mesa/src/compiler/glsl/opt_constant_variable.cpp
Timothy Arceri 77f5221233 glsl: pass mem_ctx to constant_expression_value(...) and friends
The main motivation for this is that threaded compilation can fall
over if we were to allocate IR inside constant_expression_value()
when calling it on a builtin. This is because builtins are shared
across the whole OpenGL context.

f81ede4699 worked around the problem by cloning the entire
builtin before constant_expression_value() could be called on
it. However cloning the whole function each time we referenced
it lead to a significant reduction in the GLSL IR compiler
performance. This change along with the following patch
helps fix that performance regression.

Other advantages are that we reduce the number of calls to
ralloc_parent(), and for loop unrolling we free constants after
they are used rather than leaving them hanging around.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2017-08-11 15:44:08 +10:00

225 lines
6.5 KiB
C++

/*
* Copyright © 2010 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 opt_constant_variable.cpp
*
* Marks variables assigned a single constant value over the course
* of the program as constant.
*
* The goal here is to trigger further constant folding and then dead
* code elimination. This is common with vector/matrix constructors
* and calls to builtin functions.
*/
#include "ir.h"
#include "ir_visitor.h"
#include "ir_optimization.h"
#include "compiler/glsl_types.h"
#include "util/hash_table.h"
namespace {
struct assignment_entry {
int assignment_count;
ir_variable *var;
ir_constant *constval;
bool our_scope;
};
class ir_constant_variable_visitor : public ir_hierarchical_visitor {
public:
virtual ir_visitor_status visit_enter(ir_dereference_variable *);
virtual ir_visitor_status visit(ir_variable *);
virtual ir_visitor_status visit_enter(ir_assignment *);
virtual ir_visitor_status visit_enter(ir_call *);
struct hash_table *ht;
};
} /* unnamed namespace */
static struct assignment_entry *
get_assignment_entry(ir_variable *var, struct hash_table *ht)
{
struct hash_entry *hte = _mesa_hash_table_search(ht, var);
struct assignment_entry *entry;
if (hte) {
entry = (struct assignment_entry *) hte->data;
} else {
entry = (struct assignment_entry *) calloc(1, sizeof(*entry));
entry->var = var;
_mesa_hash_table_insert(ht, var, entry);
}
return entry;
}
ir_visitor_status
ir_constant_variable_visitor::visit(ir_variable *ir)
{
struct assignment_entry *entry = get_assignment_entry(ir, this->ht);
entry->our_scope = true;
return visit_continue;
}
/* Skip derefs of variables so that we can detect declarations. */
ir_visitor_status
ir_constant_variable_visitor::visit_enter(ir_dereference_variable *ir)
{
(void)ir;
return visit_continue_with_parent;
}
ir_visitor_status
ir_constant_variable_visitor::visit_enter(ir_assignment *ir)
{
ir_constant *constval;
struct assignment_entry *entry;
entry = get_assignment_entry(ir->lhs->variable_referenced(), this->ht);
assert(entry);
entry->assignment_count++;
/* If there's more than one assignment, don't bother - we won't do anything
* with this variable anyway, and continuing just wastes memory cloning
* constant expressions.
*/
if (entry->assignment_count > 1)
return visit_continue;
/* If it's already constant, don't do the work. */
if (entry->var->constant_value)
return visit_continue;
/* OK, now find if we actually have all the right conditions for
* this to be a constant value assigned to the var.
*/
if (ir->condition)
return visit_continue;
ir_variable *var = ir->whole_variable_written();
if (!var)
return visit_continue;
/* Ignore buffer variables, since the underlying storage is shared
* and we can't be sure that this variable won't be written by another
* thread.
*/
if (var->data.mode == ir_var_shader_storage ||
var->data.mode == ir_var_shader_shared)
return visit_continue;
constval = ir->rhs->constant_expression_value(ralloc_parent(ir));
if (!constval)
return visit_continue;
/* Mark this entry as having a constant assignment (if the
* assignment count doesn't go >1). do_constant_variable will fix
* up the variable with the constant value later.
*/
entry->constval = constval;
return visit_continue;
}
ir_visitor_status
ir_constant_variable_visitor::visit_enter(ir_call *ir)
{
/* Mark any out parameters as assigned to */
foreach_two_lists(formal_node, &ir->callee->parameters,
actual_node, &ir->actual_parameters) {
ir_rvalue *param_rval = (ir_rvalue *) actual_node;
ir_variable *param = (ir_variable *) formal_node;
if (param->data.mode == ir_var_function_out ||
param->data.mode == ir_var_function_inout) {
ir_variable *var = param_rval->variable_referenced();
struct assignment_entry *entry;
assert(var);
entry = get_assignment_entry(var, this->ht);
entry->assignment_count++;
}
}
/* Mark the return storage as having been assigned to */
if (ir->return_deref != NULL) {
ir_variable *var = ir->return_deref->variable_referenced();
struct assignment_entry *entry;
assert(var);
entry = get_assignment_entry(var, this->ht);
entry->assignment_count++;
}
return visit_continue;
}
/**
* Does a copy propagation pass on the code present in the instruction stream.
*/
bool
do_constant_variable(exec_list *instructions)
{
bool progress = false;
ir_constant_variable_visitor v;
v.ht = _mesa_hash_table_create(NULL, _mesa_hash_pointer,
_mesa_key_pointer_equal);
v.run(instructions);
struct hash_entry *hte;
hash_table_foreach(v.ht, hte) {
struct assignment_entry *entry = (struct assignment_entry *) hte->data;
if (entry->assignment_count == 1 && entry->constval && entry->our_scope) {
entry->var->constant_value = entry->constval;
progress = true;
}
hte->data = NULL;
free(entry);
}
_mesa_hash_table_destroy(v.ht, NULL);
return progress;
}
bool
do_constant_variable_unlinked(exec_list *instructions)
{
bool progress = false;
foreach_in_list(ir_instruction, ir, instructions) {
ir_function *f = ir->as_function();
if (f) {
foreach_in_list(ir_function_signature, sig, &f->signatures) {
if (do_constant_variable(&sig->body))
progress = true;
}
}
}
return progress;
}