diff options
author | Dave Airlie <[email protected]> | 2014-02-11 09:41:44 +1000 |
---|---|---|
committer | Dave Airlie <[email protected]> | 2014-02-25 13:28:57 +1000 |
commit | 122c3b9486d3dbf8cb6b796c9a25f43817f006d3 (patch) | |
tree | 06adcc7da94ddd3c12568e445768dab569021e5a /src/glsl | |
parent | 945d87f95877e198fb2203f47a4ebbccae883978 (diff) |
glsl/i965: move lower_offset_array up to GLSL compiler level.
This lowering pass will be useful for gallium drivers as well, in order to support
the GL TG4 oddity that is textureGatherOffsets.
Reviewed-by: Chris Forbes <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
Diffstat (limited to 'src/glsl')
-rw-r--r-- | src/glsl/Makefile.sources | 1 | ||||
-rw-r--r-- | src/glsl/ir_optimization.h | 1 | ||||
-rw-r--r-- | src/glsl/lower_offset_array.cpp | 90 |
3 files changed, 92 insertions, 0 deletions
diff --git a/src/glsl/Makefile.sources b/src/glsl/Makefile.sources index e69c1ac612b..5945590a57e 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 055d65547ba..24d019b718e 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 00000000000..0c235eda395 --- /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 <[email protected]> + */ + +#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; +} |