUE_SAMPLE_ID)
+#define SYSTEM_BIT_SAMPLE_POS ((uint64_t)1 << SYSTEM_VALUE_SAMPLE_POS)
+#define SYSTEM_BIT_SAMPLE_MASK_IN ((uint64_t)1 << SYSTEM_VALUE_SAMPLE_MASK_IN)
+/**
+ * If the gl_register_file is PROGRAM_SYSTEM_VALUE, the register index will be
+ * one of these values.  If a NIR variable's mode is nir_var_system_value, it
+ * will be one of these values.
+ */
+typedef enum
+{
+   /**
+    * \name Vertex shader system values
+    */
+   /*@{*/
+   /**
+    * OpenGL-style vertex ID.
+    *
+    * Section 2.11.7 (Shader Execution), subsection Shader Inputs, of the
+    * OpenGL 3.3 core profile spec says:
+    *
+    *     "gl_VertexID holds the integer index i implicitly passed by
+    *     DrawArrays or one of the other drawing commands defined in section
+    *     2.8.3."
+    *
+    * Section 2.8.3 (Drawing Commands) of the same spec says:
+    *
+    *     "The commands....are equivalent to the commands with the same base
+    *     name (without the BaseVertex suffix), except that the ith element
+    *     transferred by the corresponding draw call will be taken from
+    *     element indices[i] + basevertex of each enabled array."
+    *
+    * Additionally, the overview in the GL_ARB_shader_draw_parameters spec
+    * says:
+    *
+    *     "In unextended GL, vertex shaders have inputs named gl_VertexID and
+    *     gl_InstanceID, which contain, respectively the index of the vertex
+    *     and instance. The value of gl_VertexID is the implicitly passed
+    *     index of the vertex being processed, which includes the value of
+    *     baseVertex, for those commands that accept it."
+    *
+    * gl_VertexID gets basevertex added in.  This differs from DirectX where
+    * SV_VertexID does \b not get basevertex added in.
+    *
+    * \note
+    * If all system values are available, \c SYSTEM_VALUE_VERTEX_ID will be
+    * equal to \c SYSTEM_VALUE_VERTEX_ID_ZERO_BASE plus
+    * \c SYSTEM_VALUE_BASE_VERTEX.
+    *
+    * \sa SYSTEM_VALUE_VERTEX_ID_ZERO_BASE, SYSTEM_VALUE_BASE_VERTEX
+    */
+   SYSTEM_VALUE_VERTEX_ID,
+
+   /**
+    * Instanced ID as supplied to gl_InstanceID
+    *
+    * Values assigned to gl_InstanceID always begin with zero, regardless of
+    * the value of baseinstance.
+    *
+    * Section 11.1.3.9 (Shader Inputs) of the OpenGL 4.4 core profile spec
+    * says:
+    *
+    *     "gl_InstanceID holds the integer instance number of the current
+    *     primitive in an instanced draw call (see section 10.5)."
+    *
+    * Through a big chain of pseudocode, section 10.5 describes that
+    * baseinstance is not counted by gl_InstanceID.  In that section, notice
+    *
+    *     "If an enabled vertex attribute array is instanced (it has a
+    *     non-zero divisor as specified by VertexAttribDivisor), the element
+    *     index that is transferred to the GL, for all vertices, is given by
+    *
+    *         floor(instance/divisor) + baseinstance
+    *
+    *     If an array corresponding to an attribute required by a vertex
+    *     shader is not enabled, then the corresponding element is taken from
+    *     the current attribute state (see section 10.2)."
+    *
+    * Note that baseinstance is \b not included in the value of instance.
+    */
+   SYSTEM_VALUE_INSTANCE_ID,
+
+   /**
+    * DirectX-style vertex ID.
+    *
+    * Unlike \c SYSTEM_VALUE_VERTEX_ID, this system value does \b not include
+    * the value of basevertex.
+    *
+    * \sa SYSTEM_VALUE_VERTEX_ID, SYSTEM_VALUE_BASE_VERTEX
+    */
+   SYSTEM_VALUE_VERTEX_ID_ZERO_BASE,
+
+   /**
+    * Value of \c basevertex passed to \c glDrawElementsBaseVertex and similar
+    * functions.
+    *
+    * \sa SYSTEM_VALUE_VERTEX_ID, SYSTEM_VALUE_VERTEX_ID_ZERO_BASE
+    */
+   SYSTEM_VALUE_BASE_VERTEX,
+   /*@}*/
+
+   /**
+    * \name Geometry shader system values
+    */
+   /*@{*/
+   SYSTEM_VALUE_INVOCATION_ID,
+   /*@}*/
+
+   /**
+    * \name Fragment shader system values
+    */
+   /*@{*/
+   SYSTEM_VALUE_FRONT_FACE,     /**< (not done yet) */
+   SYSTEM_VALUE_SAMPLE_ID,
+   SYSTEM_VALUE_SAMPLE_POS,
+   SYSTEM_VALUE_SAMPLE_MASK_IN,
+   /*@}*/
+
+   SYSTEM_VALUE_MAX             /**< Number of values */
+} gl_system_value;
+
+
+/**
+ * The possible interpolation qualifiers that can be applied to a fragment
+ * shader input in GLSL.
+ *
+ * Note: INTERP_QUALIFIER_NONE must be 0 so that memsetting the
+ * gl_fragment_program data structure to 0 causes the default behavior.
+ */
+enum glsl_interp_qualifier
+{
+   INTERP_QUALIFIER_NONE = 0,
+   INTERP_QUALIFIER_SMOOTH,
+   INTERP_QUALIFIER_FLAT,
+   INTERP_QUALIFIER_NOPERSPECTIVE,
+   INTERP_QUALIFIER_COUNT /**< Number of interpolation qualifiers */
+};
+
+
+#endif /* SHADER_ENUMS_H */
diff --git a/mesalib/src/loader/Makefile.am b/mesalib/src/loader/Makefile.am
index 3d32279ea..36ddba82b 100755
--- a/mesalib/src/loader/Makefile.am
+++ b/mesalib/src/loader/Makefile.am
@@ -30,7 +30,6 @@ libloader_la_CPPFLAGS = \
 	-I$(top_srcdir)/include \
 	-I$(top_srcdir)/src \
 	$(VISIBILITY_CFLAGS) \
-	$(MSVC2013_COMPAT_CFLAGS) \
 	$(LIBUDEV_CFLAGS)
 
 libloader_la_SOURCES = $(LOADER_C_FILES)
diff --git a/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml b/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
index 641e68f71..8a092d697 100644
--- a/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
+++ b/mesalib/src/mapi/glapi/gen/ARB_direct_state_access.xml
@@ -7,6 +7,47 @@
    <enum name="QUERY_TARGET"    value="0x82EA"/>
    <enum name="TEXTURE_BINDING" value="0x82EB"/>
 
+   <!-- Transform Feedback object functions -->
+
+  <function name="CreateTransformFeedbacks" offset="assign">
+      <param name="n" type="GLsizei" />
+      <param name="ids" type="GLuint *" />
+   </function>
+
+   <function name="TransformFeedbackBufferBase" offset="assign">
+      <param name="xfb" type="GLuint" />
+      <param name="index" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+   </function>
+
+   <function name="TransformFeedbackBufferRange" offset="assign">
+      <param name="xfb" type="GLuint" />
+      <param name="index" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+      <param name="offset" type="GLintptr" />
+      <param name="size" type="GLsizeiptr" />
+   </function>
+
+   <function name="GetTransformFeedbackiv" offset="assign">
+      <param name="xfb" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="param" type="GLint *" />
+   </function>
+
+   <function name="GetTransformFeedbacki_v" offset="assign">
+      <param name="xfb" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="index" type="GLuint" />
+      <param name="param" type="GLint *" />
+   </function>
+
+   <function name="GetTransformFeedbacki64_v" offset="assign">
+      <param name="xfb" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="index" type="GLuint" />
+      <param name="param" type="GLint64 *" />
+   </function>
+
    <!-- Buffer object functions -->
 
    <function name="CreateBuffers" offset="assign">
@@ -111,6 +152,34 @@
       <param name="data" type="GLvoid *" />
    </function>
 
+   <!-- Renderbuffer object functions -->
+
+   <function name="CreateRenderbuffers" offset="assign">
+      <param name="n" type="GLsizei" />
+      <param name="renderbuffers" type="GLuint *" />
+   </function>
+
+   <function name="NamedRenderbufferStorage" offset="assign">
+      <param name="renderbuffer" type="GLuint" />
+      <param name="internalformat" type="GLenum" />
+      <param name="width" type="GLsizei" />
+      <param name="height" type="GLsizei" />
+   </function>
+
+   <function name="NamedRenderbufferStorageMultisample" offset="assign">
+      <param name="renderbuffer" type="GLuint" />
+      <param name="samples" type="GLsizei" />
+      <param name="internalformat" type="GLenum" />
+      <param name="width" type="GLsizei" />
+      <param name="height" type="GLsizei" />
+   </function>
+
+   <function name="GetNamedRenderbufferParameteriv" offset="assign">
+      <param name="renderbuffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="params" type="GLint *" />
+   </function>
+
    <!-- Texture object functions -->
 
    <function name="CreateTextures" offset="assign">
@@ -379,5 +448,54 @@
       <param name="params" type="GLint *" />
    </function>
 
+   <!-- Sampler object functions -->
+
+   <function name="CreateSamplers" offset="assign">
+      <param name="n" type="GLsizei" />
+      <param name="samplers" type="GLuint *" />
+   </function>
+
+   <!-- Program Pipeline object functions -->
+
+   <function name="CreateProgramPipelines" offset="assign">
+      <param name="n" type="GLsizei" />
+      <param name="pipelines" type="GLuint *" />
+   </function>
+
+   <!-- Query object functions -->
+
+   <function name="CreateQueries" offset="assign">
+      <param name="target" type="GLenum" />
+      <param name="n" type="GLsizei" />
+      <param name="ids" type="GLuint *" />
+   </function>
+
+   <function name="GetQueryBufferObjectiv" offset="assign">
+      <param name="id" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="offset" type="GLintptr" />
+   </function>
+
+   <function name="GetQueryBufferObjectuiv" offset="assign">
+      <param name="id" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="offset" type="GLintptr" />
+   </function>
+
+   <function name="GetQueryBufferObjecti64v" offset="assign">
+      <param name="id" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="offset" type="GLintptr" />
+   </function>
+
+   <function name="GetQueryBufferObjectui64v" offset="assign">
+      <param name="id" type="GLuint" />
+      <param name="buffer" type="GLuint" />
+      <param name="pname" type="GLenum" />
+      <param name="offset" type="GLintptr" />
+   </function>
 </category>
 </OpenGLAPI>
diff --git a/mesalib/src/mapi/glapi/gen/ARB_program_interface_query.xml b/mesalib/src/mapi/glapi/gen/ARB_program_interface_query.xml
new file mode 100644
index 000000000..59eb59c64
--- /dev/null
+++ b/mesalib/src/mapi/glapi/gen/ARB_program_interface_query.xml
@@ -0,0 +1,109 @@
+<?xml version="1.0"?>
+<!DOCTYPE OpenGLAPI SYSTEM "gl_API.dtd">
+
+<OpenGLAPI>
+
+<category name="GL_ARB_program_interface_query" number="134">
+
+    <enum name="UNIFORM"                                         value="0x92E1"/>
+    <enum name="UNIFORM_BLOCK"                                   value="0x92E2"/>
+    <enum name="PROGRAM_INPUT"                                   value="0x92E3"/>
+    <enum name="PROGRAM_OUTPUT"                                  value="0x92E4"/>
+    <enum name="BUFFER_VARIABLE"                                 value="0x92E5"/>
+    <enum name="SHADER_STORAGE_BLOCK"                            value="0x92E6"/>
+    <enum name="ATOMIC_COUNTER_BUFFER"                           value="0x92C0"/>
+    <enum name="VERTEX_SUBROUTINE"                               value="0x92E8"/>
+    <enum name="TESS_CONTROL_SUBROUTINE"                         value="0x92E9"/>
+    <enum name="TESS_EVALUATION_SUBROUTINE"                      value="0x92EA"/>
+    <enum name="GEOMETRY_SUBROUTINE"                             value="0x92EB"/>
+    <enum name="FRAGMENT_SUBROUTINE"                             value="0x92EC"/>
+    <enum name="COMPUTE_SUBROUTINE"                              value="0x92ED"/>
+    <enum name="VERTEX_SUBROUTINE_UNIFORM"                       value="0x92EE"/>
+    <enum name="TESS_CONTROL_SUBROUTINE_UNIFORM"                 value="0x92EF"/>
+    <enum name="TESS_EVALUATION_SUBROUTINE_UNIFORM"              value="0x92F0"/>
+    <enum name="GEOMETRY_SUBROUTINE_UNIFORM"                     value="0x92F1"/>
+    <enum name="FRAGMENT_SUBROUTINE_UNIFORM"                     value="0x92F2"/>
+    <enum name="COMPUTE_SUBROUTINE_UNIFORM"                      value="0x92F3"/>
+    <enum name="TRANSFORM_FEEDBACK_VARYING"                      value="0x92F4"/>
+    <enum name="ACTIVE_RESOURCES"                                value="0x92F5"/>
+    <enum name="MAX_NAME_LENGTH"                                 value="0x92F6"/>
+    <enum name="MAX_NUM_ACTIVE_VARIABLES"                        value="0x92F7"/>
+    <enum name="MAX_NUM_COMPATIBLE_SUBROUTINES"                  value="0x92F8"/>
+    <enum name="NAME_LENGTH"                                     value="0x92F9"/>
+    <enum name="TYPE"                                            value="0x92FA"/>
+    <enum name="ARRAY_SIZE"                                      value="0x92FB"/>
+    <enum name="OFFSET"                                          value="0x92FC"/>
+    <enum name="BLOCK_INDEX"                                     value="0x92FD"/>
+    <enum name="ARRAY_STRIDE"                                    value="0x92FE"/>
+    <enum name="MATRIX_STRIDE"                                   value="0x92FF"/>
+    <enum name="IS_ROW_MAJOR"                                    value="0x9300"/>
+    <enum name="ATOMIC_COUNTER_BUFFER_INDEX"                     value="0x9301"/>
+    <enum name="BUFFER_BINDING"                                  value="0x9302"/>
+    <enum name="BUFFER_DATA_SIZE"                                value="0x9303"/>
+    <enum name="NUM_ACTIVE_VARIABLES"                            value="0x9304"/>
+    <enum name="ACTIVE_VARIABLES"                                value="0x9305"/>
+    <enum name="REFERENCED_BY_VERTEX_SHADER"                     value="0x9306"/>
+    <enum name="REFERENCED_BY_TESS_CONTROL_SHADER"               value="0x9307"/>
+    <enum name="REFERENCED_BY_TESS_EVALUATION_SHADER"            value="0x9308"/>
+    <enum name="REFERENCED_BY_GEOMETRY_SHADER"                   value="0x9309"/>
+    <enum name="REFERENCED_BY_FRAGMENT_SHADER"                   value="0x930A"/>
+    <enum name="REFERENCED_BY_COMPUTE_SHADER"                    value="0x930B"/>
+    <enum name="TOP_LEVEL_ARRAY_SIZE"                            value="0x930C"/>
+    <enum name="TOP_LEVEL_ARRAY_STRIDE"                          value="0x930D"/>
+    <enum name="LOCATION"                                        value="0x930E"/>
+    <enum name="LOCATION_INDEX"                                  value="0x930F"/>
+    <enum name="IS_PER_PATCH"                                    value="0x92E7"/>
+    <enum name="NUM_COMPATIBLE_SUBROUTINES"                      value="0x8E4A"/>
+    <enum name="COMPATIBLE_SUBROUTINES"                          value="0x8E4B"/>
+
+    <function name="GetProgramInterfaceiv" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="pname" type="GLenum"/>
+        <param name="params" type="GLint *" output="true"/>
+    </function>
+
+    <function name="GetProgramResourceIndex" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="name" type="const GLchar *"/>
+        <return type="GLuint"/>
+    </function>
+
+    <function name="GetProgramResourceName" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="index" type="GLuint"/>
+        <param name="bufSize" type="GLsizei "/>
+        <param name="length" type="GLsizei *" output="true"/>
+        <param name="name" type="GLchar *" output="true"/>
+    </function>
+
+    <function name="GetProgramResourceiv" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="index" type="GLuint"/>
+        <param name="propCount" type="GLsizei "/>
+        <param name="props" type="const GLenum *" output="true"/>
+        <param name="bufSize" type="GLsizei "/>
+        <param name="length" type="GLsizei *" output="true"/>
+        <param name="params" type="GLint *" output="true"/>
+    </function>
+
+    <function name="GetProgramResourceLocation" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="name" type="const GLchar *"/>
+        <return type="GLint"/>
+    </function>
+
+    <function name="GetProgramResourceLocationIndex" offset="assign">
+        <param name="program" type="GLuint"/>
+        <param name="programInterface" type="GLenum"/>
+        <param name="name" type="const GLchar *"/>
+        <return type="GLint"/>
+    </function>
+
+</category>
+
+</OpenGLAPI>
diff --git a/mesalib/src/mapi/glapi/gen/gl_API.xml b/mesalib/src/mapi/glapi/gen/gl_API.xml
index a640241e9..5fa2e14a1 100755
--- a/mesalib/src/mapi/glapi/gen/gl_API.xml
+++ b/mesalib/src/mapi/glapi/gen/gl_API.xml
@@ -8337,7 +8337,9 @@
 <!-- ARB extension #133 is ARB_multi_draw_indirect, defined in the same
      file as ARB_draw_indirect -->
 
-<!-- ARB extensions #134...#138 -->
+<xi:include href="ARB_program_interface_query.xml" xmlns:xi="http://www.w3.org/2001/XInclude"/>
+
+<!-- ARB extensions #135...#138 -->
 
 <xi:include href="ARB_texture_buffer_range.xml" xmlns:xi="http://www.w3.org/2001/XInclude"/>
 
diff --git a/mesalib/src/mapi/glapi/gen/gl_genexec.py b/mesalib/src/mapi/glapi/gen/gl_genexec.py
index d479e66da..7151f0de1 100644
--- a/mesalib/src/mapi/glapi/gen/gl_genexec.py
+++ b/mesalib/src/mapi/glapi/gen/gl_genexec.py
@@ -92,6 +92,7 @@ header = """/**
 #include "main/pixelstore.h"
 #include "main/points.h"
 #include "main/polygon.h"
+#include "main/program_resource.h"
 #include "main/querymatrix.h"
 #include "main/queryobj.h"
 #include "main/readpix.h"
diff --git a/mesalib/src/mapi/stub.c b/mesalib/src/mapi/stub.c
index 05436bab6..45e4f7dc6 100644
--- a/mesalib/src/mapi/stub.c
+++ b/mesalib/src/mapi/stub.c
@@ -102,7 +102,7 @@ stub_add_dynamic(const char *name)
    if (!stub->addr)
       return NULL;
 
-   stub->name = (const void *) name;
+   stub->name = (const void *) strdup(name);
    /* to be fixed later */
    stub->slot = -1;
 
diff --git a/mesalib/src/mesa/Makefile.am b/mesalib/src/mesa/Makefile.am
index 3dab8f03f..60114e4f6 100644
--- a/mesalib/src/mesa/Makefile.am
+++ b/mesalib/src/mesa/Makefile.am
@@ -174,6 +174,7 @@ endif
 libmesa_la_SOURCES = \
 	$(MESA_FILES) \
 	$(PROGRAM_FILES) \
+	$(PROGRAM_NIR_FILES) \
 	$(MESA_ASM_FILES_FOR_ARCH)
 
 libmesa_la_LIBADD = \
@@ -183,6 +184,7 @@ libmesa_la_LIBADD = \
 libmesagallium_la_SOURCES = \
 	$(MESA_GALLIUM_FILES) \
 	$(PROGRAM_FILES) \
+	$(PROGRAM_NIR_FILES) \
 	$(MESA_ASM_FILES_FOR_ARCH)
 
 libmesagallium_la_LIBADD = \
diff --git a/mesalib/src/mesa/Makefile.sources b/mesalib/src/mesa/Makefile.sources
index 217be9ab3..1293d4135 100644
--- a/mesalib/src/mesa/Makefile.sources
+++ b/mesalib/src/mesa/Makefile.sources
@@ -152,6 +152,8 @@ MAIN_FILES = \
 	main/points.h \
 	main/polygon.c \
 	main/polygon.h \
+	main/program_resource.c \
+	main/program_resource.h \
 	main/querymatrix.c \
 	main/querymatrix.h \
 	main/queryobj.c \
@@ -520,6 +522,10 @@ PROGRAM_FILES = \
 	program/symbol_table.c \
 	program/symbol_table.h
 
+PROGRAM_NIR_FILES = \
+	program/prog_to_nir.c \
+	program/prog_to_nir.h
+
 ASM_C_FILES =	\
 	x86/common_x86.c \
 	x86/x86_xform.c \
@@ -608,6 +614,7 @@ INCLUDE_DIRS = \
 	-I$(top_srcdir)/src \
 	-I$(top_srcdir)/src/glsl \
 	-I$(top_builddir)/src/glsl \
+	-I$(top_builddir)/src/glsl/nir \
 	-I$(top_srcdir)/src/glsl/glcpp \
 	-I$(top_srcdir)/src/mesa \
 	-I$(top_builddir)/src/mesa \
diff --git a/mesalib/src/mesa/SConscript b/mesalib/src/mesa/SConscript
index a563fd2ff..5b80a216f 100644
--- a/mesalib/src/mesa/SConscript
+++ b/mesalib/src/mesa/SConscript
@@ -130,9 +130,10 @@ def write_git_sha1_h_file(filename):
         (commit, foo) = subprocess.Popen(args, stdout=subprocess.PIPE).communicate()
     except:
         # git log command didn't work
-        dirname = os.path.dirname(filename)
-        if not os.path.exists(dirname):
-            os.makedirs(dirname)
+        if not os.path.exists(filename):
+            dirname = os.path.dirname(filename)
+            if not os.path.exists(dirname):
+                os.makedirs(dirname)
             # create an empty file if none already exists
             f = open(filename, "w")
             f.close()
diff --git a/mesalib/src/mesa/drivers/SConscript b/mesalib/src/mesa/drivers/SConscript
index 17813da96..db656780c 100644
--- a/mesalib/src/mesa/drivers/SConscript
+++ b/mesalib/src/mesa/drivers/SConscript
@@ -1,7 +1,5 @@
 Import('*')
 
-SConscript('osmesa/SConscript')
-
 if env['x11']:
     SConscript('x11/SConscript')
 
diff --git a/mesalib/src/mesa/drivers/dri/common/utils.c b/mesalib/src/mesa/drivers/dri/common/utils.c
index bb22107e9..70d34e8ce 100644
--- a/mesalib/src/mesa/drivers/dri/common/utils.c
+++ b/mesalib/src/mesa/drivers/dri/common/utils.c
@@ -227,7 +227,7 @@ driCreateConfigs(mesa_format format,
       break;
    default:
       fprintf(stderr, "[%s:%u] Unknown framebuffer type %s (%d).\n",
-              __FUNCTION__, __LINE__,
+              __func__, __LINE__,
               _mesa_get_format_name(format), format);
       return NULL;
    }
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlconfig.c b/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
index a90a97fa9..7ae38d07a 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
+++ b/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
@@ -56,7 +56,7 @@ extern char *program_invocation_name, *program_invocation_short_name;
 #        include <stdlib.h>
 #        define GET_PROGRAM_NAME() getprogname()
 #    endif
-#elif defined(__NetBSD__) && defined(__NetBSD_Version) && (__NetBSD_Version >= 106000100)
+#elif defined(__NetBSD__) && defined(__NetBSD_Version__) && (__NetBSD_Version__ >= 106000100)
 #    include <stdlib.h>
 #    define GET_PROGRAM_NAME() getprogname()
 #elif defined(__APPLE__)
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
index 555771685..a6f1652d1 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
+++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
@@ -52,14 +52,19 @@ POT=xmlpool.pot
 
 .PHONY: all clean pot po mo
 
-EXTRA_DIST = gen_xmlpool.py options.h t_options.h $(POS) SConscript
-BUILT_SOURCES = options.h
-CLEANFILES = $(MOS) options.h
+EXTRA_DIST = \
+	gen_xmlpool.py \
+	options.h \
+	t_options.h \
+	$(POS) \
+	$(MOS) \
+	SConscript
 
-# All generated files are cleaned up.
-clean:
-	-rm -f $(POT) options.h *~
-	-rm -rf $(LANGS)
+BUILT_SOURCES = options.h
+CLEANFILES = \
+	options.h
+	$(POS) \
+	$(MOS)
 
 # Default target options.h
 options.h: LOCALEDIR := .
diff --git a/mesalib/src/mesa/drivers/dri/swrast/swrast.c b/mesalib/src/mesa/drivers/dri/swrast/swrast.c
index e84059073..d07988a40 100755
--- a/mesalib/src/mesa/drivers/dri/swrast/swrast.c
+++ b/mesalib/src/mesa/drivers/dri/swrast/swrast.c
@@ -335,7 +335,7 @@ choose_pixel_format(const struct gl_config *v)
 	     && v->blueMask  == 0xc0)
 	return PF_R3G3B2;
 
-    _mesa_problem( NULL, "unexpected format in %s", __FUNCTION__ );
+    _mesa_problem( NULL, "unexpected format in %s", __func__ );
     return 0;
 }
 
diff --git a/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h b/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
index 8dccc2265..8182e7677 100644
--- a/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
+++ b/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
@@ -48,13 +48,13 @@
 #define DEBUG_SPAN	0
 
 #if DEBUG_CORE
-#define TRACE printf("--> %s\n", __FUNCTION__)
+#define TRACE printf("--> %s\n", __func__)
 #else
 #define TRACE
 #endif
 
 #if DEBUG_SPAN
-#define TRACE_SPAN printf("--> %s\n", __FUNCTION__)
+#define TRACE_SPAN printf("--> %s\n", __func__)
 #else
 #define TRACE_SPAN
 #endif
diff --git a/mesalib/src/mesa/main/.gitignore b/mesalib/src/mesa/main/.gitignore
index 54885405f..58c73d333 100755
--- a/mesalib/src/mesa/main/.gitignore
+++ b/mesalib/src/mesa/main/.gitignore
@@ -8,5 +8,6 @@ get_hash.h
 get_hash.h.tmp

 format_info.h

 glapitable.h

+format_info.c

 format_pack.c

 format_unpack.c

diff --git a/mesalib/src/mesa/main/atifragshader.c b/mesalib/src/mesa/main/atifragshader.c
index 9d967b9e6..9fc35520a 100644
--- a/mesalib/src/mesa/main/atifragshader.c
+++ b/mesalib/src/mesa/main/atifragshader.c
@@ -476,7 +476,7 @@ _mesa_PassTexCoordATI(GLuint dst, GLuint coord, GLenum swizzle)
    curI->swizzle = swizzle;
 
 #if MESA_DEBUG_ATI_FS
-   _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__,
+   _mesa_debug(ctx, "%s(%s, %s, %s)\n", __func__,
 	       _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(coord),
 	       _mesa_lookup_enum_by_nr(swizzle));
 #endif
@@ -549,7 +549,7 @@ _mesa_SampleMapATI(GLuint dst, GLuint interp, GLenum swizzle)
    curI->swizzle = swizzle;
 
 #if MESA_DEBUG_ATI_FS
-   _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__,
+   _mesa_debug(ctx, "%s(%s, %s, %s)\n", __func__,
 	       _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(interp),
 	       _mesa_lookup_enum_by_nr(swizzle));
 #endif
diff --git a/mesalib/src/mesa/main/attrib.c b/mesalib/src/mesa/main/attrib.c
index 20216a87d..b163c0aa6 100644
--- a/mesalib/src/mesa/main/attrib.c
+++ b/mesalib/src/mesa/main/attrib.c
@@ -32,7 +32,6 @@
 #include "buffers.h"
 #include "bufferobj.h"
 #include "clear.h"
-#include "colormac.h"
 #include "context.h"
 #include "depth.h"
 #include "enable.h"
diff --git a/mesalib/src/mesa/main/bufferobj.c b/mesalib/src/mesa/main/bufferobj.c
index 25110c33b..4a116acfe 100755
--- a/mesalib/src/mesa/main/bufferobj.c
+++ b/mesalib/src/mesa/main/bufferobj.c
@@ -1333,6 +1333,7 @@ create_buffers(GLsizei n, GLuint *buffers, bool dsa)
          buf = ctx->Driver.NewBufferObject(ctx, buffers[i]);
          if (!buf) {
             _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
+            mtx_unlock(&ctx->Shared->Mutex);
             return;
          }
       }
@@ -3885,8 +3886,10 @@ _mesa_BindBufferRange(GLenum target, GLuint index,
 
    switch (target) {
    case GL_TRANSFORM_FEEDBACK_BUFFER:
-      _mesa_bind_buffer_range_transform_feedback(ctx, index, bufObj,
-						 offset, size);
+      _mesa_bind_buffer_range_transform_feedback(ctx,
+                                                 ctx->TransformFeedback.CurrentObject,
+                                                 index, bufObj, offset, size,
+                                                 false);
       return;
    case GL_UNIFORM_BUFFER:
       bind_buffer_range_uniform_buffer(ctx, index, bufObj, offset, size);
@@ -3950,7 +3953,9 @@ _mesa_BindBufferBase(GLenum target, GLuint index, GLuint buffer)
 
    switch (target) {
    case GL_TRANSFORM_FEEDBACK_BUFFER:
-      _mesa_bind_buffer_base_transform_feedback(ctx, index, bufObj);
+      _mesa_bind_buffer_base_transform_feedback(ctx,
+                                                ctx->TransformFeedback.CurrentObject,
+                                                index, bufObj, false);
       return;
    case GL_UNIFORM_BUFFER:
       bind_buffer_base_uniform_buffer(ctx, index, bufObj);
diff --git a/mesalib/src/mesa/main/colormac.h b/mesalib/src/mesa/main/colormac.h
index bc69f4673..33ca5af07 100644
--- a/mesalib/src/mesa/main/colormac.h
+++ b/mesalib/src/mesa/main/colormac.h
@@ -52,70 +52,25 @@ _mesa_unclamped_float_rgba_to_ubyte(GLubyte dst[4], const GLfloat src[4])
 
 /**
  * \name Generic color packing macros.  All inputs should be GLubytes.
- *
- * \todo We may move these into texstore.h at some point.
  */
 /*@{*/
 
 #define PACK_COLOR_8888( X, Y, Z, W ) \
    (((X) << 24) | ((Y) << 16) | ((Z) << 8) | (W))
 
-#define PACK_COLOR_8888_REV( X, Y, Z, W ) \
-   (((W) << 24) | ((Z) << 16) | ((Y) << 8) | (X))
-
-#define PACK_COLOR_888( X, Y, Z ) \
-   (((X) << 16) | ((Y) << 8) | (Z))
-
 #define PACK_COLOR_565( X, Y, Z )                                  \
    ((((X) & 0xf8) << 8) | (((Y) & 0xfc) << 3) | (((Z) & 0xf8) >> 3))
 
-#define PACK_COLOR_5551( R, G, B, A )					\
-   ((((R) & 0xf8) << 8) | (((G) & 0xf8) << 3) | (((B) & 0xf8) >> 2) |	\
-    ((A) >> 7))
-
 #define PACK_COLOR_1555( A, B, G, R )					\
    ((((B) & 0xf8) << 7) | (((G) & 0xf8) << 2) | (((R) & 0xf8) >> 3) |	\
     (((A) & 0x80) << 8))
 
-#define PACK_COLOR_1555_REV( A, B, G, R )					\
-   ((((B) & 0xf8) >> 1) | (((G) & 0xc0) >> 6) | (((G) & 0x38) << 10) | (((R) & 0xf8) << 5) |	\
-    ((A) ? 0x80 : 0))
-
-#define PACK_COLOR_2101010_UB( A, B, G, R )					\
-   (((B) << 22) | ((G) << 12) | ((R) << 2) |	\
-    (((A) & 0xc0) << 24))
-
-#define PACK_COLOR_2101010_US( A, B, G, R )					\
-   ((((B) >> 6) << 20) | (((G) >> 6) << 10) | ((R) >> 6) |	\
-    (((A) >> 14) << 30))
-
 #define PACK_COLOR_4444( R, G, B, A )					\
    ((((R) & 0xf0) << 8) | (((G) & 0xf0) << 4) | ((B) & 0xf0) | ((A) >> 4))
 
-#define PACK_COLOR_4444_REV( R, G, B, A )				\
-   ((((B) & 0xf0) << 8) | (((A) & 0xf0) << 4) | ((R) & 0xf0) | ((G) >> 4))
-
-#define PACK_COLOR_44( L, A )						\
-   (((L) & 0xf0) | (((A) & 0xf0) >> 4))
-
 #define PACK_COLOR_88( L, A )						\
    (((L) << 8) | (A))
 
-#define PACK_COLOR_88_REV( L, A )					\
-   (((A) << 8) | (L))
-
-#define PACK_COLOR_1616( L, A )						\
-   (((L) << 16) | (A))
-
-#define PACK_COLOR_1616_REV( L, A )					\
-   (((A) << 16) | (L))
-
-#define PACK_COLOR_332( R, G, B )					\
-   (((R) & 0xe0) | (((G) & 0xe0) >> 3) | (((B) & 0xc0) >> 6))
-
-#define PACK_COLOR_233( B, G, R )					\
-   (((B) & 0xc0) | (((G) & 0xe0) >> 2) | (((R) & 0xe0) >> 5))
-
 /*@}*/
 
 
diff --git a/mesalib/src/mesa/main/context.c b/mesalib/src/mesa/main/context.c
index c1acda980..adf64976e 100644
--- a/mesalib/src/mesa/main/context.c
+++ b/mesalib/src/mesa/main/context.c
@@ -1288,7 +1288,6 @@ _mesa_free_context_data( struct gl_context *ctx )
    _mesa_free_eval_data( ctx );
    _mesa_free_texture_data( ctx );
    _mesa_free_matrix_data( ctx );
-   _mesa_free_viewport_data( ctx );
    _mesa_free_pipeline_data(ctx);
    _mesa_free_program_data(ctx);
    _mesa_free_shader_state(ctx);
@@ -1449,17 +1448,10 @@ _mesa_copy_context( const struct gl_context *src, struct gl_context *dst,
       dst->Transform = src->Transform;
    }
    if (mask & GL_VIEWPORT_BIT) {
-      /* Cannot use memcpy, because of pointers in GLmatrix _WindowMap */
       unsigned i;
       for (i = 0; i < src->Const.MaxViewports; i++) {
-         dst->ViewportArray[i].X = src->ViewportArray[i].X;
-         dst->ViewportArray[i].Y = src->ViewportArray[i].Y;
-         dst->ViewportArray[i].Width = src->ViewportArray[i].Width;
-         dst->ViewportArray[i].Height = src->ViewportArray[i].Height;
-         dst->ViewportArray[i].Near = src->ViewportArray[i].Near;
-         dst->ViewportArray[i].Far = src->ViewportArray[i].Far;
-         _math_matrix_copy(&dst->ViewportArray[i]._WindowMap,
-                           &src->ViewportArray[i]._WindowMap);
+         /* OK to memcpy */
+         dst->ViewportArray[i] = src->ViewportArray[i];
       }
    }
 
diff --git a/mesalib/src/mesa/main/debug.c b/mesalib/src/mesa/main/debug.c
index b50d79e0c..c93e84a04 100644
--- a/mesalib/src/mesa/main/debug.c
+++ b/mesalib/src/mesa/main/debug.c
@@ -26,11 +26,11 @@
 #include <stdio.h>
 #include "mtypes.h"
 #include "attrib.h"
-#include "colormac.h"
 #include "enums.h"
 #include "formats.h"
 #include "hash.h"
 #include "imports.h"
+#include "macros.h"
 #include "debug.h"
 #include "get.h"
 #include "pixelstore.h"
diff --git a/mesalib/src/mesa/main/dlopen.h b/mesalib/src/mesa/main/dlopen.h
index 93d1258cf..93d1258cf 100644..100755
--- a/mesalib/src/mesa/main/dlopen.h
+++ b/mesalib/src/mesa/main/dlopen.h
diff --git a/mesalib/src/mesa/main/errors.c b/mesalib/src/mesa/main/errors.c
index 8ffbf413c..2aa1deb63 100644
--- a/mesalib/src/mesa/main/errors.c
+++ b/mesalib/src/mesa/main/errors.c
@@ -1232,12 +1232,14 @@ _mesa_free_errors_data(struct gl_context *ctx)
 /** \name Diagnostics */
 /*@{*/
 
+static FILE *LogFile = NULL;
+
+
 static void
 output_if_debug(const char *prefixString, const char *outputString,
                 GLboolean newline)
 {
    static int debug = -1;
-   static FILE *fout = NULL;
 
    /* Init the local 'debug' var once.
     * Note: the _mesa_init_debug() function should have been called
@@ -1249,9 +1251,9 @@ output_if_debug(const char *prefixString, const char *outputString,
        */
       const char *logFile = getenv("MESA_LOG_FILE");
       if (logFile)
-         fout = fopen(logFile, "w");
-      if (!fout)
-         fout = stderr;
+         LogFile = fopen(logFile, "w");
+      if (!LogFile)
+         LogFile = stderr;
 #ifdef DEBUG
       /* in debug builds, print messages unless MESA_DEBUG="silent" */
       if (MESA_DEBUG_FLAGS & DEBUG_SILENT)
@@ -1266,10 +1268,13 @@ output_if_debug(const char *prefixString, const char *outputString,
 
    /* Now only print the string if we're required to do so. */
    if (debug) {
-      fprintf(fout, "%s: %s", prefixString, outputString);
+      if (prefixString)
+         fprintf(LogFile, "%s: %s", prefixString, outputString);
+      else
+         fprintf(LogFile, "%s", outputString);
       if (newline)
-         fprintf(fout, "\n");
-      fflush(fout);
+         fprintf(LogFile, "\n");
+      fflush(LogFile);
 
 #if defined(_WIN32)
       /* stderr from windows applications without console is not usually 
@@ -1285,6 +1290,18 @@ output_if_debug(const char *prefixString, const char *outputString,
 
 
 /**
+ * Return the file handle to use for debug/logging.  Defaults to stderr
+ * unless MESA_LOG_FILE is defined.
+ */
+FILE *
+_mesa_get_log_file(void)
+{
+   assert(LogFile);
+   return LogFile;
+}
+
+
+/**
  * When a new type of error is recorded, print a message describing
  * previous errors which were accumulated.
  */
@@ -1525,6 +1542,18 @@ _mesa_debug( const struct gl_context *ctx, const char *fmtString, ... )
 }
 
 
+void
+_mesa_log(const char *fmtString, ...)
+{
+   char s[MAX_DEBUG_MESSAGE_LENGTH];
+   va_list args;
+   va_start(args, fmtString);
+   _mesa_vsnprintf(s, MAX_DEBUG_MESSAGE_LENGTH, fmtString, args);
+   va_end(args);
+   output_if_debug("", s, GL_FALSE);
+}
+
+
 /**
  * Report debug information from the shader compiler via GL_ARB_debug_output.
  *
diff --git a/mesalib/src/mesa/main/errors.h b/mesalib/src/mesa/main/errors.h
index 0c521c0d0..e6dc9b5f1 100644
--- a/mesalib/src/mesa/main/errors.h
+++ b/mesalib/src/mesa/main/errors.h
@@ -36,6 +36,7 @@
 #define ERRORS_H
 
 
+#include <stdio.h>
 #include "compiler.h"
 #include "glheader.h"
 #include "mtypes.h"
@@ -69,6 +70,12 @@ extern void
 _mesa_debug( const struct gl_context *ctx, const char *fmtString, ... ) PRINTFLIKE(2, 3);
 
 extern void
+_mesa_log(const char *fmtString, ...) PRINTFLIKE(1, 2);
+
+extern FILE *
+_mesa_get_log_file(void);
+
+extern void
 _mesa_gl_debug(struct gl_context *ctx,
                GLuint *id,
                enum mesa_debug_source source,
diff --git a/mesalib/src/mesa/main/extensions.c b/mesalib/src/mesa/main/extensions.c
index f21201538..861b15006 100644
--- a/mesalib/src/mesa/main/extensions.c
+++ b/mesalib/src/mesa/main/extensions.c
@@ -139,6 +139,7 @@ static const struct extension extension_table[] = {
    { "GL_ARB_pixel_buffer_object",                 o(EXT_pixel_buffer_object),                 GL,             2004 },
    { "GL_ARB_point_parameters",                    o(EXT_point_parameters),                    GLL,            1997 },
    { "GL_ARB_point_sprite",                        o(ARB_point_sprite),                        GL,             2003 },
+   { "GL_ARB_program_interface_query",             o(dummy_true),                              GL,             2012 },
    { "GL_ARB_provoking_vertex",                    o(EXT_provoking_vertex),                    GL,             2009 },
    { "GL_ARB_robustness",                          o(dummy_true),                              GL,             2010 },
    { "GL_ARB_sample_shading",                      o(ARB_sample_shading),                      GL,             2009 },
diff --git a/mesalib/src/mesa/main/fbobject.c b/mesalib/src/mesa/main/fbobject.c
index f8d0d92bf..8032585ab 100644
--- a/mesalib/src/mesa/main/fbobject.c
+++ b/mesalib/src/mesa/main/fbobject.c
@@ -212,13 +212,13 @@ get_attachment(struct gl_context *ctx, struct gl_framebuffer *fb,
        */
       i = attachment - GL_COLOR_ATTACHMENT0_EXT;
       if (i >= ctx->Const.MaxColorAttachments
-	  || (i > 0 && ctx->API == API_OPENGLES)) {
-	 return NULL;
+          || (i > 0 && ctx->API == API_OPENGLES)) {
+         return NULL;
       }
       return &fb->Attachment[BUFFER_COLOR0 + i];
    case GL_DEPTH_STENCIL_ATTACHMENT:
       if (!_mesa_is_desktop_gl(ctx) && !_mesa_is_gles3(ctx))
-	 return NULL;
+         return NULL;
       /* fall-through */
    case GL_DEPTH_ATTACHMENT_EXT:
       return &fb->Attachment[BUFFER_DEPTH];
@@ -1122,28 +1122,28 @@ _mesa_test_framebuffer_completeness(struct gl_context *ctx,
    if (_mesa_is_desktop_gl(ctx) && !ctx->Extensions.ARB_ES2_compatibility) {
       /* Check that all DrawBuffers are present */
       for (j = 0; j < ctx->Const.MaxDrawBuffers; j++) {
-	 if (fb->ColorDrawBuffer[j] != GL_NONE) {
-	    const struct gl_renderbuffer_attachment *att
-	       = get_attachment(ctx, fb, fb->ColorDrawBuffer[j]);
-	    assert(att);
-	    if (att->Type == GL_NONE) {
-	       fb->_Status = GL_FRAMEBUFFER_INCOMPLETE_DRAW_BUFFER_EXT;
-	       fbo_incomplete(ctx, "missing drawbuffer", j);
-	       return;
-	    }
-	 }
+         if (fb->ColorDrawBuffer[j] != GL_NONE) {
+            const struct gl_renderbuffer_attachment *att
+               = get_attachment(ctx, fb, fb->ColorDrawBuffer[j]);
+            assert(att);
+            if (att->Type == GL_NONE) {
+               fb->_Status = GL_FRAMEBUFFER_INCOMPLETE_DRAW_BUFFER_EXT;
+               fbo_incomplete(ctx, "missing drawbuffer", j);
+               return;
+            }
+         }
       }
 
       /* Check that the ReadBuffer is present */
       if (fb->ColorReadBuffer != GL_NONE) {
-	 const struct gl_renderbuffer_attachment *att
-	    = get_attachment(ctx, fb, fb->ColorReadBuffer);
-	 assert(att);
-	 if (att->Type == GL_NONE) {
-	    fb->_Status = GL_FRAMEBUFFER_INCOMPLETE_READ_BUFFER_EXT;
+         const struct gl_renderbuffer_attachment *att
+            = get_attachment(ctx, fb, fb->ColorReadBuffer);
+         assert(att);
+         if (att->Type == GL_NONE) {
+            fb->_Status = GL_FRAMEBUFFER_INCOMPLETE_READ_BUFFER_EXT;
             fbo_incomplete(ctx, "missing readbuffer", -1);
-	    return;
-	 }
+            return;
+         }
       }
    }
 
@@ -1205,6 +1205,28 @@ _mesa_IsRenderbuffer(GLuint renderbuffer)
 }
 
 
+static struct gl_renderbuffer *
+allocate_renderbuffer(struct gl_context *ctx, GLuint renderbuffer,
+                      const char *func)
+{
+   struct gl_renderbuffer *newRb;
+
+   /* create new renderbuffer object */
+   newRb = ctx->Driver.NewRenderbuffer(ctx, renderbuffer);
+   if (!newRb) {
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
+      return NULL;
+   }
+   assert(newRb->AllocStorage);
+   mtx_lock(&ctx->Shared->Mutex);
+   _mesa_HashInsert(ctx->Shared->RenderBuffers, renderbuffer, newRb);
+   newRb->RefCount = 1; /* referenced by hash table */
+   mtx_unlock(&ctx->Shared->Mutex);
+
+   return newRb;
+}
+
+
 static void
 bind_renderbuffer(GLenum target, GLuint renderbuffer, bool allow_user_names)
 {
@@ -1233,15 +1255,7 @@ bind_renderbuffer(GLenum target, GLuint renderbuffer, bool allow_user_names)
       }
 
       if (!newRb) {
-	 /* create new renderbuffer object */
-	 newRb = ctx->Driver.NewRenderbuffer(ctx, renderbuffer);
-	 if (!newRb) {
-	    _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindRenderbufferEXT");
-	    return;
-	 }
-         assert(newRb->AllocStorage);
-         _mesa_HashInsert(ctx->Shared->RenderBuffers, renderbuffer, newRb);
-         newRb->RefCount = 1; /* referenced by hash table */
+         newRb = allocate_renderbuffer(ctx, renderbuffer, "glBindRenderbufferEXT");
       }
    }
    else {
@@ -1333,9 +1347,9 @@ _mesa_DeleteRenderbuffers(GLsizei n, const GLuint *renderbuffers)
 
    for (i = 0; i < n; i++) {
       if (renderbuffers[i] > 0) {
-	 struct gl_renderbuffer *rb;
-	 rb = _mesa_lookup_renderbuffer(ctx, renderbuffers[i]);
-	 if (rb) {
+         struct gl_renderbuffer *rb;
+         rb = _mesa_lookup_renderbuffer(ctx, renderbuffers[i]);
+         if (rb) {
             /* check if deleting currently bound renderbuffer object */
             if (rb == ctx->CurrentRenderbuffer) {
                /* bind default */
@@ -1368,31 +1382,31 @@ _mesa_DeleteRenderbuffers(GLsizei n, const GLuint *renderbuffers)
                _mesa_detach_renderbuffer(ctx, ctx->ReadBuffer, rb);
             }
 
-	    /* Remove from hash table immediately, to free the ID.
+            /* Remove from hash table immediately, to free the ID.
              * But the object will not be freed until it's no longer
              * referenced anywhere else.
              */
-	    _mesa_HashRemove(ctx->Shared->RenderBuffers, renderbuffers[i]);
+            _mesa_HashRemove(ctx->Shared->RenderBuffers, renderbuffers[i]);
 
             if (rb != &DummyRenderbuffer) {
                /* no longer referenced by hash table */
                _mesa_reference_renderbuffer(&rb, NULL);
-	    }
-	 }
+            }
+         }
       }
    }
 }
 
-
-void GLAPIENTRY
-_mesa_GenRenderbuffers(GLsizei n, GLuint *renderbuffers)
+static void
+create_render_buffers(struct gl_context *ctx, GLsizei n, GLuint *renderbuffers,
+                      bool dsa)
 {
-   GET_CURRENT_CONTEXT(ctx);
+   const char *func = dsa ? "glCreateRenderbuffers" : "glGenRenderbuffers";
    GLuint first;
    GLint i;
 
    if (n < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenRenderbuffersEXT(n)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(n<0)", func);
       return;
    }
 
@@ -1404,14 +1418,35 @@ _mesa_GenRenderbuffers(GLsizei n, GLuint *renderbuffers)
    for (i = 0; i < n; i++) {
       GLuint name = first + i;
       renderbuffers[i] = name;
-      /* insert dummy placeholder into hash table */
-      mtx_lock(&ctx->Shared->Mutex);
-      _mesa_HashInsert(ctx->Shared->RenderBuffers, name, &DummyRenderbuffer);
-      mtx_unlock(&ctx->Shared->Mutex);
+
+      if (dsa) {
+         allocate_renderbuffer(ctx, name, func);
+      } else {
+         /* insert a dummy renderbuffer into the hash table */
+         mtx_lock(&ctx->Shared->Mutex);
+         _mesa_HashInsert(ctx->Shared->RenderBuffers, name, &DummyRenderbuffer);
+         mtx_unlock(&ctx->Shared->Mutex);
+      }
    }
 }
 
 
+void GLAPIENTRY
+_mesa_GenRenderbuffers(GLsizei n, GLuint *renderbuffers)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   create_render_buffers(ctx, n, renderbuffers, false);
+}
+
+
+void GLAPIENTRY
+_mesa_CreateRenderbuffers(GLsizei n, GLuint *renderbuffers)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   create_render_buffers(ctx, n, renderbuffers, true);
+}
+
+
 /**
  * Given an internal format token for a render buffer, return the
  * corresponding base format (one of GL_RGB, GL_RGBA, GL_STENCIL_INDEX,
@@ -1749,40 +1784,17 @@ invalidate_rb(GLuint key, void *data, void *userData)
 
 
 /**
- * Helper function used by _mesa_RenderbufferStorage() and
- * _mesa_RenderbufferStorageMultisample().
- * samples will be NO_SAMPLES if called by _mesa_RenderbufferStorage().
+ * Helper function used by renderbuffer_storage_direct() and
+ * renderbuffer_storage_target().
+ * samples will be NO_SAMPLES if called by a non-multisample function.
  */
 static void
-renderbuffer_storage(GLenum target, GLenum internalFormat,
-                     GLsizei width, GLsizei height, GLsizei samples)
+renderbuffer_storage(struct gl_context *ctx, struct gl_renderbuffer *rb,
+                     GLenum internalFormat, GLsizei width,
+                     GLsizei height, GLsizei samples, const char *func)
 {
-   const char *func = samples == NO_SAMPLES ?
-      "glRenderbufferStorage" : "glRenderbufferStorageMultisample";
-   struct gl_renderbuffer *rb;
    GLenum baseFormat;
    GLenum sample_count_error;
-   GET_CURRENT_CONTEXT(ctx);
-
-   if (MESA_VERBOSE & VERBOSE_API) {
-      if (samples == NO_SAMPLES)
-         _mesa_debug(ctx, "%s(%s, %s, %d, %d)\n",
-                     func,
-                     _mesa_lookup_enum_by_nr(target),
-                     _mesa_lookup_enum_by_nr(internalFormat),
-                     width, height);
-      else
-         _mesa_debug(ctx, "%s(%s, %s, %d, %d, %d)\n",
-                     func,
-                     _mesa_lookup_enum_by_nr(target),
-                     _mesa_lookup_enum_by_nr(internalFormat),
-                     width, height, samples);
-   }
-
-   if (target != GL_RENDERBUFFER_EXT) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "%s(target)", func);
-      return;
-   }
 
    baseFormat = _mesa_base_fbo_format(ctx, internalFormat);
    if (baseFormat == 0) {
@@ -1792,12 +1804,14 @@ renderbuffer_storage(GLenum target, GLenum internalFormat,
    }
 
    if (width < 0 || width > (GLsizei) ctx->Const.MaxRenderbufferSize) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "%s(width)", func);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(invalid width %d)", func,
+                  width);
       return;
    }
 
    if (height < 0 || height > (GLsizei) ctx->Const.MaxRenderbufferSize) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "%s(height)", func);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(invalid height %d)", func,
+                  height);
       return;
    }
 
@@ -1809,7 +1823,7 @@ renderbuffer_storage(GLenum target, GLenum internalFormat,
       /* check the sample count;
        * note: driver may choose to use more samples than what's requested
        */
-      sample_count_error = _mesa_check_sample_count(ctx, target,
+      sample_count_error = _mesa_check_sample_count(ctx, GL_RENDERBUFFER,
             internalFormat, samples);
       if (sample_count_error != GL_NO_ERROR) {
          _mesa_error(ctx, sample_count_error, "%s(samples)", func);
@@ -1817,12 +1831,6 @@ renderbuffer_storage(GLenum target, GLenum internalFormat,
       }
    }
 
-   rb = ctx->CurrentRenderbuffer;
-   if (!rb) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "%s", func);
-      return;
-   }
-
    FLUSH_VERTICES(ctx, _NEW_BUFFERS);
 
    if (rb->InternalFormat == internalFormat &&
@@ -1864,6 +1872,83 @@ renderbuffer_storage(GLenum target, GLenum internalFormat,
    }
 }
 
+/**
+ * Helper function used by _mesa_NamedRenderbufferStorage*().
+ * samples will be NO_SAMPLES if called by a non-multisample function.
+ */
+static void
+renderbuffer_storage_named(GLuint renderbuffer, GLenum internalFormat,
+                           GLsizei width, GLsizei height, GLsizei samples,
+                           const char *func)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   if (MESA_VERBOSE & VERBOSE_API) {
+      if (samples == NO_SAMPLES)
+         _mesa_debug(ctx, "%s(%u, %s, %d, %d)\n",
+                     func, renderbuffer,
+                     _mesa_lookup_enum_by_nr(internalFormat),
+                     width, height);
+      else
+         _mesa_debug(ctx, "%s(%u, %s, %d, %d, %d)\n",
+                     func, renderbuffer,
+                     _mesa_lookup_enum_by_nr(internalFormat),
+                     width, height, samples);
+   }
+
+   struct gl_renderbuffer *rb = _mesa_lookup_renderbuffer(ctx, renderbuffer);
+   if (!rb || rb == &DummyRenderbuffer) {
+      /* ID was reserved, but no real renderbuffer object made yet */
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(invalid renderbuffer %u)",
+                  func, renderbuffer);
+      return;
+   }
+
+   renderbuffer_storage(ctx, rb, internalFormat, width, height, samples, func);
+}
+
+/**
+ * Helper function used by _mesa_RenderbufferStorage() and
+ * _mesa_RenderbufferStorageMultisample().
+ * samples will be NO_SAMPLES if called by _mesa_RenderbufferStorage().
+ */
+static void
+renderbuffer_storage_target(GLenum target, GLenum internalFormat,
+                            GLsizei width, GLsizei height, GLsizei samples,
+                            const char *func)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   if (MESA_VERBOSE & VERBOSE_API) {
+      if (samples == NO_SAMPLES)
+         _mesa_debug(ctx, "%s(%s, %s, %d, %d)\n",
+                     func,
+                     _mesa_lookup_enum_by_nr(target),
+                     _mesa_lookup_enum_by_nr(internalFormat),
+                     width, height);
+      else
+         _mesa_debug(ctx, "%s(%s, %s, %d, %d, %d)\n",
+                     func,
+                     _mesa_lookup_enum_by_nr(target),
+                     _mesa_lookup_enum_by_nr(internalFormat),
+                     width, height, samples);
+   }
+
+   if (target != GL_RENDERBUFFER_EXT) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(target)", func);
+      return;
+   }
+
+   if (!ctx->CurrentRenderbuffer) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(no renderbuffer bound)",
+                  func);
+      return;
+   }
+
+   renderbuffer_storage(ctx, ctx->CurrentRenderbuffer, internalFormat, width,
+                        height, samples, func);
+}
+
 
 void GLAPIENTRY
 _mesa_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image)
@@ -1923,7 +2008,8 @@ _mesa_RenderbufferStorage(GLenum target, GLenum internalFormat,
     * glRenderbufferStorageMultisample() with samples=0.  We pass in
     * a token value here just for error reporting purposes.
     */
-   renderbuffer_storage(target, internalFormat, width, height, NO_SAMPLES);
+   renderbuffer_storage_target(target, internalFormat, width, height,
+                               NO_SAMPLES, "glRenderbufferStorage");
 }
 
 
@@ -1932,7 +2018,8 @@ _mesa_RenderbufferStorageMultisample(GLenum target, GLsizei samples,
                                      GLenum internalFormat,
                                      GLsizei width, GLsizei height)
 {
-   renderbuffer_storage(target, internalFormat, width, height, samples);
+   renderbuffer_storage_target(target, internalFormat, width, height,
+                               samples, "glRenderbufferStorageMultisample");
 }
 
 
@@ -1941,7 +2028,7 @@ _mesa_RenderbufferStorageMultisample(GLenum target, GLsizei samples,
  */
 void GLAPIENTRY
 _es_RenderbufferStorageEXT(GLenum target, GLenum internalFormat,
-			   GLsizei width, GLsizei height)
+                           GLsizei width, GLsizei height)
 {
    switch (internalFormat) {
    case GL_RGB565:
@@ -1953,29 +2040,38 @@ _es_RenderbufferStorageEXT(GLenum target, GLenum internalFormat,
       break;
    }
 
-   renderbuffer_storage(target, internalFormat, width, height, 0);
+   renderbuffer_storage_target(target, internalFormat, width, height, 0,
+                               "glRenderbufferStorageEXT");
 }
 
-
 void GLAPIENTRY
-_mesa_GetRenderbufferParameteriv(GLenum target, GLenum pname, GLint *params)
+_mesa_NamedRenderbufferStorage(GLuint renderbuffer, GLenum internalformat,
+                               GLsizei width, GLsizei height)
 {
-   struct gl_renderbuffer *rb;
-   GET_CURRENT_CONTEXT(ctx);
+   /* GL_ARB_fbo says calling this function is equivalent to calling
+    * glRenderbufferStorageMultisample() with samples=0.  We pass in
+    * a token value here just for error reporting purposes.
+    */
+   renderbuffer_storage_named(renderbuffer, internalformat, width, height,
+                              NO_SAMPLES, "glNamedRenderbufferStorage");
+}
 
-   if (target != GL_RENDERBUFFER_EXT) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glGetRenderbufferParameterivEXT(target)");
-      return;
-   }
+void GLAPIENTRY
+_mesa_NamedRenderbufferStorageMultisample(GLuint renderbuffer, GLsizei samples,
+                                          GLenum internalformat,
+                                          GLsizei width, GLsizei height)
+{
+   renderbuffer_storage_named(renderbuffer, internalformat, width, height,
+                              samples,
+                              "glNamedRenderbufferStorageMultisample");
+}
 
-   rb = ctx->CurrentRenderbuffer;
-   if (!rb) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glGetRenderbufferParameterivEXT");
-      return;
-   }
 
+static void
+get_render_buffer_parameteriv(struct gl_context *ctx,
+                              struct gl_renderbuffer *rb, GLenum pname,
+                              GLint *params, const char *func)
+{
    /* No need to flush here since we're just quering state which is
     * not effected by rendering.
     */
@@ -2006,10 +2102,51 @@ _mesa_GetRenderbufferParameteriv(GLenum target, GLenum pname, GLint *params)
       }
       /* fallthrough */
    default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid pname=%s)", func,
+                  _mesa_lookup_enum_by_nr(pname));
+      return;
+   }
+}
+
+
+void GLAPIENTRY
+_mesa_GetRenderbufferParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   if (target != GL_RENDERBUFFER_EXT) {
       _mesa_error(ctx, GL_INVALID_ENUM,
                   "glGetRenderbufferParameterivEXT(target)");
       return;
    }
+
+   if (!ctx->CurrentRenderbuffer) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glGetRenderbufferParameterivEXT"
+                  "(no renderbuffer bound)");
+      return;
+   }
+
+   get_render_buffer_parameteriv(ctx, ctx->CurrentRenderbuffer, pname,
+                                 params, "glGetRenderbufferParameteriv");
+}
+
+
+void GLAPIENTRY
+_mesa_GetNamedRenderbufferParameteriv(GLuint renderbuffer, GLenum pname,
+                                      GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   struct gl_renderbuffer *rb = _mesa_lookup_renderbuffer(ctx, renderbuffer);
+   if (!rb || rb == &DummyRenderbuffer) {
+      /* ID was reserved, but no real renderbuffer object made yet */
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glGetNamedRenderbufferParameteriv"
+                  "(invalid renderbuffer %i)", renderbuffer);
+      return;
+   }
+
+   get_render_buffer_parameteriv(ctx, rb, pname, params,
+                                 "glGetNamedRenderbufferParameteriv");
 }
 
 
@@ -2116,12 +2253,12 @@ bind_framebuffer(GLenum target, GLuint framebuffer, bool allow_user_names)
       }
 
       if (!newDrawFb) {
-	 /* create new framebuffer object */
-	 newDrawFb = ctx->Driver.NewFramebuffer(ctx, framebuffer);
-	 if (!newDrawFb) {
-	    _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindFramebufferEXT");
-	    return;
-	 }
+         /* create new framebuffer object */
+         newDrawFb = ctx->Driver.NewFramebuffer(ctx, framebuffer);
+         if (!newDrawFb) {
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindFramebufferEXT");
+            return;
+         }
          _mesa_HashInsert(ctx->Shared->FrameBuffers, framebuffer, newDrawFb);
       }
       newReadFb = newDrawFb;
@@ -2224,9 +2361,9 @@ _mesa_DeleteFramebuffers(GLsizei n, const GLuint *framebuffers)
 
    for (i = 0; i < n; i++) {
       if (framebuffers[i] > 0) {
-	 struct gl_framebuffer *fb;
-	 fb = _mesa_lookup_framebuffer(ctx, framebuffers[i]);
-	 if (fb) {
+         struct gl_framebuffer *fb;
+         fb = _mesa_lookup_framebuffer(ctx, framebuffers[i]);
+         if (fb) {
             assert(fb == &DummyFramebuffer || fb->Name == framebuffers[i]);
 
             /* check if deleting currently bound framebuffer object */
@@ -2241,16 +2378,16 @@ _mesa_DeleteFramebuffers(GLsizei n, const GLuint *framebuffers)
                _mesa_BindFramebuffer(GL_READ_FRAMEBUFFER, 0);
             }
 
-	    /* remove from hash table immediately, to free the ID */
-	    _mesa_HashRemove(ctx->Shared->FrameBuffers, framebuffers[i]);
+            /* remove from hash table immediately, to free the ID */
+            _mesa_HashRemove(ctx->Shared->FrameBuffers, framebuffers[i]);
 
             if (fb != &DummyFramebuffer) {
                /* But the object will not be freed until it's no longer
                 * bound in any context.
                 */
                _mesa_reference_framebuffer(&fb, NULL);
-	    }
-	 }
+            }
+         }
       }
    }
 }
@@ -2498,34 +2635,34 @@ framebuffer_texture(struct gl_context *ctx, const char *caller, GLenum target,
           _mesa_tex_target_to_face(textarget) ==
           fb->Attachment[BUFFER_STENCIL].CubeMapFace &&
           zoffset == fb->Attachment[BUFFER_STENCIL].Zoffset) {
-	 /* The texture object is already attached to the stencil attachment
-	  * point. Don't create a new renderbuffer; just reuse the stencil
-	  * attachment's. This is required to prevent a GL error in
-	  * glGetFramebufferAttachmentParameteriv(GL_DEPTH_STENCIL).
-	  */
-	 reuse_framebuffer_texture_attachment(fb, BUFFER_DEPTH,
-	                                      BUFFER_STENCIL);
+         /* The texture object is already attached to the stencil attachment
+          * point. Don't create a new renderbuffer; just reuse the stencil
+          * attachment's. This is required to prevent a GL error in
+          * glGetFramebufferAttachmentParameteriv(GL_DEPTH_STENCIL).
+          */
+         reuse_framebuffer_texture_attachment(fb, BUFFER_DEPTH,
+                                              BUFFER_STENCIL);
       } else if (attachment == GL_STENCIL_ATTACHMENT &&
-	         texObj == fb->Attachment[BUFFER_DEPTH].Texture &&
+                 texObj == fb->Attachment[BUFFER_DEPTH].Texture &&
                  level == fb->Attachment[BUFFER_DEPTH].TextureLevel &&
                  _mesa_tex_target_to_face(textarget) ==
                  fb->Attachment[BUFFER_DEPTH].CubeMapFace &&
                  zoffset == fb->Attachment[BUFFER_DEPTH].Zoffset) {
-	 /* As above, but with depth and stencil transposed. */
-	 reuse_framebuffer_texture_attachment(fb, BUFFER_STENCIL,
-	                                      BUFFER_DEPTH);
+         /* As above, but with depth and stencil transposed. */
+         reuse_framebuffer_texture_attachment(fb, BUFFER_STENCIL,
+                                              BUFFER_DEPTH);
       } else {
-	 set_texture_attachment(ctx, fb, att, texObj, textarget,
-	                              level, zoffset, layered);
-	 if (attachment == GL_DEPTH_STENCIL_ATTACHMENT) {
-	    /* Above we created a new renderbuffer and attached it to the
-	     * depth attachment point. Now attach it to the stencil attachment
-	     * point too.
-	     */
-	    assert(att == &fb->Attachment[BUFFER_DEPTH]);
-	    reuse_framebuffer_texture_attachment(fb,BUFFER_STENCIL,
-	                                         BUFFER_DEPTH);
-	 }
+         set_texture_attachment(ctx, fb, att, texObj, textarget,
+                                      level, zoffset, layered);
+         if (attachment == GL_DEPTH_STENCIL_ATTACHMENT) {
+            /* Above we created a new renderbuffer and attached it to the
+             * depth attachment point. Now attach it to the stencil attachment
+             * point too.
+             */
+            assert(att == &fb->Attachment[BUFFER_DEPTH]);
+            reuse_framebuffer_texture_attachment(fb,BUFFER_STENCIL,
+                                                 BUFFER_DEPTH);
+         }
       }
 
       /* Set the render-to-texture flag.  We'll check this flag in
@@ -2541,8 +2678,8 @@ framebuffer_texture(struct gl_context *ctx, const char *caller, GLenum target,
    else {
       remove_attachment(ctx, att);
       if (attachment == GL_DEPTH_STENCIL_ATTACHMENT) {
-	 assert(att == &fb->Attachment[BUFFER_DEPTH]);
-	 remove_attachment(ctx, &fb->Attachment[BUFFER_STENCIL]);
+         assert(att == &fb->Attachment[BUFFER_DEPTH]);
+         remove_attachment(ctx, &fb->Attachment[BUFFER_STENCIL]);
       }
    }
 
@@ -2721,16 +2858,16 @@ _mesa_FramebufferRenderbuffer(GLenum target, GLenum attachment,
    if (renderbuffer) {
       rb = _mesa_lookup_renderbuffer(ctx, renderbuffer);
       if (!rb) {
-	 _mesa_error(ctx, GL_INVALID_OPERATION,
-		     "glFramebufferRenderbuffer(non-existant"
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glFramebufferRenderbuffer(non-existant"
                      " renderbuffer %u)", renderbuffer);
-	 return;
+         return;
       }
       else if (rb == &DummyRenderbuffer) {
-	 _mesa_error(ctx, GL_INVALID_OPERATION,
-		     "glFramebufferRenderbuffer(renderbuffer %u)",
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glFramebufferRenderbuffer(renderbuffer %u)",
                      renderbuffer);
-	 return;
+         return;
       }
    }
    else {
@@ -2795,9 +2932,9 @@ _mesa_GetFramebufferAttachmentParameteriv(GLenum target, GLenum attachment,
       if ((!_mesa_is_desktop_gl(ctx) ||
            !ctx->Extensions.ARB_framebuffer_object)
           && !_mesa_is_gles3(ctx)) {
-	 _mesa_error(ctx, GL_INVALID_OPERATION,
-		     "glGetFramebufferAttachmentParameteriv(bound FBO = 0)");
-	 return;
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glGetFramebufferAttachmentParameteriv(bound FBO = 0)");
+         return;
       }
 
       if (_mesa_is_gles3(ctx) && attachment != GL_BACK &&
@@ -2855,10 +2992,10 @@ _mesa_GetFramebufferAttachmentParameteriv(GLenum target, GLenum attachment,
       return;
    case GL_FRAMEBUFFER_ATTACHMENT_OBJECT_NAME_EXT:
       if (att->Type == GL_RENDERBUFFER_EXT) {
-	 *params = att->Renderbuffer->Name;
+         *params = att->Renderbuffer->Name;
       }
       else if (att->Type == GL_TEXTURE) {
-	 *params = att->Texture->Name;
+         *params = att->Texture->Name;
       }
       else {
          assert(att->Type == GL_NONE);
@@ -2871,7 +3008,7 @@ _mesa_GetFramebufferAttachmentParameteriv(GLenum target, GLenum attachment,
       return;
    case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_LEVEL_EXT:
       if (att->Type == GL_TEXTURE) {
-	 *params = att->TextureLevel;
+         *params = att->TextureLevel;
       }
       else if (att->Type == GL_NONE) {
          _mesa_error(ctx, err,
diff --git a/mesalib/src/mesa/main/fbobject.h b/mesalib/src/mesa/main/fbobject.h
index 77fdef415..61aa1f503 100644
--- a/mesalib/src/mesa/main/fbobject.h
+++ b/mesalib/src/mesa/main/fbobject.h
@@ -115,6 +115,9 @@ extern void GLAPIENTRY
 _mesa_GenRenderbuffers(GLsizei n, GLuint *renderbuffers);
 
 extern void GLAPIENTRY
+_mesa_CreateRenderbuffers(GLsizei n, GLuint *renderbuffers);
+
+extern void GLAPIENTRY
 _mesa_RenderbufferStorage(GLenum target, GLenum internalformat,
                              GLsizei width, GLsizei height);
 
@@ -128,12 +131,25 @@ _es_RenderbufferStorageEXT(GLenum target, GLenum internalFormat,
 			   GLsizei width, GLsizei height);
 
 extern void GLAPIENTRY
+_mesa_NamedRenderbufferStorage(GLuint renderbuffer, GLenum internalformat,
+                               GLsizei width, GLsizei height);
+
+extern void GLAPIENTRY
+_mesa_NamedRenderbufferStorageMultisample(GLuint renderbuffer, GLsizei samples,
+                                          GLenum internalformat,
+                                          GLsizei width, GLsizei height);
+
+extern void GLAPIENTRY
 _mesa_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image);
 
 extern void GLAPIENTRY
 _mesa_GetRenderbufferParameteriv(GLenum target, GLenum pname,
                                     GLint *params);
 
+void GLAPIENTRY
+_mesa_GetNamedRenderbufferParameteriv(GLuint renderbuffer, GLenum pname,
+                                      GLint *params);
+
 extern GLboolean GLAPIENTRY
 _mesa_IsFramebuffer(GLuint framebuffer);
 
diff --git a/mesalib/src/mesa/main/ffvertex_prog.c b/mesalib/src/mesa/main/ffvertex_prog.c
index 395b00e15..edf7e3390 100644
--- a/mesalib/src/mesa/main/ffvertex_prog.c
+++ b/mesalib/src/mesa/main/ffvertex_prog.c
@@ -619,13 +619,13 @@ static void emit_op3fn(struct tnl_program *p,
 
 
 #define emit_op3(p, op, dst, mask, src0, src1, src2) \
-   emit_op3fn(p, op, dst, mask, src0, src1, src2, __FUNCTION__, __LINE__)
+   emit_op3fn(p, op, dst, mask, src0, src1, src2, __func__, __LINE__)
 
 #define emit_op2(p, op, dst, mask, src0, src1) \
-    emit_op3fn(p, op, dst, mask, src0, src1, undef, __FUNCTION__, __LINE__)
+    emit_op3fn(p, op, dst, mask, src0, src1, undef, __func__, __LINE__)
 
 #define emit_op1(p, op, dst, mask, src0) \
-    emit_op3fn(p, op, dst, mask, src0, undef, undef, __FUNCTION__, __LINE__)
+    emit_op3fn(p, op, dst, mask, src0, undef, undef, __func__, __LINE__)
 
 
 static struct ureg make_temp( struct tnl_program *p, struct ureg reg )
diff --git a/mesalib/src/mesa/main/format_pack.py b/mesalib/src/mesa/main/format_pack.py
index f141da83c..2f43a30dd 100644
--- a/mesalib/src/mesa/main/format_pack.py
+++ b/mesalib/src/mesa/main/format_pack.py
@@ -43,7 +43,6 @@ string = """/*
 
 #include <stdint.h>
 
-#include "colormac.h"
 #include "format_pack.h"
 #include "format_utils.h"
 #include "macros.h"
diff --git a/mesalib/src/mesa/main/format_unpack.h b/mesalib/src/mesa/main/format_unpack.h
index eba3c6650..964c6077a 100644
--- a/mesalib/src/mesa/main/format_unpack.h
+++ b/mesalib/src/mesa/main/format_unpack.h
@@ -25,6 +25,8 @@
 #ifndef FORMAT_UNPACK_H
 #define FORMAT_UNPACK_H
 
+#include "formats.h"
+
 extern void
 _mesa_unpack_rgba_row(mesa_format format, GLuint n,
                       const void *src, GLfloat dst[][4]);
diff --git a/mesalib/src/mesa/main/format_unpack.py b/mesalib/src/mesa/main/format_unpack.py
index 53bdf641d..0ae73b897 100644
--- a/mesalib/src/mesa/main/format_unpack.py
+++ b/mesalib/src/mesa/main/format_unpack.py
@@ -43,7 +43,6 @@ string = """/*
 
 #include <stdint.h>
 
-#include "colormac.h"
 #include "format_unpack.h"
 #include "format_utils.h"
 #include "macros.h"
@@ -333,7 +332,7 @@ _mesa_unpack_rgba_row(mesa_format format, GLuint n,
       unpack_float_ycbcr_rev(src, dst, n);
       break;
    default:
-      _mesa_problem(NULL, "%s: bad format %s", __FUNCTION__,
+      _mesa_problem(NULL, "%s: bad format %s", __func__,
                     _mesa_get_format_name(format));
       return;
    }
@@ -402,7 +401,7 @@ _mesa_unpack_uint_rgba_row(mesa_format format, GLuint n,
       break;
 %endfor
    default:
-      _mesa_problem(NULL, "%s: bad format %s", __FUNCTION__,
+      _mesa_problem(NULL, "%s: bad format %s", __func__,
                     _mesa_get_format_name(format));
       return;
    }
diff --git a/mesalib/src/mesa/main/glformats.c b/mesalib/src/mesa/main/glformats.c
index 4e05229cf..8ced57949 100644
--- a/mesalib/src/mesa/main/glformats.c
+++ b/mesalib/src/mesa/main/glformats.c
@@ -1393,7 +1393,7 @@ _mesa_base_format_has_channel(GLenum base_format, GLenum pname)
       return GL_FALSE;
    default:
       _mesa_warning(NULL, "%s: Unexpected channel token 0x%x\n",
-		    __FUNCTION__, pname);
+		    __func__, pname);
       return GL_FALSE;
    }
 
diff --git a/mesalib/src/mesa/main/imports.h b/mesalib/src/mesa/main/imports.h
index 8877966ec..27f54d585 100755
--- a/mesalib/src/mesa/main/imports.h
+++ b/mesalib/src/mesa/main/imports.h
@@ -432,6 +432,30 @@ _mesa_fls(unsigned int n)
 #endif
 }
 
+/**
+ * Find the last (most significant) bit set in a uint64_t value.
+ *
+ * Essentially ffsll() in the reverse direction.
+ */
+static inline unsigned int
+_mesa_flsll(uint64_t n)
+{
+#ifdef HAVE___BUILTIN_CLZLL
+   return n == 0 ? 0 : 64 - __builtin_clzll(n);
+#else
+   unsigned int v = 1;
+
+   if (n == 0)
+      return 0;
+
+   while (n >>= 1)
+       v++;
+
+   return v;
+#endif
+}
+
+
 extern GLhalfARB
 _mesa_float_to_half(float f);
 
diff --git a/mesalib/src/mesa/main/macros.h b/mesalib/src/mesa/main/macros.h
index b22510965..5680b2ee1 100755
--- a/mesalib/src/mesa/main/macros.h
+++ b/mesalib/src/mesa/main/macros.h
@@ -789,7 +789,14 @@ NORMALIZE_3FV(GLfloat v[3])
 static inline GLboolean
 DIFFERENT_SIGNS(GLfloat x, GLfloat y)
 {
+#ifdef _MSC_VER
+#pragma warning( push )
+#pragma warning( disable : 6334 ) /* sizeof operator applied to an expression with an operator may yield unexpected results */
+#endif
    return signbit(x) != signbit(y);
+#ifdef _MSC_VER
+#pragma warning( pop )
+#endif
 }
 
 
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index 8e1dba6f0..1c751cfff 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -42,6 +42,7 @@
 #include "main/config.h"
 #include "glapi/glapi.h"
 #include "math/m_matrix.h"	/* GLmatrix */
+#include "glsl/shader_enums.h"
 #include "util/simple_list.h"	/* struct simple_node */
 #include "main/formats.h"       /* MESA_FORMAT_COUNT */
 
@@ -280,13 +281,6 @@ typedef enum
 /*@}*/
 
 /**
- * Bitflags for system values.
- */
-#define SYSTEM_BIT_SAMPLE_ID BITFIELD64_BIT(SYSTEM_VALUE_SAMPLE_ID)
-#define SYSTEM_BIT_SAMPLE_POS BITFIELD64_BIT(SYSTEM_VALUE_SAMPLE_POS)
-#define SYSTEM_BIT_SAMPLE_MASK_IN BITFIELD64_BIT(SYSTEM_VALUE_SAMPLE_MASK_IN)
-
-/**
  * Determine if the given gl_varying_slot appears in the fragment shader.
  */
 static inline GLboolean
@@ -395,25 +389,6 @@ typedef enum
                             BUFFER_BIT_COLOR6 | \
                             BUFFER_BIT_COLOR7)
 
-
-/**
- * Shader stages. Note that these will become 5 with tessellation.
- *
- * The order must match how shaders are ordered in the pipeline.
- * The GLSL linker assumes that if i<j, then the j-th shader is
- * executed later than the i-th shader.
- */
-typedef enum
-{
-   MESA_SHADER_VERTEX = 0,
-   MESA_SHADER_GEOMETRY = 1,
-   MESA_SHADER_FRAGMENT = 2,
-   MESA_SHADER_COMPUTE = 3,
-} gl_shader_stage;
-
-#define MESA_SHADER_STAGES (MESA_SHADER_COMPUTE + 1)
-
-
 /**
  * Framebuffer configuration (aka visual / pixelformat)
  * Note: some of these fields should be boolean, but it appears that
@@ -1456,7 +1431,6 @@ struct gl_viewport_attrib
    GLfloat X, Y;		/**< position */
    GLfloat Width, Height;	/**< size */
    GLdouble Near, Far;		/**< Depth buffer range */
-   GLmatrix _WindowMap;		/**< Mapping transformation as a matrix. */
 };
 
 
@@ -2082,140 +2056,6 @@ typedef enum
 
 
 /**
- * If the register file is PROGRAM_SYSTEM_VALUE, the register index will be
- * one of these values.
- */
-typedef enum
-{
-   /**
-    * \name Vertex shader system values
-    */
-   /*@{*/
-   /**
-    * OpenGL-style vertex ID.
-    *
-    * Section 2.11.7 (Shader Execution), subsection Shader Inputs, of the
-    * OpenGL 3.3 core profile spec says:
-    *
-    *     "gl_VertexID holds the integer index i implicitly passed by
-    *     DrawArrays or one of the other drawing commands defined in section
-    *     2.8.3."
-    *
-    * Section 2.8.3 (Drawing Commands) of the same spec says:
-    *
-    *     "The commands....are equivalent to the commands with the same base
-    *     name (without the BaseVertex suffix), except that the ith element
-    *     transferred by the corresponding draw call will be taken from
-    *     element indices[i] + basevertex of each enabled array."
-    *
-    * Additionally, the overview in the GL_ARB_shader_draw_parameters spec
-    * says:
-    *
-    *     "In unextended GL, vertex shaders have inputs named gl_VertexID and
-    *     gl_InstanceID, which contain, respectively the index of the vertex
-    *     and instance. The value of gl_VertexID is the implicitly passed
-    *     index of the vertex being processed, which includes the value of
-    *     baseVertex, for those commands that accept it."
-    *
-    * gl_VertexID gets basevertex added in.  This differs from DirectX where
-    * SV_VertexID does \b not get basevertex added in.
-    *
-    * \note
-    * If all system values are available, \c SYSTEM_VALUE_VERTEX_ID will be
-    * equal to \c SYSTEM_VALUE_VERTEX_ID_ZERO_BASE plus
-    * \c SYSTEM_VALUE_BASE_VERTEX.
-    *
-    * \sa SYSTEM_VALUE_VERTEX_ID_ZERO_BASE, SYSTEM_VALUE_BASE_VERTEX
-    */
-   SYSTEM_VALUE_VERTEX_ID,
-
-   /**
-    * Instanced ID as supplied to gl_InstanceID
-    *
-    * Values assigned to gl_InstanceID always begin with zero, regardless of
-    * the value of baseinstance.
-    *
-    * Section 11.1.3.9 (Shader Inputs) of the OpenGL 4.4 core profile spec
-    * says:
-    *
-    *     "gl_InstanceID holds the integer instance number of the current
-    *     primitive in an instanced draw call (see section 10.5)."
-    *
-    * Through a big chain of pseudocode, section 10.5 describes that
-    * baseinstance is not counted by gl_InstanceID.  In that section, notice
-    *
-    *     "If an enabled vertex attribute array is instanced (it has a
-    *     non-zero divisor as specified by VertexAttribDivisor), the element
-    *     index that is transferred to the GL, for all vertices, is given by
-    *
-    *         floor(instance/divisor) + baseinstance
-    *
-    *     If an array corresponding to an attribute required by a vertex
-    *     shader is not enabled, then the corresponding element is taken from
-    *     the current attribute state (see section 10.2)."
-    *
-    * Note that baseinstance is \b not included in the value of instance.
-    */
-   SYSTEM_VALUE_INSTANCE_ID,
-
-   /**
-    * DirectX-style vertex ID.
-    *
-    * Unlike \c SYSTEM_VALUE_VERTEX_ID, this system value does \b not include
-    * the value of basevertex.
-    *
-    * \sa SYSTEM_VALUE_VERTEX_ID, SYSTEM_VALUE_BASE_VERTEX
-    */
-   SYSTEM_VALUE_VERTEX_ID_ZERO_BASE,
-
-   /**
-    * Value of \c basevertex passed to \c glDrawElementsBaseVertex and similar
-    * functions.
-    *
-    * \sa SYSTEM_VALUE_VERTEX_ID, SYSTEM_VALUE_VERTEX_ID_ZERO_BASE
-    */
-   SYSTEM_VALUE_BASE_VERTEX,
-   /*@}*/
-
-   /**
-    * \name Geometry shader system values
-    */
-   /*@{*/
-   SYSTEM_VALUE_INVOCATION_ID,
-   /*@}*/
-
-   /**
-    * \name Fragment shader system values
-    */
-   /*@{*/
-   SYSTEM_VALUE_FRONT_FACE,     /**< (not done yet) */
-   SYSTEM_VALUE_SAMPLE_ID,
-   SYSTEM_VALUE_SAMPLE_POS,
-   SYSTEM_VALUE_SAMPLE_MASK_IN,
-   /*@}*/
-
-   SYSTEM_VALUE_MAX             /**< Number of values */
-} gl_system_value;
-
-
-/**
- * The possible interpolation qualifiers that can be applied to a fragment
- * shader input in GLSL.
- *
- * Note: INTERP_QUALIFIER_NONE must be 0 so that memsetting the
- * gl_fragment_program data structure to 0 causes the default behavior.
- */
-enum glsl_interp_qualifier
-{
-   INTERP_QUALIFIER_NONE = 0,
-   INTERP_QUALIFIER_SMOOTH,
-   INTERP_QUALIFIER_FLAT,
-   INTERP_QUALIFIER_NOPERSPECTIVE,
-   INTERP_QUALIFIER_COUNT /**< Number of interpolation qualifiers */
-};
-
-
-/**
  * \brief Layout qualifiers for gl_FragDepth.
  *
  * Extension AMD_conservative_depth allows gl_FragDepth to be redeclared with
@@ -2247,6 +2087,8 @@ struct gl_program
 
    struct prog_instruction *Instructions;
 
+   struct nir_shader *nir;
+
    GLbitfield64 InputsRead;     /**< Bitmask of which input regs are read */
    GLbitfield64 OutputsWritten; /**< Bitmask of which output regs are written */
    GLbitfield SystemValuesRead;   /**< Bitmask of SYSTEM_VALUE_x inputs used */
@@ -2762,6 +2604,16 @@ struct gl_active_atomic_buffer
 };
 
 /**
+ * Active resource in a gl_shader_program
+ */
+struct gl_program_resource
+{
+   GLenum Type; /** Program interface type. */
+   const void *Data; /** Pointer to resource associated data structure. */
+   uint8_t StageReferences; /** Bitmask of shader stage references. */
+};
+
+/**
  * A GLSL program object.
  * Basically a linked collection of vertex and fragment shaders.
  */
@@ -2935,6 +2787,10 @@ struct gl_shader_program
     */
    struct gl_shader *_LinkedShaders[MESA_SHADER_STAGES];
 
+   /** List of all active resources after linking. */
+   struct gl_program_resource *ProgramResourceList;
+   unsigned NumProgramResourceList;
+
    /* True if any of the fragment shaders attached to this program use:
     * #extension ARB_fragment_coord_conventions: enable
     */
@@ -3527,8 +3383,8 @@ struct gl_constants
    GLboolean ForceGLSLExtensionsWarn;
 
    /**
-    * If non-zero, forces GLSL shaders without the #version directive to behave
-    * as if they began with "#version ForceGLSLVersion".
+    * If non-zero, forces GLSL shaders to behave as if they began
+    * with "#version ForceGLSLVersion".
     */
    GLuint ForceGLSLVersion;
 
@@ -4523,7 +4379,7 @@ struct gl_context
 #ifdef DEBUG
 extern int MESA_VERBOSE;
 extern int MESA_DEBUG_FLAGS;
-# define MESA_FUNCTION __FUNCTION__
+# define MESA_FUNCTION __func__
 #else
 # define MESA_VERBOSE 0
 # define MESA_DEBUG_FLAGS 0
diff --git a/mesalib/src/mesa/main/multisample.c b/mesalib/src/mesa/main/multisample.c
index 0e9207bc1..816837b95 100644
--- a/mesalib/src/mesa/main/multisample.c
+++ b/mesalib/src/mesa/main/multisample.c
@@ -150,6 +150,15 @@ GLenum
 _mesa_check_sample_count(struct gl_context *ctx, GLenum target,
                          GLenum internalFormat, GLsizei samples)
 {
+   /* Section 2.5 (GL Errors) of OpenGL 3.0 specification, page 16:
+    *
+    * "If a negative number is provided where an argument of type sizei or
+    * sizeiptr is specified, the error INVALID VALUE is generated."
+    */
+   if (samples < 0) {
+      return GL_INVALID_VALUE;
+   }
+
    /* Section 4.4 (Framebuffer objects), page 198 of the OpenGL ES 3.0.0
     * specification says:
     *
diff --git a/mesalib/src/mesa/main/pipelineobj.c b/mesalib/src/mesa/main/pipelineobj.c
index fb241af1e..0fefa7d56 100644
--- a/mesalib/src/mesa/main/pipelineobj.c
+++ b/mesalib/src/mesa/main/pipelineobj.c
@@ -498,16 +498,18 @@ _mesa_DeleteProgramPipelines(GLsizei n, const GLuint *pipelines)
  * \param n       Number of IDs to generate.
  * \param pipelines  pipeline of \c n locations to store the IDs.
  */
-void GLAPIENTRY
-_mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines)
+static void
+create_program_pipelines(struct gl_context *ctx, GLsizei n, GLuint *pipelines,
+                         bool dsa)
 {
-   GET_CURRENT_CONTEXT(ctx);
-
+   const char *func;
    GLuint first;
    GLint i;
 
+   func = dsa ? "glCreateProgramPipelines" : "glGenProgramPipelines";
+
    if (n < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenProgramPipelines(n<0)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s (n < 0)", func);
       return;
    }
 
@@ -523,16 +525,37 @@ _mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines)
 
       obj = _mesa_new_pipeline_object(ctx, name);
       if (!obj) {
-         _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenProgramPipelines");
+         _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
          return;
       }
 
+      if (dsa) {
+         /* make dsa-allocated objects behave like program objects */
+         obj->EverBound = GL_TRUE;
+      }
+
       save_pipeline_object(ctx, obj);
       pipelines[i] = first + i;
    }
 
 }
 
+void GLAPIENTRY
+_mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   create_program_pipelines(ctx, n, pipelines, false);
+}
+
+void GLAPIENTRY
+_mesa_CreateProgramPipelines(GLsizei n, GLuint *pipelines)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   create_program_pipelines(ctx, n, pipelines, true);
+}
+
 /**
  * Determine if ID is the name of an pipeline object.
  *
diff --git a/mesalib/src/mesa/main/pipelineobj.h b/mesalib/src/mesa/main/pipelineobj.h
index 7285a78f1..b57bcb99e 100644
--- a/mesalib/src/mesa/main/pipelineobj.h
+++ b/mesalib/src/mesa/main/pipelineobj.h
@@ -82,6 +82,9 @@ _mesa_DeleteProgramPipelines(GLsizei n, const GLuint *pipelines);
 extern void GLAPIENTRY
 _mesa_GenProgramPipelines(GLsizei n, GLuint *pipelines);
 
+void GLAPIENTRY
+_mesa_CreateProgramPipelines(GLsizei n, GLuint *pipelines);
+
 extern GLboolean GLAPIENTRY
 _mesa_IsProgramPipeline(GLuint pipeline);
 
diff --git a/mesalib/src/mesa/main/pixel.c b/mesalib/src/mesa/main/pixel.c
index 7162c6fef..ecda2694f 100644
--- a/mesalib/src/mesa/main/pixel.c
+++ b/mesalib/src/mesa/main/pixel.c
@@ -30,7 +30,6 @@
 
 #include "glheader.h"
 #include "bufferobj.h"
-#include "colormac.h"
 #include "context.h"
 #include "macros.h"
 #include "pixel.h"
diff --git a/mesalib/src/mesa/main/pixeltransfer.c b/mesalib/src/mesa/main/pixeltransfer.c
index 8bbeeb853..94464ea67 100644
--- a/mesalib/src/mesa/main/pixeltransfer.c
+++ b/mesalib/src/mesa/main/pixeltransfer.c
@@ -31,7 +31,7 @@
 
 
 #include "glheader.h"
-#include "colormac.h"
+#include "macros.h"
 #include "pixeltransfer.h"
 #include "imports.h"
 #include "mtypes.h"
diff --git a/mesalib/src/mesa/main/program_resource.c b/mesalib/src/mesa/main/program_resource.c
new file mode 100644
index 000000000..b15a13210
--- /dev/null
+++ b/mesalib/src/mesa/main/program_resource.c
@@ -0,0 +1,417 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright (C) 2015 Intel Corporation.  All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "main/enums.h"
+#include "main/macros.h"
+#include "main/mtypes.h"
+#include "main/shaderapi.h"
+#include "main/shaderobj.h"
+#include "program_resource.h"
+
+static bool
+supported_interface_enum(GLenum iface)
+{
+   switch (iface) {
+   case GL_UNIFORM:
+   case GL_UNIFORM_BLOCK:
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+   case GL_TRANSFORM_FEEDBACK_VARYING:
+   case GL_ATOMIC_COUNTER_BUFFER:
+      return true;
+   case GL_VERTEX_SUBROUTINE:
+   case GL_TESS_CONTROL_SUBROUTINE:
+   case GL_TESS_EVALUATION_SUBROUTINE:
+   case GL_GEOMETRY_SUBROUTINE:
+   case GL_FRAGMENT_SUBROUTINE:
+   case GL_COMPUTE_SUBROUTINE:
+   case GL_VERTEX_SUBROUTINE_UNIFORM:
+   case GL_TESS_CONTROL_SUBROUTINE_UNIFORM:
+   case GL_TESS_EVALUATION_SUBROUTINE_UNIFORM:
+   case GL_GEOMETRY_SUBROUTINE_UNIFORM:
+   case GL_FRAGMENT_SUBROUTINE_UNIFORM:
+   case GL_COMPUTE_SUBROUTINE_UNIFORM:
+   case GL_BUFFER_VARIABLE:
+   case GL_SHADER_STORAGE_BLOCK:
+   default:
+      return false;
+   }
+}
+
+void GLAPIENTRY
+_mesa_GetProgramInterfaceiv(GLuint program, GLenum programInterface,
+                            GLenum pname, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   unsigned i;
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+                                      "glGetProgramInterfaceiv");
+   if (!shProg)
+      return;
+
+   if (!params) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glGetProgramInterfaceiv(params NULL)");
+      return;
+   }
+
+   /* Validate interface. */
+   if (!supported_interface_enum(programInterface)) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glGetProgramInterfaceiv(%s)",
+                  _mesa_lookup_enum_by_nr(programInterface));
+      return;
+   }
+
+   /* Validate pname against interface. */
+   switch(pname) {
+   case GL_ACTIVE_RESOURCES:
+      for (i = 0, *params = 0; i < shProg->NumProgramResourceList; i++)
+         if (shProg->ProgramResourceList[i].Type == programInterface)
+            (*params)++;
+      break;
+   case GL_MAX_NAME_LENGTH:
+      if (programInterface == GL_ATOMIC_COUNTER_BUFFER) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glGetProgramInterfaceiv(%s pname %s)",
+                     _mesa_lookup_enum_by_nr(programInterface),
+                     _mesa_lookup_enum_by_nr(pname));
+         return;
+      }
+      /* Name length consists of base name, 3 additional chars '[0]' if
+       * resource is an array and finally 1 char for string terminator.
+       */
+      for (i = 0, *params = 0; i < shProg->NumProgramResourceList; i++) {
+         if (shProg->ProgramResourceList[i].Type != programInterface)
+            continue;
+         const char *name =
+            _mesa_program_resource_name(&shProg->ProgramResourceList[i]);
+         unsigned array_size =
+            _mesa_program_resource_array_size(&shProg->ProgramResourceList[i]);
+         *params = MAX2(*params, strlen(name) + (array_size ? 3 : 0) + 1);
+      }
+      break;
+   case GL_MAX_NUM_ACTIVE_VARIABLES:
+      switch (programInterface) {
+      case GL_UNIFORM_BLOCK:
+         for (i = 0, *params = 0; i < shProg->NumProgramResourceList; i++) {
+            if (shProg->ProgramResourceList[i].Type == programInterface) {
+               struct gl_uniform_block *block =
+                  (struct gl_uniform_block *)
+                  shProg->ProgramResourceList[i].Data;
+               *params = MAX2(*params, block->NumUniforms);
+            }
+         }
+         break;
+      case GL_ATOMIC_COUNTER_BUFFER:
+         for (i = 0, *params = 0; i < shProg->NumProgramResourceList; i++) {
+            if (shProg->ProgramResourceList[i].Type == programInterface) {
+               struct gl_active_atomic_buffer *buffer =
+                  (struct gl_active_atomic_buffer *)
+                  shProg->ProgramResourceList[i].Data;
+               *params = MAX2(*params, buffer->NumUniforms);
+            }
+         }
+         break;
+      default:
+        _mesa_error(ctx, GL_INVALID_OPERATION,
+                    "glGetProgramInterfaceiv(%s pname %s)",
+                    _mesa_lookup_enum_by_nr(programInterface),
+                    _mesa_lookup_enum_by_nr(pname));
+      };
+      break;
+   case GL_MAX_NUM_COMPATIBLE_SUBROUTINES:
+   default:
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "glGetProgramInterfaceiv(pname %s)",
+                  _mesa_lookup_enum_by_nr(pname));
+   }
+}
+
+static bool
+is_xfb_marker(const char *str)
+{
+   static const char *markers[] = {
+      "gl_NextBuffer",
+      "gl_SkipComponents1",
+      "gl_SkipComponents2",
+      "gl_SkipComponents3",
+      "gl_SkipComponents4",
+      NULL
+   };
+   const char **m = markers;
+
+   if (strncmp(str, "gl_", 3) != 0)
+      return false;
+
+   for (; *m; m++)
+      if (strcmp(*m, str) == 0)
+         return true;
+
+   return false;
+}
+
+/**
+ * Checks if given name index is legal for GetProgramResourceIndex,
+ * check is written to be compatible with GL_ARB_array_of_arrays.
+ */
+static bool
+valid_program_resource_index_name(const GLchar *name)
+{
+   const char *array = strstr(name, "[");
+   const char *close = strrchr(name, ']');
+
+   /* Not array, no need for the check. */
+   if (!array)
+      return true;
+
+   /* Last array index has to be zero. */
+   if (!close || *--close != '0')
+      return false;
+
+   return true;
+}
+
+GLuint GLAPIENTRY
+_mesa_GetProgramResourceIndex(GLuint program, GLenum programInterface,
+                              const GLchar *name)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_program_resource *res;
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+                                      "glGetProgramResourceIndex");
+   if (!shProg || !name)
+      return GL_INVALID_INDEX;
+
+   /*
+    * For the interface TRANSFORM_FEEDBACK_VARYING, the value INVALID_INDEX
+    * should be returned when querying the index assigned to the special names
+    * "gl_NextBuffer", "gl_SkipComponents1", "gl_SkipComponents2",
+    * "gl_SkipComponents3", and "gl_SkipComponents4".
+    */
+   if (programInterface == GL_TRANSFORM_FEEDBACK_VARYING &&
+       is_xfb_marker(name))
+      return GL_INVALID_INDEX;
+
+   switch (programInterface) {
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+   case GL_UNIFORM:
+   case GL_UNIFORM_BLOCK:
+   case GL_TRANSFORM_FEEDBACK_VARYING:
+      /* Validate name syntax for arrays. */
+      if (!valid_program_resource_index_name(name))
+         return GL_INVALID_INDEX;
+
+      res = _mesa_program_resource_find_name(shProg, programInterface, name);
+      if (!res)
+         return GL_INVALID_INDEX;
+
+      return _mesa_program_resource_index(shProg, res);
+   case GL_ATOMIC_COUNTER_BUFFER:
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetProgramResourceIndex(%s)",
+                  _mesa_lookup_enum_by_nr(programInterface));
+   }
+
+   return GL_INVALID_INDEX;
+}
+
+void GLAPIENTRY
+_mesa_GetProgramResourceName(GLuint program, GLenum programInterface,
+                             GLuint index, GLsizei bufSize, GLsizei *length,
+                             GLchar *name)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program,
+                                      "glGetProgramResourceName");
+
+   /* Set user friendly return values in case of errors. */
+   if (name)
+      *name = '\0';
+   if (length)
+      *length = 0;
+
+   if (!shProg || !name)
+      return;
+
+   if (programInterface == GL_ATOMIC_COUNTER_BUFFER ||
+       !supported_interface_enum(programInterface)) {
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetProgramResourceName(%s)",
+                  _mesa_lookup_enum_by_nr(programInterface));
+      return;
+   }
+
+   _mesa_get_program_resource_name(shProg, programInterface, index, bufSize,
+                                   length, name, "glGetProgramResourceName");
+}
+
+void GLAPIENTRY
+_mesa_GetProgramResourceiv(GLuint program, GLenum programInterface,
+                           GLuint index, GLsizei propCount,
+                           const GLenum *props, GLsizei bufSize,
+                           GLsizei *length, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      _mesa_lookup_shader_program_err(ctx, program, "glGetProgramResourceiv");
+
+   if (!shProg || !params)
+      return;
+
+   /* The error INVALID_VALUE is generated if <propCount> is zero.
+    * Note that we check < 0 here because it makes sense to bail early.
+    */
+   if (propCount <= 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "glGetProgramResourceiv(propCount <= 0)");
+      return;
+   }
+
+   /* No need to write any properties, user requested none. */
+   if (bufSize == 0)
+      return;
+
+   _mesa_get_program_resourceiv(shProg, programInterface, index,
+                                propCount, props, bufSize, length, params);
+}
+
+/**
+ * Function verifies syntax of given name for GetProgramResourceLocation
+ * and GetProgramResourceLocationIndex for the following cases:
+ *
+ * "array element portion of a string passed to GetProgramResourceLocation
+ * or GetProgramResourceLocationIndex must not have, a "+" sign, extra
+ * leading zeroes, or whitespace".
+ *
+ * Check is written to be compatible with GL_ARB_array_of_arrays.
+ */
+static bool
+invalid_array_element_syntax(const GLchar *name)
+{
+   char *first = strchr(name, '[');
+   char *last = strrchr(name, '[');
+
+   if (!first)
+      return false;
+
+   /* No '+' or ' ' allowed anywhere. */
+   if (strchr(first, '+') || strchr(first, ' '))
+      return true;
+
+   /* Check that last array index is 0. */
+   if (last[1] == '0' && last[2] != ']')
+      return true;
+
+   return false;
+}
+
+static struct gl_shader_program *
+lookup_linked_program(GLuint program, const char *caller)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *prog =
+      _mesa_lookup_shader_program_err(ctx, program, caller);
+
+   if (!prog)
+      return NULL;
+
+   if (prog->LinkStatus == GL_FALSE) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(program not linked)",
+                  caller);
+      return NULL;
+   }
+   return prog;
+}
+
+GLint GLAPIENTRY
+_mesa_GetProgramResourceLocation(GLuint program, GLenum programInterface,
+                                 const GLchar *name)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      lookup_linked_program(program, "glGetProgramResourceLocation");
+
+   if (!shProg || !name || invalid_array_element_syntax(name))
+      return -1;
+
+   /* Validate programInterface. */
+   switch (programInterface) {
+   case GL_UNIFORM:
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+      break;
+
+   /* For reference valid cases requiring additional extension support:
+    * GL_ARB_shader_subroutine
+    * GL_ARB_tessellation_shader
+    * GL_ARB_compute_shader
+    */
+   case GL_VERTEX_SUBROUTINE_UNIFORM:
+   case GL_TESS_CONTROL_SUBROUTINE_UNIFORM:
+   case GL_TESS_EVALUATION_SUBROUTINE_UNIFORM:
+   case GL_GEOMETRY_SUBROUTINE_UNIFORM:
+   case GL_FRAGMENT_SUBROUTINE_UNIFORM:
+   case GL_COMPUTE_SUBROUTINE_UNIFORM:
+
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "glGetProgramResourceLocation(%s %s)",
+                  _mesa_lookup_enum_by_nr(programInterface), name);
+   }
+
+   return _mesa_program_resource_location(shProg, programInterface, name);
+}
+
+/**
+ * Returns output index for dual source blending.
+ */
+GLint GLAPIENTRY
+_mesa_GetProgramResourceLocationIndex(GLuint program, GLenum programInterface,
+                                      const GLchar *name)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg =
+      lookup_linked_program(program, "glGetProgramResourceLocationIndex");
+
+   if (!shProg || !name || invalid_array_element_syntax(name))
+      return -1;
+
+   /* From the GL_ARB_program_interface_query spec:
+    *
+    * "For GetProgramResourceLocationIndex, <programInterface> must be
+    * PROGRAM_OUTPUT."
+    */
+   if (programInterface != GL_PROGRAM_OUTPUT) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glGetProgramResourceLocationIndex(%s)",
+                  _mesa_lookup_enum_by_nr(programInterface));
+      return -1;
+   }
+
+   return _mesa_program_resource_location_index(shProg, GL_PROGRAM_OUTPUT,
+                                                name);
+}
diff --git a/mesalib/src/mesa/main/program_resource.h b/mesalib/src/mesa/main/program_resource.h
new file mode 100644
index 000000000..326ae1f93
--- /dev/null
+++ b/mesalib/src/mesa/main/program_resource.h
@@ -0,0 +1,58 @@
+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright (C) 2015 Intel Corporation.  All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef PROGRAM_RESOURCE_H
+#define PROGRAM_RESOURCE_H
+
+#include "glheader.h"
+
+extern void GLAPIENTRY
+_mesa_GetProgramInterfaceiv(GLuint program, GLenum programInterface,
+                            GLenum pname, GLint *params);
+
+extern GLuint GLAPIENTRY
+_mesa_GetProgramResourceIndex(GLuint program, GLenum programInterface,
+                              const GLchar *name);
+
+extern void GLAPIENTRY
+_mesa_GetProgramResourceName(GLuint program, GLenum programInterface,
+                             GLuint index, GLsizei bufSize, GLsizei *length,
+                             GLchar *name);
+
+extern void GLAPIENTRY
+_mesa_GetProgramResourceiv(GLuint program, GLenum programInterface,
+                           GLuint index, GLsizei propCount,
+                           const GLenum *props, GLsizei bufSize,
+                           GLsizei *length, GLint *params);
+
+extern GLint GLAPIENTRY
+_mesa_GetProgramResourceLocation(GLuint program, GLenum programInterface,
+                                 const GLchar *name);
+
+extern GLint GLAPIENTRY
+_mesa_GetProgramResourceLocationIndex(GLuint program, GLenum programInterface,
+                                      const GLchar *name);
+
+#endif
diff --git a/mesalib/src/mesa/main/queryobj.c b/mesalib/src/mesa/main/queryobj.c
index 9ecc7d3b5..12363de7b 100755
--- a/mesalib/src/mesa/main/queryobj.c
+++ b/mesalib/src/mesa/main/queryobj.c
@@ -233,18 +233,22 @@ get_query_binding_point(struct gl_context *ctx, GLenum target, GLuint index)
    }
 }
 
-
-void GLAPIENTRY
-_mesa_GenQueries(GLsizei n, GLuint *ids)
+/**
+ * Create $n query objects and store them in *ids. Make them of type $target
+ * if dsa is set. Called from _mesa_GenQueries() and _mesa_CreateQueries().
+ */
+static void
+create_queries(struct gl_context *ctx, GLenum target, GLsizei n, GLuint *ids,
+               bool dsa)
 {
+   const char *func = dsa ? "glGenQueries" : "glCreateQueries";
    GLuint first;
-   GET_CURRENT_CONTEXT(ctx);
 
    if (MESA_VERBOSE & VERBOSE_API)
-      _mesa_debug(ctx, "glGenQueries(%d)\n", n);
+      _mesa_debug(ctx, "%s(%d)\n", func, n);
 
    if (n < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenQueriesARB(n < 0)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(n < 0)", func);
       return;
    }
 
@@ -255,8 +259,12 @@ _mesa_GenQueries(GLsizei n, GLuint *ids)
          struct gl_query_object *q
             = ctx->Driver.NewQueryObject(ctx, first + i);
          if (!q) {
-            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenQueriesARB");
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
             return;
+         } else if (dsa) {
+            /* Do the equivalent of binding the buffer with a target */
+            q->Target = target;
+            q->EverBound = GL_TRUE;
          }
          ids[i] = first + i;
          _mesa_HashInsert(ctx->Query.QueryObjects, first + i, q);
@@ -264,6 +272,36 @@ _mesa_GenQueries(GLsizei n, GLuint *ids)
    }
 }
 
+void GLAPIENTRY
+_mesa_GenQueries(GLsizei n, GLuint *ids)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   create_queries(ctx, 0, n, ids, false);
+}
+
+void GLAPIENTRY
+_mesa_CreateQueries(GLenum target, GLsizei n, GLuint *ids)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   switch (target) {
+   case GL_SAMPLES_PASSED:
+   case GL_ANY_SAMPLES_PASSED:
+   case GL_ANY_SAMPLES_PASSED_CONSERVATIVE:
+   case GL_TIME_ELAPSED:
+   case GL_TIMESTAMP:
+   case GL_PRIMITIVES_GENERATED:
+   case GL_TRANSFORM_FEEDBACK_PRIMITIVES_WRITTEN:
+      break;
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "glCreateQueries(invalid target = %s)",
+                  _mesa_lookup_enum_by_nr(target));
+      return;
+   }
+
+   create_queries(ctx, target, n, ids, true);
+}
+
 
 void GLAPIENTRY
 _mesa_DeleteQueries(GLsizei n, const GLuint *ids)
@@ -424,6 +462,18 @@ _mesa_BeginQueryIndexed(GLenum target, GLuint index, GLuint id)
       }
    }
 
+   /* This possibly changes the target of a buffer allocated by
+    * CreateQueries. Issue 39) in the ARB_direct_state_access extension states
+    * the following:
+    *
+    * "CreateQueries adds a <target>, so strictly speaking the <target>
+    * command isn't needed for BeginQuery/EndQuery, but in the end, this also
+    * isn't a selector, so we decided not to change it."
+    *
+    * Updating the target of the query object should be acceptable, so let's
+    * do that.
+    */
+
    q->Target = target;
    q->Active = GL_TRUE;
    q->Result = 0;
@@ -541,6 +591,18 @@ _mesa_QueryCounter(GLuint id, GLenum target)
       return;
    }
 
+   /* This possibly changes the target of a buffer allocated by
+    * CreateQueries. Issue 39) in the ARB_direct_state_access extension states
+    * the following:
+    *
+    * "CreateQueries adds a <target>, so strictly speaking the <target>
+    * command isn't needed for BeginQuery/EndQuery, but in the end, this also
+    * isn't a selector, so we decided not to change it."
+    *
+    * Updating the target of the query object should be acceptable, so let's
+    * do that.
+    */
+
    q->Target = target;
    q->Result = 0;
    q->Ready = GL_FALSE;
@@ -710,8 +772,8 @@ _mesa_GetQueryObjectiv(GLuint id, GLenum pname, GLint *params)
          }
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
-	 if (!q->Ready)
-	    ctx->Driver.CheckQuery( ctx, q );
+         if (!q->Ready)
+            ctx->Driver.CheckQuery( ctx, q );
          *params = q->Ready;
          break;
       default:
@@ -761,8 +823,8 @@ _mesa_GetQueryObjectuiv(GLuint id, GLenum pname, GLuint *params)
          }
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
-	 if (!q->Ready)
-	    ctx->Driver.CheckQuery( ctx, q );
+         if (!q->Ready)
+            ctx->Driver.CheckQuery( ctx, q );
          *params = q->Ready;
          break;
       default:
@@ -801,8 +863,8 @@ _mesa_GetQueryObjecti64v(GLuint id, GLenum pname, GLint64EXT *params)
          *params = q->Result;
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
-	 if (!q->Ready)
-	    ctx->Driver.CheckQuery( ctx, q );
+         if (!q->Ready)
+            ctx->Driver.CheckQuery( ctx, q );
          *params = q->Ready;
          break;
       default:
@@ -841,8 +903,8 @@ _mesa_GetQueryObjectui64v(GLuint id, GLenum pname, GLuint64EXT *params)
          *params = q->Result;
          break;
       case GL_QUERY_RESULT_AVAILABLE_ARB:
-	 if (!q->Ready)
-	    ctx->Driver.CheckQuery( ctx, q );
+         if (!q->Ready)
+            ctx->Driver.CheckQuery( ctx, q );
          *params = q->Ready;
          break;
       default:
@@ -852,6 +914,45 @@ _mesa_GetQueryObjectui64v(GLuint id, GLenum pname, GLuint64EXT *params)
 }
 
 /**
+ * New with GL_ARB_query_buffer_object
+ */
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectiv(GLuint id, GLuint buffer, GLenum pname,
+                             GLintptr offset)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   _mesa_error(ctx, GL_INVALID_OPERATION, "glGetQueryBufferObjectiv");
+}
+
+
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectuiv(GLuint id, GLuint buffer, GLenum pname,
+                              GLintptr offset)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   _mesa_error(ctx, GL_INVALID_OPERATION, "glGetQueryBufferObjectuiv");
+}
+
+
+void GLAPIENTRY
+_mesa_GetQueryBufferObjecti64v(GLuint id, GLuint buffer, GLenum pname,
+                               GLintptr offset)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   _mesa_error(ctx, GL_INVALID_OPERATION, "glGetQueryBufferObjecti64v");
+}
+
+
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectui64v(GLuint id, GLuint buffer, GLenum pname,
+                                GLintptr offset)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   _mesa_error(ctx, GL_INVALID_OPERATION, "glGetQueryBufferObjectui64v");
+}
+
+
+/**
  * Allocate/init the context state related to query objects.
  */
 void
diff --git a/mesalib/src/mesa/main/queryobj.h b/mesalib/src/mesa/main/queryobj.h
index 6cbcabd46..d1036fcce 100644
--- a/mesalib/src/mesa/main/queryobj.h
+++ b/mesalib/src/mesa/main/queryobj.h
@@ -51,6 +51,8 @@ _mesa_free_queryobj_data(struct gl_context *ctx);
 void GLAPIENTRY
 _mesa_GenQueries(GLsizei n, GLuint *ids);
 void GLAPIENTRY
+_mesa_CreateQueries(GLenum target, GLsizei n, GLuint *ids);
+void GLAPIENTRY
 _mesa_DeleteQueries(GLsizei n, const GLuint *ids);
 GLboolean GLAPIENTRY
 _mesa_IsQuery(GLuint id);
@@ -77,5 +79,17 @@ void GLAPIENTRY
 _mesa_GetQueryObjecti64v(GLuint id, GLenum pname, GLint64EXT *params);
 void GLAPIENTRY
 _mesa_GetQueryObjectui64v(GLuint id, GLenum pname, GLuint64EXT *params);
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectiv(GLuint id, GLuint buffer, GLenum pname,
+                             GLintptr offset);
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectuiv(GLuint id, GLuint buffer, GLenum pname,
+                              GLintptr offset);
+void GLAPIENTRY
+_mesa_GetQueryBufferObjecti64v(GLuint id, GLuint buffer, GLenum pname,
+                               GLintptr offset);
+void GLAPIENTRY
+_mesa_GetQueryBufferObjectui64v(GLuint id, GLuint buffer, GLenum pname,
+                                GLintptr offset);
 
 #endif /* QUERYOBJ_H */
diff --git a/mesalib/src/mesa/main/samplerobj.c b/mesalib/src/mesa/main/samplerobj.c
index d66b0b5e8..a3aacc66a 100644
--- a/mesalib/src/mesa/main/samplerobj.c
+++ b/mesalib/src/mesa/main/samplerobj.c
@@ -181,19 +181,18 @@ _mesa_delete_sampler_object(struct gl_context *ctx,
    free(sampObj);
 }
 
-
-void GLAPIENTRY
-_mesa_GenSamplers(GLsizei count, GLuint *samplers)
+static void
+create_samplers(struct gl_context *ctx, GLsizei count, GLuint *samplers,
+                const char *caller)
 {
-   GET_CURRENT_CONTEXT(ctx);
    GLuint first;
    GLint i;
 
    if (MESA_VERBOSE & VERBOSE_API)
-      _mesa_debug(ctx, "glGenSamplers(%d)\n", count);
+      _mesa_debug(ctx, "%s(%d)\n", caller, count);
 
    if (count < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenSamplers");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(n<0)", caller);
       return;
    }
 
@@ -211,6 +210,20 @@ _mesa_GenSamplers(GLsizei count, GLuint *samplers)
    }
 }
 
+void GLAPIENTRY
+_mesa_GenSamplers(GLsizei count, GLuint *samplers)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   create_samplers(ctx, count, samplers, "glGenSamplers");
+}
+
+void GLAPIENTRY
+_mesa_CreateSamplers(GLsizei count, GLuint *samplers)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   create_samplers(ctx, count, samplers, "glCreateSamplers");
+}
+
 
 void GLAPIENTRY
 _mesa_DeleteSamplers(GLsizei count, const GLuint *samplers)
diff --git a/mesalib/src/mesa/main/samplerobj.h b/mesalib/src/mesa/main/samplerobj.h
index 1bb3193e4..7bea91114 100644
--- a/mesalib/src/mesa/main/samplerobj.h
+++ b/mesalib/src/mesa/main/samplerobj.h
@@ -32,6 +32,9 @@ extern "C" {
 #endif
 
 
+#include "mtypes.h"
+
+
 struct dd_function_table;
 
 static inline struct gl_sampler_object *
@@ -80,6 +83,8 @@ _mesa_init_sampler_object_functions(struct dd_function_table *driver);
 void GLAPIENTRY
 _mesa_GenSamplers(GLsizei count, GLuint *samplers);
 void GLAPIENTRY
+_mesa_CreateSamplers(GLsizei count, GLuint *samplers);
+void GLAPIENTRY
 _mesa_DeleteSamplers(GLsizei count, const GLuint *samplers);
 GLboolean GLAPIENTRY
 _mesa_IsSampler(GLuint sampler);
diff --git a/mesalib/src/mesa/main/shader_query.cpp b/mesalib/src/mesa/main/shader_query.cpp
index df9081b73..b5f1d082c 100644
--- a/mesalib/src/mesa/main/shader_query.cpp
+++ b/mesalib/src/mesa/main/shader_query.cpp
@@ -34,11 +34,33 @@
 #include "shaderobj.h"
 #include "program/hash_table.h"
 #include "../glsl/program.h"
+#include "uniforms.h"
+#include "main/enums.h"
 
 extern "C" {
 #include "shaderapi.h"
 }
 
+static GLint
+program_resource_location(struct gl_shader_program *shProg,
+                          struct gl_program_resource *res, const char *name);
+
+/**
+ * Declare convenience functions to return resource data in a given type.
+ * Warning! this is not type safe so be *very* careful when using these.
+ */
+#define DECL_RESOURCE_FUNC(name, type) \
+const type * RESOURCE_ ## name (gl_program_resource *res) { \
+   assert(res->Data); \
+   return (type *) res->Data; \
+}
+
+DECL_RESOURCE_FUNC(VAR, ir_variable);
+DECL_RESOURCE_FUNC(UBO, gl_uniform_block);
+DECL_RESOURCE_FUNC(UNI, gl_uniform_storage);
+DECL_RESOURCE_FUNC(ATC, gl_active_atomic_buffer);
+DECL_RESOURCE_FUNC(XFB, gl_transform_feedback_varying_info);
+
 void GLAPIENTRY
 _mesa_BindAttribLocation(GLhandleARB program, GLuint index,
                             const GLcharARB *name)
@@ -129,45 +151,41 @@ _mesa_GetActiveAttrib(GLhandleARB program, GLuint desired_index,
       return;
    }
 
-   exec_list *const ir = shProg->_LinkedShaders[MESA_SHADER_VERTEX]->ir;
-   unsigned current_index = 0;
-
-   foreach_in_list(ir_instruction, node, ir) {
-      const ir_variable *const var = node->as_variable();
-
-      if (!is_active_attrib(var))
-         continue;
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_index(shProg, GL_PROGRAM_INPUT,
+                                        desired_index);
 
-      if (current_index == desired_index) {
-         const char *var_name = var->name;
+   /* User asked for index that does not exist. */
+   if (!res) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveAttrib(index)");
+      return;
+   }
 
-         /* Since gl_VertexID may be lowered to gl_VertexIDMESA, we need to
-          * consider gl_VertexIDMESA as gl_VertexID for purposes of checking
-          * active attributes.
-          */
-         if (var->data.mode == ir_var_system_value &&
-             var->data.location == SYSTEM_VALUE_VERTEX_ID_ZERO_BASE) {
-            var_name = "gl_VertexID";
-         }
+   const ir_variable *const var = RESOURCE_VAR(res);
 
-	 _mesa_copy_string(name, maxLength, length, var_name);
+   if (!is_active_attrib(var))
+      return;
 
-	 if (size)
-	    *size = (var->type->is_array()) ? var->type->length : 1;
+   const char *var_name = var->name;
 
-	 if (type)
-	    *type = var->type->gl_type;
+   /* Since gl_VertexID may be lowered to gl_VertexIDMESA, we need to
+    * consider gl_VertexIDMESA as gl_VertexID for purposes of checking
+    * active attributes.
+    */
+   if (var->data.mode == ir_var_system_value &&
+       var->data.location == SYSTEM_VALUE_VERTEX_ID_ZERO_BASE) {
+      var_name = "gl_VertexID";
+   }
 
-	 return;
-      }
+   _mesa_copy_string(name, maxLength, length, var_name);
 
-      current_index++;
-   }
+   if (size)
+      _mesa_program_resource_prop(shProg, res, desired_index, GL_ARRAY_SIZE,
+                                  size, "glGetActiveAttrib");
 
-   /* If the loop did not return early, the caller must have asked for
-    * an index that did not exit.  Set an error.
-    */
-   _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveAttrib(index)");
+   if (type)
+      _mesa_program_resource_prop(shProg, res, desired_index, GL_TYPE,
+                                  (GLint *) type, "glGetActiveAttrib");
 }
 
 /* Locations associated with shader variables (array or non-array) can be
@@ -252,31 +270,25 @@ _mesa_GetAttribLocation(GLhandleARB program, const GLcharARB * name)
    if (shProg->_LinkedShaders[MESA_SHADER_VERTEX] == NULL)
       return -1;
 
-   exec_list *ir = shProg->_LinkedShaders[MESA_SHADER_VERTEX]->ir;
-   foreach_in_list(ir_instruction, node, ir) {
-      const ir_variable *const var = node->as_variable();
-
-      /* The extra check against VERT_ATTRIB_GENERIC0 is because
-       * glGetAttribLocation cannot be used on "conventional" attributes.
-       *
-       * From page 95 of the OpenGL 3.0 spec:
-       *
-       *     "If name is not an active attribute, if name is a conventional
-       *     attribute, or if an error occurs, -1 will be returned."
-       */
-      if (var == NULL
-	  || var->data.mode != ir_var_shader_in
-	  || var->data.location == -1
-	  || var->data.location < VERT_ATTRIB_GENERIC0)
-	 continue;
-
-      int index = get_matching_index(var, (const char *) name);
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_name(shProg, GL_PROGRAM_INPUT, name);
 
-      if (index >= 0)
-         return var->data.location + index - VERT_ATTRIB_GENERIC0;
-   }
+   if (!res)
+      return -1;
 
-   return -1;
+   GLint loc = program_resource_location(shProg, res, name);
+
+   /* The extra check against against 0 is made because of builtin-attribute
+    * locations that have offset applied. Function program_resource_location
+    * can return built-in attribute locations < 0 and glGetAttribLocation
+    * cannot be used on "conventional" attributes.
+    *
+    * From page 95 of the OpenGL 3.0 spec:
+    *
+    *     "If name is not an active attribute, if name is a conventional
+    *     attribute, or if an error occurs, -1 will be returned."
+    */
+   return (loc >= 0) ? loc : -1;
 }
 
 
@@ -416,29 +428,8 @@ _mesa_GetFragDataIndex(GLuint program, const GLchar *name)
    if (shProg->_LinkedShaders[MESA_SHADER_FRAGMENT] == NULL)
       return -1;
 
-   exec_list *ir = shProg->_LinkedShaders[MESA_SHADER_FRAGMENT]->ir;
-   foreach_in_list(ir_instruction, node, ir) {
-      const ir_variable *const var = node->as_variable();
-
-      /* The extra check against FRAG_RESULT_DATA0 is because
-       * glGetFragDataLocation cannot be used on "conventional" attributes.
-       *
-       * From page 95 of the OpenGL 3.0 spec:
-       *
-       *     "If name is not an active attribute, if name is a conventional
-       *     attribute, or if an error occurs, -1 will be returned."
-       */
-      if (var == NULL
-          || var->data.mode != ir_var_shader_out
-          || var->data.location == -1
-          || var->data.location < FRAG_RESULT_DATA0)
-         continue;
-
-      if (get_matching_index(var, (const char *) name) >= 0)
-         return var->data.index;
-   }
-
-   return -1;
+   return _mesa_program_resource_location_index(shProg, GL_PROGRAM_OUTPUT,
+                                                name);
 }
 
 GLint GLAPIENTRY
@@ -472,29 +463,612 @@ _mesa_GetFragDataLocation(GLuint program, const GLchar *name)
    if (shProg->_LinkedShaders[MESA_SHADER_FRAGMENT] == NULL)
       return -1;
 
-   exec_list *ir = shProg->_LinkedShaders[MESA_SHADER_FRAGMENT]->ir;
-   foreach_in_list(ir_instruction, node, ir) {
-      const ir_variable *const var = node->as_variable();
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_name(shProg, GL_PROGRAM_OUTPUT, name);
 
-      /* The extra check against FRAG_RESULT_DATA0 is because
-       * glGetFragDataLocation cannot be used on "conventional" attributes.
-       *
-       * From page 95 of the OpenGL 3.0 spec:
+   if (!res)
+      return -1;
+
+   GLint loc = program_resource_location(shProg, res, name);
+
+   /* The extra check against against 0 is made because of builtin-attribute
+    * locations that have offset applied. Function program_resource_location
+    * can return built-in attribute locations < 0 and glGetFragDataLocation
+    * cannot be used on "conventional" attributes.
+    *
+    * From page 95 of the OpenGL 3.0 spec:
+    *
+    *     "If name is not an active attribute, if name is a conventional
+    *     attribute, or if an error occurs, -1 will be returned."
+    */
+   return (loc >= 0) ? loc : -1;
+}
+
+const char*
+_mesa_program_resource_name(struct gl_program_resource *res)
+{
+   switch (res->Type) {
+   case GL_UNIFORM_BLOCK:
+      return RESOURCE_UBO(res)->Name;
+   case GL_TRANSFORM_FEEDBACK_VARYING:
+      return RESOURCE_XFB(res)->Name;
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+      return RESOURCE_VAR(res)->name;
+   case GL_UNIFORM:
+      return RESOURCE_UNI(res)->name;
+   default:
+      assert(!"support for resource type not implemented");
+   }
+   return NULL;
+}
+
+
+unsigned
+_mesa_program_resource_array_size(struct gl_program_resource *res)
+{
+   switch (res->Type) {
+   case GL_TRANSFORM_FEEDBACK_VARYING:
+      return RESOURCE_XFB(res)->Size > 1 ?
+             RESOURCE_XFB(res)->Size : 0;
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+      return RESOURCE_VAR(res)->data.max_array_access;
+   case GL_UNIFORM:
+      return RESOURCE_UNI(res)->array_elements;
+   case GL_ATOMIC_COUNTER_BUFFER:
+   case GL_UNIFORM_BLOCK:
+      return 0;
+   default:
+      assert(!"support for resource type not implemented");
+   }
+   return 0;
+}
+
+static int
+array_index_of_resource(struct gl_program_resource *res,
+                        const char *name)
+{
+   assert(res->Data);
+
+   switch (res->Type) {
+   case GL_PROGRAM_INPUT:
+   case GL_PROGRAM_OUTPUT:
+      return get_matching_index(RESOURCE_VAR(res), name);
+   default:
+      assert(!"support for resource type not implemented");
+   }
+}
+
+/* Find a program resource with specific name in given interface.
+ */
+struct gl_program_resource *
+_mesa_program_resource_find_name(struct gl_shader_program *shProg,
+                                 GLenum programInterface, const char *name)
+{
+   struct gl_program_resource *res = shProg->ProgramResourceList;
+   for (unsigned i = 0; i < shProg->NumProgramResourceList; i++, res++) {
+      if (res->Type != programInterface)
+         continue;
+
+      /* Resource basename. */
+      const char *rname = _mesa_program_resource_name(res);
+      unsigned baselen = strlen(rname);
+
+      switch (programInterface) {
+      case GL_TRANSFORM_FEEDBACK_VARYING:
+      case GL_UNIFORM_BLOCK:
+      case GL_UNIFORM:
+         if (strncmp(rname, name, baselen) == 0) {
+            /* Basename match, check if array or struct. */
+            if (name[baselen] == '\0' ||
+                name[baselen] == '[' ||
+                name[baselen] == '.') {
+               return res;
+            }
+         }
+         break;
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+         if (array_index_of_resource(res, name) >= 0)
+            return res;
+         break;
+      default:
+         assert(!"not implemented for given interface");
+      }
+   }
+   return NULL;
+}
+
+static GLuint
+calc_resource_index(struct gl_shader_program *shProg,
+                    struct gl_program_resource *res)
+{
+   unsigned i;
+   GLuint index = 0;
+   for (i = 0; i < shProg->NumProgramResourceList; i++) {
+      if (&shProg->ProgramResourceList[i] == res)
+         return index;
+      if (shProg->ProgramResourceList[i].Type == res->Type)
+         index++;
+   }
+   return GL_INVALID_INDEX;
+}
+
+/**
+ * Calculate index for the given resource.
+ */
+GLuint
+_mesa_program_resource_index(struct gl_shader_program *shProg,
+                             struct gl_program_resource *res)
+{
+   if (!res)
+      return GL_INVALID_INDEX;
+
+   switch (res->Type) {
+   case GL_UNIFORM_BLOCK:
+      return RESOURCE_UBO(res)- shProg->UniformBlocks;
+   case GL_ATOMIC_COUNTER_BUFFER:
+      return RESOURCE_ATC(res) - shProg->AtomicBuffers;
+   case GL_TRANSFORM_FEEDBACK_VARYING:
+   default:
+      return calc_resource_index(shProg, res);
+   }
+}
+
+/* Find a program resource with specific index in given interface.
+ */
+struct gl_program_resource *
+_mesa_program_resource_find_index(struct gl_shader_program *shProg,
+                                  GLenum programInterface, GLuint index)
+{
+   struct gl_program_resource *res = shProg->ProgramResourceList;
+   int idx = -1;
+
+   for (unsigned i = 0; i < shProg->NumProgramResourceList; i++, res++) {
+      if (res->Type != programInterface)
+         continue;
+
+      switch (res->Type) {
+      case GL_UNIFORM_BLOCK:
+      case GL_ATOMIC_COUNTER_BUFFER:
+         if (_mesa_program_resource_index(shProg, res) == index)
+            return res;
+
+      case GL_TRANSFORM_FEEDBACK_VARYING:
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+      case GL_UNIFORM:
+         if (++idx == (int) index)
+            return res;
+         break;
+      default:
+         assert(!"not implemented for given interface");
+      }
+   }
+   return NULL;
+}
+
+/* Get full name of a program resource.
+ */
+bool
+_mesa_get_program_resource_name(struct gl_shader_program *shProg,
+                                GLenum programInterface, GLuint index,
+                                GLsizei bufSize, GLsizei *length,
+                                GLchar *name, const char *caller)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   /* Find resource with given interface and index. */
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_index(shProg, programInterface, index);
+
+   /* The error INVALID_VALUE is generated if <index> is greater than
+   * or equal to the number of entries in the active resource list for
+   * <programInterface>.
+   */
+   if (!res) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(index %u)", caller, index);
+      return false;
+   }
+
+   if (bufSize < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(bufSize %d)", caller, bufSize);
+      return false;
+   }
+
+   GLsizei localLength;
+
+   if (length == NULL)
+      length = &localLength;
+
+   _mesa_copy_string(name, bufSize, length, _mesa_program_resource_name(res));
+
+   /* Page 61 (page 73 of the PDF) in section 2.11 of the OpenGL ES 3.0
+    * spec says:
+    *
+    *     "If the active uniform is an array, the uniform name returned in
+    *     name will always be the name of the uniform array appended with
+    *     "[0]"."
+    *
+    * The same text also appears in the OpenGL 4.2 spec.  It does not,
+    * however, appear in any previous spec.  Previous specifications are
+    * ambiguous in this regard.  However, either name can later be passed
+    * to glGetUniformLocation (and related APIs), so there shouldn't be any
+    * harm in always appending "[0]" to uniform array names.
+    *
+    * Geometry shader stage has different naming convention where the 'normal'
+    * condition is an array, therefore for variables referenced in geometry
+    * stage we do not add '[0]'.
+    *
+    * Note, that TCS outputs and TES inputs should not have index appended
+    * either.
+    */
+   bool add_index = !(((programInterface == GL_PROGRAM_INPUT) &&
+                       res->StageReferences & (1 << MESA_SHADER_GEOMETRY)));
+
+   if (add_index && _mesa_program_resource_array_size(res)) {
+      int i;
+
+      /* The comparison is strange because *length does *NOT* include the
+       * terminating NUL, but maxLength does.
+       */
+      for (i = 0; i < 3 && (*length + i + 1) < bufSize; i++)
+         name[*length + i] = "[0]"[i];
+
+      name[*length + i] = '\0';
+      *length += i;
+   }
+   return true;
+}
+
+static GLint
+program_resource_location(struct gl_shader_program *shProg,
+                          struct gl_program_resource *res, const char *name)
+{
+   unsigned index, offset;
+   int array_index = -1;
+
+   if (res->Type == GL_PROGRAM_INPUT || res->Type == GL_PROGRAM_OUTPUT) {
+      array_index = array_index_of_resource(res, name);
+      if (array_index < 0)
+         return -1;
+   }
+
+   /* VERT_ATTRIB_GENERIC0 and FRAG_RESULT_DATA0 are decremented as these
+    * offsets are used internally to differentiate between built-in attributes
+    * and user-defined attributes.
+    */
+   switch (res->Type) {
+   case GL_PROGRAM_INPUT:
+      return RESOURCE_VAR(res)->data.location + array_index - VERT_ATTRIB_GENERIC0;
+   case GL_PROGRAM_OUTPUT:
+      return RESOURCE_VAR(res)->data.location + array_index - FRAG_RESULT_DATA0;
+   case GL_UNIFORM:
+      index = _mesa_get_uniform_location(shProg, name, &offset);
+
+      if (index == GL_INVALID_INDEX)
+         return -1;
+
+      /* From the GL_ARB_uniform_buffer_object spec:
        *
-       *     "If name is not an active attribute, if name is a conventional
-       *     attribute, or if an error occurs, -1 will be returned."
+       *     "The value -1 will be returned if <name> does not correspond to an
+       *     active uniform variable name in <program>, if <name> is associated
+       *     with a named uniform block, or if <name> starts with the reserved
+       *     prefix "gl_"."
        */
-      if (var == NULL
-	  || var->data.mode != ir_var_shader_out
-	  || var->data.location == -1
-	  || var->data.location < FRAG_RESULT_DATA0)
-	 continue;
+      if (RESOURCE_UNI(res)->block_index != -1 ||
+          RESOURCE_UNI(res)->atomic_buffer_index != -1)
+         return -1;
 
-      int index = get_matching_index(var, (const char *) name);
+      /* location in remap table + array element offset */
+      return RESOURCE_UNI(res)->remap_location + offset;
 
-      if (index >= 0)
-         return var->data.location + index - FRAG_RESULT_DATA0;
+   default:
+      return -1;
    }
+}
 
-   return -1;
+/**
+ * Function implements following location queries:
+ *    glGetAttribLocation
+ *    glGetFragDataLocation
+ *    glGetUniformLocation
+ */
+GLint
+_mesa_program_resource_location(struct gl_shader_program *shProg,
+                                GLenum programInterface, const char *name)
+{
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_name(shProg, programInterface, name);
+
+   /* Resource not found. */
+   if (!res)
+      return -1;
+
+   return program_resource_location(shProg, res, name);
+}
+
+/**
+ * Function implements following index queries:
+ *    glGetFragDataIndex
+ */
+GLint
+_mesa_program_resource_location_index(struct gl_shader_program *shProg,
+                                      GLenum programInterface, const char *name)
+{
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_name(shProg, programInterface, name);
+
+   /* Non-existent variable or resource is not referenced by fragment stage. */
+   if (!res || !(res->StageReferences & (1 << MESA_SHADER_FRAGMENT)))
+      return -1;
+
+   return RESOURCE_VAR(res)->data.index;
+}
+
+static uint8_t
+stage_from_enum(GLenum ref)
+{
+   switch (ref) {
+   case GL_REFERENCED_BY_VERTEX_SHADER:
+      return MESA_SHADER_VERTEX;
+   case GL_REFERENCED_BY_GEOMETRY_SHADER:
+      return MESA_SHADER_GEOMETRY;
+   case GL_REFERENCED_BY_FRAGMENT_SHADER:
+      return MESA_SHADER_FRAGMENT;
+   default:
+      assert(!"shader stage not supported");
+      return MESA_SHADER_STAGES;
+   }
+}
+
+/**
+ * Check if resource is referenced by given 'referenced by' stage enum.
+ * ATC and UBO resources hold stage references of their own.
+ */
+static bool
+is_resource_referenced(struct gl_shader_program *shProg,
+                       struct gl_program_resource *res,
+                       GLuint index, uint8_t stage)
+{
+   if (res->Type == GL_ATOMIC_COUNTER_BUFFER)
+      return RESOURCE_ATC(res)->StageReferences[stage];
+
+   if (res->Type == GL_UNIFORM_BLOCK)
+      return shProg->UniformBlockStageIndex[stage][index] != -1;
+
+   return res->StageReferences & (1 << stage);
+}
+
+static unsigned
+get_buffer_property(struct gl_shader_program *shProg,
+                    struct gl_program_resource *res, const GLenum prop,
+                    GLint *val, const char *caller)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   if (res->Type != GL_UNIFORM_BLOCK &&
+       res->Type != GL_ATOMIC_COUNTER_BUFFER)
+      goto invalid_operation;
+
+   if (res->Type == GL_UNIFORM_BLOCK) {
+      switch (prop) {
+      case GL_BUFFER_BINDING:
+         *val = RESOURCE_UBO(res)->Binding;
+         return 1;
+      case GL_BUFFER_DATA_SIZE:
+         *val = RESOURCE_UBO(res)->UniformBufferSize;
+         return 1;
+      case GL_NUM_ACTIVE_VARIABLES:
+         *val = RESOURCE_UBO(res)->NumUniforms;
+         return 1;
+      case GL_ACTIVE_VARIABLES:
+         for (unsigned i = 0; i < RESOURCE_UBO(res)->NumUniforms; i++) {
+            const char *iname = RESOURCE_UBO(res)->Uniforms[i].IndexName;
+            struct gl_program_resource *uni =
+               _mesa_program_resource_find_name(shProg, GL_UNIFORM, iname);
+            *val++ =
+               _mesa_program_resource_index(shProg, uni);
+         }
+         return RESOURCE_UBO(res)->NumUniforms;
+      }
+   } else if (res->Type == GL_ATOMIC_COUNTER_BUFFER) {
+      switch (prop) {
+      case GL_BUFFER_BINDING:
+         *val = RESOURCE_ATC(res)->Binding;
+         return 1;
+      case GL_BUFFER_DATA_SIZE:
+         *val = RESOURCE_ATC(res)->MinimumSize;
+         return 1;
+      case GL_NUM_ACTIVE_VARIABLES:
+         *val = RESOURCE_ATC(res)->NumUniforms;
+         return 1;
+      case GL_ACTIVE_VARIABLES:
+         for (unsigned i = 0; i < RESOURCE_ATC(res)->NumUniforms; i++)
+            *val++ = RESOURCE_ATC(res)->Uniforms[i];
+         return RESOURCE_ATC(res)->NumUniforms;
+      }
+   }
+   assert(!"support for property type not implemented");
+
+invalid_operation:
+   _mesa_error(ctx, GL_INVALID_OPERATION, "%s(%s prop %s)", caller,
+               _mesa_lookup_enum_by_nr(res->Type),
+               _mesa_lookup_enum_by_nr(prop));
+
+   return 0;
+}
+
+unsigned
+_mesa_program_resource_prop(struct gl_shader_program *shProg,
+                            struct gl_program_resource *res, GLuint index,
+                            const GLenum prop, GLint *val, const char *caller)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+#define VALIDATE_TYPE(type)\
+   if (res->Type != type)\
+      goto invalid_operation;
+
+   switch(prop) {
+   case GL_NAME_LENGTH:
+      if (res->Type == GL_ATOMIC_COUNTER_BUFFER)
+         goto invalid_operation;
+      /* Base name +3 if array '[0]' + terminator. */
+      *val = strlen(_mesa_program_resource_name(res)) +
+         (_mesa_program_resource_array_size(res) > 0 ? 3 : 0) + 1;
+      return 1;
+   case GL_TYPE:
+      switch (res->Type) {
+      case GL_UNIFORM:
+         *val = RESOURCE_UNI(res)->type->gl_type;
+         return 1;
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+         *val = RESOURCE_VAR(res)->type->gl_type;
+         return 1;
+      case GL_TRANSFORM_FEEDBACK_VARYING:
+         *val = RESOURCE_XFB(res)->Type;
+         return 1;
+      default:
+         goto invalid_operation;
+      }
+   case GL_ARRAY_SIZE:
+      switch (res->Type) {
+      case GL_UNIFORM:
+            *val = MAX2(RESOURCE_UNI(res)->array_elements, 1);
+            return 1;
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+         *val = MAX2(RESOURCE_VAR(res)->type->length, 1);
+         return 1;
+      case GL_TRANSFORM_FEEDBACK_VARYING:
+         *val = MAX2(RESOURCE_XFB(res)->Size, 1);
+         return 1;
+      default:
+         goto invalid_operation;
+      }
+   case GL_OFFSET:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->offset;
+      return 1;
+   case GL_BLOCK_INDEX:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->block_index;
+      return 1;
+   case GL_ARRAY_STRIDE:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->array_stride;
+      return 1;
+   case GL_MATRIX_STRIDE:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->matrix_stride;
+      return 1;
+   case GL_IS_ROW_MAJOR:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->row_major;
+      return 1;
+   case GL_ATOMIC_COUNTER_BUFFER_INDEX:
+      VALIDATE_TYPE(GL_UNIFORM);
+      *val = RESOURCE_UNI(res)->atomic_buffer_index;
+      return 1;
+   case GL_BUFFER_BINDING:
+   case GL_BUFFER_DATA_SIZE:
+   case GL_NUM_ACTIVE_VARIABLES:
+   case GL_ACTIVE_VARIABLES:
+      return get_buffer_property(shProg, res, prop, val, caller);
+   case GL_REFERENCED_BY_VERTEX_SHADER:
+   case GL_REFERENCED_BY_GEOMETRY_SHADER:
+   case GL_REFERENCED_BY_FRAGMENT_SHADER:
+      switch (res->Type) {
+      case GL_UNIFORM:
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+      case GL_UNIFORM_BLOCK:
+      case GL_ATOMIC_COUNTER_BUFFER:
+         *val = is_resource_referenced(shProg, res, index,
+                                       stage_from_enum(prop));
+         return 1;
+      default:
+         goto invalid_operation;
+      }
+   case GL_LOCATION:
+      switch (res->Type) {
+      case GL_UNIFORM:
+      case GL_PROGRAM_INPUT:
+      case GL_PROGRAM_OUTPUT:
+         *val = program_resource_location(shProg, res,
+                                          _mesa_program_resource_name(res));
+         return 1;
+      default:
+         goto invalid_operation;
+      }
+   case GL_LOCATION_INDEX:
+      if (res->Type != GL_PROGRAM_OUTPUT)
+         goto invalid_operation;
+      *val = RESOURCE_VAR(res)->data.index;
+      return 1;
+
+   /* GL_ARB_tessellation_shader */
+   case GL_IS_PER_PATCH:
+   case GL_REFERENCED_BY_TESS_CONTROL_SHADER:
+   case GL_REFERENCED_BY_TESS_EVALUATION_SHADER:
+   /* GL_ARB_compute_shader */
+   case GL_REFERENCED_BY_COMPUTE_SHADER:
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(%s prop %s)", caller,
+                  _mesa_lookup_enum_by_nr(res->Type),
+                  _mesa_lookup_enum_by_nr(prop));
+      return 0;
+   }
+
+#undef VALIDATE_TYPE
+
+invalid_operation:
+   _mesa_error(ctx, GL_INVALID_OPERATION, "%s(%s prop %s)", caller,
+               _mesa_lookup_enum_by_nr(res->Type),
+               _mesa_lookup_enum_by_nr(prop));
+   return 0;
+}
+
+extern void
+_mesa_get_program_resourceiv(struct gl_shader_program *shProg,
+                             GLenum programInterface, GLuint index, GLsizei propCount,
+                             const GLenum *props, GLsizei bufSize,
+                             GLsizei *length, GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint *val = (GLint *) params;
+   const GLenum *prop = props;
+   GLsizei amount = 0;
+
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_index(shProg, programInterface, index);
+
+   /* No such resource found or bufSize negative. */
+   if (!res || bufSize < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "glGetProgramResourceiv(%s index %d bufSize %d)",
+                  _mesa_lookup_enum_by_nr(programInterface), index, bufSize);
+      return;
+   }
+
+   /* Write propCount values until error occurs or bufSize reached. */
+   for (int i = 0; i < propCount && i < bufSize; i++, val++, prop++) {
+      int props_written =
+         _mesa_program_resource_prop(shProg, res, index, *prop, val,
+                                     "glGetProgramResourceiv");
+
+      /* Error happened. */
+      if (props_written == 0)
+         return;
+
+      amount += props_written;
+   }
+
+   /* If <length> is not NULL, the actual number of integer values
+    * written to <params> will be written to <length>.
+    */
+   if (length)
+      *length = amount;
 }
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index 30716f5e3..77e2b8745 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -882,10 +882,9 @@ compile_shader(struct gl_context *ctx, GLuint shaderObj)
       sh->CompileStatus = GL_FALSE;
    } else {
       if (ctx->_Shader->Flags & GLSL_DUMP) {
-         fprintf(stderr, "GLSL source for %s shader %d:\n",
+         _mesa_log("GLSL source for %s shader %d:\n",
                  _mesa_shader_stage_to_string(sh->Stage), sh->Name);
-         fprintf(stderr, "%s\n", sh->Source);
-         fflush(stderr);
+         _mesa_log("%s\n", sh->Source);
       }
 
       /* this call will set the shader->CompileStatus field to indicate if
@@ -899,27 +898,25 @@ compile_shader(struct gl_context *ctx, GLuint shaderObj)
 
       if (ctx->_Shader->Flags & GLSL_DUMP) {
          if (sh->CompileStatus) {
-            fprintf(stderr, "GLSL IR for shader %d:\n", sh->Name);
-            _mesa_print_ir(stderr, sh->ir, NULL);
-            fprintf(stderr, "\n\n");
+            _mesa_log("GLSL IR for shader %d:\n", sh->Name);
+            _mesa_print_ir(_mesa_get_log_file(), sh->ir, NULL);
+            _mesa_log("\n\n");
          } else {
-            fprintf(stderr, "GLSL shader %d failed to compile.\n", sh->Name);
+            _mesa_log("GLSL shader %d failed to compile.\n", sh->Name);
          }
          if (sh->InfoLog && sh->InfoLog[0] != 0) {
-            fprintf(stderr, "GLSL shader %d info log:\n", sh->Name);
-            fprintf(stderr, "%s\n", sh->InfoLog);
+            _mesa_log("GLSL shader %d info log:\n", sh->Name);
+            _mesa_log("%s\n", sh->InfoLog);
          }
-         fflush(stderr);
       }
    }
 
    if (!sh->CompileStatus) {
       if (ctx->_Shader->Flags & GLSL_DUMP_ON_ERROR) {
-         fprintf(stderr, "GLSL source for %s shader %d:\n",
+         _mesa_log("GLSL source for %s shader %d:\n",
                  _mesa_shader_stage_to_string(sh->Stage), sh->Name);
-         fprintf(stderr, "%s\n", sh->Source);
-         fprintf(stderr, "Info Log:\n%s\n", sh->InfoLog);
-         fflush(stderr);
+         _mesa_log("%s\n", sh->Source);
+         _mesa_log("Info Log:\n%s\n", sh->InfoLog);
       }
 
       if (ctx->_Shader->Flags & GLSL_REPORT_ERRORS) {
diff --git a/mesalib/src/mesa/main/shaderapi.h b/mesalib/src/mesa/main/shaderapi.h
index 047d2562d..aba6d5d83 100644
--- a/mesalib/src/mesa/main/shaderapi.h
+++ b/mesalib/src/mesa/main/shaderapi.h
@@ -219,6 +219,51 @@ extern GLuint GLAPIENTRY
 _mesa_CreateShaderProgramv(GLenum type, GLsizei count,
                            const GLchar* const *strings);
 
+/* GL_ARB_program_resource_query */
+extern const char*
+_mesa_program_resource_name(struct gl_program_resource *res);
+
+extern unsigned
+_mesa_program_resource_array_size(struct gl_program_resource *res);
+
+extern GLuint
+_mesa_program_resource_index(struct gl_shader_program *shProg,
+                             struct gl_program_resource *res);
+
+extern struct gl_program_resource *
+_mesa_program_resource_find_name(struct gl_shader_program *shProg,
+                                 GLenum programInterface, const char *name);
+
+extern struct gl_program_resource *
+_mesa_program_resource_find_index(struct gl_shader_program *shProg,
+                                  GLenum programInterface, GLuint index);
+
+extern bool
+_mesa_get_program_resource_name(struct gl_shader_program *shProg,
+                                GLenum programInterface, GLuint index,
+                                GLsizei bufSize, GLsizei *length,
+                                GLchar *name, const char *caller);
+
+extern GLint
+_mesa_program_resource_location(struct gl_shader_program *shProg,
+                                GLenum programInterface, const char *name);
+
+extern GLint
+_mesa_program_resource_location_index(struct gl_shader_program *shProg,
+                                      GLenum programInterface, const char *name);
+
+extern unsigned
+_mesa_program_resource_prop(struct gl_shader_program *shProg,
+                            struct gl_program_resource *res, GLuint index,
+                            const GLenum prop, GLint *val, const char *caller);
+
+extern void
+_mesa_get_program_resourceiv(struct gl_shader_program *shProg,
+                             GLenum programInterface, GLuint index,
+                             GLsizei propCount, const GLenum *props,
+                             GLsizei bufSize, GLsizei *length,
+                             GLint *params);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/mesalib/src/mesa/main/shaderobj.c b/mesalib/src/mesa/main/shaderobj.c
index d7620c8ef..e42896036 100644
--- a/mesalib/src/mesa/main/shaderobj.c
+++ b/mesalib/src/mesa/main/shaderobj.c
@@ -315,6 +315,12 @@ _mesa_clear_shader_program_data(struct gl_shader_program *shProg)
    ralloc_free(shProg->AtomicBuffers);
    shProg->AtomicBuffers = NULL;
    shProg->NumAtomicBuffers = 0;
+
+   if (shProg->ProgramResourceList) {
+      ralloc_free(shProg->ProgramResourceList);
+      shProg->ProgramResourceList = NULL;
+      shProg->NumProgramResourceList = 0;
+   }
 }
 
 
diff --git a/mesalib/src/mesa/main/state.c b/mesalib/src/mesa/main/state.c
index dadfb3c8c..cc84c6148 100644
--- a/mesalib/src/mesa/main/state.c
+++ b/mesalib/src/mesa/main/state.c
@@ -269,28 +269,6 @@ update_program_constants(struct gl_context *ctx)
 
 
 
-static void
-update_viewport_matrix(struct gl_context *ctx)
-{
-   const GLfloat depthMax = ctx->DrawBuffer->_DepthMaxF;
-   unsigned i;
-
-   assert(depthMax > 0);
-
-   /* Compute scale and bias values. This is really driver-specific
-    * and should be maintained elsewhere if at all.
-    * NOTE: RasterPos uses this.
-    */
-   for (i = 0; i < ctx->Const.MaxViewports; i++) {
-      double scale[3], translate[3];
-
-      _mesa_get_viewport_xform(ctx, i, scale, translate);
-      _math_matrix_viewport(&ctx->ViewportArray[i]._WindowMap,
-                            scale, translate, depthMax);
-   }
-}
-
-
 /**
  * Update the ctx->Polygon._FrontBit flag.
  */
@@ -407,9 +385,6 @@ _mesa_update_state_locked( struct gl_context *ctx )
    if (new_state & _NEW_PIXEL)
       _mesa_update_pixel( ctx, new_state );
 
-   if (new_state & (_NEW_BUFFERS | _NEW_VIEWPORT))
-      update_viewport_matrix(ctx);
-
    if (new_state & (_NEW_MULTISAMPLE | _NEW_BUFFERS))
       update_multisample( ctx );
 
@@ -507,7 +482,7 @@ _mesa_set_varying_vp_inputs( struct gl_context *ctx,
           ctx->FragmentProgram._TexEnvProgram) {
          ctx->NewState |= _NEW_VARYING_VP_INPUTS;
       }
-      /*printf("%s %x\n", __FUNCTION__, varying_inputs);*/
+      /*printf("%s %x\n", __func__, varying_inputs);*/
    }
 }
 
diff --git a/mesalib/src/mesa/main/texcompress.c b/mesalib/src/mesa/main/texcompress.c
index 00234d4cc..0fd1a3683 100644
--- a/mesalib/src/mesa/main/texcompress.c
+++ b/mesalib/src/mesa/main/texcompress.c
@@ -32,7 +32,6 @@
 
 #include "glheader.h"
 #include "imports.h"
-#include "colormac.h"
 #include "context.h"
 #include "formats.h"
 #include "mtypes.h"
diff --git a/mesalib/src/mesa/main/texcompress_fxt1.c b/mesalib/src/mesa/main/texcompress_fxt1.c
index 562359410..f06f04882 100644
--- a/mesalib/src/mesa/main/texcompress_fxt1.c
+++ b/mesalib/src/mesa/main/texcompress_fxt1.c
@@ -31,7 +31,6 @@
 
 #include "glheader.h"
 #include "imports.h"
-#include "colormac.h"
 #include "image.h"
 #include "macros.h"
 #include "mipmap.h"
diff --git a/mesalib/src/mesa/main/texcompress_rgtc.c b/mesalib/src/mesa/main/texcompress_rgtc.c
index f40e4e6c0..66de1f17d 100644
--- a/mesalib/src/mesa/main/texcompress_rgtc.c
+++ b/mesalib/src/mesa/main/texcompress_rgtc.c
@@ -35,7 +35,6 @@
 
 #include "glheader.h"
 #include "imports.h"
-#include "colormac.h"
 #include "image.h"
 #include "macros.h"
 #include "mipmap.h"
diff --git a/mesalib/src/mesa/main/texcompress_s3tc.c b/mesalib/src/mesa/main/texcompress_s3tc.c
index 41d00d4e9..7ce3cb88e 100644
--- a/mesalib/src/mesa/main/texcompress_s3tc.c
+++ b/mesalib/src/mesa/main/texcompress_s3tc.c
@@ -29,13 +29,8 @@
  * GL_EXT_texture_compression_s3tc support.
  */
 
-#ifndef USE_EXTERNAL_DXTN_LIB
-#define USE_EXTERNAL_DXTN_LIB 1
-#endif
-
 #include "glheader.h"
 #include "imports.h"
-#include "colormac.h"
 #include "dlopen.h"
 #include "image.h"
 #include "macros.h"
@@ -77,7 +72,6 @@ _mesa_init_texture_s3tc( struct gl_context *ctx )
 {
    /* called during context initialization */
    ctx->Mesa_DXTn = GL_FALSE;
-#if USE_EXTERNAL_DXTN_LIB
    if (!dxtlibhandle) {
       dxtlibhandle = _mesa_dlopen(DXTN_LIBNAME, 0);
       if (!dxtlibhandle) {
@@ -118,9 +112,6 @@ _mesa_init_texture_s3tc( struct gl_context *ctx )
    if (dxtlibhandle) {
       ctx->Mesa_DXTn = GL_TRUE;
    }
-#else
-   (void) ctx;
-#endif
 }
 
 /**
diff --git a/mesalib/src/mesa/main/texparam.c b/mesalib/src/mesa/main/texparam.c
index 1ef728660..b5d42d304 100644
--- a/mesalib/src/mesa/main/texparam.c
+++ b/mesalib/src/mesa/main/texparam.c
@@ -32,7 +32,6 @@
 #include <stdbool.h>
 #include "main/glheader.h"
 #include "main/blend.h"
-#include "main/colormac.h"
 #include "main/context.h"
 #include "main/enums.h"
 #include "main/formats.h"
diff --git a/mesalib/src/mesa/main/texstate.c b/mesalib/src/mesa/main/texstate.c
index 0a7f983fb..1af9d47f0 100644
--- a/mesalib/src/mesa/main/texstate.c
+++ b/mesalib/src/mesa/main/texstate.c
@@ -31,8 +31,6 @@
 #include <stdio.h>
 #include "glheader.h"
 #include "bufferobj.h"
-#include "colormac.h"
-#include "colortab.h"
 #include "context.h"
 #include "enums.h"
 #include "macros.h"
diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c
index 7ad9d2b68..152520598 100644
--- a/mesalib/src/mesa/main/texstore.c
+++ b/mesalib/src/mesa/main/texstore.c
@@ -53,7 +53,6 @@
 
 #include "glheader.h"
 #include "bufferobj.h"
-#include "colormac.h"
 #include "format_pack.h"
 #include "format_utils.h"
 #include "image.h"
diff --git a/mesalib/src/mesa/main/transformfeedback.c b/mesalib/src/mesa/main/transformfeedback.c
index ce678c864..103011ce5 100644
--- a/mesalib/src/mesa/main/transformfeedback.c
+++ b/mesalib/src/mesa/main/transformfeedback.c
@@ -514,22 +514,24 @@ _mesa_EndTransformFeedback(void)
  * Helper used by BindBufferRange() and BindBufferBase().
  */
 static void
-bind_buffer_range(struct gl_context *ctx, GLuint index,
+bind_buffer_range(struct gl_context *ctx,
+                  struct gl_transform_feedback_object *obj,
+                  GLuint index,
                   struct gl_buffer_object *bufObj,
-                  GLintptr offset, GLsizeiptr size)
+                  GLintptr offset, GLsizeiptr size,
+                  bool dsa)
 {
-   struct gl_transform_feedback_object *obj =
-      ctx->TransformFeedback.CurrentObject;
-
    /* Note: no need to FLUSH_VERTICES or flag NewTransformFeedback, because
     * transform feedback buffers can't be changed while transform feedback is
     * active.
     */
 
-   /* The general binding point */
-   _mesa_reference_buffer_object(ctx,
-                                 &ctx->TransformFeedback.CurrentBuffer,
-                                 bufObj);
+   if (!dsa) {
+      /* The general binding point */
+      _mesa_reference_buffer_object(ctx,
+                                    &ctx->TransformFeedback.CurrentBuffer,
+                                    bufObj);
+   }
 
    /* The per-attribute binding point */
    _mesa_set_transform_feedback_binding(ctx, obj, index, bufObj, offset, size);
@@ -539,75 +541,209 @@ bind_buffer_range(struct gl_context *ctx, GLuint index,
 /**
  * Specify a buffer object to receive transform feedback results.  Plus,
  * specify the starting offset to place the results, and max size.
- * Called from the glBindBufferRange() function.
+ * Called from the glBindBufferRange() and glTransformFeedbackBufferRange
+ * functions.
  */
 void
 _mesa_bind_buffer_range_transform_feedback(struct gl_context *ctx,
-					   GLuint index,
-					   struct gl_buffer_object *bufObj,
-					   GLintptr offset,
-					   GLsizeiptr size)
+                                           struct gl_transform_feedback_object *obj,
+                                           GLuint index,
+                                           struct gl_buffer_object *bufObj,
+                                           GLintptr offset,
+                                           GLsizeiptr size,
+                                           bool dsa)
 {
-   struct gl_transform_feedback_object *obj;
+   const char *gl_methd_name;
+   if (dsa)
+      gl_methd_name = "glTransformFeedbackBufferRange";
+   else
+      gl_methd_name = "glBindBufferRange";
 
-   obj = ctx->TransformFeedback.CurrentObject;
 
    if (obj->Active) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glBindBufferRange(transform feedback active)");
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(transform feedback active)",
+                  gl_methd_name);
       return;
    }
 
    if (index >= ctx->Const.MaxTransformFeedbackBuffers) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBindBufferRange(index=%d)", index);
+      /* OpenGL 4.5 core profile, 6.1, pdf page 82: "An INVALID_VALUE error is
+       * generated if index is greater than or equal to the number of binding
+       * points for transform feedback, as described in section 6.7.1."
+       */
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(index=%d out of bounds)",
+                  gl_methd_name, index);
       return;
    }
 
    if (size & 0x3) {
-      /* must a multiple of four */
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBindBufferRange(size=%d)", (int) size);
+      /* OpenGL 4.5 core profile, 6.7, pdf page 103: multiple of 4 */
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(size=%d must be a multiple of "
+                  "four)", gl_methd_name, (int) size);
       return;
    }  
 
    if (offset & 0x3) {
-      /* must be multiple of four */
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glBindBufferRange(offset=%d)", (int) offset);
+      /* OpenGL 4.5 core profile, 6.7, pdf page 103: multiple of 4 */
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(offset=%d must be a multiple "
+                  "of four)", gl_methd_name, (int) offset);
       return;
-   }  
+   }
+
+   if (offset < 0) {
+      /* OpenGL 4.5 core profile, 6.1, pdf page 82: "An INVALID_VALUE error is
+       * generated by BindBufferRange if offset is negative."
+       *
+       * OpenGL 4.5 core profile, 13.2, pdf page 445: "An INVALID_VALUE error
+       * is generated by TransformFeedbackBufferRange if offset is negative."
+       */
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(offset=%d must be >= 0)",
+                  gl_methd_name,
+                  (int) offset);
+      return;
+   }
 
-   bind_buffer_range(ctx, index, bufObj, offset, size);
+   if (size <= 0 && (dsa || bufObj != ctx->Shared->NullBufferObj)) {
+      /* OpenGL 4.5 core profile, 6.1, pdf page 82: "An INVALID_VALUE error is
+       * generated by BindBufferRange if buffer is non-zero and size is less
+       * than or equal to zero."
+       *
+       * OpenGL 4.5 core profile, 13.2, pdf page 445: "An INVALID_VALUE error
+       * is generated by TransformFeedbackBufferRange if size is less than or
+       * equal to zero."
+       */
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(size=%d must be > 0)",
+                  gl_methd_name, (int) size);
+      return;
+   }
+
+   bind_buffer_range(ctx, obj, index, bufObj, offset, size, dsa);
 }
 
 
 /**
  * Specify a buffer object to receive transform feedback results.
  * As above, but start at offset = 0.
- * Called from the glBindBufferBase() function.
+ * Called from the glBindBufferBase() and glTransformFeedbackBufferBase()
+ * functions.
  */
 void
 _mesa_bind_buffer_base_transform_feedback(struct gl_context *ctx,
-					  GLuint index,
-					  struct gl_buffer_object *bufObj)
+                                          struct gl_transform_feedback_object *obj,
+                                          GLuint index,
+                                          struct gl_buffer_object *bufObj,
+                                          bool dsa)
 {
-   struct gl_transform_feedback_object *obj;
-
-   obj = ctx->TransformFeedback.CurrentObject;
-
    if (obj->Active) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glBindBufferBase(transform feedback active)");
+                  "%s(transform feedback active)",
+                  dsa ? "glTransformFeedbackBufferBase" : "glBindBufferBase");
       return;
    }
 
    if (index >= ctx->Const.MaxTransformFeedbackBuffers) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glBindBufferBase(index=%d)", index);
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(index=%d out of bounds)",
+                  dsa ? "glTransformFeedbackBufferBase" : "glBindBufferBase",
+                  index);
+      return;
+   }
+
+   bind_buffer_range(ctx, obj, index, bufObj, 0, 0, dsa);
+}
+
+/**
+ * Wrapper around lookup_transform_feedback_object that throws
+ * GL_INVALID_OPERATION if id is not in the hash table. After calling
+ * _mesa_error, it returns NULL.
+ */
+static struct gl_transform_feedback_object *
+lookup_transform_feedback_object_err(struct gl_context *ctx,
+                                     GLuint xfb, const char* func)
+{
+   struct gl_transform_feedback_object *obj;
+
+   obj = _mesa_lookup_transform_feedback_object(ctx, xfb);
+   if (!obj) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "%s(xfb=%u: non-generated object name)", func, xfb);
+   }
+
+   return obj;
+}
+
+/**
+ * Wrapper around _mesa_lookup_bufferobj that throws GL_INVALID_VALUE if id
+ * is not in the hash table. Specialised version for the
+ * transform-feedback-related functions. After calling _mesa_error, it
+ * returns NULL.
+ */
+static struct gl_buffer_object *
+lookup_transform_feedback_bufferobj_err(struct gl_context *ctx,
+                                        GLuint buffer, const char* func)
+{
+   struct gl_buffer_object *bufObj;
+
+   /* OpenGL 4.5 core profile, 13.2, pdf page 444: buffer must be zero or the
+    * name of an existing buffer object.
+    */
+   if (buffer == 0) {
+      bufObj = ctx->Shared->NullBufferObj;
+   } else {
+      bufObj = _mesa_lookup_bufferobj(ctx, buffer);
+      if (!bufObj) {
+         _mesa_error(ctx, GL_INVALID_VALUE, "%s(invalid buffer=%u)", func,
+                     buffer);
+      }
+   }
+
+   return bufObj;
+}
+
+void GLAPIENTRY
+_mesa_TransformFeedbackBufferBase(GLuint xfb, GLuint index, GLuint buffer)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_transform_feedback_object *obj;
+   struct gl_buffer_object *bufObj;
+
+   obj = lookup_transform_feedback_object_err(ctx, xfb,
+                                              "glTransformFeedbackBufferBase");
+   if(!obj) {
+      return;
+   }
+
+   bufObj = lookup_transform_feedback_bufferobj_err(ctx, buffer,
+                                              "glTransformFeedbackBufferBase");
+   if(!bufObj) {
       return;
    }
 
-   bind_buffer_range(ctx, index, bufObj, 0, 0);
+   _mesa_bind_buffer_base_transform_feedback(ctx, obj, index, bufObj, true);
 }
 
+void GLAPIENTRY
+_mesa_TransformFeedbackBufferRange(GLuint xfb, GLuint index, GLuint buffer,
+                                   GLintptr offset, GLsizeiptr size)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_transform_feedback_object *obj;
+   struct gl_buffer_object *bufObj;
+
+   obj = lookup_transform_feedback_object_err(ctx, xfb,
+                                              "glTransformFeedbackBufferRange");
+   if(!obj) {
+      return;
+   }
+
+   bufObj = lookup_transform_feedback_bufferobj_err(ctx, buffer,
+                                              "glTransformFeedbackBufferRange");
+   if(!bufObj) {
+      return;
+   }
+
+   _mesa_bind_buffer_range_transform_feedback(ctx, obj, index, bufObj, offset,
+                                              size, true);
+}
 
 /**
  * Specify a buffer object to receive transform feedback results, plus the
@@ -660,7 +796,7 @@ _mesa_BindBufferOffsetEXT(GLenum target, GLuint index, GLuint buffer,
       return;
    }
 
-   bind_buffer_range(ctx, index, bufObj, offset, 0);
+   bind_buffer_range(ctx, obj, index, bufObj, offset, 0, false);
 }
 
 
@@ -784,7 +920,7 @@ _mesa_GetTransformFeedbackVarying(GLuint program, GLuint index,
                                   GLsizei *size, GLenum *type, GLchar *name)
 {
    const struct gl_shader_program *shProg;
-   const struct gl_transform_feedback_info *linked_xfb_info;
+   struct gl_program_resource *res;
    GET_CURRENT_CONTEXT(ctx);
 
    shProg = _mesa_lookup_shader_program(ctx, program);
@@ -794,22 +930,27 @@ _mesa_GetTransformFeedbackVarying(GLuint program, GLuint index,
       return;
    }
 
-   linked_xfb_info = &shProg->LinkedTransformFeedback;
-   if (index >= (GLuint) linked_xfb_info->NumVarying) {
+   res = _mesa_program_resource_find_index((struct gl_shader_program *) shProg,
+                                           GL_TRANSFORM_FEEDBACK_VARYING,
+                                           index);
+   if (!res) {
       _mesa_error(ctx, GL_INVALID_VALUE,
                   "glGetTransformFeedbackVarying(index=%u)", index);
       return;
    }
 
    /* return the varying's name and length */
-   _mesa_copy_string(name, bufSize, length,
-		     linked_xfb_info->Varyings[index].Name);
+   _mesa_copy_string(name, bufSize, length, _mesa_program_resource_name(res));
 
    /* return the datatype and value's size (in datatype units) */
    if (type)
-      *type = linked_xfb_info->Varyings[index].Type;
+      _mesa_program_resource_prop((struct gl_shader_program *) shProg,
+                                  res, index, GL_TYPE, (GLint*) type,
+                                  "glGetTransformFeedbackVarying");
    if (size)
-      *size = linked_xfb_info->Varyings[index].Size;
+      _mesa_program_resource_prop((struct gl_shader_program *) shProg,
+                                  res, index, GL_ARRAY_SIZE, (GLint*) size,
+                                  "glGetTransformFeedbackVarying");
 }
 
 
@@ -817,6 +958,10 @@ _mesa_GetTransformFeedbackVarying(GLuint program, GLuint index,
 struct gl_transform_feedback_object *
 _mesa_lookup_transform_feedback_object(struct gl_context *ctx, GLuint name)
 {
+   /* OpenGL 4.5 core profile, 13.2 pdf page 444: "xfb must be zero, indicating
+    * the default transform feedback object, or the name of an existing
+    * transform feedback object."
+    */
    if (name == 0) {
       return ctx->TransformFeedback.DefaultObject;
    }
@@ -825,25 +970,24 @@ _mesa_lookup_transform_feedback_object(struct gl_context *ctx, GLuint name)
          _mesa_HashLookup(ctx->TransformFeedback.Objects, name);
 }
 
-
-/**
- * Create new transform feedback objects.   Transform feedback objects
- * encapsulate the state related to transform feedback to allow quickly
- * switching state (and drawing the results, below).
- * Part of GL_ARB_transform_feedback2.
- */
-void GLAPIENTRY
-_mesa_GenTransformFeedbacks(GLsizei n, GLuint *names)
+static void
+create_transform_feedbacks(struct gl_context *ctx, GLsizei n, GLuint *ids,
+                           bool dsa)
 {
    GLuint first;
-   GET_CURRENT_CONTEXT(ctx);
+   const char* func;
+
+   if (dsa)
+      func = "glCreateTransformFeedbacks";
+   else
+      func = "glGenTransformFeedbacks";
 
    if (n < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGenTransformFeedbacks(n < 0)");
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(n < 0)", func);
       return;
    }
 
-   if (!names)
+   if (!ids)
       return;
 
    /* we don't need contiguous IDs, but this might be faster */
@@ -854,18 +998,56 @@ _mesa_GenTransformFeedbacks(GLsizei n, GLuint *names)
          struct gl_transform_feedback_object *obj
             = ctx->Driver.NewTransformFeedback(ctx, first + i);
          if (!obj) {
-            _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenTransformFeedbacks");
+            _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
             return;
          }
-         names[i] = first + i;
+         ids[i] = first + i;
          _mesa_HashInsert(ctx->TransformFeedback.Objects, first + i, obj);
+         if (dsa) {
+            /* this is normally done at bind time in the non-dsa case */
+            obj->EverBound = GL_TRUE;
+         }
       }
    }
    else {
-      _mesa_error(ctx, GL_OUT_OF_MEMORY, "glGenTransformFeedbacks");
+      _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", func);
    }
 }
 
+/**
+ * Create new transform feedback objects.   Transform feedback objects
+ * encapsulate the state related to transform feedback to allow quickly
+ * switching state (and drawing the results, below).
+ * Part of GL_ARB_transform_feedback2.
+ */
+void GLAPIENTRY
+_mesa_GenTransformFeedbacks(GLsizei n, GLuint *names)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   /* GenTransformFeedbacks should just reserve the object names that a
+    * subsequent call to BindTransformFeedback should actively create. For
+    * the sake of simplicity, we reserve the names and create the objects
+    * straight away.
+    */
+
+   create_transform_feedbacks(ctx, n, names, false);
+}
+
+/**
+ * Create new transform feedback objects.   Transform feedback objects
+ * encapsulate the state related to transform feedback to allow quickly
+ * switching state (and drawing the results, below).
+ * Part of GL_ARB_direct_state_access.
+ */
+void GLAPIENTRY
+_mesa_CreateTransformFeedbacks(GLsizei n, GLuint *names)
+{
+   GET_CURRENT_CONTEXT(ctx);
+
+   create_transform_feedbacks(ctx, n, names, true);
+}
+
 
 /**
  * Is the given ID a transform feedback object?
@@ -1026,3 +1208,89 @@ _mesa_ResumeTransformFeedback(void)
    assert(ctx->Driver.ResumeTransformFeedback);
    ctx->Driver.ResumeTransformFeedback(ctx, obj);
 }
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbackiv(GLuint xfb, GLenum pname, GLint *param)
+{
+    struct gl_transform_feedback_object *obj;
+    GET_CURRENT_CONTEXT(ctx);
+
+    obj = lookup_transform_feedback_object_err(ctx, xfb,
+                                               "glGetTransformFeedbackiv");
+    if(!obj) {
+       return;
+    }
+
+    switch(pname) {
+    case GL_TRANSFORM_FEEDBACK_PAUSED:
+       *param = obj->Paused;
+       break;
+    case GL_TRANSFORM_FEEDBACK_ACTIVE:
+       *param = obj->Active;
+       break;
+    default:
+       _mesa_error(ctx, GL_INVALID_ENUM,
+                   "glGetTransformFeedbackiv(pname=%i)", pname);
+    }
+}
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbacki_v(GLuint xfb, GLenum pname, GLuint index,
+                              GLint *param)
+{
+   struct gl_transform_feedback_object *obj;
+   GET_CURRENT_CONTEXT(ctx);
+
+   obj = lookup_transform_feedback_object_err(ctx, xfb,
+                                              "glGetTransformFeedbacki_v");
+   if(!obj) {
+      return;
+   }
+
+   if (index >= ctx->Const.MaxTransformFeedbackBuffers) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "glGetTransformFeedbacki_v(index=%i)", index);
+      return;
+   }
+
+   switch(pname) {
+   case GL_TRANSFORM_FEEDBACK_BUFFER_BINDING:
+      *param = obj->BufferNames[index];
+      break;
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glGetTransformFeedbacki_v(pname=%i)", pname);
+   }
+}
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbacki64_v(GLuint xfb, GLenum pname, GLuint index,
+                                GLint64 *param)
+{
+   struct gl_transform_feedback_object *obj;
+   GET_CURRENT_CONTEXT(ctx);
+
+   obj = lookup_transform_feedback_object_err(ctx, xfb,
+                                              "glGetTransformFeedbacki64_v");
+   if(!obj) {
+      return;
+   }
+
+   if (index >= ctx->Const.MaxTransformFeedbackBuffers) {
+      _mesa_error(ctx, GL_INVALID_VALUE,
+                  "glGetTransformFeedbacki64_v(index=%i)", index);
+      return;
+   }
+
+   switch(pname) {
+   case GL_TRANSFORM_FEEDBACK_BUFFER_START:
+      *param = obj->Offset[index];
+      break;
+   case GL_TRANSFORM_FEEDBACK_BUFFER_SIZE:
+      *param = obj->RequestedSize[index];
+      break;
+   default:
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glGetTransformFeedbacki64_v(pname=%i)", pname);
+   }
+}
diff --git a/mesalib/src/mesa/main/transformfeedback.h b/mesalib/src/mesa/main/transformfeedback.h
index 87f4080e0..bb9729cdb 100644
--- a/mesalib/src/mesa/main/transformfeedback.h
+++ b/mesalib/src/mesa/main/transformfeedback.h
@@ -65,15 +65,18 @@ _mesa_EndTransformFeedback(void);
 
 extern void
 _mesa_bind_buffer_range_transform_feedback(struct gl_context *ctx,
+					   struct gl_transform_feedback_object *obj,
 					   GLuint index,
 					   struct gl_buffer_object *bufObj,
 					   GLintptr offset,
-					   GLsizeiptr size);
+					   GLsizeiptr size, bool dsa);
 
 extern void
 _mesa_bind_buffer_base_transform_feedback(struct gl_context *ctx,
+					  struct gl_transform_feedback_object *obj,
 					  GLuint index,
-					  struct gl_buffer_object *bufObj);
+					  struct gl_buffer_object *bufObj,
+					  bool dsa);
 
 extern void GLAPIENTRY
 _mesa_BindBufferOffsetEXT(GLenum target, GLuint index, GLuint buffer,
@@ -102,6 +105,9 @@ _mesa_lookup_transform_feedback_object(struct gl_context *ctx, GLuint name);
 extern void GLAPIENTRY
 _mesa_GenTransformFeedbacks(GLsizei n, GLuint *names);
 
+extern void GLAPIENTRY
+_mesa_CreateTransformFeedbacks(GLsizei n, GLuint *names);
+
 extern GLboolean GLAPIENTRY
 _mesa_IsTransformFeedback(GLuint name);
 
@@ -141,4 +147,24 @@ _mesa_set_transform_feedback_binding(struct gl_context *ctx,
    tfObj->RequestedSize[index] = size;
 }
 
+/*** GL_ARB_direct_state_access ***/
+
+extern void GLAPIENTRY
+_mesa_TransformFeedbackBufferBase(GLuint xfb, GLuint index, GLuint buffer);
+
+extern void GLAPIENTRY
+_mesa_TransformFeedbackBufferRange(GLuint xfb, GLuint index, GLuint buffer,
+                                   GLintptr offset, GLsizeiptr size);
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbackiv(GLuint xfb, GLenum pname, GLint *param);
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbacki_v(GLuint xfb, GLenum pname, GLuint index,
+                              GLint *param);
+
+extern void GLAPIENTRY
+_mesa_GetTransformFeedbacki64_v(GLuint xfb, GLenum pname, GLuint index,
+                                GLint64 *param);
+
 #endif /* TRANSFORM_FEEDBACK_H */
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index 2ab5528c3..4e77b3284 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -46,6 +46,7 @@ _mesa_GetActiveUniform(GLuint program, GLuint index,
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_shader_program *shProg;
+   struct gl_program_resource *res;
 
    if (maxLength < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveUniform(maxLength < 0)");
@@ -56,26 +57,51 @@ _mesa_GetActiveUniform(GLuint program, GLuint index,
    if (!shProg)
       return;
 
-   if (index >= shProg->NumUserUniformStorage) {
+   res = _mesa_program_resource_find_index((struct gl_shader_program *) shProg,
+                                           GL_UNIFORM, index);
+
+   if (!res) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveUniform(index)");
       return;
    }
 
-   const struct gl_uniform_storage *const uni = &shProg->UniformStorage[index];
-
-   if (nameOut) {
-      _mesa_get_uniform_name(uni, maxLength, length, nameOut);
-   }
-
-   if (size) {
-      /* array_elements is zero for non-arrays, but the API requires that 1 be
-       * returned.
-       */
-      *size = MAX2(1, uni->array_elements);
-   }
+   if (nameOut)
+      _mesa_get_program_resource_name(shProg, GL_UNIFORM, index, maxLength,
+                                      length, nameOut, "glGetActiveUniform");
+   if (type)
+      _mesa_program_resource_prop((struct gl_shader_program *) shProg,
+                                  res, index, GL_TYPE, (GLint*) type,
+                                  "glGetActiveUniform");
+   if (size)
+      _mesa_program_resource_prop((struct gl_shader_program *) shProg,
+                                  res, index, GL_ARRAY_SIZE, (GLint*) size,
+                                  "glGetActiveUniform");
+}
 
-   if (type) {
-      *type = uni->type->gl_type;
+static GLenum
+resource_prop_from_uniform_prop(GLenum uni_prop)
+{
+   switch (uni_prop) {
+   case GL_UNIFORM_TYPE:
+      return GL_TYPE;
+   case GL_UNIFORM_SIZE:
+      return GL_ARRAY_SIZE;
+   case GL_UNIFORM_NAME_LENGTH:
+      return GL_NAME_LENGTH;
+   case GL_UNIFORM_BLOCK_INDEX:
+      return GL_BLOCK_INDEX;
+   case GL_UNIFORM_OFFSET:
+      return GL_OFFSET;
+   case GL_UNIFORM_ARRAY_STRIDE:
+      return GL_ARRAY_STRIDE;
+   case GL_UNIFORM_MATRIX_STRIDE:
+      return GL_MATRIX_STRIDE;
+   case GL_UNIFORM_IS_ROW_MAJOR:
+      return GL_IS_ROW_MAJOR;
+   case GL_UNIFORM_ATOMIC_COUNTER_BUFFER_INDEX:
+      return GL_ATOMIC_COUNTER_BUFFER_INDEX;
+   default:
+      return 0;
    }
 }
 
@@ -88,7 +114,8 @@ _mesa_GetActiveUniformsiv(GLuint program,
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_shader_program *shProg;
-   GLsizei i;
+   struct gl_program_resource *res;
+   GLenum res_prop;
 
    if (uniformCount < 0) {
       _mesa_error(ctx, GL_INVALID_VALUE,
@@ -100,80 +127,21 @@ _mesa_GetActiveUniformsiv(GLuint program,
    if (!shProg)
       return;
 
-   for (i = 0; i < uniformCount; i++) {
-      GLuint index = uniformIndices[i];
+   res_prop = resource_prop_from_uniform_prop(pname);
 
-      if (index >= shProg->NumUserUniformStorage) {
-	 _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveUniformsiv(index)");
-	 return;
+   for (int i = 0; i < uniformCount; i++) {
+      res = _mesa_program_resource_find_index(shProg, GL_UNIFORM,
+                                              uniformIndices[i]);
+      if (!res) {
+         _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveUniformsiv(index)");
+         break;
       }
-   }
-
-   for (i = 0; i < uniformCount; i++) {
-      GLuint index = uniformIndices[i];
-      const struct gl_uniform_storage *uni = &shProg->UniformStorage[index];
-
-      switch (pname) {
-      case GL_UNIFORM_TYPE:
-	 params[i] = uni->type->gl_type;
-	 break;
-
-      case GL_UNIFORM_SIZE:
-	 /* array_elements is zero for non-arrays, but the API requires that 1 be
-	  * returned.
-	  */
-	 params[i] = MAX2(1, uni->array_elements);
-	 break;
 
-      case GL_UNIFORM_NAME_LENGTH:
-	 params[i] = strlen(uni->name) + 1;
-
-         /* Page 61 (page 73 of the PDF) in section 2.11 of the OpenGL ES 3.0
-          * spec says:
-          *
-          *     "If the active uniform is an array, the uniform name returned
-          *     in name will always be the name of the uniform array appended
-          *     with "[0]"."
-          */
-         if (uni->array_elements != 0)
-            params[i] += 3;
-	 break;
-
-      case GL_UNIFORM_BLOCK_INDEX:
-	 params[i] = uni->block_index;
-	 break;
-
-      case GL_UNIFORM_OFFSET:
-	 params[i] = uni->offset;
-	 break;
-
-      case GL_UNIFORM_ARRAY_STRIDE:
-	 params[i] = uni->array_stride;
-	 break;
-
-      case GL_UNIFORM_MATRIX_STRIDE:
-	 params[i] = uni->matrix_stride;
-	 break;
-
-      case GL_UNIFORM_IS_ROW_MAJOR:
-	 params[i] = uni->row_major;
-	 break;
-
-      case GL_UNIFORM_ATOMIC_COUNTER_BUFFER_INDEX:
-         if (!ctx->Extensions.ARB_shader_atomic_counters)
-            goto invalid_enum;
-         params[i] = uni->atomic_buffer_index;
+      if (!_mesa_program_resource_prop(shProg, res, uniformIndices[i],
+                                       res_prop, &params[i],
+                                       "glGetActiveUniformsiv"))
          break;
-
-      default:
-         goto invalid_enum;
-      }
    }
-
-   return;
-
- invalid_enum:
-   _mesa_error(ctx, GL_INVALID_ENUM, "glGetActiveUniformsiv(pname)");
 }
 
 static struct gl_uniform_storage *
diff --git a/mesalib/src/mesa/main/uniforms.c b/mesalib/src/mesa/main/uniforms.c
index fb1482f9c..5548d1d02 100644
--- a/mesalib/src/mesa/main/uniforms.c
+++ b/mesalib/src/mesa/main/uniforms.c
@@ -911,7 +911,6 @@ GLint GLAPIENTRY
 _mesa_GetUniformLocation(GLuint programObj, const GLcharARB *name)
 {
    struct gl_shader_program *shProg;
-   GLuint index, offset;
 
    GET_CURRENT_CONTEXT(ctx);
 
@@ -931,23 +930,7 @@ _mesa_GetUniformLocation(GLuint programObj, const GLcharARB *name)
       return -1;
    }
 
-   index = _mesa_get_uniform_location(shProg, name, &offset);
-   if (index == GL_INVALID_INDEX)
-      return -1;
-
-   /* From the GL_ARB_uniform_buffer_object spec:
-    *
-    *     "The value -1 will be returned if <name> does not correspond to an
-    *      active uniform variable name in <program>, if <name> is associated
-    *      with a named uniform block, or if <name> starts with the reserved
-    *      prefix "gl_"."
-    */
-   if (shProg->UniformStorage[index].block_index != -1 ||
-       shProg->UniformStorage[index].atomic_buffer_index != -1)
-      return -1;
-
-   /* location in remap table + array element offset */
-   return shProg->UniformStorage[index].remap_location + offset;
+   return _mesa_program_resource_location(shProg, GL_UNIFORM, name);
 }
 
 GLuint GLAPIENTRY
@@ -955,7 +938,6 @@ _mesa_GetUniformBlockIndex(GLuint program,
 			   const GLchar *uniformBlockName)
 {
    GET_CURRENT_CONTEXT(ctx);
-   GLuint i;
    struct gl_shader_program *shProg;
 
    if (!ctx->Extensions.ARB_uniform_buffer_object) {
@@ -968,12 +950,13 @@ _mesa_GetUniformBlockIndex(GLuint program,
    if (!shProg)
       return GL_INVALID_INDEX;
 
-   for (i = 0; i < shProg->NumUniformBlocks; i++) {
-      if (!strcmp(shProg->UniformBlocks[i].Name, uniformBlockName))
-	 return i;
-   }
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_name(shProg, GL_UNIFORM_BLOCK,
+                                       uniformBlockName);
+   if (!res)
+      return GL_INVALID_INDEX;
 
-   return GL_INVALID_INDEX;
+   return _mesa_program_resource_index(shProg, res);
 }
 
 void GLAPIENTRY
@@ -1003,9 +986,9 @@ _mesa_GetUniformIndices(GLuint program,
    }
 
    for (i = 0; i < uniformCount; i++) {
-      unsigned offset;
-      uniformIndices[i] = _mesa_get_uniform_location(shProg,
-						     uniformNames[i], &offset);
+      struct gl_program_resource *res =
+         _mesa_program_resource_find_name(shProg, GL_UNIFORM, uniformNames[i]);
+      uniformIndices[i] = _mesa_program_resource_index(shProg, res);
    }
 }
 
@@ -1061,104 +1044,104 @@ _mesa_UniformBlockBinding(GLuint program,
    }
 }
 
-void GLAPIENTRY
-_mesa_GetActiveUniformBlockiv(GLuint program,
-			      GLuint uniformBlockIndex,
-			      GLenum pname,
-			      GLint *params)
+
+/**
+ * Generic program resource property query.
+ */
+static void
+mesa_bufferiv(struct gl_shader_program *shProg, GLenum type,
+              GLuint index, GLenum pname, GLint *params, const char *caller)
 {
    GET_CURRENT_CONTEXT(ctx);
-   struct gl_shader_program *shProg;
-   struct gl_uniform_block *block;
-   unsigned i;
+   struct gl_program_resource *res =
+      _mesa_program_resource_find_index(shProg, type, index);
 
-   if (!ctx->Extensions.ARB_uniform_buffer_object) {
-      _mesa_error(ctx, GL_INVALID_OPERATION, "glGetActiveUniformBlockiv");
+   if (!res) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(bufferindex %d)", caller, index);
       return;
    }
 
-   shProg = _mesa_lookup_shader_program_err(ctx, program,
-					    "glGetActiveUniformBlockiv");
-   if (!shProg)
-      return;
-
-   if (uniformBlockIndex >= shProg->NumUniformBlocks) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-		  "glGetActiveUniformBlockiv(block index %u >= %u)",
-		  uniformBlockIndex, shProg->NumUniformBlocks);
-      return;
-   }
-
-   block = &shProg->UniformBlocks[uniformBlockIndex];
-
    switch (pname) {
    case GL_UNIFORM_BLOCK_BINDING:
-      params[0] = block->Binding;
+   case GL_ATOMIC_COUNTER_BUFFER_BINDING:
+      _mesa_program_resource_prop(shProg, res, index, GL_BUFFER_BINDING,
+                                  params, caller);
       return;
-
    case GL_UNIFORM_BLOCK_DATA_SIZE:
-      params[0] = block->UniformBufferSize;
+   case GL_ATOMIC_COUNTER_BUFFER_DATA_SIZE:
+      _mesa_program_resource_prop(shProg, res, index, GL_BUFFER_DATA_SIZE,
+                                  params, caller);
       return;
-
    case GL_UNIFORM_BLOCK_NAME_LENGTH:
-      params[0] = strlen(block->Name) + 1;
+      _mesa_program_resource_prop(shProg, res, index, GL_NAME_LENGTH,
+                                  params, caller);
       return;
-
-   case GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS: {
-      unsigned count = 0;
-
-      for (i = 0; i < block->NumUniforms; i++) {
-	 unsigned offset;
-         const int idx =
-            _mesa_get_uniform_location(shProg,
-                                       block->Uniforms[i].IndexName,
-                                       &offset);
-         if (idx != -1)
-            count++;
-      }
-
-      params[0] = count;
+   case GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS:
+   case GL_ATOMIC_COUNTER_BUFFER_ACTIVE_ATOMIC_COUNTERS:
+      _mesa_program_resource_prop(shProg, res, index, GL_NUM_ACTIVE_VARIABLES,
+                                  params, caller);
       return;
-   }
-
-   case GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES: {
-      unsigned count = 0;
-
-      for (i = 0; i < block->NumUniforms; i++) {
-	 unsigned offset;
-         const int idx =
-            _mesa_get_uniform_location(shProg,
-                                       block->Uniforms[i].IndexName,
-                                       &offset);
-
-         if (idx != -1)
-            params[count++] = idx;
-      }
+   case GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES:
+   case GL_ATOMIC_COUNTER_BUFFER_ACTIVE_ATOMIC_COUNTER_INDICES:
+      _mesa_program_resource_prop(shProg, res, index, GL_ACTIVE_VARIABLES,
+                                  params, caller);
       return;
-   }
-
    case GL_UNIFORM_BLOCK_REFERENCED_BY_VERTEX_SHADER:
-      params[0] = shProg->UniformBlockStageIndex[MESA_SHADER_VERTEX][uniformBlockIndex] != -1;
+   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_VERTEX_SHADER:
+      _mesa_program_resource_prop(shProg, res, index,
+                                  GL_REFERENCED_BY_VERTEX_SHADER, params,
+                                  caller);
       return;
-
    case GL_UNIFORM_BLOCK_REFERENCED_BY_GEOMETRY_SHADER:
-      if (!_mesa_has_geometry_shaders(ctx))
-         break;
-      params[0] = shProg->UniformBlockStageIndex[MESA_SHADER_GEOMETRY][uniformBlockIndex] != -1;
+   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_GEOMETRY_SHADER:
+      _mesa_program_resource_prop(shProg, res, index,
+                                  GL_REFERENCED_BY_GEOMETRY_SHADER, params,
+                                  caller);
       return;
-
    case GL_UNIFORM_BLOCK_REFERENCED_BY_FRAGMENT_SHADER:
-      params[0] = shProg->UniformBlockStageIndex[MESA_SHADER_FRAGMENT][uniformBlockIndex] != -1;
+   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_FRAGMENT_SHADER:
+      _mesa_program_resource_prop(shProg, res, index,
+                                  GL_REFERENCED_BY_FRAGMENT_SHADER, params,
+                                  caller);
+      return;
+   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_TESS_CONTROL_SHADER:
+      params[0] = GL_FALSE;
+      return;
+   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_TESS_EVALUATION_SHADER:
+      params[0] = GL_FALSE;
       return;
-
    default:
       _mesa_error(ctx, GL_INVALID_ENUM,
-		  "glGetActiveUniformBlockiv(pname 0x%x (%s))",
-		  pname, _mesa_lookup_enum_by_nr(pname));
+                  "%s(pname 0x%x (%s))", caller, pname,
+                  _mesa_lookup_enum_by_nr(pname));
       return;
    }
 }
 
+
+void GLAPIENTRY
+_mesa_GetActiveUniformBlockiv(GLuint program,
+			      GLuint uniformBlockIndex,
+			      GLenum pname,
+			      GLint *params)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   struct gl_shader_program *shProg;
+
+   if (!ctx->Extensions.ARB_uniform_buffer_object) {
+      _mesa_error(ctx, GL_INVALID_OPERATION, "glGetActiveUniformBlockiv");
+      return;
+   }
+
+   shProg = _mesa_lookup_shader_program_err(ctx, program,
+					    "glGetActiveUniformBlockiv");
+   if (!shProg)
+      return;
+
+   mesa_bufferiv(shProg, GL_UNIFORM_BLOCK, uniformBlockIndex, pname, params,
+                 "glGetActiveUniformBlockiv");
+}
+
 void GLAPIENTRY
 _mesa_GetActiveUniformBlockName(GLuint program,
 				GLuint uniformBlockIndex,
@@ -1168,7 +1151,6 @@ _mesa_GetActiveUniformBlockName(GLuint program,
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_shader_program *shProg;
-   struct gl_uniform_block *block;
 
    if (!ctx->Extensions.ARB_uniform_buffer_object) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "glGetActiveUniformBlockiv");
@@ -1187,18 +1169,11 @@ _mesa_GetActiveUniformBlockName(GLuint program,
    if (!shProg)
       return;
 
-   if (uniformBlockIndex >= shProg->NumUniformBlocks) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-		  "glGetActiveUniformBlockiv(block index %u >= %u)",
-		  uniformBlockIndex, shProg->NumUniformBlocks);
-      return;
-   }
-
-   block = &shProg->UniformBlocks[uniformBlockIndex];
-
-   if (uniformBlockName) {
-      _mesa_copy_string(uniformBlockName, bufSize, length, block->Name);
-   }
+   if (uniformBlockName)
+      _mesa_get_program_resource_name(shProg, GL_UNIFORM_BLOCK,
+                                      uniformBlockIndex, bufSize, length,
+                                      uniformBlockName,
+                                      "glGetActiveUniformBlockName");
 }
 
 void GLAPIENTRY
@@ -1226,54 +1201,8 @@ _mesa_GetActiveUniformName(GLuint program, GLuint uniformIndex,
    if (!shProg)
       return;
 
-   if (uniformIndex >= shProg->NumUserUniformStorage) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGetActiveUniform(index)");
-      return;
-   }
-
-   if (uniformName) {
-      _mesa_get_uniform_name(& shProg->UniformStorage[uniformIndex],
-                             bufSize, length, uniformName);
-   }
-}
-
-void
-_mesa_get_uniform_name(const struct gl_uniform_storage *uni,
-                       GLsizei maxLength, GLsizei *length,
-                       GLchar *nameOut)
-{
-   GLsizei localLength;
-
-   if (length == NULL)
-      length = &localLength;
-
-   _mesa_copy_string(nameOut, maxLength, length, uni->name);
-
-   /* Page 61 (page 73 of the PDF) in section 2.11 of the OpenGL ES 3.0
-    * spec says:
-    *
-    *     "If the active uniform is an array, the uniform name returned in
-    *     name will always be the name of the uniform array appended with
-    *     "[0]"."
-    *
-    * The same text also appears in the OpenGL 4.2 spec.  It does not,
-    * however, appear in any previous spec.  Previous specifications are
-    * ambiguous in this regard.  However, either name can later be passed
-    * to glGetUniformLocation (and related APIs), so there shouldn't be any
-    * harm in always appending "[0]" to uniform array names.
-    */
-   if (uni->array_elements != 0) {
-      int i;
-
-      /* The comparison is strange because *length does *NOT* include the
-       * terminating NUL, but maxLength does.
-       */
-      for (i = 0; i < 3 && (*length + i + 1) < maxLength; i++)
-         nameOut[*length + i] = "[0]"[i];
-
-      nameOut[*length + i] = '\0';
-      *length += i;
-   }
+   _mesa_get_program_resource_name(shProg, GL_UNIFORM, uniformIndex, bufSize,
+                                   length, uniformName, "glGetActiveUniformName");
 }
 
 void GLAPIENTRY
@@ -1282,8 +1211,6 @@ _mesa_GetActiveAtomicCounterBufferiv(GLuint program, GLuint bufferIndex,
 {
    GET_CURRENT_CONTEXT(ctx);
    struct gl_shader_program *shProg;
-   struct gl_active_atomic_buffer *ab;
-   GLuint i;
 
    if (!ctx->Extensions.ARB_shader_atomic_counters) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
@@ -1296,49 +1223,8 @@ _mesa_GetActiveAtomicCounterBufferiv(GLuint program, GLuint bufferIndex,
    if (!shProg)
       return;
 
-   if (bufferIndex >= shProg->NumAtomicBuffers) {
-      _mesa_error(ctx, GL_INVALID_VALUE,
-                  "glGetActiveAtomicCounterBufferiv(bufferIndex)");
-      return;
-   }
-
-   ab = &shProg->AtomicBuffers[bufferIndex];
-
-   switch (pname) {
-   case GL_ATOMIC_COUNTER_BUFFER_BINDING:
-      params[0] = ab->Binding;
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_DATA_SIZE:
-      params[0] = ab->MinimumSize;
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_ACTIVE_ATOMIC_COUNTERS:
-      params[0] = ab->NumUniforms;
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_ACTIVE_ATOMIC_COUNTER_INDICES:
-      for (i = 0; i < ab->NumUniforms; ++i)
-         params[i] = ab->Uniforms[i];
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_VERTEX_SHADER:
-      params[0] = ab->StageReferences[MESA_SHADER_VERTEX];
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_GEOMETRY_SHADER:
-      params[0] = ab->StageReferences[MESA_SHADER_GEOMETRY];
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_FRAGMENT_SHADER:
-      params[0] = ab->StageReferences[MESA_SHADER_FRAGMENT];
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_TESS_CONTROL_SHADER:
-      params[0] = GL_FALSE;
-      return;
-   case GL_ATOMIC_COUNTER_BUFFER_REFERENCED_BY_TESS_EVALUATION_SHADER:
-      params[0] = GL_FALSE;
-      return;
-   default:
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glGetActiveAtomicCounterBufferiv(pname 0x%x (%s))",
-                  pname, _mesa_lookup_enum_by_nr(pname));
-      return;
-   }
+   mesa_bufferiv(shProg, GL_ATOMIC_COUNTER_BUFFER, bufferIndex, pname, params,
+                 "glGetActiveAtomicCounterBufferiv");
 }
 
 void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/uniforms.h b/mesalib/src/mesa/main/uniforms.h
index 0e6113fe9..55fa2357e 100644
--- a/mesalib/src/mesa/main/uniforms.h
+++ b/mesalib/src/mesa/main/uniforms.h
@@ -398,11 +398,6 @@ _mesa_sampler_uniforms_pipeline_are_valid(struct gl_pipeline_object *);
 extern const struct gl_program_parameter *
 get_uniform_parameter(struct gl_shader_program *shProg, GLint index);
 
-extern void
-_mesa_get_uniform_name(const struct gl_uniform_storage *uni,
-                       GLsizei maxLength, GLsizei *length,
-                       GLchar *nameOut);
-
 struct gl_builtin_uniform_element {
    const char *field;
    int tokens[STATE_LENGTH];
diff --git a/mesalib/src/mesa/main/version.c b/mesalib/src/mesa/main/version.c
index d8e22d438..5df5a2922 100644
--- a/mesalib/src/mesa/main/version.c
+++ b/mesalib/src/mesa/main/version.c
@@ -382,6 +382,7 @@ compute_version_es2(const struct gl_extensions *extensions)
                               extensions->ARB_shader_texture_lod &&
                               extensions->ARB_texture_float &&
                               extensions->ARB_texture_rg &&
+                              extensions->ARB_depth_buffer_float &&
                               extensions->EXT_draw_buffers2 &&
                               /* extensions->ARB_framebuffer_object && */
                               extensions->EXT_framebuffer_sRGB &&
diff --git a/mesalib/src/mesa/main/viewport.c b/mesalib/src/mesa/main/viewport.c
index 0adce9c78..b27063031 100644
--- a/mesalib/src/mesa/main/viewport.c
+++ b/mesalib/src/mesa/main/viewport.c
@@ -40,8 +40,6 @@ set_viewport_no_notify(struct gl_context *ctx, unsigned idx,
                        GLfloat x, GLfloat y,
                        GLfloat width, GLfloat height)
 {
-   double scale[3], translate[3];
-
    /* clamp width and height to the implementation dependent range */
    width  = MIN2(width, (GLfloat) ctx->Const.MaxViewportWidth);
    height = MIN2(height, (GLfloat) ctx->Const.MaxViewportHeight);
@@ -72,16 +70,6 @@ set_viewport_no_notify(struct gl_context *ctx, unsigned idx,
    ctx->ViewportArray[idx].Y = y;
    ctx->ViewportArray[idx].Height = height;
    ctx->NewState |= _NEW_VIEWPORT;
-
-#if 1
-   /* XXX remove this someday.  Currently the DRI drivers rely on
-    * the WindowMap matrix being up to date in the driver's Viewport
-    * and DepthRange functions.
-    */
-   _mesa_get_viewport_xform(ctx, idx, scale, translate);
-   _math_matrix_viewport(&ctx->ViewportArray[idx]._WindowMap,
-                         scale, translate, ctx->DrawBuffer->_DepthMaxF);
-#endif
 }
 
 struct gl_viewport_inputs {
@@ -140,8 +128,8 @@ _mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height)
 
 
 /**
- * Set new viewport parameters and update derived state (the _WindowMap
- * matrix).  Usually called from _mesa_Viewport().
+ * Set new viewport parameters and update derived state.
+ * Usually called from _mesa_Viewport().
  * 
  * \param ctx GL context.
  * \param idx    Index of the viewport to be updated.
@@ -246,8 +234,6 @@ static void
 set_depth_range_no_notify(struct gl_context *ctx, unsigned idx,
                           GLclampd nearval, GLclampd farval)
 {
-   double scale[3], translate[3];
-
    if (ctx->ViewportArray[idx].Near == nearval &&
        ctx->ViewportArray[idx].Far == farval)
       return;
@@ -255,16 +241,6 @@ set_depth_range_no_notify(struct gl_context *ctx, unsigned idx,
    ctx->ViewportArray[idx].Near = CLAMP(nearval, 0.0, 1.0);
    ctx->ViewportArray[idx].Far = CLAMP(farval, 0.0, 1.0);
    ctx->NewState |= _NEW_VIEWPORT;
-
-#if 1
-   /* XXX remove this someday.  Currently the DRI drivers rely on
-    * the WindowMap matrix being up to date in the driver's Viewport
-    * and DepthRange functions.
-    */
-   _mesa_get_viewport_xform(ctx, idx, scale, translate);
-   _math_matrix_viewport(&ctx->ViewportArray[idx]._WindowMap,
-                         scale, translate, ctx->DrawBuffer->_DepthMaxF);
-#endif
 }
 
 void
@@ -388,7 +364,6 @@ _mesa_DepthRangeIndexed(GLuint index, GLclampd nearval, GLclampd farval)
  */
 void _mesa_init_viewport(struct gl_context *ctx)
 {
-   GLfloat depthMax = 65535.0F; /* sorf of arbitrary */
    unsigned i;
 
    ctx->Transform.ClipOrigin = GL_LOWER_LEFT;
@@ -398,8 +373,6 @@ void _mesa_init_viewport(struct gl_context *ctx)
     * so just initialize all of them.
     */
    for (i = 0; i < MAX_VIEWPORTS; i++) {
-      double scale[3], translate[3];
-
       /* Viewport group */
       ctx->ViewportArray[i].X = 0;
       ctx->ViewportArray[i].Y = 0;
@@ -407,27 +380,10 @@ void _mesa_init_viewport(struct gl_context *ctx)
       ctx->ViewportArray[i].Height = 0;
       ctx->ViewportArray[i].Near = 0.0;
       ctx->ViewportArray[i].Far = 1.0;
-      _math_matrix_ctr(&ctx->ViewportArray[i]._WindowMap);
-
-      _mesa_get_viewport_xform(ctx, i, scale, translate);
-      _math_matrix_viewport(&ctx->ViewportArray[i]._WindowMap,
-                            scale, translate, depthMax);
    }
 }
 
 
-/** 
- * Free the context viewport attribute group data.
- * \param ctx  the GL context.
- */
-void _mesa_free_viewport_data(struct gl_context *ctx)
-{
-   unsigned i;
-
-   for (i = 0; i < MAX_VIEWPORTS; i++)
-      _math_matrix_dtr(&ctx->ViewportArray[i]._WindowMap);
-}
-
 extern void GLAPIENTRY
 _mesa_ClipControl(GLenum origin, GLenum depth)
 {
diff --git a/mesalib/src/mesa/main/viewport.h b/mesalib/src/mesa/main/viewport.h
index 426e194bd..899dc2d0b 100644
--- a/mesalib/src/mesa/main/viewport.h
+++ b/mesalib/src/mesa/main/viewport.h
@@ -68,9 +68,6 @@ extern void
 _mesa_init_viewport(struct gl_context *ctx);
 
 
-extern void 
-_mesa_free_viewport_data(struct gl_context *ctx);
-
 extern void GLAPIENTRY
 _mesa_ClipControl(GLenum origin, GLenum depth);
 
diff --git a/mesalib/src/mesa/program/ir_to_mesa.cpp b/mesalib/src/mesa/program/ir_to_mesa.cpp
index 39790ec8e..3dcb53702 100644
--- a/mesalib/src/mesa/program/ir_to_mesa.cpp
+++ b/mesalib/src/mesa/program/ir_to_mesa.cpp
@@ -303,9 +303,6 @@ public:
    void emit_scalar(ir_instruction *ir, enum prog_opcode op,
 		    dst_reg dst, src_reg src0, src_reg src1);
 
-   void emit_scs(ir_instruction *ir, enum prog_opcode op,
-		 dst_reg dst, const src_reg &src);
-
    bool try_emit_mad(ir_expression *ir,
 			  int mul_operand);
    bool try_emit_mad_for_and_not(ir_expression *ir,
@@ -479,101 +476,6 @@ ir_to_mesa_visitor::emit_scalar(ir_instruction *ir, enum prog_opcode op,
    emit_scalar(ir, op, dst, src0, undef);
 }
 
-/**
- * Emit an OPCODE_SCS instruction
- *
- * The \c SCS opcode functions a bit differently than the other Mesa (or
- * ARB_fragment_program) opcodes.  Instead of splatting its result across all
- * four components of the destination, it writes one value to the \c x
- * component and another value to the \c y component.
- *
- * \param ir        IR instruction being processed
- * \param op        Either \c OPCODE_SIN or \c OPCODE_COS depending on which
- *                  value is desired.
- * \param dst       Destination register
- * \param src       Source register
- */
-void
-ir_to_mesa_visitor::emit_scs(ir_instruction *ir, enum prog_opcode op,
-			     dst_reg dst,
-			     const src_reg &src)
-{
-   /* Vertex programs cannot use the SCS opcode.
-    */
-   if (this->prog->Target == GL_VERTEX_PROGRAM_ARB) {
-      emit_scalar(ir, op, dst, src);
-      return;
-   }
-
-   const unsigned component = (op == OPCODE_SIN) ? 0 : 1;
-   const unsigned scs_mask = (1U << component);
-   int done_mask = ~dst.writemask;
-   src_reg tmp;
-
-   assert(op == OPCODE_SIN || op == OPCODE_COS);
-
-   /* If there are compnents in the destination that differ from the component
-    * that will be written by the SCS instrution, we'll need a temporary.
-    */
-   if (scs_mask != unsigned(dst.writemask)) {
-      tmp = get_temp(glsl_type::vec4_type);
-   }
-
-   for (unsigned i = 0; i < 4; i++) {
-      unsigned this_mask = (1U << i);
-      src_reg src0 = src;
-
-      if ((done_mask & this_mask) != 0)
-	 continue;
-
-      /* The source swizzle specified which component of the source generates
-       * sine / cosine for the current component in the destination.  The SCS
-       * instruction requires that this value be swizzle to the X component.
-       * Replace the current swizzle with a swizzle that puts the source in
-       * the X component.
-       */
-      unsigned src0_swiz = GET_SWZ(src.swizzle, i);
-
-      src0.swizzle = MAKE_SWIZZLE4(src0_swiz, src0_swiz,
-				   src0_swiz, src0_swiz);
-      for (unsigned j = i + 1; j < 4; j++) {
-	 /* If there is another enabled component in the destination that is
-	  * derived from the same inputs, generate its value on this pass as
-	  * well.
-	  */
-	 if (!(done_mask & (1 << j)) &&
-	     GET_SWZ(src0.swizzle, j) == src0_swiz) {
-	    this_mask |= (1 << j);
-	 }
-      }
-
-      if (this_mask != scs_mask) {
-	 ir_to_mesa_instruction *inst;
-	 dst_reg tmp_dst = dst_reg(tmp);
-
-	 /* Emit the SCS instruction.
-	  */
-	 inst = emit(ir, OPCODE_SCS, tmp_dst, src0);
-	 inst->dst.writemask = scs_mask;
-
-	 /* Move the result of the SCS instruction to the desired location in
-	  * the destination.
-	  */
-	 tmp.swizzle = MAKE_SWIZZLE4(component, component,
-				     component, component);
-	 inst = emit(ir, OPCODE_SCS, dst, tmp);
-	 inst->dst.writemask = this_mask;
-      } else {
-	 /* Emit the SCS instruction to write directly to the destination.
-	  */
-	 ir_to_mesa_instruction *inst = emit(ir, OPCODE_SCS, dst, src0);
-	 inst->dst.writemask = scs_mask;
-      }
-
-      done_mask |= this_mask;
-   }
-}
-
 src_reg
 ir_to_mesa_visitor::src_reg_for_float(float val)
 {
@@ -1122,12 +1024,6 @@ ir_to_mesa_visitor::visit(ir_expression *ir)
    case ir_unop_cos:
       emit_scalar(ir, OPCODE_COS, result_dst, op[0]);
       break;
-   case ir_unop_sin_reduced:
-      emit_scs(ir, OPCODE_SIN, result_dst, op[0]);
-      break;
-   case ir_unop_cos_reduced:
-      emit_scs(ir, OPCODE_COS, result_dst, op[0]);
-      break;
 
    case ir_unop_dFdx:
       emit(ir, OPCODE_DDX, result_dst, op[0]);
diff --git a/mesalib/src/mesa/program/prog_execute.c b/mesalib/src/mesa/program/prog_execute.c
index dc4919ae8..16e8e340d 100644
--- a/mesalib/src/mesa/program/prog_execute.c
+++ b/mesalib/src/mesa/program/prog_execute.c
@@ -37,7 +37,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "prog_execute.h"
 #include "prog_instruction.h"
diff --git a/mesalib/src/mesa/program/prog_instruction.h b/mesalib/src/mesa/program/prog_instruction.h
index ab3acbc02..96da198f8 100644
--- a/mesalib/src/mesa/program/prog_instruction.h
+++ b/mesalib/src/mesa/program/prog_instruction.h
@@ -59,6 +59,8 @@
 #define SWIZZLE_NOOP           MAKE_SWIZZLE4(0,1,2,3)
 #define GET_SWZ(swz, idx)      (((swz) >> ((idx)*3)) & 0x7)
 #define GET_BIT(msk, idx)      (((msk) >> (idx)) & 0x1)
+/** Determine if swz contains SWIZZLE_ZERO/ONE/NIL for any components. */
+#define HAS_EXTENDED_SWIZZLE(swz) (swz & 0x924)
 
 #define SWIZZLE_XYZW MAKE_SWIZZLE4(SWIZZLE_X, SWIZZLE_Y, SWIZZLE_Z, SWIZZLE_W)
 #define SWIZZLE_XXXX MAKE_SWIZZLE4(SWIZZLE_X, SWIZZLE_X, SWIZZLE_X, SWIZZLE_X)
diff --git a/mesalib/src/mesa/program/prog_parameter.c b/mesalib/src/mesa/program/prog_parameter.c
index cdfe25145..53e9813e6 100644
--- a/mesalib/src/mesa/program/prog_parameter.c
+++ b/mesalib/src/mesa/program/prog_parameter.c
@@ -190,40 +190,6 @@ _mesa_add_parameter(struct gl_program_parameter_list *paramList,
 
 
 /**
- * Add a new named constant to the parameter list.
- * This will be used when the program contains something like this:
- *    PARAM myVals = { 0, 1, 2, 3 };
- *
- * \param paramList  the parameter list
- * \param name  the name for the constant
- * \param values  four float values
- * \return index/position of the new parameter in the parameter list
- */
-GLint
-_mesa_add_named_constant(struct gl_program_parameter_list *paramList,
-                         const char *name, const gl_constant_value values[4],
-                         GLuint size)
-{
-   /* first check if this is a duplicate constant */
-   GLint pos;
-   for (pos = 0; pos < (GLint)paramList->NumParameters; pos++) {
-      const gl_constant_value *pvals = paramList->ParameterValues[pos];
-      if (pvals[0].u == values[0].u &&
-          pvals[1].u == values[1].u &&
-          pvals[2].u == values[2].u &&
-          pvals[3].u == values[3].u &&
-          strcmp(paramList->Parameters[pos].Name, name) == 0) {
-         /* Same name and value is already in the param list - reuse it */
-         return pos;
-      }
-   }
-   /* not found, add new parameter */
-   return _mesa_add_parameter(paramList, PROGRAM_CONSTANT, name,
-                              size, GL_NONE, values, NULL);
-}
-
-
-/**
  * Add a new unnamed constant to the parameter list.  This will be used
  * when a fragment/vertex program contains something like this:
  *    MOV r, { 0, 1, 2, 3 };
@@ -303,28 +269,6 @@ _mesa_add_unnamed_constant(struct gl_program_parameter_list *paramList,
                                            swizzleOut);
 }
 
-#if 0 /* not used yet */
-/**
- * Returns the number of 4-component registers needed to store a piece
- * of GL state.  For matrices this may be as many as 4 registers,
- * everything else needs
- * just 1 register.
- */
-static GLuint
-sizeof_state_reference(const GLint *stateTokens)
-{
-   if (stateTokens[0] == STATE_MATRIX) {
-      GLuint rows = stateTokens[4] - stateTokens[3] + 1;
-      assert(rows >= 1);
-      assert(rows <= 4);
-      return rows;
-   }
-   else {
-      return 1;
-   }
-}
-#endif
-
 
 /**
  * Add a new state reference to the parameter list.
@@ -365,22 +309,6 @@ _mesa_add_state_reference(struct gl_program_parameter_list *paramList,
 
 
 /**
- * Lookup a parameter value by name in the given parameter list.
- * \return pointer to the float[4] values.
- */
-gl_constant_value *
-_mesa_lookup_parameter_value(const struct gl_program_parameter_list *paramList,
-                             GLsizei nameLen, const char *name)
-{
-   GLint i = _mesa_lookup_parameter_index(paramList, nameLen, name);
-   if (i < 0)
-      return NULL;
-   else
-      return paramList->ParameterValues[i];
-}
-
-
-/**
  * Given a program parameter name, find its position in the list of parameters.
  * \param paramList  the parameter list to search
  * \param nameLen  length of name (in chars).
diff --git a/mesalib/src/mesa/program/prog_parameter.h b/mesalib/src/mesa/program/prog_parameter.h
index 6b3b3c262..74a5fd918 100644
--- a/mesalib/src/mesa/program/prog_parameter.h
+++ b/mesalib/src/mesa/program/prog_parameter.h
@@ -120,11 +120,6 @@ _mesa_add_parameter(struct gl_program_parameter_list *paramList,
                     const gl_state_index state[STATE_LENGTH]);
 
 extern GLint
-_mesa_add_named_constant(struct gl_program_parameter_list *paramList,
-                         const char *name, const gl_constant_value values[4],
-                         GLuint size);
-
-extern GLint
 _mesa_add_typed_unnamed_constant(struct gl_program_parameter_list *paramList,
                            const gl_constant_value values[4], GLuint size,
                            GLenum datatype, GLuint *swizzleOut);
@@ -138,10 +133,6 @@ extern GLint
 _mesa_add_state_reference(struct gl_program_parameter_list *paramList,
                           const gl_state_index stateTokens[STATE_LENGTH]);
 
-extern gl_constant_value *
-_mesa_lookup_parameter_value(const struct gl_program_parameter_list *paramList,
-                             GLsizei nameLen, const char *name);
-
 extern GLint
 _mesa_lookup_parameter_index(const struct gl_program_parameter_list *paramList,
                              GLsizei nameLen, const char *name);
diff --git a/mesalib/src/mesa/program/prog_to_nir.c b/mesalib/src/mesa/program/prog_to_nir.c
new file mode 100644
index 000000000..c738f5073
--- /dev/null
+++ b/mesalib/src/mesa/program/prog_to_nir.c
@@ -0,0 +1,1096 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ * Copyright © 2014-2015 Broadcom
+ * Copyright (C) 2014 Rob Clark <robclark@freedesktop.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "nir/nir.h"
+#include "nir/nir_builder.h"
+#include "glsl/list.h"
+#include "main/imports.h"
+#include "util/ralloc.h"
+
+#include "prog_to_nir.h"
+#include "prog_instruction.h"
+#include "prog_parameter.h"
+#include "prog_print.h"
+
+/**
+ * \file prog_to_nir.c
+ *
+ * A translator from Mesa IR (prog_instruction.h) to NIR.  This is primarily
+ * intended to support ARB_vertex_program, ARB_fragment_program, and fixed-function
+ * vertex processing.  Full GLSL support should use glsl_to_nir instead.
+ */
+
+struct ptn_compile {
+   const struct gl_program *prog;
+   nir_builder build;
+   bool error;
+
+   nir_variable *input_vars[VARYING_SLOT_MAX];
+   nir_variable *output_vars[VARYING_SLOT_MAX];
+   nir_register **output_regs;
+   nir_register **temp_regs;
+
+   nir_register *addr_reg;
+};
+
+#define SWIZ(X, Y, Z, W) \
+   (unsigned[4]){ SWIZZLE_##X, SWIZZLE_##Y, SWIZZLE_##Z, SWIZZLE_##W }
+#define ptn_swizzle(b, src, x, y, z, w) nir_swizzle(b, src, SWIZ(x, y, z, w), 4, true)
+#define ptn_channel(b, src, ch) nir_swizzle(b, src, SWIZ(ch, ch, ch, ch), 1, true)
+
+static nir_ssa_def *
+ptn_src_for_dest(struct ptn_compile *c, nir_alu_dest *dest)
+{
+   nir_builder *b = &c->build;
+
+   nir_alu_src src;
+   memset(&src, 0, sizeof(src));
+
+   if (dest->dest.is_ssa)
+      src.src = nir_src_for_ssa(&dest->dest.ssa);
+   else {
+      assert(!dest->dest.reg.indirect);
+      src.src = nir_src_for_reg(dest->dest.reg.reg);
+      src.src.reg.base_offset = dest->dest.reg.base_offset;
+   }
+
+   for (int i = 0; i < 4; i++)
+      src.swizzle[i] = i;
+
+   return nir_fmov_alu(b, src, 4);
+}
+
+static nir_alu_dest
+ptn_get_dest(struct ptn_compile *c, const struct prog_dst_register *prog_dst)
+{
+   nir_alu_dest dest;
+
+   memset(&dest, 0, sizeof(dest));
+
+   switch (prog_dst->File) {
+   case PROGRAM_TEMPORARY:
+      dest.dest.reg.reg = c->temp_regs[prog_dst->Index];
+      break;
+   case PROGRAM_OUTPUT:
+      dest.dest.reg.reg = c->output_regs[prog_dst->Index];
+      break;
+   case PROGRAM_ADDRESS:
+      assert(prog_dst->Index == 0);
+      dest.dest.reg.reg = c->addr_reg;
+      break;
+   case PROGRAM_UNDEFINED:
+      break;
+   }
+
+   dest.write_mask = prog_dst->WriteMask;
+   dest.saturate = false;
+
+   assert(!prog_dst->RelAddr);
+
+   return dest;
+}
+
+/**
+ * Multiply the contents of the ADDR register by 4 to convert from the number
+ * of vec4s to the number of floating point components.
+ */
+static nir_ssa_def *
+ptn_addr_reg_value(struct ptn_compile *c)
+{
+   nir_builder *b = &c->build;
+   nir_alu_src src;
+   memset(&src, 0, sizeof(src));
+   src.src = nir_src_for_reg(c->addr_reg);
+
+   return nir_imul(b, nir_fmov_alu(b, src, 1), nir_imm_int(b, 4));
+}
+
+static nir_ssa_def *
+ptn_get_src(struct ptn_compile *c, const struct prog_src_register *prog_src)
+{
+   nir_builder *b = &c->build;
+   nir_alu_src src;
+
+   memset(&src, 0, sizeof(src));
+
+   switch (prog_src->File) {
+   case PROGRAM_UNDEFINED:
+      return nir_imm_float(b, 0.0);
+   case PROGRAM_TEMPORARY:
+      assert(!prog_src->RelAddr && prog_src->Index >= 0);
+      src.src.reg.reg = c->temp_regs[prog_src->Index];
+      break;
+   case PROGRAM_INPUT: {
+      /* ARB_vertex_program doesn't allow relative addressing on vertex
+       * attributes; ARB_fragment_program has no relative addressing at all.
+       */
+      assert(!prog_src->RelAddr);
+
+      assert(prog_src->Index >= 0 && prog_src->Index < VARYING_SLOT_MAX);
+
+      nir_intrinsic_instr *load =
+         nir_intrinsic_instr_create(b->shader, nir_intrinsic_load_var);
+      load->num_components = 4;
+      load->variables[0] = nir_deref_var_create(load, c->input_vars[prog_src->Index]);
+
+      nir_ssa_dest_init(&load->instr, &load->dest, 4, NULL);
+      nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+
+      src.src = nir_src_for_ssa(&load->dest.ssa);
+      break;
+   }
+   case PROGRAM_STATE_VAR:
+   case PROGRAM_CONSTANT: {
+      /* We actually want to look at the type in the Parameters list for this,
+       * because it lets us upload constant builtin uniforms as actual
+       * constants.
+       */
+      struct gl_program_parameter_list *plist = c->prog->Parameters;
+      gl_register_file file = prog_src->RelAddr ? prog_src->File :
+         plist->Parameters[prog_src->Index].Type;
+
+      switch (file) {
+      case PROGRAM_CONSTANT:
+         if ((c->prog->IndirectRegisterFiles & (1 << PROGRAM_CONSTANT)) == 0) {
+            float *v = (float *) plist->ParameterValues[prog_src->Index];
+            src.src = nir_src_for_ssa(nir_imm_vec4(b, v[0], v[1], v[2], v[3]));
+            break;
+         }
+         /* FALLTHROUGH */
+      case PROGRAM_STATE_VAR: {
+         nir_intrinsic_op load_op =
+            prog_src->RelAddr ? nir_intrinsic_load_uniform_indirect :
+                                nir_intrinsic_load_uniform;
+         nir_intrinsic_instr *load = nir_intrinsic_instr_create(b->shader, load_op);
+         nir_ssa_dest_init(&load->instr, &load->dest, 4, NULL);
+         load->num_components = 4;
+
+         /* Multiply src->Index by 4 to scale from # of vec4s to components. */
+         load->const_index[0] = 4 * prog_src->Index;
+         load->const_index[1] = 1;
+
+         if (prog_src->RelAddr) {
+            nir_ssa_def *reladdr = ptn_addr_reg_value(c);
+            if (prog_src->Index < 0) {
+               /* This is a negative offset which should be added to the address
+                * register's value.
+                */
+               reladdr = nir_iadd(b, reladdr, nir_imm_int(b, load->const_index[0]));
+               load->const_index[0] = 0;
+            }
+            load->src[0] = nir_src_for_ssa(reladdr);
+         }
+
+         nir_instr_insert_after_cf_list(b->cf_node_list, &load->instr);
+
+         src.src = nir_src_for_ssa(&load->dest.ssa);
+         break;
+      }
+      default:
+         fprintf(stderr, "bad uniform src register file: %s (%d)\n",
+                 _mesa_register_file_name(file), file);
+         abort();
+      }
+      break;
+   }
+   default:
+      fprintf(stderr, "unknown src register file: %s (%d)\n",
+              _mesa_register_file_name(prog_src->File), prog_src->File);
+      abort();
+   }
+
+   nir_ssa_def *def;
+   if (!HAS_EXTENDED_SWIZZLE(prog_src->Swizzle)) {
+      for (int i = 0; i < 4; i++)
+         src.swizzle[i] = GET_SWZ(prog_src->Swizzle, i);
+
+      def = nir_fmov_alu(b, src, 4);
+   } else {
+      nir_ssa_def *chans[4];
+      for (int i = 0; i < 4; i++) {
+         int swizzle = GET_SWZ(prog_src->Swizzle, i);
+         if (swizzle == SWIZZLE_ZERO) {
+            chans[i] = nir_imm_float(b, 0.0);
+         } else if (swizzle == SWIZZLE_ONE) {
+            chans[i] = nir_imm_float(b, 1.0);
+         } else {
+            assert(swizzle != SWIZZLE_NIL);
+            nir_alu_instr *mov = nir_alu_instr_create(b->shader, nir_op_fmov);
+            nir_ssa_dest_init(&mov->instr, &mov->dest.dest, 1, NULL);
+            mov->dest.write_mask = 0x1;
+            mov->src[0] = src;
+            mov->src[0].swizzle[0] = swizzle;
+            nir_instr_insert_after_cf_list(b->cf_node_list, &mov->instr);
+
+            chans[i] = &mov->dest.dest.ssa;
+         }
+      }
+      def = nir_vec4(b, chans[0], chans[1], chans[2], chans[3]);
+   }
+
+   if (prog_src->Abs)
+      def = nir_fabs(b, def);
+
+   if (prog_src->Negate)
+      def = nir_fneg(b, def);
+
+   return def;
+}
+
+static void
+ptn_alu(nir_builder *b, nir_op op, nir_alu_dest dest, nir_ssa_def **src)
+{
+   unsigned num_srcs = nir_op_infos[op].num_inputs;
+   nir_alu_instr *instr = nir_alu_instr_create(b->shader, op);
+   unsigned i;
+
+   for (i = 0; i < num_srcs; i++)
+      instr->src[i].src = nir_src_for_ssa(src[i]);
+
+   instr->dest = dest;
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+}
+
+static void
+ptn_move_dest_masked(nir_builder *b, nir_alu_dest dest,
+                     nir_ssa_def *def, unsigned write_mask)
+{
+   if (!(dest.write_mask & write_mask))
+      return;
+
+   nir_alu_instr *mov = nir_alu_instr_create(b->shader, nir_op_fmov);
+   if (!mov)
+      return;
+
+   mov->dest = dest;
+   mov->dest.write_mask &= write_mask;
+   mov->src[0].src = nir_src_for_ssa(def);
+   for (unsigned i = def->num_components; i < 4; i++)
+      mov->src[0].swizzle[i] = def->num_components - 1;
+   nir_instr_insert_after_cf_list(b->cf_node_list, &mov->instr);
+}
+
+static void
+ptn_move_dest(nir_builder *b, nir_alu_dest dest, nir_ssa_def *def)
+{
+   ptn_move_dest_masked(b, dest, def, WRITEMASK_XYZW);
+}
+
+static void
+ptn_arl(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest(b, dest, nir_f2i(b, nir_ffloor(b, src[0])));
+}
+
+/* EXP - Approximate Exponential Base 2
+ *  dst.x = 2^{\lfloor src.x\rfloor}
+ *  dst.y = src.x - \lfloor src.x\rfloor
+ *  dst.z = 2^{src.x}
+ *  dst.w = 1.0
+ */
+static void
+ptn_exp(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *srcx = ptn_channel(b, src[0], X);
+
+   ptn_move_dest_masked(b, dest, nir_fexp2(b, nir_ffloor(b, srcx)), WRITEMASK_X);
+   ptn_move_dest_masked(b, dest, nir_fsub(b, srcx, nir_ffloor(b, srcx)), WRITEMASK_Y);
+   ptn_move_dest_masked(b, dest, nir_fexp2(b, srcx), WRITEMASK_Z);
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_W);
+}
+
+/* LOG - Approximate Logarithm Base 2
+ *  dst.x = \lfloor\log_2{|src.x|}\rfloor
+ *  dst.y = |src.x| * 2^{-\lfloor\log_2{|src.x|}\rfloor}}
+ *  dst.z = \log_2{|src.x|}
+ *  dst.w = 1.0
+ */
+static void
+ptn_log(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *abs_srcx = nir_fabs(b, ptn_channel(b, src[0], X));
+   nir_ssa_def *log2 = nir_flog2(b, abs_srcx);
+   nir_ssa_def *floor_log2 = nir_ffloor(b, log2);
+
+   ptn_move_dest_masked(b, dest, floor_log2, WRITEMASK_X);
+   ptn_move_dest_masked(b, dest,
+                        nir_fmul(b, abs_srcx,
+                                 nir_fexp2(b, nir_fneg(b, floor_log2))),
+                        WRITEMASK_Y);
+   ptn_move_dest_masked(b, dest, log2, WRITEMASK_Z);
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_W);
+}
+
+/* DST - Distance Vector
+ *   dst.x = 1.0
+ *   dst.y = src0.y \times src1.y
+ *   dst.z = src0.z
+ *   dst.w = src1.w
+ */
+static void
+ptn_dst(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_X);
+   ptn_move_dest_masked(b, dest, nir_fmul(b, src[0], src[1]), WRITEMASK_Y);
+   ptn_move_dest_masked(b, dest, nir_fmov(b, src[0]), WRITEMASK_Z);
+   ptn_move_dest_masked(b, dest, nir_fmov(b, src[1]), WRITEMASK_W);
+}
+
+/* LIT - Light Coefficients
+ *  dst.x = 1.0
+ *  dst.y = max(src.x, 0.0)
+ *  dst.z = (src.x > 0.0) ? max(src.y, 0.0)^{clamp(src.w, -128.0, 128.0))} : 0
+ *  dst.w = 1.0
+ */
+static void
+ptn_lit(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_XW);
+
+   ptn_move_dest_masked(b, dest, nir_fmax(b, ptn_channel(b, src[0], X),
+                                          nir_imm_float(b, 0.0)), WRITEMASK_Y);
+
+   if (dest.write_mask & WRITEMASK_Z) {
+      nir_ssa_def *src0_y = ptn_channel(b, src[0], Y);
+      nir_ssa_def *wclamp = nir_fmax(b, nir_fmin(b, ptn_channel(b, src[0], W),
+                                                 nir_imm_float(b, 128.0)),
+                                     nir_imm_float(b, -128.0));
+      nir_ssa_def *pow = nir_fpow(b, nir_fmax(b, src0_y, nir_imm_float(b, 0.0)),
+                                  wclamp);
+
+      nir_ssa_def *z;
+      if (b->shader->options->native_integers) {
+         z = nir_bcsel(b,
+                       nir_fge(b, nir_imm_float(b, 0.0), ptn_channel(b, src[0], X)),
+                       nir_imm_float(b, 0.0),
+                       pow);
+      } else {
+         z = nir_fcsel(b,
+                       nir_sge(b, nir_imm_float(b, 0.0), ptn_channel(b, src[0], X)),
+                       nir_imm_float(b, 0.0),
+                       pow);
+      }
+
+      ptn_move_dest_masked(b, dest, z, WRITEMASK_Z);
+   }
+}
+
+/* SCS - Sine Cosine
+ *   dst.x = \cos{src.x}
+ *   dst.y = \sin{src.x}
+ *   dst.z = 0.0
+ *   dst.w = 1.0
+ */
+static void
+ptn_scs(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest_masked(b, dest, nir_fcos(b, ptn_channel(b, src[0], X)),
+                        WRITEMASK_X);
+   ptn_move_dest_masked(b, dest, nir_fsin(b, ptn_channel(b, src[0], X)),
+                        WRITEMASK_Y);
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 0.0), WRITEMASK_Z);
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_W);
+}
+
+/**
+ * Emit SLT.  For platforms with integers, prefer b2f(flt(...)).
+ */
+static void
+ptn_slt(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   if (b->shader->options->native_integers) {
+      ptn_move_dest(b, dest, nir_b2f(b, nir_flt(b, src[0], src[1])));
+   } else {
+      ptn_move_dest(b, dest, nir_slt(b, src[0], src[1]));
+   }
+}
+
+/**
+ * Emit SGE.  For platforms with integers, prefer b2f(fge(...)).
+ */
+static void
+ptn_sge(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   if (b->shader->options->native_integers) {
+      ptn_move_dest(b, dest, nir_b2f(b, nir_fge(b, src[0], src[1])));
+   } else {
+      ptn_move_dest(b, dest, nir_sge(b, src[0], src[1]));
+   }
+}
+
+static void
+ptn_sle(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *commuted[] = { src[1], src[0] };
+   ptn_sge(b, dest, commuted);
+}
+
+static void
+ptn_sgt(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *commuted[] = { src[1], src[0] };
+   ptn_slt(b, dest, commuted);
+}
+
+/**
+ * Emit SEQ.  For platforms with integers, prefer b2f(feq(...)).
+ */
+static void
+ptn_seq(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   if (b->shader->options->native_integers) {
+      ptn_move_dest(b, dest, nir_b2f(b, nir_feq(b, src[0], src[1])));
+   } else {
+      ptn_move_dest(b, dest, nir_seq(b, src[0], src[1]));
+   }
+}
+
+/**
+ * Emit SNE.  For platforms with integers, prefer b2f(fne(...)).
+ */
+static void
+ptn_sne(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   if (b->shader->options->native_integers) {
+      ptn_move_dest(b, dest, nir_b2f(b, nir_fne(b, src[0], src[1])));
+   } else {
+      ptn_move_dest(b, dest, nir_sne(b, src[0], src[1]));
+   }
+}
+
+static void
+ptn_xpd(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest_masked(b, dest,
+                        nir_fsub(b,
+                                 nir_fmul(b,
+                                          ptn_swizzle(b, src[0], Y, Z, X, X),
+                                          ptn_swizzle(b, src[1], Z, X, Y, X)),
+                                 nir_fmul(b,
+                                          ptn_swizzle(b, src[1], Y, Z, X, X),
+                                          ptn_swizzle(b, src[0], Z, X, Y, X))),
+                        WRITEMASK_XYZ);
+   ptn_move_dest_masked(b, dest, nir_imm_float(b, 1.0), WRITEMASK_W);
+}
+
+static void
+ptn_dp2(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest(b, dest, nir_fdot2(b, src[0], src[1]));
+}
+
+static void
+ptn_dp3(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest(b, dest, nir_fdot3(b, src[0], src[1]));
+}
+
+static void
+ptn_dp4(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest(b, dest, nir_fdot4(b, src[0], src[1]));
+}
+
+static void
+ptn_dph(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *dp3 = nir_fdot3(b, src[0], src[1]);
+   ptn_move_dest(b, dest, nir_fadd(b, dp3, ptn_channel(b, src[1], W)));
+}
+
+static void
+ptn_cmp(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   if (b->shader->options->native_integers) {
+      ptn_move_dest(b, dest, nir_bcsel(b,
+                                       nir_flt(b, src[0], nir_imm_float(b, 0.0)),
+                                       src[1], src[2]));
+   } else {
+      ptn_move_dest(b, dest, nir_fcsel(b,
+                                       nir_slt(b, src[0], nir_imm_float(b, 0.0)),
+                                       src[1], src[2]));
+   }
+}
+
+static void
+ptn_lrp(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   ptn_move_dest(b, dest, nir_flrp(b, src[2], src[1], src[0]));
+}
+
+static void
+ptn_kil(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src)
+{
+   nir_ssa_def *cmp = b->shader->options->native_integers ?
+      nir_bany4(b, nir_flt(b, src[0], nir_imm_float(b, 0.0))) :
+      nir_fany4(b, nir_slt(b, src[0], nir_imm_float(b, 0.0)));
+
+   nir_intrinsic_instr *discard =
+      nir_intrinsic_instr_create(b->shader, nir_intrinsic_discard_if);
+   discard->src[0] = nir_src_for_ssa(cmp);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &discard->instr);
+}
+
+static void
+ptn_tex(nir_builder *b, nir_alu_dest dest, nir_ssa_def **src,
+        struct prog_instruction *prog_inst)
+{
+   nir_tex_instr *instr;
+   nir_texop op;
+   unsigned num_srcs;
+
+   switch (prog_inst->Opcode) {
+   case OPCODE_TEX:
+      op = nir_texop_tex;
+      num_srcs = 1;
+      break;
+   case OPCODE_TXB:
+      op = nir_texop_txb;
+      num_srcs = 2;
+      break;
+   case OPCODE_TXD:
+      op = nir_texop_txd;
+      num_srcs = 3;
+      break;
+   case OPCODE_TXL:
+      op = nir_texop_txl;
+      num_srcs = 2;
+      break;
+   case OPCODE_TXP:
+      op = nir_texop_tex;
+      num_srcs = 2;
+      break;
+   case OPCODE_TXP_NV:
+      assert(!"not handled");
+      op = nir_texop_tex;
+      num_srcs = 2;
+      break;
+   default:
+      fprintf(stderr, "unknown tex op %d\n", prog_inst->Opcode);
+      abort();
+   }
+
+   if (prog_inst->TexShadow)
+      num_srcs++;
+
+   instr = nir_tex_instr_create(b->shader, num_srcs);
+   instr->op = op;
+   instr->dest_type = nir_type_float;
+   instr->is_shadow = prog_inst->TexShadow;
+   instr->sampler_index = prog_inst->TexSrcUnit;
+
+   switch (prog_inst->TexSrcTarget) {
+   case TEXTURE_1D_INDEX:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_1D;
+      break;
+   case TEXTURE_2D_INDEX:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_2D;
+      break;
+   case TEXTURE_3D_INDEX:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_3D;
+      break;
+   case TEXTURE_CUBE_INDEX:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_CUBE;
+      break;
+   case TEXTURE_RECT_INDEX:
+      instr->sampler_dim = GLSL_SAMPLER_DIM_RECT;
+      break;
+   default:
+      fprintf(stderr, "Unknown texture target %d\n", prog_inst->TexSrcTarget);
+      abort();
+   }
+
+   switch (instr->sampler_dim) {
+   case GLSL_SAMPLER_DIM_1D:
+   case GLSL_SAMPLER_DIM_BUF:
+      instr->coord_components = 1;
+      break;
+   case GLSL_SAMPLER_DIM_2D:
+   case GLSL_SAMPLER_DIM_RECT:
+   case GLSL_SAMPLER_DIM_EXTERNAL:
+   case GLSL_SAMPLER_DIM_MS:
+      instr->coord_components = 2;
+      break;
+   case GLSL_SAMPLER_DIM_3D:
+   case GLSL_SAMPLER_DIM_CUBE:
+      instr->coord_components = 3;
+      break;
+   }
+
+   unsigned src_number = 0;
+
+   instr->src[src_number].src =
+      nir_src_for_ssa(ptn_swizzle(b, src[0], X, Y, Z, W));
+   instr->src[src_number].src_type = nir_tex_src_coord;
+   src_number++;
+
+   if (prog_inst->Opcode == OPCODE_TXP) {
+      instr->src[src_number].src = nir_src_for_ssa(ptn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_projector;
+      src_number++;
+   }
+
+   if (prog_inst->Opcode == OPCODE_TXB) {
+      instr->src[src_number].src = nir_src_for_ssa(ptn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_bias;
+      src_number++;
+   }
+
+   if (prog_inst->Opcode == OPCODE_TXL) {
+      instr->src[src_number].src = nir_src_for_ssa(ptn_channel(b, src[0], W));
+      instr->src[src_number].src_type = nir_tex_src_lod;
+      src_number++;
+   }
+
+   if (instr->is_shadow) {
+      if (instr->coord_components < 3)
+         instr->src[src_number].src = nir_src_for_ssa(ptn_channel(b, src[0], Z));
+      else
+         instr->src[src_number].src = nir_src_for_ssa(ptn_channel(b, src[0], W));
+
+      instr->src[src_number].src_type = nir_tex_src_comparitor;
+      src_number++;
+   }
+
+   assert(src_number == num_srcs);
+
+   nir_ssa_dest_init(&instr->instr, &instr->dest, 4, NULL);
+   nir_instr_insert_after_cf_list(b->cf_node_list, &instr->instr);
+
+   /* Resolve the writemask on the texture op. */
+   ptn_move_dest(b, dest, &instr->dest.ssa);
+}
+
+static const nir_op op_trans[MAX_OPCODE] = {
+   [OPCODE_NOP] = 0,
+   [OPCODE_ABS] = nir_op_fabs,
+   [OPCODE_ADD] = nir_op_fadd,
+   [OPCODE_ARL] = 0,
+   [OPCODE_CMP] = 0,
+   [OPCODE_COS] = nir_op_fcos,
+   [OPCODE_DDX] = nir_op_fddx,
+   [OPCODE_DDY] = nir_op_fddy,
+   [OPCODE_DP2] = 0,
+   [OPCODE_DP3] = 0,
+   [OPCODE_DP4] = 0,
+   [OPCODE_DPH] = 0,
+   [OPCODE_DST] = 0,
+   [OPCODE_END] = 0,
+   [OPCODE_EX2] = nir_op_fexp2,
+   [OPCODE_EXP] = nir_op_fexp,
+   [OPCODE_FLR] = nir_op_ffloor,
+   [OPCODE_FRC] = nir_op_ffract,
+   [OPCODE_LG2] = nir_op_flog2,
+   [OPCODE_LIT] = 0,
+   [OPCODE_LOG] = 0,
+   [OPCODE_LRP] = 0,
+   [OPCODE_MAD] = nir_op_ffma,
+   [OPCODE_MAX] = nir_op_fmax,
+   [OPCODE_MIN] = nir_op_fmin,
+   [OPCODE_MOV] = nir_op_fmov,
+   [OPCODE_MUL] = nir_op_fmul,
+   [OPCODE_POW] = nir_op_fpow,
+   [OPCODE_RCP] = nir_op_frcp,
+
+   [OPCODE_RSQ] = nir_op_frsq,
+   [OPCODE_SCS] = 0,
+   [OPCODE_SEQ] = 0,
+   [OPCODE_SGE] = 0,
+   [OPCODE_SGT] = 0,
+   [OPCODE_SIN] = nir_op_fsin,
+   [OPCODE_SLE] = 0,
+   [OPCODE_SLT] = 0,
+   [OPCODE_SNE] = 0,
+   [OPCODE_SSG] = nir_op_fsign,
+   [OPCODE_SUB] = nir_op_fsub,
+   [OPCODE_SWZ] = 0,
+   [OPCODE_TEX] = 0,
+   [OPCODE_TRUNC] = nir_op_ftrunc,
+   [OPCODE_TXB] = 0,
+   [OPCODE_TXD] = 0,
+   [OPCODE_TXL] = 0,
+   [OPCODE_TXP] = 0,
+   [OPCODE_TXP_NV] = 0,
+   [OPCODE_XPD] = 0,
+};
+
+static void
+ptn_emit_instruction(struct ptn_compile *c, struct prog_instruction *prog_inst)
+{
+   nir_builder *b = &c->build;
+   unsigned i;
+   const unsigned op = prog_inst->Opcode;
+
+   if (op == OPCODE_END)
+      return;
+
+   nir_ssa_def *src[3];
+   for (i = 0; i < 3; i++) {
+      src[i] = ptn_get_src(c, &prog_inst->SrcReg[i]);
+   }
+   nir_alu_dest dest = ptn_get_dest(c, &prog_inst->DstReg);
+   if (c->error)
+      return;
+
+   switch (op) {
+   case OPCODE_RSQ:
+      ptn_move_dest(b, dest, nir_frsq(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_RCP:
+      ptn_move_dest(b, dest, nir_frcp(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_EX2:
+      ptn_move_dest(b, dest, nir_fexp2(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_LG2:
+      ptn_move_dest(b, dest, nir_flog2(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_POW:
+      ptn_move_dest(b, dest, nir_fpow(b,
+                                      ptn_channel(b, src[0], X),
+                                      ptn_channel(b, src[1], X)));
+      break;
+
+   case OPCODE_COS:
+      ptn_move_dest(b, dest, nir_fcos(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_SIN:
+      ptn_move_dest(b, dest, nir_fsin(b, ptn_channel(b, src[0], X)));
+      break;
+
+   case OPCODE_ARL:
+      ptn_arl(b, dest, src);
+      break;
+
+   case OPCODE_EXP:
+      ptn_exp(b, dest, src);
+      break;
+
+   case OPCODE_LOG:
+      ptn_log(b, dest, src);
+      break;
+
+   case OPCODE_LRP:
+      ptn_lrp(b, dest, src);
+      break;
+
+   case OPCODE_DST:
+      ptn_dst(b, dest, src);
+      break;
+
+   case OPCODE_LIT:
+      ptn_lit(b, dest, src);
+      break;
+
+   case OPCODE_XPD:
+      ptn_xpd(b, dest, src);
+      break;
+
+   case OPCODE_DP2:
+      ptn_dp2(b, dest, src);
+      break;
+
+   case OPCODE_DP3:
+      ptn_dp3(b, dest, src);
+      break;
+
+   case OPCODE_DP4:
+      ptn_dp4(b, dest, src);
+      break;
+
+   case OPCODE_DPH:
+      ptn_dph(b, dest, src);
+      break;
+
+   case OPCODE_KIL:
+      ptn_kil(b, dest, src);
+      break;
+
+   case OPCODE_CMP:
+      ptn_cmp(b, dest, src);
+      break;
+
+   case OPCODE_SCS:
+      ptn_scs(b, dest, src);
+      break;
+
+   case OPCODE_SLT:
+      ptn_slt(b, dest, src);
+      break;
+
+   case OPCODE_SGT:
+      ptn_sgt(b, dest, src);
+      break;
+
+   case OPCODE_SLE:
+      ptn_sle(b, dest, src);
+      break;
+
+   case OPCODE_SGE:
+      ptn_sge(b, dest, src);
+      break;
+
+   case OPCODE_SEQ:
+      ptn_seq(b, dest, src);
+      break;
+
+   case OPCODE_SNE:
+      ptn_sne(b, dest, src);
+      break;
+
+   case OPCODE_TEX:
+   case OPCODE_TXB:
+   case OPCODE_TXD:
+   case OPCODE_TXL:
+   case OPCODE_TXP:
+   case OPCODE_TXP_NV:
+      ptn_tex(b, dest, src, prog_inst);
+      break;
+
+   case OPCODE_SWZ:
+      /* Extended swizzles were already handled in ptn_get_src(). */
+      ptn_alu(b, nir_op_fmov, dest, src);
+      break;
+
+   case OPCODE_NOP:
+      break;
+
+   default:
+      if (op_trans[op] != 0 || op == OPCODE_MOV) {
+         ptn_alu(b, op_trans[op], dest, src);
+      } else {
+         fprintf(stderr, "unknown opcode: %s\n", _mesa_opcode_string(op));
+         abort();
+      }
+      break;
+   }
+
+   if (prog_inst->SaturateMode) {
+      assert(prog_inst->SaturateMode == SATURATE_ZERO_ONE);
+      assert(!dest.dest.is_ssa);
+      ptn_move_dest(b, dest, nir_fsat(b, ptn_src_for_dest(c, &dest)));
+   }
+}
+
+/**
+ * Puts a NIR intrinsic to store of each PROGRAM_OUTPUT value to the output
+ * variables at the end of the shader.
+ *
+ * We don't generate these incrementally as the PROGRAM_OUTPUT values are
+ * written, because there's no output load intrinsic, which means we couldn't
+ * handle writemasks.
+ */
+static void
+ptn_add_output_stores(struct ptn_compile *c)
+{
+   nir_builder *b = &c->build;
+
+   foreach_list_typed(nir_variable, var, node, &b->shader->outputs) {
+      nir_intrinsic_instr *store =
+         nir_intrinsic_instr_create(b->shader, nir_intrinsic_store_var);
+      store->num_components = 4;
+      store->variables[0] =
+         nir_deref_var_create(store, c->output_vars[var->data.location]);
+      store->src[0].reg.reg = c->output_regs[var->data.location];
+      nir_instr_insert_after_cf_list(c->build.cf_node_list, &store->instr);
+   }
+}
+
+static void
+setup_registers_and_variables(struct ptn_compile *c)
+{
+   nir_builder *b = &c->build;
+   struct nir_shader *shader = b->shader;
+
+   /* Create input variables. */
+   const int num_inputs = _mesa_flsll(c->prog->InputsRead);
+   for (int i = 0; i < num_inputs; i++) {
+      if (!(c->prog->InputsRead & BITFIELD64_BIT(i)))
+         continue;
+      nir_variable *var = rzalloc(shader, nir_variable);
+      var->type = glsl_vec4_type();
+      var->data.read_only = true;
+      var->data.mode = nir_var_shader_in;
+      var->name = ralloc_asprintf(var, "in_%d", i);
+      var->data.location = i;
+      var->data.index = 0;
+
+      if (c->prog->Target == GL_FRAGMENT_PROGRAM_ARB) {
+         struct gl_fragment_program *fp =
+            (struct gl_fragment_program *) c->prog;
+
+         var->data.interpolation = fp->InterpQualifier[i];
+
+         if (i == VARYING_SLOT_POS) {
+            var->data.origin_upper_left = fp->OriginUpperLeft;
+            var->data.pixel_center_integer = fp->PixelCenterInteger;
+         } else if (i == VARYING_SLOT_FOGC) {
+            /* fogcoord is defined as <f, 0.0, 0.0, 1.0>.  Make the actual
+             * input variable a float, and create a local containing the
+             * full vec4 value.
+             */
+            var->type = glsl_float_type();
+
+            nir_intrinsic_instr *load_x =
+               nir_intrinsic_instr_create(shader, nir_intrinsic_load_var);
+            load_x->num_components = 1;
+            load_x->variables[0] = nir_deref_var_create(load_x, var);
+            nir_ssa_dest_init(&load_x->instr, &load_x->dest, 1, NULL);
+            nir_instr_insert_after_cf_list(b->cf_node_list, &load_x->instr);
+
+            nir_ssa_def *f001 = nir_vec4(b, &load_x->dest.ssa, nir_imm_float(b, 0.0),
+                                         nir_imm_float(b, 0.0), nir_imm_float(b, 1.0));
+
+            nir_variable *fullvar = rzalloc(shader, nir_variable);
+            fullvar->type = glsl_vec4_type();
+            fullvar->data.mode = nir_var_local;
+            fullvar->name = "fogcoord_tmp";
+            exec_list_push_tail(&b->impl->locals, &fullvar->node);
+
+            nir_intrinsic_instr *store =
+               nir_intrinsic_instr_create(shader, nir_intrinsic_store_var);
+            store->num_components = 4;
+            store->variables[0] = nir_deref_var_create(store, fullvar);
+            store->src[0] = nir_src_for_ssa(f001);
+            nir_instr_insert_after_cf_list(b->cf_node_list, &store->instr);
+
+            /* Insert the real input into the list so the driver has real
+             * inputs, but set c->input_vars[i] to the temporary so we use
+             * the splatted value.
+             */
+            exec_list_push_tail(&shader->inputs, &var->node);
+            c->input_vars[i] = fullvar;
+            continue;
+         }
+      }
+
+      exec_list_push_tail(&shader->inputs, &var->node);
+      c->input_vars[i] = var;
+   }
+
+   /* Create output registers and variables. */
+   int max_outputs = _mesa_fls(c->prog->OutputsWritten);
+   c->output_regs = rzalloc_array(c, nir_register *, max_outputs);
+
+   for (int i = 0; i < max_outputs; i++) {
+      if (!(c->prog->OutputsWritten & BITFIELD64_BIT(i)))
+         continue;
+
+      /* Since we can't load from outputs in the IR, we make temporaries
+       * for the outputs and emit stores to the real outputs at the end of
+       * the shader.
+       */
+      nir_register *reg = nir_local_reg_create(b->impl);
+      reg->num_components = 4;
+
+      nir_variable *var = rzalloc(shader, nir_variable);
+      var->type = glsl_vec4_type();
+      var->data.mode = nir_var_shader_out;
+      var->name = ralloc_asprintf(var, "out_%d", i);
+
+      var->data.location = i;
+      var->data.index = 0;
+
+      c->output_regs[i] = reg;
+
+      exec_list_push_tail(&shader->outputs, &var->node);
+      c->output_vars[i] = var;
+   }
+
+   /* Create temporary registers. */
+   c->temp_regs = rzalloc_array(c, nir_register *, c->prog->NumTemporaries);
+
+   nir_register *reg;
+   for (int i = 0; i < c->prog->NumTemporaries; i++) {
+      reg = nir_local_reg_create(b->impl);
+      if (!reg) {
+         c->error = true;
+         return;
+      }
+      reg->num_components = 4;
+      c->temp_regs[i] = reg;
+   }
+
+   /* Create the address register (for ARB_vertex_program). */
+   reg = nir_local_reg_create(b->impl);
+   if (!reg) {
+      c->error = true;
+      return;
+   }
+   reg->num_components = 1;
+   c->addr_reg = reg;
+
+   /* Set the number of uniforms */
+   shader->num_uniforms = 4 * c->prog->Parameters->NumParameters;
+}
+
+struct nir_shader *
+prog_to_nir(const struct gl_program *prog, const nir_shader_compiler_options *options)
+{
+   struct ptn_compile *c;
+   struct nir_shader *s;
+
+   c = rzalloc(NULL, struct ptn_compile);
+   if (!c)
+      return NULL;
+   s = nir_shader_create(NULL, options);
+   if (!s)
+      goto fail;
+   c->prog = prog;
+
+   nir_function *func = nir_function_create(s, "main");
+   nir_function_overload *overload = nir_function_overload_create(func);
+   nir_function_impl *impl = nir_function_impl_create(overload);
+
+   c->build.shader = s;
+   c->build.impl = impl;
+   c->build.cf_node_list = &impl->body;
+
+   setup_registers_and_variables(c);
+   if (unlikely(c->error))
+      goto fail;
+
+   for (unsigned int i = 0; i < prog->NumInstructions; i++) {
+      ptn_emit_instruction(c, &prog->Instructions[i]);
+
+      if (unlikely(c->error))
+         break;
+   }
+
+   ptn_add_output_stores(c);
+
+fail:
+   if (c->error) {
+      ralloc_free(s);
+      s = NULL;
+   }
+   ralloc_free(c);
+   return s;
+}
diff --git a/mesalib/src/mesa/program/prog_to_nir.h b/mesalib/src/mesa/program/prog_to_nir.h
new file mode 100644
index 000000000..34e4cd104
--- /dev/null
+++ b/mesalib/src/mesa/program/prog_to_nir.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#pragma once
+#ifndef PROG_TO_NIR_H
+#define PROG_TO_NIR_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct nir_shader *prog_to_nir(const struct gl_program *prog,
+                               const nir_shader_compiler_options *options);
+
+#ifdef __cplusplus
+}
+#endif
+#endif
diff --git a/mesalib/src/mesa/program/program.c b/mesalib/src/mesa/program/program.c
index 3c214d5e3..4f28e2a3b 100644
--- a/mesalib/src/mesa/program/program.c
+++ b/mesalib/src/mesa/program/program.c
@@ -37,6 +37,7 @@
 #include "prog_cache.h"
 #include "prog_parameter.h"
 #include "prog_instruction.h"
+#include "util/ralloc.h"
 
 
 /**
@@ -380,6 +381,10 @@ _mesa_delete_program(struct gl_context *ctx, struct gl_program *prog)
       _mesa_free_parameter_list(prog->Parameters);
    }
 
+   if (prog->nir) {
+      ralloc_free(prog->nir);
+   }
+
    free(prog);
 }
 
diff --git a/mesalib/src/mesa/state_tracker/st_atom.c b/mesalib/src/mesa/state_tracker/st_atom.c
index f0fe11ffa..428f2d9d7 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.c
+++ b/mesalib/src/mesa/state_tracker/st_atom.c
@@ -183,7 +183,7 @@ void st_validate_state( struct st_context *st )
    if (state->st == 0)
       return;
 
-   /*printf("%s %x/%x\n", __FUNCTION__, state->mesa, state->st);*/
+   /*printf("%s %x/%x\n", __func__, state->mesa, state->st);*/
 
 #ifdef DEBUG
    if (1) {
diff --git a/mesalib/src/mesa/state_tracker/st_atom_array.c b/mesalib/src/mesa/state_tracker/st_atom_array.c
index 9b52f9779..d4fb8b862 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_array.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_array.c
@@ -598,7 +598,7 @@ static void update_array(struct st_context *st)
 const struct st_tracked_state st_update_array = {
    "st_update_array",					/* name */
    {							/* dirty */
-      0,                                                /* mesa */
+      _NEW_CURRENT_ATTRIB,                              /* mesa */
       ST_NEW_VERTEX_ARRAYS | ST_NEW_VERTEX_PROGRAM,     /* st */
    },
    update_array						/* update */
diff --git a/mesalib/src/mesa/state_tracker/st_atom_constbuf.c b/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
index 7984bf742..a54e0d9db 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_constbuf.c
@@ -92,7 +92,7 @@ void st_upload_constants( struct st_context *st,
 
       if (ST_DEBUG & DEBUG_CONSTANTS) {
          debug_printf("%s(shader=%d, numParams=%d, stateFlags=0x%x)\n",
-                      __FUNCTION__, shader_type, params->NumParameters,
+                      __func__, shader_type, params->NumParameters,
                       params->StateFlags);
          _mesa_print_parameter_list(params);
       }
diff --git a/mesalib/src/mesa/state_tracker/st_atom_shader.c b/mesalib/src/mesa/state_tracker/st_atom_shader.c
index 73768ed12..629f54f25 100644
--- a/mesalib/src/mesa/state_tracker/st_atom_shader.c
+++ b/mesalib/src/mesa/state_tracker/st_atom_shader.c
@@ -40,9 +40,8 @@
 #include "program/program.h"
 
 #include "pipe/p_context.h"
-
+#include "pipe/p_shader_tokens.h"
 #include "util/u_simple_shaders.h"
-
 #include "cso_cache/cso_context.h"
 
 #include "st_context.h"
diff --git a/mesalib/src/mesa/state_tracker/st_cb_clear.c b/mesalib/src/mesa/state_tracker/st_cb_clear.c
index dd81a6273..f10e9063a 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_clear.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_clear.c
@@ -247,7 +247,7 @@ clear_with_quad(struct gl_context *ctx, unsigned clear_buffers)
       util_framebuffer_get_num_layers(&st->state.framebuffer);
 
    /*
-   printf("%s %s%s%s %f,%f %f,%f\n", __FUNCTION__, 
+   printf("%s %s%s%s %f,%f %f,%f\n", __func__,
 	  color ? "color, " : "",
 	  depth ? "depth, " : "",
 	  stencil ? "stencil" : "",
diff --git a/mesalib/src/mesa/state_tracker/st_cb_program.c b/mesalib/src/mesa/state_tracker/st_cb_program.c
index aa301d830..c382d7d2c 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_program.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_program.c
@@ -41,6 +41,7 @@
 #include "draw/draw_context.h"
 
 #include "st_context.h"
+#include "st_debug.h"
 #include "st_program.h"
 #include "st_mesa_to_tgsi.h"
 #include "st_cb_program.h"
@@ -214,6 +215,9 @@ st_program_string_notify( struct gl_context *ctx,
 	 st->dirty.st |= ST_NEW_VERTEX_PROGRAM;
    }
 
+   if (ST_DEBUG & DEBUG_PRECOMPILE)
+      st_precompile_shader_variant(st, prog);
+
    /* XXX check if program is legal, within limits */
    return GL_TRUE;
 }
diff --git a/mesalib/src/mesa/state_tracker/st_cb_texture.c b/mesalib/src/mesa/state_tracker/st_cb_texture.c
index 5c520b44f..bdf236e82 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_texture.c
@@ -123,7 +123,7 @@ gl_target_to_pipe(GLenum target)
 static struct gl_texture_image *
 st_NewTextureImage(struct gl_context * ctx)
 {
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
    (void) ctx;
    return (struct gl_texture_image *) ST_CALLOC_STRUCT(st_texture_image);
 }
@@ -144,7 +144,7 @@ st_NewTextureObject(struct gl_context * ctx, GLuint name, GLenum target)
 {
    struct st_texture_object *obj = ST_CALLOC_STRUCT(st_texture_object);
 
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
    _mesa_initialize_texture_object(ctx, &obj->base, name, target);
 
    return &obj->base;
@@ -172,7 +172,7 @@ st_FreeTextureImageBuffer(struct gl_context *ctx,
 {
    struct st_texture_image *stImage = st_texture_image(texImage);
 
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
 
    if (stImage->pt) {
       pipe_resource_reference(&stImage->pt, NULL);
@@ -405,7 +405,7 @@ guess_and_alloc_texture(struct st_context *st,
    GLuint ptWidth, ptHeight, ptDepth, ptLayers;
    enum pipe_format fmt;
 
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
 
    assert(!stObj->pt);
 
@@ -473,7 +473,7 @@ guess_and_alloc_texture(struct st_context *st,
 
    stObj->lastLevel = lastLevel;
 
-   DBG("%s returning %d\n", __FUNCTION__, (stObj->pt != NULL));
+   DBG("%s returning %d\n", __func__, (stObj->pt != NULL));
 
    return stObj->pt != NULL;
 }
@@ -496,7 +496,7 @@ st_AllocTextureImageBuffer(struct gl_context *ctx,
    GLuint height = texImage->Height;
    GLuint depth = texImage->Depth;
 
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
 
    assert(!stImage->pt); /* xxx this might be wrong */
 
@@ -738,6 +738,11 @@ st_TexSubImage(struct gl_context *ctx, GLuint dims,
    if (gl_target == GL_TEXTURE_CUBE_MAP) {
       gl_target = GL_TEXTURE_2D;
    }
+   /* TexSubImage can specify subsets of cube map array faces
+    * so we need to upload via 2D array instead */
+   if (gl_target == GL_TEXTURE_CUBE_MAP_ARRAY) {
+      gl_target = GL_TEXTURE_2D_ARRAY;
+   }
 
    /* Initialize the source texture description. */
    memset(&src_templ, 0, sizeof(src_templ));
@@ -1148,7 +1153,7 @@ st_GetTexImage(struct gl_context * ctx,
       }
 
       if (ST_DEBUG & DEBUG_FALLBACK)
-         debug_printf("%s: fallback format translation\n", __FUNCTION__);
+         debug_printf("%s: fallback format translation\n", __func__);
 
       dstMesaFormat = _mesa_format_from_format_and_type(format, type);
       dstStride = _mesa_image_row_stride(&ctx->Pack, width, format, type);
@@ -1234,7 +1239,7 @@ fallback_copy_texsubimage(struct gl_context *ctx,
    struct pipe_transfer *transfer;
 
    if (ST_DEBUG & DEBUG_FALLBACK)
-      debug_printf("%s: fallback processing\n", __FUNCTION__);
+      debug_printf("%s: fallback processing\n", __func__);
 
    if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
       srcY = strb->Base.Height - srcY - height;
@@ -1389,7 +1394,7 @@ st_CopyTexSubImage(struct gl_context *ctx, GLuint dims,
           texImage->TexFormat != MESA_FORMAT_ETC1_RGB8);
 
    if (!strb || !strb->surface || !stImage->pt) {
-      debug_printf("%s: null strb or stImage\n", __FUNCTION__);
+      debug_printf("%s: null strb or stImage\n", __func__);
       return;
    }
 
diff --git a/mesalib/src/mesa/state_tracker/st_debug.c b/mesalib/src/mesa/state_tracker/st_debug.c
index de3e3a9cd..50891c112 100644
--- a/mesalib/src/mesa/state_tracker/st_debug.c
+++ b/mesalib/src/mesa/state_tracker/st_debug.c
@@ -56,6 +56,7 @@ static const struct debug_named_value st_debug_flags[] = {
    { "draw",     DEBUG_DRAW, NULL },
    { "buffer",   DEBUG_BUFFER, NULL },
    { "wf",       DEBUG_WIREFRAME, NULL },
+   { "precompile",  DEBUG_PRECOMPILE, NULL },
    DEBUG_NAMED_VALUE_END
 };
 
diff --git a/mesalib/src/mesa/state_tracker/st_debug.h b/mesalib/src/mesa/state_tracker/st_debug.h
index cc8197836..288eccf9f 100644
--- a/mesalib/src/mesa/state_tracker/st_debug.h
+++ b/mesalib/src/mesa/state_tracker/st_debug.h
@@ -47,6 +47,7 @@ st_print_current(void);
 #define DEBUG_DRAW      0x100
 #define DEBUG_BUFFER    0x200
 #define DEBUG_WIREFRAME 0x400
+#define DEBUG_PRECOMPILE   0x800
 
 #ifdef DEBUG
 extern int ST_DEBUG;
diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index efee4b258..93671ba9c 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -30,40 +30,25 @@
  * Translate GLSL IR to TGSI.
  */
 
-#include <stdio.h>
-#include "main/compiler.h"
-#include "ir.h"
-#include "ir_visitor.h"
-#include "ir_expression_flattening.h"
-#include "glsl_types.h"
+#include "st_glsl_to_tgsi.h"
+
 #include "glsl_parser_extras.h"
-#include "../glsl/program.h"
 #include "ir_optimization.h"
-#include "ast.h"
 
-#include "main/mtypes.h"
+#include "main/errors.h"
 #include "main/shaderobj.h"
 #include "main/uniforms.h"
 #include "main/shaderapi.h"
-#include "program/hash_table.h"
 #include "program/prog_instruction.h"
-#include "program/prog_optimize.h"
-#include "program/prog_print.h"
-#include "program/program.h"
-#include "program/prog_parameter.h"
 #include "program/sampler.h"
 
-#include "pipe/p_compiler.h"
 #include "pipe/p_context.h"
 #include "pipe/p_screen.h"
-#include "pipe/p_shader_tokens.h"
-#include "pipe/p_state.h"
-#include "util/u_math.h"
 #include "tgsi/tgsi_ureg.h"
 #include "tgsi/tgsi_info.h"
-#include "st_context.h"
+#include "util/u_math.h"
+#include "util/u_memory.h"
 #include "st_program.h"
-#include "st_glsl_to_tgsi.h"
 #include "st_mesa_to_tgsi.h"
 
 
@@ -328,6 +313,7 @@ public:
    int num_address_regs;
    int samplers_used;
    bool indirect_addr_consts;
+   int wpos_transform_const;
 
    int glsl_version;
    bool native_integers;
@@ -441,9 +427,6 @@ public:
 
    void emit_arl(ir_instruction *ir, st_dst_reg dst, st_src_reg src0);
 
-   void emit_scs(ir_instruction *ir, unsigned op,
-                 st_dst_reg dst, const st_src_reg &src);
-
    bool try_emit_mad(ir_expression *ir,
               int mul_operand);
    bool try_emit_mad_for_and_not(ir_expression *ir,
@@ -966,101 +949,6 @@ glsl_to_tgsi_visitor::emit_arl(ir_instruction *ir,
    emit(NULL, op, dst, src0);
 }
 
-/**
- * Emit an TGSI_OPCODE_SCS instruction
- *
- * The \c SCS opcode functions a bit differently than the other TGSI opcodes.
- * Instead of splatting its result across all four components of the
- * destination, it writes one value to the \c x component and another value to
- * the \c y component.
- *
- * \param ir        IR instruction being processed
- * \param op        Either \c TGSI_OPCODE_SIN or \c TGSI_OPCODE_COS depending
- *                  on which value is desired.
- * \param dst       Destination register
- * \param src       Source register
- */
-void
-glsl_to_tgsi_visitor::emit_scs(ir_instruction *ir, unsigned op,
-                               st_dst_reg dst,
-                               const st_src_reg &src)
-{
-   /* Vertex programs cannot use the SCS opcode.
-    */
-   if (this->prog->Target == GL_VERTEX_PROGRAM_ARB) {
-      emit_scalar(ir, op, dst, src);
-      return;
-   }
-
-   const unsigned component = (op == TGSI_OPCODE_SIN) ? 0 : 1;
-   const unsigned scs_mask = (1U << component);
-   int done_mask = ~dst.writemask;
-   st_src_reg tmp;
-
-   assert(op == TGSI_OPCODE_SIN || op == TGSI_OPCODE_COS);
-
-   /* If there are compnents in the destination that differ from the component
-    * that will be written by the SCS instrution, we'll need a temporary.
-    */
-   if (scs_mask != unsigned(dst.writemask)) {
-      tmp = get_temp(glsl_type::vec4_type);
-   }
-
-   for (unsigned i = 0; i < 4; i++) {
-      unsigned this_mask = (1U << i);
-      st_src_reg src0 = src;
-
-      if ((done_mask & this_mask) != 0)
-         continue;
-
-      /* The source swizzle specified which component of the source generates
-       * sine / cosine for the current component in the destination.  The SCS
-       * instruction requires that this value be swizzle to the X component.
-       * Replace the current swizzle with a swizzle that puts the source in
-       * the X component.
-       */
-      unsigned src0_swiz = GET_SWZ(src.swizzle, i);
-
-      src0.swizzle = MAKE_SWIZZLE4(src0_swiz, src0_swiz,
-                                   src0_swiz, src0_swiz);
-      for (unsigned j = i + 1; j < 4; j++) {
-         /* If there is another enabled component in the destination that is
-          * derived from the same inputs, generate its value on this pass as
-          * well.
-          */
-         if (!(done_mask & (1 << j)) &&
-             GET_SWZ(src0.swizzle, j) == src0_swiz) {
-            this_mask |= (1 << j);
-         }
-      }
-
-      if (this_mask != scs_mask) {
-         glsl_to_tgsi_instruction *inst;
-         st_dst_reg tmp_dst = st_dst_reg(tmp);
-
-         /* Emit the SCS instruction.
-          */
-         inst = emit(ir, TGSI_OPCODE_SCS, tmp_dst, src0);
-         inst->dst[0].writemask = scs_mask;
-
-         /* Move the result of the SCS instruction to the desired location in
-          * the destination.
-          */
-         tmp.swizzle = MAKE_SWIZZLE4(component, component,
-                                     component, component);
-         inst = emit(ir, TGSI_OPCODE_SCS, dst, tmp);
-         inst->dst[0].writemask = this_mask;
-      } else {
-         /* Emit the SCS instruction to write directly to the destination.
-          */
-         glsl_to_tgsi_instruction *inst = emit(ir, TGSI_OPCODE_SCS, dst, src0);
-         inst->dst[0].writemask = scs_mask;
-      }
-
-      done_mask |= this_mask;
-   }
-}
-
 int
 glsl_to_tgsi_visitor::add_constant(gl_register_file file,
                                    gl_constant_value values[8], int size, int datatype,
@@ -1611,12 +1499,6 @@ glsl_to_tgsi_visitor::visit(ir_expression *ir)
    case ir_unop_cos:
       emit_scalar(ir, TGSI_OPCODE_COS, result_dst, op[0]);
       break;
-   case ir_unop_sin_reduced:
-      emit_scs(ir, TGSI_OPCODE_SIN, result_dst, op[0]);
-      break;
-   case ir_unop_cos_reduced:
-      emit_scs(ir, TGSI_OPCODE_COS, result_dst, op[0]);
-      break;
    case ir_unop_saturate: {
       glsl_to_tgsi_instruction *inst;
       inst = emit(ir, TGSI_OPCODE_MOV, result_dst, op[0]);
@@ -3134,7 +3016,7 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
       break;
    case ir_query_levels:
       opcode = TGSI_OPCODE_TXQ;
-      lod_info = st_src_reg(PROGRAM_IMMEDIATE, 0, GLSL_TYPE_INT);
+      lod_info = undef_src;
       levels_src = get_temp(ir->type);
       break;
    case ir_txf:
@@ -3456,6 +3338,7 @@ glsl_to_tgsi_visitor::glsl_to_tgsi_visitor()
    num_address_regs = 0;
    samplers_used = 0;
    indirect_addr_consts = false;
+   wpos_transform_const = -1;
    glsl_version = 0;
    native_integers = false;
    mem_ctx = ralloc_context(NULL);
@@ -3465,6 +3348,7 @@ glsl_to_tgsi_visitor::glsl_to_tgsi_visitor()
    shader = NULL;
    options = NULL;
    have_sqrt = false;
+   have_fma = false;
 }
 
 glsl_to_tgsi_visitor::~glsl_to_tgsi_visitor()
@@ -3937,6 +3821,7 @@ glsl_to_tgsi_visitor::copy_propagate(void)
              inst->dst[0].index == inst->src[0].index) &&
           !inst->dst[0].reladdr &&
           !inst->saturate &&
+          inst->src[0].file != PROGRAM_ARRAY &&
           !inst->src[0].reladdr &&
           !inst->src[0].reladdr2 &&
           !inst->src[0].negate) {
@@ -4464,7 +4349,9 @@ struct st_translate {
 
    struct ureg_dst arrays[MAX_ARRAYS];
    struct ureg_src *constants;
+   int num_constants;
    struct ureg_src *immediates;
+   int num_immediates;
    struct ureg_dst outputs[PIPE_MAX_SHADER_OUTPUTS];
    struct ureg_src inputs[PIPE_MAX_SHADER_INPUTS];
    struct ureg_dst address[3];
@@ -4666,7 +4553,7 @@ src_register(struct st_translate *t, const st_src_reg *reg)
 {
    switch(reg->file) {
    case PROGRAM_UNDEFINED:
-      return ureg_src_undef();
+      return ureg_imm4f(t->ureg, 0, 0, 0, 0);
 
    case PROGRAM_TEMPORARY:
    case PROGRAM_ARRAY:
@@ -4674,17 +4561,18 @@ src_register(struct st_translate *t, const st_src_reg *reg)
 
    case PROGRAM_UNIFORM:
       assert(reg->index >= 0);
-      return t->constants[reg->index];
+      return reg->index < t->num_constants ?
+               t->constants[reg->index] : ureg_imm4f(t->ureg, 0, 0, 0, 0);
    case PROGRAM_STATE_VAR:
    case PROGRAM_CONSTANT:       /* ie, immediate */
       if (reg->has_index2)
          return ureg_src_register(TGSI_FILE_CONSTANT, reg->index);
-      else if (reg->index < 0)
-         return ureg_DECL_constant(t->ureg, 0);
       else
-         return t->constants[reg->index];
+         return reg->index >= 0 && reg->index < t->num_constants ?
+                  t->constants[reg->index] : ureg_imm4f(t->ureg, 0, 0, 0, 0);
 
    case PROGRAM_IMMEDIATE:
+      assert(reg->index >= 0 && reg->index < t->num_immediates);
       return t->immediates[reg->index];
 
    case PROGRAM_INPUT:
@@ -4805,6 +4693,7 @@ translate_tex_offset(struct st_translate *t,
 
    switch (in_offset->file) {
    case PROGRAM_IMMEDIATE:
+      assert(in_offset->index >= 0 && in_offset->index < t->num_immediates);
       imm_src = t->immediates[in_offset->index];
 
       offset.File = imm_src.File;
@@ -4867,10 +4756,8 @@ compile_tgsi_instruction(struct st_translate *t,
                              inst->saturate,
                              clamp_dst_color_output);
 
-   for (i = 0; i < num_src; i++) {
-      assert(inst->src[i].file != PROGRAM_UNDEFINED);
+   for (i = 0; i < num_src; i++)
       src[i] = translate_src(t, &inst->src[i]);
-   }
 
    switch(inst->op) {
    case TGSI_OPCODE_BGNLOOP:
@@ -4939,28 +4826,19 @@ compile_tgsi_instruction(struct st_translate *t,
  */
 static void
 emit_wpos_adjustment( struct st_translate *t,
-                      const struct gl_program *program,
+                      int wpos_transform_const,
                       boolean invert,
                       GLfloat adjX, GLfloat adjY[2])
 {
    struct ureg_program *ureg = t->ureg;
 
+   assert(wpos_transform_const >= 0);
+
    /* Fragment program uses fragment position input.
     * Need to replace instances of INPUT[WPOS] with temp T
-    * where T = INPUT[WPOS] by y is inverted.
+    * where T = INPUT[WPOS] is inverted by Y.
     */
-   static const gl_state_index wposTransformState[STATE_LENGTH]
-      = { STATE_INTERNAL, STATE_FB_WPOS_Y_TRANSFORM,
-          (gl_state_index)0, (gl_state_index)0, (gl_state_index)0 };
-
-   /* XXX: note we are modifying the incoming shader here!  Need to
-    * do this before emitting the constant decls below, or this
-    * will be missed:
-    */
-   unsigned wposTransConst = _mesa_add_state_reference(program->Parameters,
-                                                       wposTransformState);
-
-   struct ureg_src wpostrans = ureg_DECL_constant( ureg, wposTransConst );
+   struct ureg_src wpostrans = ureg_DECL_constant(ureg, wpos_transform_const);
    struct ureg_dst wpos_temp = ureg_DECL_temporary( ureg );
    struct ureg_src wpos_input = t->inputs[t->inputMapping[VARYING_SLOT_POS]];
 
@@ -5024,7 +4902,8 @@ static void
 emit_wpos(struct st_context *st,
           struct st_translate *t,
           const struct gl_program *program,
-          struct ureg_program *ureg)
+          struct ureg_program *ureg,
+          int wpos_transform_const)
 {
    const struct gl_fragment_program *fp =
       (const struct gl_fragment_program *) program;
@@ -5121,7 +5000,7 @@ emit_wpos(struct st_context *st,
 
    /* we invert after adjustment so that we avoid the MOV to temporary,
     * and reuse the adjustment ADD instead */
-   emit_wpos_adjustment(t, program, invert, adjX, adjY);
+   emit_wpos_adjustment(t, wpos_transform_const, invert, adjX, adjY);
 }
 
 /**
@@ -5238,15 +5117,6 @@ st_translate_program(
    t->outputMapping = outputMapping;
    t->ureg = ureg;
 
-   if (program->shader_program) {
-      for (i = 0; i < program->shader_program->NumUserUniformStorage; i++) {
-         struct gl_uniform_storage *const storage =
-               &program->shader_program->UniformStorage[i];
-
-         _mesa_uniform_detach_all_driver_storage(storage);
-      }
-   }
-
    /*
     * Declare input attributes.
     */
@@ -5260,10 +5130,9 @@ st_translate_program(
       }
 
       if (proginfo->InputsRead & VARYING_BIT_POS) {
-         /* Must do this after setting up t->inputs, and before
-          * emitting constant references, below:
-          */
-          emit_wpos(st_context(ctx), t, proginfo, ureg);
+          /* Must do this after setting up t->inputs. */
+          emit_wpos(st_context(ctx), t, proginfo, ureg,
+                    program->wpos_transform_const);
       }
 
       if (proginfo->InputsRead & VARYING_BIT_FACE)
@@ -5400,6 +5269,7 @@ st_translate_program(
          ret = PIPE_ERROR_OUT_OF_MEMORY;
          goto out;
       }
+      t->num_constants = proginfo->Parameters->NumParameters;
 
       for (i = 0; i < proginfo->Parameters->NumParameters; i++) {
          switch (proginfo->Parameters->Parameters[i].Type) {
@@ -5451,6 +5321,8 @@ st_translate_program(
       ret = PIPE_ERROR_OUT_OF_MEMORY;
       goto out;
    }
+   t->num_immediates = program->num_immediates;
+
    i = 0;
    foreach_in_list(immediate_storage, imm, &program->immediates) {
       assert(i < program->num_immediates);
@@ -5479,33 +5351,21 @@ st_translate_program(
                        t->insn[t->labels[i].branch_target]);
    }
 
-   if (program->shader_program) {
-      /* This has to be done last.  Any operation the can cause
-       * prog->ParameterValues to get reallocated (e.g., anything that adds a
-       * program constant) has to happen before creating this linkage.
-       */
-      for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
-         if (program->shader_program->_LinkedShaders[i] == NULL)
-            continue;
-
-         _mesa_associate_uniform_storage(ctx, program->shader_program,
-               program->shader_program->_LinkedShaders[i]->Program->Parameters);
-      }
-   }
-
 out:
    if (t) {
       free(t->temps);
       free(t->insn);
       free(t->labels);
       free(t->constants);
+      t->num_constants = 0;
       free(t->immediates);
+      t->num_immediates = 0;
 
       if (t->error) {
-         debug_printf("%s: translate error flag set\n", __FUNCTION__);
+         debug_printf("%s: translate error flag set\n", __func__);
       }
 
-      free(t);
+      FREE(t);
    }
 
    return ret;
@@ -5634,14 +5494,12 @@ get_mesa_program(struct gl_context *ctx,
    v->emit(NULL, TGSI_OPCODE_END);
 
    if (ctx->_Shader->Flags & GLSL_DUMP) {
-      printf("\n");
-      printf("GLSL IR for linked %s program %d:\n",
+      _mesa_log("\n");
+      _mesa_log("GLSL IR for linked %s program %d:\n",
              _mesa_shader_stage_to_string(shader->Stage),
              shader_program->Name);
-      _mesa_print_ir(stdout, shader->ir, NULL);
-      printf("\n");
-      printf("\n");
-      fflush(stdout);
+      _mesa_print_ir(_mesa_get_log_file(), shader->ir, NULL);
+      _mesa_log("\n\n");
    }
 
    prog->Instructions = NULL;
@@ -5650,6 +5508,17 @@ get_mesa_program(struct gl_context *ctx,
    do_set_program_inouts(shader->ir, prog, shader->Stage);
    count_resources(v, prog);
 
+   /* This must be done before the uniform storage is associated. */
+   if (shader->Type == GL_FRAGMENT_SHADER &&
+       prog->InputsRead & VARYING_BIT_POS){
+      static const gl_state_index wposTransformState[STATE_LENGTH] = {
+         STATE_INTERNAL, STATE_FB_WPOS_Y_TRANSFORM
+      };
+
+      v->wpos_transform_const = _mesa_add_state_reference(prog->Parameters,
+                                                          wposTransformState);
+   }
+
    _mesa_reference_program(ctx, &shader->Program, prog);
 
    /* This has to be done last.  Any operation the can cause
diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.h b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.h
index 5ed640747..2cb80bcf9 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.h
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.h
@@ -22,17 +22,18 @@
  * DEALINGS IN THE SOFTWARE.
  */
 
+#include "pipe/p_defines.h"
+#include "main/mtypes.h"
+
 #ifdef __cplusplus
 extern "C" {
 #endif
 
-#include "main/glheader.h"
-#include "tgsi/tgsi_ureg.h"
-
 struct gl_context;
 struct gl_shader;
 struct gl_shader_program;
 struct glsl_to_tgsi_visitor;
+struct ureg_program;
 
 enum pipe_error st_translate_program(
    struct gl_context *ctx,
diff --git a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
index 2f1016110..98d525c86 100644
--- a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
+++ b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
@@ -1249,7 +1249,7 @@ out:
    free(t->constants);
 
    if (t->error) {
-      debug_printf("%s: translate error flag set\n", __FUNCTION__);
+      debug_printf("%s: translate error flag set\n", __func__);
    }
 
    return ret;
diff --git a/mesalib/src/mesa/state_tracker/st_program.c b/mesalib/src/mesa/state_tracker/st_program.c
index 4cfd817ce..d93b3c7bc 100644
--- a/mesalib/src/mesa/state_tracker/st_program.c
+++ b/mesalib/src/mesa/state_tracker/st_program.c
@@ -185,9 +185,6 @@ st_prepare_vertex_program(struct gl_context *ctx,
    if (stvp->Base.IsPositionInvariant)
       _mesa_insert_mvp_code(ctx, &stvp->Base);
 
-   if (!stvp->glsl_to_tgsi)
-      assert(stvp->Base.Base.NumInstructions > 1);
-
    /*
     * Determine number of inputs, the mappings between VERT_ATTRIB_x
     * and TGSI generic input indexes, plus input attrib semantic info.
@@ -403,7 +400,7 @@ st_translate_vertex_program(struct st_context *st,
    return vpv;
 
 fail:
-   debug_printf("%s: failed to translate Mesa program:\n", __FUNCTION__);
+   debug_printf("%s: failed to translate Mesa program:\n", __func__);
    _mesa_print_program(&stvp->Base.Base);
    debug_assert(0);
 
@@ -1318,3 +1315,47 @@ st_print_current_vertex_program(void)
       }
    }
 }
+
+
+/**
+ * Compile one shader variant.
+ */
+void
+st_precompile_shader_variant(struct st_context *st,
+                             struct gl_program *prog)
+{
+   switch (prog->Target) {
+   case GL_VERTEX_PROGRAM_ARB: {
+      struct st_vertex_program *p = (struct st_vertex_program *)prog;
+      struct st_vp_variant_key key;
+
+      memset(&key, 0, sizeof(key));
+      key.st = st;
+      st_get_vp_variant(st, p, &key);
+      break;
+   }
+
+   case GL_GEOMETRY_PROGRAM_NV: {
+      struct st_geometry_program *p = (struct st_geometry_program *)prog;
+      struct st_gp_variant_key key;
+
+      memset(&key, 0, sizeof(key));
+      key.st = st;
+      st_get_gp_variant(st, p, &key);
+      break;
+   }
+
+   case GL_FRAGMENT_PROGRAM_ARB: {
+      struct st_fragment_program *p = (struct st_fragment_program *)prog;
+      struct st_fp_variant_key key;
+
+      memset(&key, 0, sizeof(key));
+      key.st = st;
+      st_get_fp_variant(st, p, &key);
+      break;
+   }
+
+   default:
+      assert(0);
+   }
+}
diff --git a/mesalib/src/mesa/state_tracker/st_program.h b/mesalib/src/mesa/state_tracker/st_program.h
index 451d7bb6a..b2c86faec 100644
--- a/mesalib/src/mesa/state_tracker/st_program.h
+++ b/mesalib/src/mesa/state_tracker/st_program.h
@@ -329,6 +329,9 @@ st_destroy_program_variants(struct st_context *st);
 extern void
 st_print_current_vertex_program(void);
 
+extern void
+st_precompile_shader_variant(struct st_context *st,
+                             struct gl_program *prog);
 
 #ifdef __cplusplus
 }
diff --git a/mesalib/src/mesa/state_tracker/st_texture.c b/mesalib/src/mesa/state_tracker/st_texture.c
index ca7c83c21..6beb21e33 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_texture.c
@@ -74,7 +74,7 @@ st_texture_create(struct st_context *st,
    if (target == PIPE_TEXTURE_CUBE)
       assert(layers == 6);
 
-   DBG("%s target %d format %s last_level %d\n", __FUNCTION__,
+   DBG("%s target %d format %s last_level %d\n", __func__,
        (int) target, util_format_name(format), last_level);
 
    assert(format);
@@ -177,7 +177,7 @@ st_gl_texture_dims_to_pipe_dims(GLenum texture,
       *widthOut = widthIn;
       *heightOut = heightIn;
       *depthOut = 1;
-      *layersOut = depthIn;
+      *layersOut = util_align_npot(depthIn, 6);
       break;
    default:
       assert(0 && "Unexpected texture in st_gl_texture_dims_to_pipe_dims()");
@@ -250,7 +250,7 @@ st_texture_image_map(struct st_context *st, struct st_texture_image *stImage,
    GLuint level;
    void *map;
 
-   DBG("%s \n", __FUNCTION__);
+   DBG("%s \n", __func__);
 
    if (!stImage->pt)
       return NULL;
@@ -304,7 +304,7 @@ st_texture_image_unmap(struct st_context *st,
       slice += stObj->base.MinLayer;
    transfer = &stImage->transfer[slice + stImage->base.Face].transfer;
 
-   DBG("%s\n", __FUNCTION__);
+   DBG("%s\n", __func__);
 
    pipe_transfer_unmap(pipe, *transfer);
    *transfer = NULL;
diff --git a/mesalib/src/mesa/swrast/s_aatriangle.c b/mesalib/src/mesa/swrast/s_aatriangle.c
index 1d076cc7d..b51098704 100644
--- a/mesalib/src/mesa/swrast/s_aatriangle.c
+++ b/mesalib/src/mesa/swrast/s_aatriangle.c
@@ -30,7 +30,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/state.h"
diff --git a/mesalib/src/mesa/swrast/s_alpha.c b/mesalib/src/mesa/swrast/s_alpha.c
index b1a7ff132..841642fa9 100644
--- a/mesalib/src/mesa/swrast/s_alpha.c
+++ b/mesalib/src/mesa/swrast/s_alpha.c
@@ -29,7 +29,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 
 #include "s_alpha.h"
diff --git a/mesalib/src/mesa/swrast/s_atifragshader.c b/mesalib/src/mesa/swrast/s_atifragshader.c
index 0bf03771f..9e029db25 100644
--- a/mesalib/src/mesa/swrast/s_atifragshader.c
+++ b/mesalib/src/mesa/swrast/s_atifragshader.c
@@ -20,7 +20,6 @@
  */
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/atifragshader.h"
 #include "main/samplerobj.h"
diff --git a/mesalib/src/mesa/swrast/s_context.c b/mesalib/src/mesa/swrast/s_context.c
index ecde292e3..af24207e5 100644
--- a/mesalib/src/mesa/swrast/s_context.c
+++ b/mesalib/src/mesa/swrast/s_context.c
@@ -27,7 +27,6 @@
 
 #include "main/imports.h"
 #include "main/bufferobj.h"
-#include "main/colormac.h"
 #include "main/mtypes.h"
 #include "main/samplerobj.h"
 #include "main/teximage.h"
diff --git a/mesalib/src/mesa/swrast/s_copypix.c b/mesalib/src/mesa/swrast/s_copypix.c
index 17140ad2d..68c83e44e 100644
--- a/mesalib/src/mesa/swrast/s_copypix.c
+++ b/mesalib/src/mesa/swrast/s_copypix.c
@@ -25,7 +25,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/condrender.h"
 #include "main/macros.h"
 #include "main/pixeltransfer.h"
diff --git a/mesalib/src/mesa/swrast/s_feedback.c b/mesalib/src/mesa/swrast/s_feedback.c
index f25b89735..71f48ce92 100644
--- a/mesalib/src/mesa/swrast/s_feedback.c
+++ b/mesalib/src/mesa/swrast/s_feedback.c
@@ -23,7 +23,6 @@
  */
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/feedback.h"
 #include "main/macros.h"
 
diff --git a/mesalib/src/mesa/swrast/s_fog.c b/mesalib/src/mesa/swrast/s_fog.c
index e270b7e24..8b0bdf8fa 100644
--- a/mesalib/src/mesa/swrast/s_fog.c
+++ b/mesalib/src/mesa/swrast/s_fog.c
@@ -25,7 +25,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 
 #include "s_context.h"
diff --git a/mesalib/src/mesa/swrast/s_fragprog.c b/mesalib/src/mesa/swrast/s_fragprog.c
index 12bcda311..175915a5a 100644
--- a/mesalib/src/mesa/swrast/s_fragprog.c
+++ b/mesalib/src/mesa/swrast/s_fragprog.c
@@ -23,7 +23,7 @@
  */
 
 #include "main/glheader.h"
-#include "main/colormac.h"
+#include "main/macros.h"
 #include "main/samplerobj.h"
 #include "main/teximage.h"
 #include "program/prog_instruction.h"
diff --git a/mesalib/src/mesa/swrast/s_lines.c b/mesalib/src/mesa/swrast/s_lines.c
index 3e626b9e0..58bd2fc72 100644
--- a/mesalib/src/mesa/swrast/s_lines.c
+++ b/mesalib/src/mesa/swrast/s_lines.c
@@ -25,7 +25,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "s_aaline.h"
 #include "s_context.h"
diff --git a/mesalib/src/mesa/swrast/s_linetemp.h b/mesalib/src/mesa/swrast/s_linetemp.h
index 352c88428..035a1e640 100644
--- a/mesalib/src/mesa/swrast/s_linetemp.h
+++ b/mesalib/src/mesa/swrast/s_linetemp.h
@@ -106,7 +106,7 @@ NAME( struct gl_context *ctx, const SWvertex *vert0, const SWvertex *vert1 )
    }
 
    /*
-   printf("%s():\n", __FUNCTION__);
+   printf("%s():\n", __func__);
    printf(" (%f, %f, %f) -> (%f, %f, %f)\n",
           vert0->attrib[VARYING_SLOT_POS][0],
           vert0->attrib[VARYING_SLOT_POS][1],
@@ -154,7 +154,7 @@ NAME( struct gl_context *ctx, const SWvertex *vert0, const SWvertex *vert1 )
       return;
 
    /*
-   printf("%s %d,%d  %g %g %g %g  %g %g %g %g\n", __FUNCTION__, dx, dy,
+   printf("%s %d,%d  %g %g %g %g  %g %g %g %g\n", __func__, dx, dy,
           vert0->attrib[VARYING_SLOT_COL1][0],
           vert0->attrib[VARYING_SLOT_COL1][1],
           vert0->attrib[VARYING_SLOT_COL1][2],
diff --git a/mesalib/src/mesa/swrast/s_points.c b/mesalib/src/mesa/swrast/s_points.c
index 8180483ba..2212c95fa 100644
--- a/mesalib/src/mesa/swrast/s_points.c
+++ b/mesalib/src/mesa/swrast/s_points.c
@@ -24,7 +24,6 @@
 
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "s_context.h"
 #include "s_feedback.h"
diff --git a/mesalib/src/mesa/swrast/s_span.c b/mesalib/src/mesa/swrast/s_span.c
index e304b6b5a..3db10e163 100644
--- a/mesalib/src/mesa/swrast/s_span.c
+++ b/mesalib/src/mesa/swrast/s_span.c
@@ -33,7 +33,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/format_pack.h"
 #include "main/format_unpack.h"
 #include "main/macros.h"
@@ -1144,7 +1143,7 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
    struct gl_framebuffer *fb = ctx->DrawBuffer;
 
    /*
-   printf("%s()  interp 0x%x  array 0x%x\n", __FUNCTION__,
+   printf("%s()  interp 0x%x  array 0x%x\n", __func__,
           span->interpMask, span->arrayMask);
    */
 
diff --git a/mesalib/src/mesa/swrast/s_texcombine.c b/mesalib/src/mesa/swrast/s_texcombine.c
index 58ff16465..0adb8e5ba 100644
--- a/mesalib/src/mesa/swrast/s_texcombine.c
+++ b/mesalib/src/mesa/swrast/s_texcombine.c
@@ -26,8 +26,8 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/imports.h"
+#include "main/macros.h"
 #include "main/pixeltransfer.h"
 #include "main/samplerobj.h"
 #include "program/prog_instruction.h"
diff --git a/mesalib/src/mesa/swrast/s_texfetch.c b/mesalib/src/mesa/swrast/s_texfetch.c
index 3c4ee15ba..1fe21c0b4 100644
--- a/mesalib/src/mesa/swrast/s_texfetch.c
+++ b/mesalib/src/mesa/swrast/s_texfetch.c
@@ -33,7 +33,6 @@
  */
 
 
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/texcompress.h"
 #include "main/texcompress_fxt1.h"
diff --git a/mesalib/src/mesa/swrast/s_texfilter.c b/mesalib/src/mesa/swrast/s_texfilter.c
index 3ade99519..abc1727cf 100644
--- a/mesalib/src/mesa/swrast/s_texfilter.c
+++ b/mesalib/src/mesa/swrast/s_texfilter.c
@@ -26,8 +26,8 @@
 #include "c99_math.h"
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/imports.h"
+#include "main/macros.h"
 #include "main/samplerobj.h"
 #include "main/teximage.h"
 #include "main/texobj.h"
diff --git a/mesalib/src/mesa/swrast/s_texrender.c b/mesalib/src/mesa/swrast/s_texrender.c
index 29bb270d6..fa853c919 100644
--- a/mesalib/src/mesa/swrast/s_texrender.c
+++ b/mesalib/src/mesa/swrast/s_texrender.c
@@ -1,6 +1,5 @@
 
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/fbobject.h"
 #include "main/macros.h"
 #include "main/teximage.h"
diff --git a/mesalib/src/mesa/swrast/s_triangle.c b/mesalib/src/mesa/swrast/s_triangle.c
index af039c359..876a74b08 100644
--- a/mesalib/src/mesa/swrast/s_triangle.c
+++ b/mesalib/src/mesa/swrast/s_triangle.c
@@ -31,7 +31,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "main/imports.h"
 #include "main/macros.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/swrast/s_tritemp.h b/mesalib/src/mesa/swrast/s_tritemp.h
index fb73b2d59..fddbbfd99 100644
--- a/mesalib/src/mesa/swrast/s_tritemp.h
+++ b/mesalib/src/mesa/swrast/s_tritemp.h
@@ -92,7 +92,7 @@
 
 
 #ifndef MAX_GLUINT
-#define MAX_GLUINT	0xffffffff
+#define MAX_GLUINT	0xffffffffu
 #endif
 
 
@@ -156,7 +156,7 @@ static void NAME(struct gl_context *ctx, const SWvertex *v0,
 #endif
 
    /*
-   printf("%s()\n", __FUNCTION__);
+   printf("%s()\n", __func__);
    printf("  %g, %g, %g\n",
           v0->attrib[VARYING_SLOT_POS][0],
           v0->attrib[VARYING_SLOT_POS][1],
diff --git a/mesalib/src/mesa/swrast/s_zoom.c b/mesalib/src/mesa/swrast/s_zoom.c
index ab22652c7..9879e2a5f 100644
--- a/mesalib/src/mesa/swrast/s_zoom.c
+++ b/mesalib/src/mesa/swrast/s_zoom.c
@@ -26,7 +26,6 @@
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/format_pack.h"
-#include "main/colormac.h"
 
 #include "s_context.h"
 #include "s_span.h"
diff --git a/mesalib/src/mesa/swrast_setup/ss_context.c b/mesalib/src/mesa/swrast_setup/ss_context.c
index 0b3b9e4df..74b1da342 100644
--- a/mesalib/src/mesa/swrast_setup/ss_context.c
+++ b/mesalib/src/mesa/swrast_setup/ss_context.c
@@ -27,7 +27,7 @@
 
 #include "main/glheader.h"
 #include "main/imports.h"
-#include "main/colormac.h"
+#include "main/macros.h"
 #include "tnl/tnl.h"
 #include "tnl/t_context.h"
 #include "tnl/t_pipeline.h"
@@ -167,7 +167,7 @@ setup_vertex_format(struct gl_context *ctx)
          EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F, pointSize );
 
       _tnl_install_attrs( ctx, map, e,
-                          ctx->ViewportArray[0]._WindowMap.m,
+                          tnl->_WindowMap.m,
                           sizeof(SWvertex) );
 
       swsetup->last_index_bitset = index_bitset;
@@ -265,7 +265,8 @@ _swsetup_Wakeup( struct gl_context *ctx )
 void 
 _swsetup_Translate( struct gl_context *ctx, const void *vertex, SWvertex *dest )
 {
-   const GLfloat *m = ctx->ViewportArray[0]._WindowMap.m;
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   const GLfloat *m = tnl->_WindowMap.m;
    GLfloat tmp[4];
    GLuint i;
 
diff --git a/mesalib/src/mesa/swrast_setup/ss_triangle.c b/mesalib/src/mesa/swrast_setup/ss_triangle.c
index 483c41592..b92c20be2 100644
--- a/mesalib/src/mesa/swrast_setup/ss_triangle.c
+++ b/mesalib/src/mesa/swrast_setup/ss_triangle.c
@@ -27,7 +27,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/mtypes.h"
 
diff --git a/mesalib/src/mesa/tnl/t_context.c b/mesalib/src/mesa/tnl/t_context.c
index bc705d7a3..5b9dd54d7 100644
--- a/mesalib/src/mesa/tnl/t_context.c
+++ b/mesalib/src/mesa/tnl/t_context.c
@@ -35,6 +35,7 @@
 #include "math/m_translate.h"
 #include "math/m_xform.h"
 #include "main/state.h"
+#include "main/viewport.h"
 
 #include "tnl.h"
 #include "t_context.h"
@@ -69,6 +70,8 @@ _tnl_CreateContext( struct gl_context *ctx )
       _tnl_install_pipeline( ctx, _tnl_default_pipeline );
    }
 
+   _math_matrix_ctr(&tnl->_WindowMap);
+
    tnl->NeedNdcCoords = GL_TRUE;
    tnl->AllowVertexFog = GL_TRUE;
    tnl->AllowPixelFog = GL_TRUE;
@@ -108,6 +111,8 @@ _tnl_DestroyContext( struct gl_context *ctx )
    struct tnl_shine_tab *s, *tmps;
    TNLcontext *tnl = TNL_CONTEXT(ctx);
 
+   _math_matrix_dtr(&tnl->_WindowMap);
+
    /* Free lighting shininess exponentiation table */
    foreach_s( s, tmps, tnl->_ShineTabList ) {
       free( s );
@@ -182,6 +187,13 @@ _tnl_InvalidateState( struct gl_context *ctx, GLuint new_state )
          }
       }
    }
+
+   if (new_state & (_NEW_VIEWPORT | _NEW_BUFFERS)) {
+      double scale[3], translate[3];
+      _mesa_get_viewport_xform(ctx, 0, scale, translate);
+      _math_matrix_viewport(&tnl->_WindowMap, scale, translate,
+                            ctx->DrawBuffer->_DepthMaxF);
+   }
 }
 
 
diff --git a/mesalib/src/mesa/tnl/t_context.h b/mesalib/src/mesa/tnl/t_context.h
index e89a7f836..e7adb5f53 100644
--- a/mesalib/src/mesa/tnl/t_context.h
+++ b/mesalib/src/mesa/tnl/t_context.h
@@ -514,6 +514,7 @@ typedef struct
    /* Clipspace/ndc/window vertex managment:
     */
    struct tnl_clipspace clipspace;
+   GLmatrix _WindowMap;
 
    /* Probably need a better configuration mechanism:
     */
diff --git a/mesalib/src/mesa/tnl/t_draw.c b/mesalib/src/mesa/tnl/t_draw.c
index 60265d60b..6adf1dce6 100644
--- a/mesalib/src/mesa/tnl/t_draw.c
+++ b/mesalib/src/mesa/tnl/t_draw.c
@@ -448,7 +448,7 @@ void _tnl_draw_prims(struct gl_context *ctx,
 
    if (0)
    {
-      printf("%s %d..%d\n", __FUNCTION__, min_index, max_index);
+      printf("%s %d..%d\n", __func__, min_index, max_index);
       for (i = 0; i < nr_prims; i++)
 	 printf("prim %d: %s start %d count %d\n", i, 
 		_mesa_lookup_enum_by_nr(prim[i].mode),
diff --git a/mesalib/src/mesa/tnl/t_rasterpos.c b/mesalib/src/mesa/tnl/t_rasterpos.c
index 9ecf947df..d4b45bac9 100644
--- a/mesalib/src/mesa/tnl/t_rasterpos.c
+++ b/mesalib/src/mesa/tnl/t_rasterpos.c
@@ -25,12 +25,12 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/feedback.h"
 #include "main/light.h"
 #include "main/macros.h"
 #include "util/simple_list.h"
 #include "main/mtypes.h"
+#include "main/viewport.h"
 
 #include "math/m_matrix.h"
 #include "tnl/tnl.h"
@@ -378,6 +378,7 @@ _tnl_RasterPos(struct gl_context *ctx, const GLfloat vObj[4])
       GLfloat eye[4], clip[4], ndc[3], d;
       GLfloat *norm, eyenorm[3];
       GLfloat *objnorm = ctx->Current.Attrib[VERT_ATTRIB_NORMAL];
+      double scale[3], translate[3];
 
       /* apply modelview matrix:  eye = MV * obj */
       TRANSFORM_POINT( eye, ctx->ModelviewMatrixStack.Top->m, vObj );
@@ -410,13 +411,10 @@ _tnl_RasterPos(struct gl_context *ctx, const GLfloat vObj[4])
       ndc[1] = clip[1] * d;
       ndc[2] = clip[2] * d;
       /* wincoord = viewport_mapping(ndc) */
-      ctx->Current.RasterPos[0] = (ndc[0] * ctx->ViewportArray[0]._WindowMap.m[MAT_SX]
-                                   + ctx->ViewportArray[0]._WindowMap.m[MAT_TX]);
-      ctx->Current.RasterPos[1] = (ndc[1] * ctx->ViewportArray[0]._WindowMap.m[MAT_SY]
-                                   + ctx->ViewportArray[0]._WindowMap.m[MAT_TY]);
-      ctx->Current.RasterPos[2] = (ndc[2] * ctx->ViewportArray[0]._WindowMap.m[MAT_SZ]
-                                   + ctx->ViewportArray[0]._WindowMap.m[MAT_TZ])
-                                  / ctx->DrawBuffer->_DepthMaxF;
+      _mesa_get_viewport_xform(ctx, 0, scale, translate);
+      ctx->Current.RasterPos[0] = ndc[0] * scale[0] + translate[0];
+      ctx->Current.RasterPos[1] = ndc[1] * scale[1] + translate[1];
+      ctx->Current.RasterPos[2] = ndc[2] * scale[2] + translate[2];
       ctx->Current.RasterPos[3] = clip[3];
 
       if (ctx->Transform.DepthClamp) {
diff --git a/mesalib/src/mesa/tnl/t_vb_fog.c b/mesalib/src/mesa/tnl/t_vb_fog.c
index 3626f1da3..1ca72f866 100644
--- a/mesalib/src/mesa/tnl/t_vb_fog.c
+++ b/mesalib/src/mesa/tnl/t_vb_fog.c
@@ -28,7 +28,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_light.c b/mesalib/src/mesa/tnl/t_vb_light.c
index 7781b6a6c..dbd57fa6b 100644
--- a/mesalib/src/mesa/tnl/t_vb_light.c
+++ b/mesalib/src/mesa/tnl/t_vb_light.c
@@ -25,7 +25,6 @@
 
 #include "c99_math.h"
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/light.h"
 #include "main/macros.h"
 #include "main/imports.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_lighttmp.h b/mesalib/src/mesa/tnl/t_vb_lighttmp.h
index 57f569bd7..f8786accb 100644
--- a/mesalib/src/mesa/tnl/t_vb_lighttmp.h
+++ b/mesalib/src/mesa/tnl/t_vb_lighttmp.h
@@ -68,7 +68,7 @@ static void TAG(light_rgba_spec)( struct gl_context *ctx,
    const GLuint nr = VB->Count;
 
 #ifdef TRACE
-   fprintf(stderr, "%s\n", __FUNCTION__ );
+   fprintf(stderr, "%s\n", __func__ );
 #endif
 
    VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
@@ -249,7 +249,7 @@ static void TAG(light_rgba)( struct gl_context *ctx,
    const GLuint nr = VB->Count;
 
 #ifdef TRACE
-   fprintf(stderr, "%s\n", __FUNCTION__ );
+   fprintf(stderr, "%s\n", __func__ );
 #endif
 
    VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
@@ -429,7 +429,7 @@ static void TAG(light_fast_rgba_single)( struct gl_context *ctx,
 #endif
 
 #ifdef TRACE
-   fprintf(stderr, "%s\n", __FUNCTION__ );
+   fprintf(stderr, "%s\n", __func__ );
 #endif
 
    (void) input;		/* doesn't refer to Eye or Obj */
@@ -533,7 +533,7 @@ static void TAG(light_fast_rgba)( struct gl_context *ctx,
    const struct gl_light *light;
 
 #ifdef TRACE
-   fprintf(stderr, "%s %d\n", __FUNCTION__, nr );
+   fprintf(stderr, "%s %d\n", __func__, nr );
 #endif
 
    (void) input;
diff --git a/mesalib/src/mesa/tnl/t_vb_normals.c b/mesalib/src/mesa/tnl/t_vb_normals.c
index b67789e42..9aee1a2fb 100644
--- a/mesalib/src/mesa/tnl/t_vb_normals.c
+++ b/mesalib/src/mesa/tnl/t_vb_normals.c
@@ -27,7 +27,6 @@
 
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_program.c b/mesalib/src/mesa/tnl/t_vb_program.c
index 464a4cddd..149434971 100644
--- a/mesalib/src/mesa/tnl/t_vb_program.c
+++ b/mesalib/src/mesa/tnl/t_vb_program.c
@@ -32,7 +32,6 @@
 
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/samplerobj.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_texgen.c b/mesalib/src/mesa/tnl/t_vb_texgen.c
index 9a61ef2fe..94066f4f6 100644
--- a/mesalib/src/mesa/tnl/t_vb_texgen.c
+++ b/mesalib/src/mesa/tnl/t_vb_texgen.c
@@ -35,7 +35,6 @@
  */
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_texmat.c b/mesalib/src/mesa/tnl/t_vb_texmat.c
index 1cc2c8116..ef034d643 100644
--- a/mesalib/src/mesa/tnl/t_vb_texmat.c
+++ b/mesalib/src/mesa/tnl/t_vb_texmat.c
@@ -27,7 +27,6 @@
 
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/tnl/t_vb_vertex.c b/mesalib/src/mesa/tnl/t_vb_vertex.c
index ea3a56cd1..b56d6803c 100644
--- a/mesalib/src/mesa/tnl/t_vb_vertex.c
+++ b/mesalib/src/mesa/tnl/t_vb_vertex.c
@@ -27,7 +27,6 @@
 
 
 #include "main/glheader.h"
-#include "main/colormac.h"
 #include "main/macros.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/tnl/t_vertex.c b/mesalib/src/mesa/tnl/t_vertex.c
index 369d6d945..c3294b007 100644
--- a/mesalib/src/mesa/tnl/t_vertex.c
+++ b/mesalib/src/mesa/tnl/t_vertex.c
@@ -28,7 +28,6 @@
 #include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "swrast/s_chan.h"
 #include "t_context.h"
 #include "t_vertex.h"
diff --git a/mesalib/src/mesa/tnl/t_vertex_generic.c b/mesalib/src/mesa/tnl/t_vertex_generic.c
index 079d473fc..2a25a9692 100644
--- a/mesalib/src/mesa/tnl/t_vertex_generic.c
+++ b/mesalib/src/mesa/tnl/t_vertex_generic.c
@@ -28,7 +28,7 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
+#include "main/macros.h"
 #include "util/simple_list.h"
 #include "swrast/s_chan.h"
 #include "t_context.h"
@@ -36,7 +36,7 @@
 
 
 #if 0
-#define DEBUG_INSERT printf("%s\n", __FUNCTION__)
+#define DEBUG_INSERT printf("%s\n", __func__)
 #else
 #define DEBUG_INSERT
 #endif
diff --git a/mesalib/src/mesa/tnl/t_vertex_sse.c b/mesalib/src/mesa/tnl/t_vertex_sse.c
index 963432c48..30dc1a720 100644
--- a/mesalib/src/mesa/tnl/t_vertex_sse.c
+++ b/mesalib/src/mesa/tnl/t_vertex_sse.c
@@ -29,7 +29,6 @@
 
 #include "main/glheader.h"
 #include "main/context.h"
-#include "main/colormac.h"
 #include "util/simple_list.h"
 #include "main/enums.h"
 #include "swrast/s_chan.h"
diff --git a/mesalib/src/mesa/vbo/vbo_exec_api.c b/mesalib/src/mesa/vbo/vbo_exec_api.c
index 02741c2bc..859078f12 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_api.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_api.c
@@ -439,7 +439,7 @@ do {									\
 } while (0)
 
 
-#define ERROR(err) _mesa_error( ctx, err, __FUNCTION__ )
+#define ERROR(err) _mesa_error( ctx, err, __func__ )
 #define TAG(x) vbo_##x
 
 #include "vbo_attrib_tmp.h"
diff --git a/mesalib/src/mesa/vbo/vbo_exec_draw.c b/mesalib/src/mesa/vbo/vbo_exec_draw.c
index 91f2ca43a..37b53a830 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_draw.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_draw.c
@@ -45,7 +45,7 @@ vbo_exec_debug_verts( struct vbo_exec_context *exec )
    GLuint i;
 
    printf("%s: %u vertices %d primitives, %d vertsize\n",
-	  __FUNCTION__,
+	  __func__,
 	  count,
 	  exec->vtx.prim_count,
 	  exec->vtx.vertex_size);
@@ -402,7 +402,7 @@ vbo_exec_vtx_flush(struct vbo_exec_context *exec, GLboolean keepUnmapped)
          }
 
          if (0)
-            printf("%s %d %d\n", __FUNCTION__, exec->vtx.prim_count,
+            printf("%s %d %d\n", __func__, exec->vtx.prim_count,
 		   exec->vtx.vert_count);
 
 	 vbo_context(ctx)->draw_prims( ctx, 
diff --git a/mesalib/src/mesa/vbo/vbo_primitive_restart.c b/mesalib/src/mesa/vbo/vbo_primitive_restart.c
index 562dedcd5..dafc4fd2a 100644
--- a/mesalib/src/mesa/vbo/vbo_primitive_restart.c
+++ b/mesalib/src/mesa/vbo/vbo_primitive_restart.c
@@ -167,6 +167,8 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
                          struct gl_buffer_object *indirect)
 {
    GLuint prim_num;
+   struct _mesa_prim new_prim;
+   struct _mesa_index_buffer new_ib;
    struct sub_primitive *sub_prims;
    struct sub_primitive *sub_prim;
    GLuint num_sub_prims;
@@ -182,8 +184,6 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
 
    /* If there is an indirect buffer, map it and extract the draw params */
    if (indirect && prims[0].is_indirect) {
-      struct _mesa_prim new_prim = *prims;
-      struct _mesa_index_buffer new_ib = *ib;
       const uint32_t *indirect_params;
       if (!ctx->Driver.MapBufferRange(ctx, 0, indirect->Size, GL_MAP_READ_BIT,
                                       indirect, MAP_INTERNAL)) {
@@ -195,6 +195,7 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
       }
 
       assert(nr_prims == 1);
+      new_prim = prims[0];
       indirect_params = (const uint32_t *)
                         ADD_POINTERS(indirect->Mappings[MAP_INTERNAL].Pointer,
                                      new_prim.indirect_offset);
@@ -206,6 +207,7 @@ vbo_sw_primitive_restart(struct gl_context *ctx,
       new_prim.basevertex = indirect_params[3];
       new_prim.base_instance = indirect_params[4];
 
+      new_ib = *ib;
       new_ib.count = new_prim.count;
 
       prims = &new_prim;
diff --git a/mesalib/src/mesa/vbo/vbo_rebase.c b/mesalib/src/mesa/vbo/vbo_rebase.c
index b06df4ab2..c3c4b64e6 100644
--- a/mesalib/src/mesa/vbo/vbo_rebase.c
+++ b/mesalib/src/mesa/vbo/vbo_rebase.c
@@ -142,7 +142,7 @@ void vbo_rebase_prims( struct gl_context *ctx,
    assert(min_index != 0);
 
    if (0)
-      printf("%s %d..%d\n", __FUNCTION__, min_index, max_index);
+      printf("%s %d..%d\n", __func__, min_index, max_index);
 
 
    /* XXX this path is disabled for now.
diff --git a/mesalib/src/mesa/vbo/vbo_save_api.c b/mesalib/src/mesa/vbo/vbo_save_api.c
index fd9a5de3d..5927beeb3 100644
--- a/mesalib/src/mesa/vbo/vbo_save_api.c
+++ b/mesalib/src/mesa/vbo/vbo_save_api.c
@@ -763,7 +763,7 @@ _save_reset_vertex(struct gl_context *ctx)
 
 
 
-#define ERROR(err)   _mesa_compile_error(ctx, err, __FUNCTION__);
+#define ERROR(err)   _mesa_compile_error(ctx, err, __func__);
 
 
 /* Only one size for each attribute may be active at once.  Eg. if
diff --git a/mesalib/src/mesa/x86/common_x86.c b/mesalib/src/mesa/x86/common_x86.c
index 86fbca91e..1c8640514 100644
--- a/mesalib/src/mesa/x86/common_x86.c
+++ b/mesalib/src/mesa/x86/common_x86.c
@@ -42,7 +42,7 @@
 #include <sys/types.h>
 #include <sys/sysctl.h>
 #endif
-#if defined(USE_SSE_ASM) && defined(__OpenBSD__)
+#if defined(USE_SSE_ASM) && (defined(__OpenBSD__) || defined(__NetBSD__))
 #include <sys/param.h>
 #include <sys/sysctl.h>
 #include <machine/cpu.h>
diff --git a/mesalib/src/util/ralloc.c b/mesalib/src/util/ralloc.c
index 36bc61fd0..01719c888 100644
--- a/mesalib/src/util/ralloc.c
+++ b/mesalib/src/util/ralloc.c
@@ -271,6 +271,32 @@ ralloc_steal(const void *new_ctx, void *ptr)
    add_child(parent, info);
 }
 
+void
+ralloc_adopt(const void *new_ctx, void *old_ctx)
+{
+   ralloc_header *new_info, *old_info, *child;
+
+   if (unlikely(old_ctx == NULL))
+      return;
+
+   old_info = get_header(old_ctx);
+   new_info = get_header(new_ctx);
+
+   /* If there are no children, bail. */
+   if (unlikely(old_info->child == NULL))
+      return;
+
+   /* Set all the children's parent to new_ctx; get a pointer to the last child. */
+   for (child = old_info->child; child->next != NULL; child = child->next) {
+      child->parent = new_info;
+   }
+
+   /* Connect the two lists together; parent them to new_ctx; make old_ctx empty. */
+   child->next = new_info->child;
+   new_info->child = old_info->child;
+   old_info->child = NULL;
+}
+
 void *
 ralloc_parent(const void *ptr)
 {
diff --git a/mesalib/src/util/ralloc.h b/mesalib/src/util/ralloc.h
index f088a3627..7587e1190 100644
--- a/mesalib/src/util/ralloc.h
+++ b/mesalib/src/util/ralloc.h
@@ -46,16 +46,16 @@
 #ifndef RALLOC_H
 #define RALLOC_H
 
-#ifdef __cplusplus
-extern "C" {
-#endif
-
 #include <stddef.h>
 #include <stdarg.h>
 #include <stdbool.h>
 
 #include "macros.h"
 
+#ifdef __cplusplus
+extern "C" {
+#endif
+
 /**
  * \def ralloc(ctx, type)
  * Allocate a new object chained off of the given context.
@@ -235,6 +235,13 @@ void ralloc_free(void *ptr);
 void ralloc_steal(const void *new_ctx, void *ptr);
 
 /**
+ * Reparent all children from one context to another.
+ *
+ * This effectively calls ralloc_steal(new_ctx, child) for all children of \p old_ctx.
+ */
+void ralloc_adopt(const void *new_ctx, void *old_ctx);
+
+/**
  * Return the given pointer's ralloc context.
  */
 void *ralloc_parent(const void *ptr);
diff --git a/mesalib/src/util/u_atomic_test.c b/mesalib/src/util/u_atomic_test.c
index 939cfe445..7844f6162 100644
--- a/mesalib/src/util/u_atomic_test.c
+++ b/mesalib/src/util/u_atomic_test.c
@@ -36,6 +36,11 @@
 
 #include "u_atomic.h"
 
+#ifdef _MSC_VER
+#pragma warning( disable : 28112 ) /* Accessing a local variable via an Interlocked function */
+#pragma warning( disable : 28113 ) /* A variable which is accessed via an Interlocked function must always be accessed via an Interlocked function */
+#endif
+
 
 /* Test only assignment-like operations, which are supported on all types */
 #define test_atomic_assign(type, ones) \