3 * Copyright © 2016 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 DEALINGS
25 #ifndef BRW_IR_ANALYSIS_H
26 #define BRW_IR_ANALYSIS_H
30 * Bitset of state categories that can influence the result of IR analysis
33 enum analysis_dependency_class
{
35 * The analysis doesn't depend on the IR, its result is effectively a
36 * constant during the compilation.
38 DEPENDENCY_NOTHING
= 0,
40 * The analysis depends on the set of instructions in the program and
41 * their naming. Note that because instructions are named sequentially
42 * by IP this implies a dependency on the control flow edges between
43 * instructions. This will be signaled whenever instructions are
44 * inserted, removed or reordered in the program.
46 DEPENDENCY_INSTRUCTION_IDENTITY
= 0x1,
48 * The analysis is sensitive to the detailed semantics of instructions
49 * in the program, where "detailed" means any change in the instruction
50 * data structures other than the linked-list pointers (which are
51 * already covered by DEPENDENCY_INSTRUCTION_IDENTITY). E.g. changing
52 * the negate or abs flags of an instruction source would signal this
53 * flag alone because it would preserve all other instruction dependency
56 DEPENDENCY_INSTRUCTION_DETAIL
= 0x2,
58 * The analysis depends on the set of data flow edges between
59 * instructions. This will be signaled whenever the dataflow relation
60 * between instructions has potentially changed, e.g. when the VGRF
61 * index of an instruction source or destination changes (in which case
62 * it will appear in combination with DEPENDENCY_INSTRUCTION_DETAIL), or
63 * when data-dependent instructions are reordered (in which case it will
64 * appear in combination with DEPENDENCY_INSTRUCTION_IDENTITY).
66 DEPENDENCY_INSTRUCTION_DATA_FLOW
= 0x4,
68 * The analysis depends on all instruction dependency classes. These
69 * will typically be signaled simultaneously when inserting or removing
70 * instructions in the program (or if you're feeling too lazy to read
71 * through your optimization pass to figure out which of the instruction
72 * dependency classes above it invalidates).
74 DEPENDENCY_INSTRUCTIONS
= 0x7,
76 * The analysis depends on the set of VGRFs in the program and their
77 * naming. This will be signaled when VGRFs are allocated or released.
79 DEPENDENCY_VARIABLES
= 0x8,
81 * The analysis depends on the set of basic blocks in the program, their
82 * control flow edges and naming.
84 DEPENDENCY_BLOCKS
= 0x10,
86 * The analysis depends on the program being literally the same (good
87 * luck...), any change in the input invalidates previous analysis
90 DEPENDENCY_EVERYTHING
= ~0
93 inline analysis_dependency_class
94 operator|(analysis_dependency_class x
, analysis_dependency_class y
)
96 return static_cast<analysis_dependency_class
>(
97 static_cast<unsigned>(x
) | static_cast<unsigned>(y
));
102 * Instantiate a program analysis class \p L which can calculate an object of
103 * type \p T as result. \p C is a closure that encapsulates whatever
104 * information is required as argument to run the analysis pass. The purpose
105 * of this class is to make sure that:
107 * - The analysis pass is executed lazily whenever it's needed and multiple
108 * executions are optimized out as long as the cached result remains marked
111 * - There is no way to access the cached analysis result without first
112 * calling L::require(), which makes sure that the analysis pass is rerun
115 * - The cached result doesn't become inconsistent with the program for as
116 * long as it remains marked up-to-date. (This is only enforced in debug
117 * builds for performance reasons)
119 * The requirements on \p T are the following:
121 * - Constructible with a single argument, as in 'x = T(c)' for \p c of type
124 * - 'x.dependency_class()' on const \p x returns a bitset of
125 * brw::analysis_dependency_class specifying the set of IR objects that are
126 * required to remain invariant for the cached analysis result to be
129 * - 'x.validate(c)' on const \p x returns a boolean result specifying
130 * whether the analysis result \p x is consistent with the input IR. This
131 * is currently only used for validation in debug builds.
133 #define BRW_ANALYSIS(L, T, C) \
137 * Construct a program analysis. \p c is an arbitrary object \
138 * passed as argument to the constructor of the analysis result \
139 * object of type \p T. \
141 L(C const &c) : c(c), p(NULL) {} \
144 * Destroy a program analysis. \
152 * Obtain the result of a program analysis. This gives a \
153 * guaranteed up-to-date result, the analysis pass will be \
154 * rerun implicitly if it has become stale. \
160 assert(p->validate(c)); \
170 return const_cast<L *>(this)->require(); \
174 * Report that dependencies of the analysis pass may have changed \
175 * since the last calculation and the cached analysis result may \
176 * have to be discarded. \
179 invalidate(brw::analysis_dependency_class c) \
181 if (p && c & p->dependency_class()) { \