[Mesa-dev] [PATCH 1/3] glsl/i965: move lower_offset_array up to GLSL compiler level.
Dave Airlie
airlied at gmail.com
Mon Feb 10 17:43:32 PST 2014
From: Dave Airlie <airlied at redhat.com>
This lowering pass will be useful for gallium drivers as well, in order to support
the GL TG4 oddity that is textureGatherOffsets.
Signed-off-by: Dave Airlie <airlied at redhat.com>
---
src/glsl/Makefile.sources | 1 +
src/glsl/ir_optimization.h | 1 +
src/glsl/lower_offset_array.cpp | 90 +++++++++++++++++++++
src/mesa/drivers/dri/i965/Makefile.sources | 1 -
src/mesa/drivers/dri/i965/brw_context.h | 1 -
.../drivers/dri/i965/brw_lower_offset_array.cpp | 93 ----------------------
src/mesa/drivers/dri/i965/brw_shader.cpp | 2 +-
7 files changed, 93 insertions(+), 96 deletions(-)
create mode 100644 src/glsl/lower_offset_array.cpp
delete mode 100644 src/mesa/drivers/dri/i965/brw_lower_offset_array.cpp
diff --git a/src/glsl/Makefile.sources b/src/glsl/Makefile.sources
index e69c1ac..5945590 100644
--- a/src/glsl/Makefile.sources
+++ b/src/glsl/Makefile.sources
@@ -67,6 +67,7 @@ LIBGLSL_FILES = \
$(GLSL_SRCDIR)/lower_jumps.cpp \
$(GLSL_SRCDIR)/lower_mat_op_to_vec.cpp \
$(GLSL_SRCDIR)/lower_noise.cpp \
+ $(GLSL_SRCDIR)/lower_offset_array.cpp \
$(GLSL_SRCDIR)/lower_packed_varyings.cpp \
$(GLSL_SRCDIR)/lower_named_interface_blocks.cpp \
$(GLSL_SRCDIR)/lower_packing_builtins.cpp \
diff --git a/src/glsl/ir_optimization.h b/src/glsl/ir_optimization.h
index 055d655..24d019b 100644
--- a/src/glsl/ir_optimization.h
+++ b/src/glsl/ir_optimization.h
@@ -120,6 +120,7 @@ bool lower_vector_insert(exec_list *instructions, bool lower_nonconstant_index);
void lower_named_interface_blocks(void *mem_ctx, gl_shader *shader);
bool optimize_redundant_jumps(exec_list *instructions);
bool optimize_split_arrays(exec_list *instructions, bool linked);
+bool lower_offset_arrays(exec_list *instructions);
ir_rvalue *
compare_index_block(exec_list *instructions, ir_variable *index,
diff --git a/src/glsl/lower_offset_array.cpp b/src/glsl/lower_offset_array.cpp
new file mode 100644
index 0000000..0c235ed
--- /dev/null
+++ b/src/glsl/lower_offset_array.cpp
@@ -0,0 +1,90 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * \file brw_lower_offset_array.cpp
+ *
+ * IR lower pass to decompose ir_texture ir_tg4 with an array of offsets
+ * into four ir_tg4s with a single ivec2 offset, select the .w component of each,
+ * and return those four values packed into a gvec4.
+ *
+ * \author Chris Forbes <chrisf at ijw.co.nz>
+ */
+
+#include "glsl_types.h"
+#include "ir.h"
+#include "ir_builder.h"
+#include "ir_optimization.h"
+#include "ir_rvalue_visitor.h"
+
+using namespace ir_builder;
+
+class brw_lower_offset_array_visitor : public ir_rvalue_visitor {
+public:
+ brw_lower_offset_array_visitor()
+ {
+ progress = false;
+ }
+
+ void handle_rvalue(ir_rvalue **rv);
+
+ bool progress;
+};
+
+void
+brw_lower_offset_array_visitor::handle_rvalue(ir_rvalue **rv)
+{
+ if (*rv == NULL || (*rv)->ir_type != ir_type_texture)
+ return;
+
+ ir_texture *ir = (ir_texture *) *rv;
+ if (ir->op != ir_tg4 || !ir->offset || !ir->offset->type->is_array())
+ return;
+
+ void *mem_ctx = ralloc_parent(ir);
+
+ ir_variable *var = new (mem_ctx) ir_variable(ir->type, "result", ir_var_auto);
+ base_ir->insert_before(var);
+
+ for (int i = 0; i < 4; i++) {
+ ir_texture *tex = ir->clone(mem_ctx, NULL);
+ tex->offset = new (mem_ctx) ir_dereference_array(tex->offset,
+ new (mem_ctx) ir_constant(i));
+
+ base_ir->insert_before(assign(var, swizzle_w(tex), 1 << i));
+ }
+
+ *rv = new (mem_ctx) ir_dereference_variable(var);
+
+ progress = true;
+}
+
+bool
+lower_offset_arrays(exec_list *instructions)
+{
+ brw_lower_offset_array_visitor v;
+
+ visit_list_elements(&v, instructions);
+
+ return v.progress;
+}
diff --git a/src/mesa/drivers/dri/i965/Makefile.sources b/src/mesa/drivers/dri/i965/Makefile.sources
index dc30eb3..1649369 100644
--- a/src/mesa/drivers/dri/i965/Makefile.sources
+++ b/src/mesa/drivers/dri/i965/Makefile.sources
@@ -73,7 +73,6 @@ i965_FILES = \
brw_gs_surface_state.c \
brw_interpolation_map.c \
brw_lower_texture_gradients.cpp \
- brw_lower_offset_array.cpp \
brw_lower_unnormalized_offset.cpp \
brw_misc_state.c \
brw_object_purgeable.c \
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 98e90e2..f055b13 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1845,7 +1845,6 @@ brw_program_reloc(struct brw_context *brw, uint32_t state_offset,
bool brw_do_cubemap_normalize(struct exec_list *instructions);
bool brw_lower_texture_gradients(struct brw_context *brw,
struct exec_list *instructions);
-bool brw_do_lower_offset_arrays(struct exec_list *instructions);
bool brw_do_lower_unnormalized_offset(struct exec_list *instructions);
struct opcode_desc {
diff --git a/src/mesa/drivers/dri/i965/brw_lower_offset_array.cpp b/src/mesa/drivers/dri/i965/brw_lower_offset_array.cpp
deleted file mode 100644
index 91c002e..0000000
--- a/src/mesa/drivers/dri/i965/brw_lower_offset_array.cpp
+++ /dev/null
@@ -1,93 +0,0 @@
-/*
- * Copyright © 2013 Intel Corporation
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file brw_lower_offset_array.cpp
- *
- * IR lower pass to decompose ir_texture ir_tg4 with an array of offsets
- * into four ir_tg4s with a single ivec2 offset, select the .w component of each,
- * and return those four values packed into a gvec4.
- *
- * \author Chris Forbes <chrisf at ijw.co.nz>
- */
-
-#include "glsl/glsl_types.h"
-#include "glsl/ir.h"
-#include "glsl/ir_builder.h"
-#include "glsl/ir_rvalue_visitor.h"
-
-using namespace ir_builder;
-
-class brw_lower_offset_array_visitor : public ir_rvalue_visitor {
-public:
- brw_lower_offset_array_visitor()
- {
- progress = false;
- }
-
- void handle_rvalue(ir_rvalue **rv);
-
- bool progress;
-};
-
-void
-brw_lower_offset_array_visitor::handle_rvalue(ir_rvalue **rv)
-{
- if (*rv == NULL || (*rv)->ir_type != ir_type_texture)
- return;
-
- ir_texture *ir = (ir_texture *) *rv;
- if (ir->op != ir_tg4 || !ir->offset || !ir->offset->type->is_array())
- return;
-
- void *mem_ctx = ralloc_parent(ir);
-
- ir_variable *var = new (mem_ctx) ir_variable(ir->type, "result", ir_var_auto);
- base_ir->insert_before(var);
-
- for (int i = 0; i < 4; i++) {
- ir_texture *tex = ir->clone(mem_ctx, NULL);
- tex->offset = new (mem_ctx) ir_dereference_array(tex->offset,
- new (mem_ctx) ir_constant(i));
-
- base_ir->insert_before(assign(var, swizzle_w(tex), 1 << i));
- }
-
- *rv = new (mem_ctx) ir_dereference_variable(var);
-
- progress = true;
-}
-
-extern "C" {
-
-bool
-brw_do_lower_offset_arrays(exec_list *instructions)
-{
- brw_lower_offset_array_visitor v;
-
- visit_list_elements(&v, instructions);
-
- return v.progress;
-}
-
-}
diff --git a/src/mesa/drivers/dri/i965/brw_shader.cpp b/src/mesa/drivers/dri/i965/brw_shader.cpp
index 6cc2595..f1ffd9a 100644
--- a/src/mesa/drivers/dri/i965/brw_shader.cpp
+++ b/src/mesa/drivers/dri/i965/brw_shader.cpp
@@ -167,7 +167,7 @@ brw_link_shader(struct gl_context *ctx, struct gl_shader_program *shProg)
do_vec_index_to_cond_assign(shader->base.ir);
lower_vector_insert(shader->base.ir, true);
brw_do_cubemap_normalize(shader->base.ir);
- brw_do_lower_offset_arrays(shader->base.ir);
+ lower_offset_arrays(shader->base.ir);
brw_do_lower_unnormalized_offset(shader->base.ir);
lower_noise(shader->base.ir);
lower_quadop_vector(shader->base.ir, false);
--
1.8.3.1
More information about the mesa-dev
mailing list