2 * Copyright © 2012 Vincent Lejeune
3 * Copyright © 2012 Intel Corporation
5 * Permission is hereby granted, free of charge, to any person obtaining a
6 * copy of this software and associated documentation files (the "Software"),
7 * to deal in the Software without restriction, including without limitation
8 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
9 * and/or sell copies of the Software, and to permit persons to whom the
10 * Software is furnished to do so, subject to the following conditions:
12 * The above copyright notice and this permission notice (including the next
13 * paragraph) shall be included in all copies or substantial portions of the
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
19 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
21 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
22 * DEALINGS IN THE SOFTWARE.
26 #include "program/hash_table.h"
29 * \file lower_output_reads.cpp
31 * In GLSL, shader output variables (such as varyings) can be both read and
32 * written. However, on some hardware, reading an output register causes
35 * This pass creates temporary shadow copies of every (used) shader output,
36 * and replaces all accesses to use those instead. It also adds code to the
37 * main() function to copy the final values to the actual shader outputs.
40 class output_read_remover : public ir_hierarchical_visitor {
43 * A hash table mapping from the original ir_variable shader outputs
44 * (ir_var_out mode) to the new temporaries to be used instead.
46 hash_table *replacements;
50 output_read_remover();
51 ~output_read_remover();
52 virtual ir_visitor_status visit(class ir_dereference_variable *);
53 virtual ir_visitor_status visit_leave(class ir_return *);
54 virtual ir_visitor_status visit_leave(class ir_function_signature *);
57 output_read_remover::output_read_remover()
59 mem_ctx = ralloc_context(NULL);
61 hash_table_ctor(0, hash_table_pointer_hash, hash_table_pointer_compare);
64 output_read_remover::~output_read_remover()
66 hash_table_dtor(replacements);
71 output_read_remover::visit(ir_dereference_variable *ir)
73 if (ir->var->mode != ir_var_out)
74 return visit_continue;
76 ir_variable *temp = (ir_variable *) hash_table_find(replacements, ir->var);
78 /* If we don't have an existing temporary, create one. */
80 void *var_ctx = ralloc_parent(ir->var);
81 temp = new(var_ctx) ir_variable(ir->var->type, ir->var->name,
83 hash_table_insert(replacements, temp, ir->var);
86 /* Update the dereference to use the temporary */
89 return visit_continue;
93 * Create an assignment to copy a temporary value back to the actual output.
95 static ir_assignment *
96 copy(void *ctx, ir_variable *output, ir_variable *temp)
98 ir_dereference_variable *lhs = new(ctx) ir_dereference_variable(output);
99 ir_dereference_variable *rhs = new(ctx) ir_dereference_variable(temp);
100 return new(ctx) ir_assignment(lhs, rhs);
103 /** Insert a copy-back assignment before a "return" statement */
105 emit_return_copy(const void *key, void *data, void *closure)
107 ir_return *ir = (ir_return *) closure;
108 ir->insert_before(copy(ir, (ir_variable *) key, (ir_variable *) data));
111 /** Insert a copy-back assignment at the end of the main() function */
113 emit_main_copy(const void *key, void *data, void *closure)
115 ir_function_signature *sig = (ir_function_signature *) closure;
116 sig->body.push_tail(copy(sig, (ir_variable *) key, (ir_variable *) data));
120 output_read_remover::visit_leave(ir_return *ir)
122 hash_table_call_foreach(replacements, emit_return_copy, ir);
123 return visit_continue;
127 output_read_remover::visit_leave(ir_function_signature *sig)
129 if (strcmp(sig->function_name(), "main") != 0)
130 return visit_continue;
132 hash_table_call_foreach(replacements, emit_main_copy, sig);
133 return visit_continue;
137 lower_output_reads(exec_list *instructions)
139 output_read_remover v;
140 visit_list_elements(&v, instructions);