2013-10-11 Marc Glisse <marc.glisse@inria.fr>
[official-gcc.git] / gcc / ira.c
blob203fbff6a269f6d6b0bbb11c14fe5c1b3bc1f628
1 /* Integrated Register Allocator (IRA) entry point.
2 Copyright (C) 2006-2013 Free Software Foundation, Inc.
3 Contributed by Vladimir Makarov <vmakarov@redhat.com>.
5 This file is part of GCC.
7 GCC is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 3, or (at your option) any later
10 version.
12 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
15 for more details.
17 You should have received a copy of the GNU General Public License
18 along with GCC; see the file COPYING3. If not see
19 <http://www.gnu.org/licenses/>. */
21 /* The integrated register allocator (IRA) is a
22 regional register allocator performing graph coloring on a top-down
23 traversal of nested regions. Graph coloring in a region is based
24 on Chaitin-Briggs algorithm. It is called integrated because
25 register coalescing, register live range splitting, and choosing a
26 better hard register are done on-the-fly during coloring. Register
27 coalescing and choosing a cheaper hard register is done by hard
28 register preferencing during hard register assigning. The live
29 range splitting is a byproduct of the regional register allocation.
31 Major IRA notions are:
33 o *Region* is a part of CFG where graph coloring based on
34 Chaitin-Briggs algorithm is done. IRA can work on any set of
35 nested CFG regions forming a tree. Currently the regions are
36 the entire function for the root region and natural loops for
37 the other regions. Therefore data structure representing a
38 region is called loop_tree_node.
40 o *Allocno class* is a register class used for allocation of
41 given allocno. It means that only hard register of given
42 register class can be assigned to given allocno. In reality,
43 even smaller subset of (*profitable*) hard registers can be
44 assigned. In rare cases, the subset can be even smaller
45 because our modification of Chaitin-Briggs algorithm requires
46 that sets of hard registers can be assigned to allocnos forms a
47 forest, i.e. the sets can be ordered in a way where any
48 previous set is not intersected with given set or is a superset
49 of given set.
51 o *Pressure class* is a register class belonging to a set of
52 register classes containing all of the hard-registers available
53 for register allocation. The set of all pressure classes for a
54 target is defined in the corresponding machine-description file
55 according some criteria. Register pressure is calculated only
56 for pressure classes and it affects some IRA decisions as
57 forming allocation regions.
59 o *Allocno* represents the live range of a pseudo-register in a
60 region. Besides the obvious attributes like the corresponding
61 pseudo-register number, allocno class, conflicting allocnos and
62 conflicting hard-registers, there are a few allocno attributes
63 which are important for understanding the allocation algorithm:
65 - *Live ranges*. This is a list of ranges of *program points*
66 where the allocno lives. Program points represent places
67 where a pseudo can be born or become dead (there are
68 approximately two times more program points than the insns)
69 and they are represented by integers starting with 0. The
70 live ranges are used to find conflicts between allocnos.
71 They also play very important role for the transformation of
72 the IRA internal representation of several regions into a one
73 region representation. The later is used during the reload
74 pass work because each allocno represents all of the
75 corresponding pseudo-registers.
77 - *Hard-register costs*. This is a vector of size equal to the
78 number of available hard-registers of the allocno class. The
79 cost of a callee-clobbered hard-register for an allocno is
80 increased by the cost of save/restore code around the calls
81 through the given allocno's life. If the allocno is a move
82 instruction operand and another operand is a hard-register of
83 the allocno class, the cost of the hard-register is decreased
84 by the move cost.
86 When an allocno is assigned, the hard-register with minimal
87 full cost is used. Initially, a hard-register's full cost is
88 the corresponding value from the hard-register's cost vector.
89 If the allocno is connected by a *copy* (see below) to
90 another allocno which has just received a hard-register, the
91 cost of the hard-register is decreased. Before choosing a
92 hard-register for an allocno, the allocno's current costs of
93 the hard-registers are modified by the conflict hard-register
94 costs of all of the conflicting allocnos which are not
95 assigned yet.
97 - *Conflict hard-register costs*. This is a vector of the same
98 size as the hard-register costs vector. To permit an
99 unassigned allocno to get a better hard-register, IRA uses
100 this vector to calculate the final full cost of the
101 available hard-registers. Conflict hard-register costs of an
102 unassigned allocno are also changed with a change of the
103 hard-register cost of the allocno when a copy involving the
104 allocno is processed as described above. This is done to
105 show other unassigned allocnos that a given allocno prefers
106 some hard-registers in order to remove the move instruction
107 corresponding to the copy.
109 o *Cap*. If a pseudo-register does not live in a region but
110 lives in a nested region, IRA creates a special allocno called
111 a cap in the outer region. A region cap is also created for a
112 subregion cap.
114 o *Copy*. Allocnos can be connected by copies. Copies are used
115 to modify hard-register costs for allocnos during coloring.
116 Such modifications reflects a preference to use the same
117 hard-register for the allocnos connected by copies. Usually
118 copies are created for move insns (in this case it results in
119 register coalescing). But IRA also creates copies for operands
120 of an insn which should be assigned to the same hard-register
121 due to constraints in the machine description (it usually
122 results in removing a move generated in reload to satisfy
123 the constraints) and copies referring to the allocno which is
124 the output operand of an instruction and the allocno which is
125 an input operand dying in the instruction (creation of such
126 copies results in less register shuffling). IRA *does not*
127 create copies between the same register allocnos from different
128 regions because we use another technique for propagating
129 hard-register preference on the borders of regions.
131 Allocnos (including caps) for the upper region in the region tree
132 *accumulate* information important for coloring from allocnos with
133 the same pseudo-register from nested regions. This includes
134 hard-register and memory costs, conflicts with hard-registers,
135 allocno conflicts, allocno copies and more. *Thus, attributes for
136 allocnos in a region have the same values as if the region had no
137 subregions*. It means that attributes for allocnos in the
138 outermost region corresponding to the function have the same values
139 as though the allocation used only one region which is the entire
140 function. It also means that we can look at IRA work as if the
141 first IRA did allocation for all function then it improved the
142 allocation for loops then their subloops and so on.
144 IRA major passes are:
146 o Building IRA internal representation which consists of the
147 following subpasses:
149 * First, IRA builds regions and creates allocnos (file
150 ira-build.c) and initializes most of their attributes.
152 * Then IRA finds an allocno class for each allocno and
153 calculates its initial (non-accumulated) cost of memory and
154 each hard-register of its allocno class (file ira-cost.c).
156 * IRA creates live ranges of each allocno, calulates register
157 pressure for each pressure class in each region, sets up
158 conflict hard registers for each allocno and info about calls
159 the allocno lives through (file ira-lives.c).
161 * IRA removes low register pressure loops from the regions
162 mostly to speed IRA up (file ira-build.c).
164 * IRA propagates accumulated allocno info from lower region
165 allocnos to corresponding upper region allocnos (file
166 ira-build.c).
168 * IRA creates all caps (file ira-build.c).
170 * Having live-ranges of allocnos and their classes, IRA creates
171 conflicting allocnos for each allocno. Conflicting allocnos
172 are stored as a bit vector or array of pointers to the
173 conflicting allocnos whatever is more profitable (file
174 ira-conflicts.c). At this point IRA creates allocno copies.
176 o Coloring. Now IRA has all necessary info to start graph coloring
177 process. It is done in each region on top-down traverse of the
178 region tree (file ira-color.c). There are following subpasses:
180 * Finding profitable hard registers of corresponding allocno
181 class for each allocno. For example, only callee-saved hard
182 registers are frequently profitable for allocnos living
183 through colors. If the profitable hard register set of
184 allocno does not form a tree based on subset relation, we use
185 some approximation to form the tree. This approximation is
186 used to figure out trivial colorability of allocnos. The
187 approximation is a pretty rare case.
189 * Putting allocnos onto the coloring stack. IRA uses Briggs
190 optimistic coloring which is a major improvement over
191 Chaitin's coloring. Therefore IRA does not spill allocnos at
192 this point. There is some freedom in the order of putting
193 allocnos on the stack which can affect the final result of
194 the allocation. IRA uses some heuristics to improve the
195 order.
197 We also use a modification of Chaitin-Briggs algorithm which
198 works for intersected register classes of allocnos. To
199 figure out trivial colorability of allocnos, the mentioned
200 above tree of hard register sets is used. To get an idea how
201 the algorithm works in i386 example, let us consider an
202 allocno to which any general hard register can be assigned.
203 If the allocno conflicts with eight allocnos to which only
204 EAX register can be assigned, given allocno is still
205 trivially colorable because all conflicting allocnos might be
206 assigned only to EAX and all other general hard registers are
207 still free.
209 To get an idea of the used trivial colorability criterion, it
210 is also useful to read article "Graph-Coloring Register
211 Allocation for Irregular Architectures" by Michael D. Smith
212 and Glen Holloway. Major difference between the article
213 approach and approach used in IRA is that Smith's approach
214 takes register classes only from machine description and IRA
215 calculate register classes from intermediate code too
216 (e.g. an explicit usage of hard registers in RTL code for
217 parameter passing can result in creation of additional
218 register classes which contain or exclude the hard
219 registers). That makes IRA approach useful for improving
220 coloring even for architectures with regular register files
221 and in fact some benchmarking shows the improvement for
222 regular class architectures is even bigger than for irregular
223 ones. Another difference is that Smith's approach chooses
224 intersection of classes of all insn operands in which a given
225 pseudo occurs. IRA can use bigger classes if it is still
226 more profitable than memory usage.
228 * Popping the allocnos from the stack and assigning them hard
229 registers. If IRA can not assign a hard register to an
230 allocno and the allocno is coalesced, IRA undoes the
231 coalescing and puts the uncoalesced allocnos onto the stack in
232 the hope that some such allocnos will get a hard register
233 separately. If IRA fails to assign hard register or memory
234 is more profitable for it, IRA spills the allocno. IRA
235 assigns the allocno the hard-register with minimal full
236 allocation cost which reflects the cost of usage of the
237 hard-register for the allocno and cost of usage of the
238 hard-register for allocnos conflicting with given allocno.
240 * Chaitin-Briggs coloring assigns as many pseudos as possible
241 to hard registers. After coloringh we try to improve
242 allocation with cost point of view. We improve the
243 allocation by spilling some allocnos and assigning the freed
244 hard registers to other allocnos if it decreases the overall
245 allocation cost.
247 * After allono assigning in the region, IRA modifies the hard
248 register and memory costs for the corresponding allocnos in
249 the subregions to reflect the cost of possible loads, stores,
250 or moves on the border of the region and its subregions.
251 When default regional allocation algorithm is used
252 (-fira-algorithm=mixed), IRA just propagates the assignment
253 for allocnos if the register pressure in the region for the
254 corresponding pressure class is less than number of available
255 hard registers for given pressure class.
257 o Spill/restore code moving. When IRA performs an allocation
258 by traversing regions in top-down order, it does not know what
259 happens below in the region tree. Therefore, sometimes IRA
260 misses opportunities to perform a better allocation. A simple
261 optimization tries to improve allocation in a region having
262 subregions and containing in another region. If the
263 corresponding allocnos in the subregion are spilled, it spills
264 the region allocno if it is profitable. The optimization
265 implements a simple iterative algorithm performing profitable
266 transformations while they are still possible. It is fast in
267 practice, so there is no real need for a better time complexity
268 algorithm.
270 o Code change. After coloring, two allocnos representing the
271 same pseudo-register outside and inside a region respectively
272 may be assigned to different locations (hard-registers or
273 memory). In this case IRA creates and uses a new
274 pseudo-register inside the region and adds code to move allocno
275 values on the region's borders. This is done during top-down
276 traversal of the regions (file ira-emit.c). In some
277 complicated cases IRA can create a new allocno to move allocno
278 values (e.g. when a swap of values stored in two hard-registers
279 is needed). At this stage, the new allocno is marked as
280 spilled. IRA still creates the pseudo-register and the moves
281 on the region borders even when both allocnos were assigned to
282 the same hard-register. If the reload pass spills a
283 pseudo-register for some reason, the effect will be smaller
284 because another allocno will still be in the hard-register. In
285 most cases, this is better then spilling both allocnos. If
286 reload does not change the allocation for the two
287 pseudo-registers, the trivial move will be removed by
288 post-reload optimizations. IRA does not generate moves for
289 allocnos assigned to the same hard register when the default
290 regional allocation algorithm is used and the register pressure
291 in the region for the corresponding pressure class is less than
292 number of available hard registers for given pressure class.
293 IRA also does some optimizations to remove redundant stores and
294 to reduce code duplication on the region borders.
296 o Flattening internal representation. After changing code, IRA
297 transforms its internal representation for several regions into
298 one region representation (file ira-build.c). This process is
299 called IR flattening. Such process is more complicated than IR
300 rebuilding would be, but is much faster.
302 o After IR flattening, IRA tries to assign hard registers to all
303 spilled allocnos. This is impelemented by a simple and fast
304 priority coloring algorithm (see function
305 ira_reassign_conflict_allocnos::ira-color.c). Here new allocnos
306 created during the code change pass can be assigned to hard
307 registers.
309 o At the end IRA calls the reload pass. The reload pass
310 communicates with IRA through several functions in file
311 ira-color.c to improve its decisions in
313 * sharing stack slots for the spilled pseudos based on IRA info
314 about pseudo-register conflicts.
316 * reassigning hard-registers to all spilled pseudos at the end
317 of each reload iteration.
319 * choosing a better hard-register to spill based on IRA info
320 about pseudo-register live ranges and the register pressure
321 in places where the pseudo-register lives.
323 IRA uses a lot of data representing the target processors. These
324 data are initilized in file ira.c.
326 If function has no loops (or the loops are ignored when
327 -fira-algorithm=CB is used), we have classic Chaitin-Briggs
328 coloring (only instead of separate pass of coalescing, we use hard
329 register preferencing). In such case, IRA works much faster
330 because many things are not made (like IR flattening, the
331 spill/restore optimization, and the code change).
333 Literature is worth to read for better understanding the code:
335 o Preston Briggs, Keith D. Cooper, Linda Torczon. Improvements to
336 Graph Coloring Register Allocation.
338 o David Callahan, Brian Koblenz. Register allocation via
339 hierarchical graph coloring.
341 o Keith Cooper, Anshuman Dasgupta, Jason Eckhardt. Revisiting Graph
342 Coloring Register Allocation: A Study of the Chaitin-Briggs and
343 Callahan-Koblenz Algorithms.
345 o Guei-Yuan Lueh, Thomas Gross, and Ali-Reza Adl-Tabatabai. Global
346 Register Allocation Based on Graph Fusion.
348 o Michael D. Smith and Glenn Holloway. Graph-Coloring Register
349 Allocation for Irregular Architectures
351 o Vladimir Makarov. The Integrated Register Allocator for GCC.
353 o Vladimir Makarov. The top-down register allocator for irregular
354 register file architectures.
359 #include "config.h"
360 #include "system.h"
361 #include "coretypes.h"
362 #include "tm.h"
363 #include "regs.h"
364 #include "rtl.h"
365 #include "tm_p.h"
366 #include "target.h"
367 #include "flags.h"
368 #include "obstack.h"
369 #include "bitmap.h"
370 #include "hard-reg-set.h"
371 #include "basic-block.h"
372 #include "df.h"
373 #include "expr.h"
374 #include "recog.h"
375 #include "params.h"
376 #include "tree-pass.h"
377 #include "output.h"
378 #include "except.h"
379 #include "reload.h"
380 #include "diagnostic-core.h"
381 #include "function.h"
382 #include "ggc.h"
383 #include "ira-int.h"
384 #include "lra.h"
385 #include "dce.h"
386 #include "dbgcnt.h"
388 struct target_ira default_target_ira;
389 struct target_ira_int default_target_ira_int;
390 #if SWITCHABLE_TARGET
391 struct target_ira *this_target_ira = &default_target_ira;
392 struct target_ira_int *this_target_ira_int = &default_target_ira_int;
393 #endif
395 /* A modified value of flag `-fira-verbose' used internally. */
396 int internal_flag_ira_verbose;
398 /* Dump file of the allocator if it is not NULL. */
399 FILE *ira_dump_file;
401 /* The number of elements in the following array. */
402 int ira_spilled_reg_stack_slots_num;
404 /* The following array contains info about spilled pseudo-registers
405 stack slots used in current function so far. */
406 struct ira_spilled_reg_stack_slot *ira_spilled_reg_stack_slots;
408 /* Correspondingly overall cost of the allocation, overall cost before
409 reload, cost of the allocnos assigned to hard-registers, cost of
410 the allocnos assigned to memory, cost of loads, stores and register
411 move insns generated for pseudo-register live range splitting (see
412 ira-emit.c). */
413 int ira_overall_cost, overall_cost_before;
414 int ira_reg_cost, ira_mem_cost;
415 int ira_load_cost, ira_store_cost, ira_shuffle_cost;
416 int ira_move_loops_num, ira_additional_jumps_num;
418 /* All registers that can be eliminated. */
420 HARD_REG_SET eliminable_regset;
422 /* Value of max_reg_num () before IRA work start. This value helps
423 us to recognize a situation when new pseudos were created during
424 IRA work. */
425 static int max_regno_before_ira;
427 /* Temporary hard reg set used for a different calculation. */
428 static HARD_REG_SET temp_hard_regset;
430 #define last_mode_for_init_move_cost \
431 (this_target_ira_int->x_last_mode_for_init_move_cost)
434 /* The function sets up the map IRA_REG_MODE_HARD_REGSET. */
435 static void
436 setup_reg_mode_hard_regset (void)
438 int i, m, hard_regno;
440 for (m = 0; m < NUM_MACHINE_MODES; m++)
441 for (hard_regno = 0; hard_regno < FIRST_PSEUDO_REGISTER; hard_regno++)
443 CLEAR_HARD_REG_SET (ira_reg_mode_hard_regset[hard_regno][m]);
444 for (i = hard_regno_nregs[hard_regno][m] - 1; i >= 0; i--)
445 if (hard_regno + i < FIRST_PSEUDO_REGISTER)
446 SET_HARD_REG_BIT (ira_reg_mode_hard_regset[hard_regno][m],
447 hard_regno + i);
452 #define no_unit_alloc_regs \
453 (this_target_ira_int->x_no_unit_alloc_regs)
455 /* The function sets up the three arrays declared above. */
456 static void
457 setup_class_hard_regs (void)
459 int cl, i, hard_regno, n;
460 HARD_REG_SET processed_hard_reg_set;
462 ira_assert (SHRT_MAX >= FIRST_PSEUDO_REGISTER);
463 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
465 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
466 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
467 CLEAR_HARD_REG_SET (processed_hard_reg_set);
468 for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
470 ira_non_ordered_class_hard_regs[cl][i] = -1;
471 ira_class_hard_reg_index[cl][i] = -1;
473 for (n = 0, i = 0; i < FIRST_PSEUDO_REGISTER; i++)
475 #ifdef REG_ALLOC_ORDER
476 hard_regno = reg_alloc_order[i];
477 #else
478 hard_regno = i;
479 #endif
480 if (TEST_HARD_REG_BIT (processed_hard_reg_set, hard_regno))
481 continue;
482 SET_HARD_REG_BIT (processed_hard_reg_set, hard_regno);
483 if (! TEST_HARD_REG_BIT (temp_hard_regset, hard_regno))
484 ira_class_hard_reg_index[cl][hard_regno] = -1;
485 else
487 ira_class_hard_reg_index[cl][hard_regno] = n;
488 ira_class_hard_regs[cl][n++] = hard_regno;
491 ira_class_hard_regs_num[cl] = n;
492 for (n = 0, i = 0; i < FIRST_PSEUDO_REGISTER; i++)
493 if (TEST_HARD_REG_BIT (temp_hard_regset, i))
494 ira_non_ordered_class_hard_regs[cl][n++] = i;
495 ira_assert (ira_class_hard_regs_num[cl] == n);
499 /* Set up global variables defining info about hard registers for the
500 allocation. These depend on USE_HARD_FRAME_P whose TRUE value means
501 that we can use the hard frame pointer for the allocation. */
502 static void
503 setup_alloc_regs (bool use_hard_frame_p)
505 #ifdef ADJUST_REG_ALLOC_ORDER
506 ADJUST_REG_ALLOC_ORDER;
507 #endif
508 COPY_HARD_REG_SET (no_unit_alloc_regs, fixed_reg_set);
509 if (! use_hard_frame_p)
510 SET_HARD_REG_BIT (no_unit_alloc_regs, HARD_FRAME_POINTER_REGNUM);
511 setup_class_hard_regs ();
516 #define alloc_reg_class_subclasses \
517 (this_target_ira_int->x_alloc_reg_class_subclasses)
519 /* Initialize the table of subclasses of each reg class. */
520 static void
521 setup_reg_subclasses (void)
523 int i, j;
524 HARD_REG_SET temp_hard_regset2;
526 for (i = 0; i < N_REG_CLASSES; i++)
527 for (j = 0; j < N_REG_CLASSES; j++)
528 alloc_reg_class_subclasses[i][j] = LIM_REG_CLASSES;
530 for (i = 0; i < N_REG_CLASSES; i++)
532 if (i == (int) NO_REGS)
533 continue;
535 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[i]);
536 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
537 if (hard_reg_set_empty_p (temp_hard_regset))
538 continue;
539 for (j = 0; j < N_REG_CLASSES; j++)
540 if (i != j)
542 enum reg_class *p;
544 COPY_HARD_REG_SET (temp_hard_regset2, reg_class_contents[j]);
545 AND_COMPL_HARD_REG_SET (temp_hard_regset2, no_unit_alloc_regs);
546 if (! hard_reg_set_subset_p (temp_hard_regset,
547 temp_hard_regset2))
548 continue;
549 p = &alloc_reg_class_subclasses[j][0];
550 while (*p != LIM_REG_CLASSES) p++;
551 *p = (enum reg_class) i;
558 /* Set up IRA_MEMORY_MOVE_COST and IRA_MAX_MEMORY_MOVE_COST. */
559 static void
560 setup_class_subset_and_memory_move_costs (void)
562 int cl, cl2, mode, cost;
563 HARD_REG_SET temp_hard_regset2;
565 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
566 ira_memory_move_cost[mode][NO_REGS][0]
567 = ira_memory_move_cost[mode][NO_REGS][1] = SHRT_MAX;
568 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
570 if (cl != (int) NO_REGS)
571 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
573 ira_max_memory_move_cost[mode][cl][0]
574 = ira_memory_move_cost[mode][cl][0]
575 = memory_move_cost ((enum machine_mode) mode,
576 (reg_class_t) cl, false);
577 ira_max_memory_move_cost[mode][cl][1]
578 = ira_memory_move_cost[mode][cl][1]
579 = memory_move_cost ((enum machine_mode) mode,
580 (reg_class_t) cl, true);
581 /* Costs for NO_REGS are used in cost calculation on the
582 1st pass when the preferred register classes are not
583 known yet. In this case we take the best scenario. */
584 if (ira_memory_move_cost[mode][NO_REGS][0]
585 > ira_memory_move_cost[mode][cl][0])
586 ira_max_memory_move_cost[mode][NO_REGS][0]
587 = ira_memory_move_cost[mode][NO_REGS][0]
588 = ira_memory_move_cost[mode][cl][0];
589 if (ira_memory_move_cost[mode][NO_REGS][1]
590 > ira_memory_move_cost[mode][cl][1])
591 ira_max_memory_move_cost[mode][NO_REGS][1]
592 = ira_memory_move_cost[mode][NO_REGS][1]
593 = ira_memory_move_cost[mode][cl][1];
596 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
597 for (cl2 = (int) N_REG_CLASSES - 1; cl2 >= 0; cl2--)
599 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
600 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
601 COPY_HARD_REG_SET (temp_hard_regset2, reg_class_contents[cl2]);
602 AND_COMPL_HARD_REG_SET (temp_hard_regset2, no_unit_alloc_regs);
603 ira_class_subset_p[cl][cl2]
604 = hard_reg_set_subset_p (temp_hard_regset, temp_hard_regset2);
605 if (! hard_reg_set_empty_p (temp_hard_regset2)
606 && hard_reg_set_subset_p (reg_class_contents[cl2],
607 reg_class_contents[cl]))
608 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
610 cost = ira_memory_move_cost[mode][cl2][0];
611 if (cost > ira_max_memory_move_cost[mode][cl][0])
612 ira_max_memory_move_cost[mode][cl][0] = cost;
613 cost = ira_memory_move_cost[mode][cl2][1];
614 if (cost > ira_max_memory_move_cost[mode][cl][1])
615 ira_max_memory_move_cost[mode][cl][1] = cost;
618 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
619 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
621 ira_memory_move_cost[mode][cl][0]
622 = ira_max_memory_move_cost[mode][cl][0];
623 ira_memory_move_cost[mode][cl][1]
624 = ira_max_memory_move_cost[mode][cl][1];
626 setup_reg_subclasses ();
631 /* Define the following macro if allocation through malloc if
632 preferable. */
633 #define IRA_NO_OBSTACK
635 #ifndef IRA_NO_OBSTACK
636 /* Obstack used for storing all dynamic data (except bitmaps) of the
637 IRA. */
638 static struct obstack ira_obstack;
639 #endif
641 /* Obstack used for storing all bitmaps of the IRA. */
642 static struct bitmap_obstack ira_bitmap_obstack;
644 /* Allocate memory of size LEN for IRA data. */
645 void *
646 ira_allocate (size_t len)
648 void *res;
650 #ifndef IRA_NO_OBSTACK
651 res = obstack_alloc (&ira_obstack, len);
652 #else
653 res = xmalloc (len);
654 #endif
655 return res;
658 /* Free memory ADDR allocated for IRA data. */
659 void
660 ira_free (void *addr ATTRIBUTE_UNUSED)
662 #ifndef IRA_NO_OBSTACK
663 /* do nothing */
664 #else
665 free (addr);
666 #endif
670 /* Allocate and returns bitmap for IRA. */
671 bitmap
672 ira_allocate_bitmap (void)
674 return BITMAP_ALLOC (&ira_bitmap_obstack);
677 /* Free bitmap B allocated for IRA. */
678 void
679 ira_free_bitmap (bitmap b ATTRIBUTE_UNUSED)
681 /* do nothing */
686 /* Output information about allocation of all allocnos (except for
687 caps) into file F. */
688 void
689 ira_print_disposition (FILE *f)
691 int i, n, max_regno;
692 ira_allocno_t a;
693 basic_block bb;
695 fprintf (f, "Disposition:");
696 max_regno = max_reg_num ();
697 for (n = 0, i = FIRST_PSEUDO_REGISTER; i < max_regno; i++)
698 for (a = ira_regno_allocno_map[i];
699 a != NULL;
700 a = ALLOCNO_NEXT_REGNO_ALLOCNO (a))
702 if (n % 4 == 0)
703 fprintf (f, "\n");
704 n++;
705 fprintf (f, " %4d:r%-4d", ALLOCNO_NUM (a), ALLOCNO_REGNO (a));
706 if ((bb = ALLOCNO_LOOP_TREE_NODE (a)->bb) != NULL)
707 fprintf (f, "b%-3d", bb->index);
708 else
709 fprintf (f, "l%-3d", ALLOCNO_LOOP_TREE_NODE (a)->loop_num);
710 if (ALLOCNO_HARD_REGNO (a) >= 0)
711 fprintf (f, " %3d", ALLOCNO_HARD_REGNO (a));
712 else
713 fprintf (f, " mem");
715 fprintf (f, "\n");
718 /* Outputs information about allocation of all allocnos into
719 stderr. */
720 void
721 ira_debug_disposition (void)
723 ira_print_disposition (stderr);
728 /* Set up ira_stack_reg_pressure_class which is the biggest pressure
729 register class containing stack registers or NO_REGS if there are
730 no stack registers. To find this class, we iterate through all
731 register pressure classes and choose the first register pressure
732 class containing all the stack registers and having the biggest
733 size. */
734 static void
735 setup_stack_reg_pressure_class (void)
737 ira_stack_reg_pressure_class = NO_REGS;
738 #ifdef STACK_REGS
740 int i, best, size;
741 enum reg_class cl;
742 HARD_REG_SET temp_hard_regset2;
744 CLEAR_HARD_REG_SET (temp_hard_regset);
745 for (i = FIRST_STACK_REG; i <= LAST_STACK_REG; i++)
746 SET_HARD_REG_BIT (temp_hard_regset, i);
747 best = 0;
748 for (i = 0; i < ira_pressure_classes_num; i++)
750 cl = ira_pressure_classes[i];
751 COPY_HARD_REG_SET (temp_hard_regset2, temp_hard_regset);
752 AND_HARD_REG_SET (temp_hard_regset2, reg_class_contents[cl]);
753 size = hard_reg_set_size (temp_hard_regset2);
754 if (best < size)
756 best = size;
757 ira_stack_reg_pressure_class = cl;
761 #endif
764 /* Find pressure classes which are register classes for which we
765 calculate register pressure in IRA, register pressure sensitive
766 insn scheduling, and register pressure sensitive loop invariant
767 motion.
769 To make register pressure calculation easy, we always use
770 non-intersected register pressure classes. A move of hard
771 registers from one register pressure class is not more expensive
772 than load and store of the hard registers. Most likely an allocno
773 class will be a subset of a register pressure class and in many
774 cases a register pressure class. That makes usage of register
775 pressure classes a good approximation to find a high register
776 pressure. */
777 static void
778 setup_pressure_classes (void)
780 int cost, i, n, curr;
781 int cl, cl2;
782 enum reg_class pressure_classes[N_REG_CLASSES];
783 int m;
784 HARD_REG_SET temp_hard_regset2;
785 bool insert_p;
787 n = 0;
788 for (cl = 0; cl < N_REG_CLASSES; cl++)
790 if (ira_class_hard_regs_num[cl] == 0)
791 continue;
792 if (ira_class_hard_regs_num[cl] != 1
793 /* A register class without subclasses may contain a few
794 hard registers and movement between them is costly
795 (e.g. SPARC FPCC registers). We still should consider it
796 as a candidate for a pressure class. */
797 && alloc_reg_class_subclasses[cl][0] < cl)
799 /* Check that the moves between any hard registers of the
800 current class are not more expensive for a legal mode
801 than load/store of the hard registers of the current
802 class. Such class is a potential candidate to be a
803 register pressure class. */
804 for (m = 0; m < NUM_MACHINE_MODES; m++)
806 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
807 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
808 AND_COMPL_HARD_REG_SET (temp_hard_regset,
809 ira_prohibited_class_mode_regs[cl][m]);
810 if (hard_reg_set_empty_p (temp_hard_regset))
811 continue;
812 ira_init_register_move_cost_if_necessary ((enum machine_mode) m);
813 cost = ira_register_move_cost[m][cl][cl];
814 if (cost <= ira_max_memory_move_cost[m][cl][1]
815 || cost <= ira_max_memory_move_cost[m][cl][0])
816 break;
818 if (m >= NUM_MACHINE_MODES)
819 continue;
821 curr = 0;
822 insert_p = true;
823 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
824 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
825 /* Remove so far added pressure classes which are subset of the
826 current candidate class. Prefer GENERAL_REGS as a pressure
827 register class to another class containing the same
828 allocatable hard registers. We do this because machine
829 dependent cost hooks might give wrong costs for the latter
830 class but always give the right cost for the former class
831 (GENERAL_REGS). */
832 for (i = 0; i < n; i++)
834 cl2 = pressure_classes[i];
835 COPY_HARD_REG_SET (temp_hard_regset2, reg_class_contents[cl2]);
836 AND_COMPL_HARD_REG_SET (temp_hard_regset2, no_unit_alloc_regs);
837 if (hard_reg_set_subset_p (temp_hard_regset, temp_hard_regset2)
838 && (! hard_reg_set_equal_p (temp_hard_regset, temp_hard_regset2)
839 || cl2 == (int) GENERAL_REGS))
841 pressure_classes[curr++] = (enum reg_class) cl2;
842 insert_p = false;
843 continue;
845 if (hard_reg_set_subset_p (temp_hard_regset2, temp_hard_regset)
846 && (! hard_reg_set_equal_p (temp_hard_regset2, temp_hard_regset)
847 || cl == (int) GENERAL_REGS))
848 continue;
849 if (hard_reg_set_equal_p (temp_hard_regset2, temp_hard_regset))
850 insert_p = false;
851 pressure_classes[curr++] = (enum reg_class) cl2;
853 /* If the current candidate is a subset of a so far added
854 pressure class, don't add it to the list of the pressure
855 classes. */
856 if (insert_p)
857 pressure_classes[curr++] = (enum reg_class) cl;
858 n = curr;
860 #ifdef ENABLE_IRA_CHECKING
862 HARD_REG_SET ignore_hard_regs;
864 /* Check pressure classes correctness: here we check that hard
865 registers from all register pressure classes contains all hard
866 registers available for the allocation. */
867 CLEAR_HARD_REG_SET (temp_hard_regset);
868 CLEAR_HARD_REG_SET (temp_hard_regset2);
869 COPY_HARD_REG_SET (ignore_hard_regs, no_unit_alloc_regs);
870 for (cl = 0; cl < LIM_REG_CLASSES; cl++)
872 /* For some targets (like MIPS with MD_REGS), there are some
873 classes with hard registers available for allocation but
874 not able to hold value of any mode. */
875 for (m = 0; m < NUM_MACHINE_MODES; m++)
876 if (contains_reg_of_mode[cl][m])
877 break;
878 if (m >= NUM_MACHINE_MODES)
880 IOR_HARD_REG_SET (ignore_hard_regs, reg_class_contents[cl]);
881 continue;
883 for (i = 0; i < n; i++)
884 if ((int) pressure_classes[i] == cl)
885 break;
886 IOR_HARD_REG_SET (temp_hard_regset2, reg_class_contents[cl]);
887 if (i < n)
888 IOR_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
890 for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
891 /* Some targets (like SPARC with ICC reg) have alocatable regs
892 for which no reg class is defined. */
893 if (REGNO_REG_CLASS (i) == NO_REGS)
894 SET_HARD_REG_BIT (ignore_hard_regs, i);
895 AND_COMPL_HARD_REG_SET (temp_hard_regset, ignore_hard_regs);
896 AND_COMPL_HARD_REG_SET (temp_hard_regset2, ignore_hard_regs);
897 ira_assert (hard_reg_set_subset_p (temp_hard_regset2, temp_hard_regset));
899 #endif
900 ira_pressure_classes_num = 0;
901 for (i = 0; i < n; i++)
903 cl = (int) pressure_classes[i];
904 ira_reg_pressure_class_p[cl] = true;
905 ira_pressure_classes[ira_pressure_classes_num++] = (enum reg_class) cl;
907 setup_stack_reg_pressure_class ();
910 /* Set up IRA_UNIFORM_CLASS_P. Uniform class is a register class
911 whose register move cost between any registers of the class is the
912 same as for all its subclasses. We use the data to speed up the
913 2nd pass of calculations of allocno costs. */
914 static void
915 setup_uniform_class_p (void)
917 int i, cl, cl2, m;
919 for (cl = 0; cl < N_REG_CLASSES; cl++)
921 ira_uniform_class_p[cl] = false;
922 if (ira_class_hard_regs_num[cl] == 0)
923 continue;
924 /* We can not use alloc_reg_class_subclasses here because move
925 cost hooks does not take into account that some registers are
926 unavailable for the subtarget. E.g. for i686, INT_SSE_REGS
927 is element of alloc_reg_class_subclasses for GENERAL_REGS
928 because SSE regs are unavailable. */
929 for (i = 0; (cl2 = reg_class_subclasses[cl][i]) != LIM_REG_CLASSES; i++)
931 if (ira_class_hard_regs_num[cl2] == 0)
932 continue;
933 for (m = 0; m < NUM_MACHINE_MODES; m++)
934 if (contains_reg_of_mode[cl][m] && contains_reg_of_mode[cl2][m])
936 ira_init_register_move_cost_if_necessary ((enum machine_mode) m);
937 if (ira_register_move_cost[m][cl][cl]
938 != ira_register_move_cost[m][cl2][cl2])
939 break;
941 if (m < NUM_MACHINE_MODES)
942 break;
944 if (cl2 == LIM_REG_CLASSES)
945 ira_uniform_class_p[cl] = true;
949 /* Set up IRA_ALLOCNO_CLASSES, IRA_ALLOCNO_CLASSES_NUM,
950 IRA_IMPORTANT_CLASSES, and IRA_IMPORTANT_CLASSES_NUM.
952 Target may have many subtargets and not all target hard regiters can
953 be used for allocation, e.g. x86 port in 32-bit mode can not use
954 hard registers introduced in x86-64 like r8-r15). Some classes
955 might have the same allocatable hard registers, e.g. INDEX_REGS
956 and GENERAL_REGS in x86 port in 32-bit mode. To decrease different
957 calculations efforts we introduce allocno classes which contain
958 unique non-empty sets of allocatable hard-registers.
960 Pseudo class cost calculation in ira-costs.c is very expensive.
961 Therefore we are trying to decrease number of classes involved in
962 such calculation. Register classes used in the cost calculation
963 are called important classes. They are allocno classes and other
964 non-empty classes whose allocatable hard register sets are inside
965 of an allocno class hard register set. From the first sight, it
966 looks like that they are just allocno classes. It is not true. In
967 example of x86-port in 32-bit mode, allocno classes will contain
968 GENERAL_REGS but not LEGACY_REGS (because allocatable hard
969 registers are the same for the both classes). The important
970 classes will contain GENERAL_REGS and LEGACY_REGS. It is done
971 because a machine description insn constraint may refers for
972 LEGACY_REGS and code in ira-costs.c is mostly base on investigation
973 of the insn constraints. */
974 static void
975 setup_allocno_and_important_classes (void)
977 int i, j, n, cl;
978 bool set_p;
979 HARD_REG_SET temp_hard_regset2;
980 static enum reg_class classes[LIM_REG_CLASSES + 1];
982 n = 0;
983 /* Collect classes which contain unique sets of allocatable hard
984 registers. Prefer GENERAL_REGS to other classes containing the
985 same set of hard registers. */
986 for (i = 0; i < LIM_REG_CLASSES; i++)
988 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[i]);
989 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
990 for (j = 0; j < n; j++)
992 cl = classes[j];
993 COPY_HARD_REG_SET (temp_hard_regset2, reg_class_contents[cl]);
994 AND_COMPL_HARD_REG_SET (temp_hard_regset2,
995 no_unit_alloc_regs);
996 if (hard_reg_set_equal_p (temp_hard_regset,
997 temp_hard_regset2))
998 break;
1000 if (j >= n)
1001 classes[n++] = (enum reg_class) i;
1002 else if (i == GENERAL_REGS)
1003 /* Prefer general regs. For i386 example, it means that
1004 we prefer GENERAL_REGS over INDEX_REGS or LEGACY_REGS
1005 (all of them consists of the same available hard
1006 registers). */
1007 classes[j] = (enum reg_class) i;
1009 classes[n] = LIM_REG_CLASSES;
1011 /* Set up classes which can be used for allocnos as classes
1012 conatining non-empty unique sets of allocatable hard
1013 registers. */
1014 ira_allocno_classes_num = 0;
1015 for (i = 0; (cl = classes[i]) != LIM_REG_CLASSES; i++)
1016 if (ira_class_hard_regs_num[cl] > 0)
1017 ira_allocno_classes[ira_allocno_classes_num++] = (enum reg_class) cl;
1018 ira_important_classes_num = 0;
1019 /* Add non-allocno classes containing to non-empty set of
1020 allocatable hard regs. */
1021 for (cl = 0; cl < N_REG_CLASSES; cl++)
1022 if (ira_class_hard_regs_num[cl] > 0)
1024 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
1025 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
1026 set_p = false;
1027 for (j = 0; j < ira_allocno_classes_num; j++)
1029 COPY_HARD_REG_SET (temp_hard_regset2,
1030 reg_class_contents[ira_allocno_classes[j]]);
1031 AND_COMPL_HARD_REG_SET (temp_hard_regset2, no_unit_alloc_regs);
1032 if ((enum reg_class) cl == ira_allocno_classes[j])
1033 break;
1034 else if (hard_reg_set_subset_p (temp_hard_regset,
1035 temp_hard_regset2))
1036 set_p = true;
1038 if (set_p && j >= ira_allocno_classes_num)
1039 ira_important_classes[ira_important_classes_num++]
1040 = (enum reg_class) cl;
1042 /* Now add allocno classes to the important classes. */
1043 for (j = 0; j < ira_allocno_classes_num; j++)
1044 ira_important_classes[ira_important_classes_num++]
1045 = ira_allocno_classes[j];
1046 for (cl = 0; cl < N_REG_CLASSES; cl++)
1048 ira_reg_allocno_class_p[cl] = false;
1049 ira_reg_pressure_class_p[cl] = false;
1051 for (j = 0; j < ira_allocno_classes_num; j++)
1052 ira_reg_allocno_class_p[ira_allocno_classes[j]] = true;
1053 setup_pressure_classes ();
1054 setup_uniform_class_p ();
1057 /* Setup translation in CLASS_TRANSLATE of all classes into a class
1058 given by array CLASSES of length CLASSES_NUM. The function is used
1059 make translation any reg class to an allocno class or to an
1060 pressure class. This translation is necessary for some
1061 calculations when we can use only allocno or pressure classes and
1062 such translation represents an approximate representation of all
1063 classes.
1065 The translation in case when allocatable hard register set of a
1066 given class is subset of allocatable hard register set of a class
1067 in CLASSES is pretty simple. We use smallest classes from CLASSES
1068 containing a given class. If allocatable hard register set of a
1069 given class is not a subset of any corresponding set of a class
1070 from CLASSES, we use the cheapest (with load/store point of view)
1071 class from CLASSES whose set intersects with given class set */
1072 static void
1073 setup_class_translate_array (enum reg_class *class_translate,
1074 int classes_num, enum reg_class *classes)
1076 int cl, mode;
1077 enum reg_class aclass, best_class, *cl_ptr;
1078 int i, cost, min_cost, best_cost;
1080 for (cl = 0; cl < N_REG_CLASSES; cl++)
1081 class_translate[cl] = NO_REGS;
1083 for (i = 0; i < classes_num; i++)
1085 aclass = classes[i];
1086 for (cl_ptr = &alloc_reg_class_subclasses[aclass][0];
1087 (cl = *cl_ptr) != LIM_REG_CLASSES;
1088 cl_ptr++)
1089 if (class_translate[cl] == NO_REGS)
1090 class_translate[cl] = aclass;
1091 class_translate[aclass] = aclass;
1093 /* For classes which are not fully covered by one of given classes
1094 (in other words covered by more one given class), use the
1095 cheapest class. */
1096 for (cl = 0; cl < N_REG_CLASSES; cl++)
1098 if (cl == NO_REGS || class_translate[cl] != NO_REGS)
1099 continue;
1100 best_class = NO_REGS;
1101 best_cost = INT_MAX;
1102 for (i = 0; i < classes_num; i++)
1104 aclass = classes[i];
1105 COPY_HARD_REG_SET (temp_hard_regset,
1106 reg_class_contents[aclass]);
1107 AND_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
1108 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
1109 if (! hard_reg_set_empty_p (temp_hard_regset))
1111 min_cost = INT_MAX;
1112 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
1114 cost = (ira_memory_move_cost[mode][aclass][0]
1115 + ira_memory_move_cost[mode][aclass][1]);
1116 if (min_cost > cost)
1117 min_cost = cost;
1119 if (best_class == NO_REGS || best_cost > min_cost)
1121 best_class = aclass;
1122 best_cost = min_cost;
1126 class_translate[cl] = best_class;
1130 /* Set up array IRA_ALLOCNO_CLASS_TRANSLATE and
1131 IRA_PRESSURE_CLASS_TRANSLATE. */
1132 static void
1133 setup_class_translate (void)
1135 setup_class_translate_array (ira_allocno_class_translate,
1136 ira_allocno_classes_num, ira_allocno_classes);
1137 setup_class_translate_array (ira_pressure_class_translate,
1138 ira_pressure_classes_num, ira_pressure_classes);
1141 /* Order numbers of allocno classes in original target allocno class
1142 array, -1 for non-allocno classes. */
1143 static int allocno_class_order[N_REG_CLASSES];
1145 /* The function used to sort the important classes. */
1146 static int
1147 comp_reg_classes_func (const void *v1p, const void *v2p)
1149 enum reg_class cl1 = *(const enum reg_class *) v1p;
1150 enum reg_class cl2 = *(const enum reg_class *) v2p;
1151 enum reg_class tcl1, tcl2;
1152 int diff;
1154 tcl1 = ira_allocno_class_translate[cl1];
1155 tcl2 = ira_allocno_class_translate[cl2];
1156 if (tcl1 != NO_REGS && tcl2 != NO_REGS
1157 && (diff = allocno_class_order[tcl1] - allocno_class_order[tcl2]) != 0)
1158 return diff;
1159 return (int) cl1 - (int) cl2;
1162 /* For correct work of function setup_reg_class_relation we need to
1163 reorder important classes according to the order of their allocno
1164 classes. It places important classes containing the same
1165 allocatable hard register set adjacent to each other and allocno
1166 class with the allocatable hard register set right after the other
1167 important classes with the same set.
1169 In example from comments of function
1170 setup_allocno_and_important_classes, it places LEGACY_REGS and
1171 GENERAL_REGS close to each other and GENERAL_REGS is after
1172 LEGACY_REGS. */
1173 static void
1174 reorder_important_classes (void)
1176 int i;
1178 for (i = 0; i < N_REG_CLASSES; i++)
1179 allocno_class_order[i] = -1;
1180 for (i = 0; i < ira_allocno_classes_num; i++)
1181 allocno_class_order[ira_allocno_classes[i]] = i;
1182 qsort (ira_important_classes, ira_important_classes_num,
1183 sizeof (enum reg_class), comp_reg_classes_func);
1184 for (i = 0; i < ira_important_classes_num; i++)
1185 ira_important_class_nums[ira_important_classes[i]] = i;
1188 /* Set up IRA_REG_CLASS_SUBUNION, IRA_REG_CLASS_SUPERUNION,
1189 IRA_REG_CLASS_SUPER_CLASSES, IRA_REG_CLASSES_INTERSECT, and
1190 IRA_REG_CLASSES_INTERSECT_P. For the meaning of the relations,
1191 please see corresponding comments in ira-int.h. */
1192 static void
1193 setup_reg_class_relations (void)
1195 int i, cl1, cl2, cl3;
1196 HARD_REG_SET intersection_set, union_set, temp_set2;
1197 bool important_class_p[N_REG_CLASSES];
1199 memset (important_class_p, 0, sizeof (important_class_p));
1200 for (i = 0; i < ira_important_classes_num; i++)
1201 important_class_p[ira_important_classes[i]] = true;
1202 for (cl1 = 0; cl1 < N_REG_CLASSES; cl1++)
1204 ira_reg_class_super_classes[cl1][0] = LIM_REG_CLASSES;
1205 for (cl2 = 0; cl2 < N_REG_CLASSES; cl2++)
1207 ira_reg_classes_intersect_p[cl1][cl2] = false;
1208 ira_reg_class_intersect[cl1][cl2] = NO_REGS;
1209 ira_reg_class_subset[cl1][cl2] = NO_REGS;
1210 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl1]);
1211 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
1212 COPY_HARD_REG_SET (temp_set2, reg_class_contents[cl2]);
1213 AND_COMPL_HARD_REG_SET (temp_set2, no_unit_alloc_regs);
1214 if (hard_reg_set_empty_p (temp_hard_regset)
1215 && hard_reg_set_empty_p (temp_set2))
1217 /* The both classes have no allocatable hard registers
1218 -- take all class hard registers into account and use
1219 reg_class_subunion and reg_class_superunion. */
1220 for (i = 0;; i++)
1222 cl3 = reg_class_subclasses[cl1][i];
1223 if (cl3 == LIM_REG_CLASSES)
1224 break;
1225 if (reg_class_subset_p (ira_reg_class_intersect[cl1][cl2],
1226 (enum reg_class) cl3))
1227 ira_reg_class_intersect[cl1][cl2] = (enum reg_class) cl3;
1229 ira_reg_class_subunion[cl1][cl2] = reg_class_subunion[cl1][cl2];
1230 ira_reg_class_superunion[cl1][cl2] = reg_class_superunion[cl1][cl2];
1231 continue;
1233 ira_reg_classes_intersect_p[cl1][cl2]
1234 = hard_reg_set_intersect_p (temp_hard_regset, temp_set2);
1235 if (important_class_p[cl1] && important_class_p[cl2]
1236 && hard_reg_set_subset_p (temp_hard_regset, temp_set2))
1238 /* CL1 and CL2 are important classes and CL1 allocatable
1239 hard register set is inside of CL2 allocatable hard
1240 registers -- make CL1 a superset of CL2. */
1241 enum reg_class *p;
1243 p = &ira_reg_class_super_classes[cl1][0];
1244 while (*p != LIM_REG_CLASSES)
1245 p++;
1246 *p++ = (enum reg_class) cl2;
1247 *p = LIM_REG_CLASSES;
1249 ira_reg_class_subunion[cl1][cl2] = NO_REGS;
1250 ira_reg_class_superunion[cl1][cl2] = NO_REGS;
1251 COPY_HARD_REG_SET (intersection_set, reg_class_contents[cl1]);
1252 AND_HARD_REG_SET (intersection_set, reg_class_contents[cl2]);
1253 AND_COMPL_HARD_REG_SET (intersection_set, no_unit_alloc_regs);
1254 COPY_HARD_REG_SET (union_set, reg_class_contents[cl1]);
1255 IOR_HARD_REG_SET (union_set, reg_class_contents[cl2]);
1256 AND_COMPL_HARD_REG_SET (union_set, no_unit_alloc_regs);
1257 for (cl3 = 0; cl3 < N_REG_CLASSES; cl3++)
1259 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl3]);
1260 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
1261 if (hard_reg_set_subset_p (temp_hard_regset, intersection_set))
1263 /* CL3 allocatable hard register set is inside of
1264 intersection of allocatable hard register sets
1265 of CL1 and CL2. */
1266 if (important_class_p[cl3])
1268 COPY_HARD_REG_SET
1269 (temp_set2,
1270 reg_class_contents
1271 [(int) ira_reg_class_intersect[cl1][cl2]]);
1272 AND_COMPL_HARD_REG_SET (temp_set2, no_unit_alloc_regs);
1273 if (! hard_reg_set_subset_p (temp_hard_regset, temp_set2)
1274 /* If the allocatable hard register sets are
1275 the same, prefer GENERAL_REGS or the
1276 smallest class for debugging
1277 purposes. */
1278 || (hard_reg_set_equal_p (temp_hard_regset, temp_set2)
1279 && (cl3 == GENERAL_REGS
1280 || ((ira_reg_class_intersect[cl1][cl2]
1281 != GENERAL_REGS)
1282 && hard_reg_set_subset_p
1283 (reg_class_contents[cl3],
1284 reg_class_contents
1285 [(int)
1286 ira_reg_class_intersect[cl1][cl2]])))))
1287 ira_reg_class_intersect[cl1][cl2] = (enum reg_class) cl3;
1289 COPY_HARD_REG_SET
1290 (temp_set2,
1291 reg_class_contents[(int) ira_reg_class_subset[cl1][cl2]]);
1292 AND_COMPL_HARD_REG_SET (temp_set2, no_unit_alloc_regs);
1293 if (! hard_reg_set_subset_p (temp_hard_regset, temp_set2)
1294 /* Ignore unavailable hard registers and prefer
1295 smallest class for debugging purposes. */
1296 || (hard_reg_set_equal_p (temp_hard_regset, temp_set2)
1297 && hard_reg_set_subset_p
1298 (reg_class_contents[cl3],
1299 reg_class_contents
1300 [(int) ira_reg_class_subset[cl1][cl2]])))
1301 ira_reg_class_subset[cl1][cl2] = (enum reg_class) cl3;
1303 if (important_class_p[cl3]
1304 && hard_reg_set_subset_p (temp_hard_regset, union_set))
1306 /* CL3 allocatbale hard register set is inside of
1307 union of allocatable hard register sets of CL1
1308 and CL2. */
1309 COPY_HARD_REG_SET
1310 (temp_set2,
1311 reg_class_contents[(int) ira_reg_class_subunion[cl1][cl2]]);
1312 AND_COMPL_HARD_REG_SET (temp_set2, no_unit_alloc_regs);
1313 if (ira_reg_class_subunion[cl1][cl2] == NO_REGS
1314 || (hard_reg_set_subset_p (temp_set2, temp_hard_regset)
1316 && (! hard_reg_set_equal_p (temp_set2,
1317 temp_hard_regset)
1318 || cl3 == GENERAL_REGS
1319 /* If the allocatable hard register sets are the
1320 same, prefer GENERAL_REGS or the smallest
1321 class for debugging purposes. */
1322 || (ira_reg_class_subunion[cl1][cl2] != GENERAL_REGS
1323 && hard_reg_set_subset_p
1324 (reg_class_contents[cl3],
1325 reg_class_contents
1326 [(int) ira_reg_class_subunion[cl1][cl2]])))))
1327 ira_reg_class_subunion[cl1][cl2] = (enum reg_class) cl3;
1329 if (hard_reg_set_subset_p (union_set, temp_hard_regset))
1331 /* CL3 allocatable hard register set contains union
1332 of allocatable hard register sets of CL1 and
1333 CL2. */
1334 COPY_HARD_REG_SET
1335 (temp_set2,
1336 reg_class_contents[(int) ira_reg_class_superunion[cl1][cl2]]);
1337 AND_COMPL_HARD_REG_SET (temp_set2, no_unit_alloc_regs);
1338 if (ira_reg_class_superunion[cl1][cl2] == NO_REGS
1339 || (hard_reg_set_subset_p (temp_hard_regset, temp_set2)
1341 && (! hard_reg_set_equal_p (temp_set2,
1342 temp_hard_regset)
1343 || cl3 == GENERAL_REGS
1344 /* If the allocatable hard register sets are the
1345 same, prefer GENERAL_REGS or the smallest
1346 class for debugging purposes. */
1347 || (ira_reg_class_superunion[cl1][cl2] != GENERAL_REGS
1348 && hard_reg_set_subset_p
1349 (reg_class_contents[cl3],
1350 reg_class_contents
1351 [(int) ira_reg_class_superunion[cl1][cl2]])))))
1352 ira_reg_class_superunion[cl1][cl2] = (enum reg_class) cl3;
1359 /* Output all unifrom and important classes into file F. */
1360 static void
1361 print_unform_and_important_classes (FILE *f)
1363 static const char *const reg_class_names[] = REG_CLASS_NAMES;
1364 int i, cl;
1366 fprintf (f, "Uniform classes:\n");
1367 for (cl = 0; cl < N_REG_CLASSES; cl++)
1368 if (ira_uniform_class_p[cl])
1369 fprintf (f, " %s", reg_class_names[cl]);
1370 fprintf (f, "\nImportant classes:\n");
1371 for (i = 0; i < ira_important_classes_num; i++)
1372 fprintf (f, " %s", reg_class_names[ira_important_classes[i]]);
1373 fprintf (f, "\n");
1376 /* Output all possible allocno or pressure classes and their
1377 translation map into file F. */
1378 static void
1379 print_translated_classes (FILE *f, bool pressure_p)
1381 int classes_num = (pressure_p
1382 ? ira_pressure_classes_num : ira_allocno_classes_num);
1383 enum reg_class *classes = (pressure_p
1384 ? ira_pressure_classes : ira_allocno_classes);
1385 enum reg_class *class_translate = (pressure_p
1386 ? ira_pressure_class_translate
1387 : ira_allocno_class_translate);
1388 static const char *const reg_class_names[] = REG_CLASS_NAMES;
1389 int i;
1391 fprintf (f, "%s classes:\n", pressure_p ? "Pressure" : "Allocno");
1392 for (i = 0; i < classes_num; i++)
1393 fprintf (f, " %s", reg_class_names[classes[i]]);
1394 fprintf (f, "\nClass translation:\n");
1395 for (i = 0; i < N_REG_CLASSES; i++)
1396 fprintf (f, " %s -> %s\n", reg_class_names[i],
1397 reg_class_names[class_translate[i]]);
1400 /* Output all possible allocno and translation classes and the
1401 translation maps into stderr. */
1402 void
1403 ira_debug_allocno_classes (void)
1405 print_unform_and_important_classes (stderr);
1406 print_translated_classes (stderr, false);
1407 print_translated_classes (stderr, true);
1410 /* Set up different arrays concerning class subsets, allocno and
1411 important classes. */
1412 static void
1413 find_reg_classes (void)
1415 setup_allocno_and_important_classes ();
1416 setup_class_translate ();
1417 reorder_important_classes ();
1418 setup_reg_class_relations ();
1423 /* Set up the array above. */
1424 static void
1425 setup_hard_regno_aclass (void)
1427 int i;
1429 for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
1431 #if 1
1432 ira_hard_regno_allocno_class[i]
1433 = (TEST_HARD_REG_BIT (no_unit_alloc_regs, i)
1434 ? NO_REGS
1435 : ira_allocno_class_translate[REGNO_REG_CLASS (i)]);
1436 #else
1437 int j;
1438 enum reg_class cl;
1439 ira_hard_regno_allocno_class[i] = NO_REGS;
1440 for (j = 0; j < ira_allocno_classes_num; j++)
1442 cl = ira_allocno_classes[j];
1443 if (ira_class_hard_reg_index[cl][i] >= 0)
1445 ira_hard_regno_allocno_class[i] = cl;
1446 break;
1449 #endif
1455 /* Form IRA_REG_CLASS_MAX_NREGS and IRA_REG_CLASS_MIN_NREGS maps. */
1456 static void
1457 setup_reg_class_nregs (void)
1459 int i, cl, cl2, m;
1461 for (m = 0; m < MAX_MACHINE_MODE; m++)
1463 for (cl = 0; cl < N_REG_CLASSES; cl++)
1464 ira_reg_class_max_nregs[cl][m]
1465 = ira_reg_class_min_nregs[cl][m]
1466 = targetm.class_max_nregs ((reg_class_t) cl, (enum machine_mode) m);
1467 for (cl = 0; cl < N_REG_CLASSES; cl++)
1468 for (i = 0;
1469 (cl2 = alloc_reg_class_subclasses[cl][i]) != LIM_REG_CLASSES;
1470 i++)
1471 if (ira_reg_class_min_nregs[cl2][m]
1472 < ira_reg_class_min_nregs[cl][m])
1473 ira_reg_class_min_nregs[cl][m] = ira_reg_class_min_nregs[cl2][m];
1479 /* Set up IRA_PROHIBITED_CLASS_MODE_REGS and IRA_CLASS_SINGLETON.
1480 This function is called once IRA_CLASS_HARD_REGS has been initialized. */
1481 static void
1482 setup_prohibited_class_mode_regs (void)
1484 int j, k, hard_regno, cl, last_hard_regno, count;
1486 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
1488 COPY_HARD_REG_SET (temp_hard_regset, reg_class_contents[cl]);
1489 AND_COMPL_HARD_REG_SET (temp_hard_regset, no_unit_alloc_regs);
1490 for (j = 0; j < NUM_MACHINE_MODES; j++)
1492 count = 0;
1493 last_hard_regno = -1;
1494 CLEAR_HARD_REG_SET (ira_prohibited_class_mode_regs[cl][j]);
1495 for (k = ira_class_hard_regs_num[cl] - 1; k >= 0; k--)
1497 hard_regno = ira_class_hard_regs[cl][k];
1498 if (! HARD_REGNO_MODE_OK (hard_regno, (enum machine_mode) j))
1499 SET_HARD_REG_BIT (ira_prohibited_class_mode_regs[cl][j],
1500 hard_regno);
1501 else if (in_hard_reg_set_p (temp_hard_regset,
1502 (enum machine_mode) j, hard_regno))
1504 last_hard_regno = hard_regno;
1505 count++;
1508 ira_class_singleton[cl][j] = (count == 1 ? last_hard_regno : -1);
1513 /* Clarify IRA_PROHIBITED_CLASS_MODE_REGS by excluding hard registers
1514 spanning from one register pressure class to another one. It is
1515 called after defining the pressure classes. */
1516 static void
1517 clarify_prohibited_class_mode_regs (void)
1519 int j, k, hard_regno, cl, pclass, nregs;
1521 for (cl = (int) N_REG_CLASSES - 1; cl >= 0; cl--)
1522 for (j = 0; j < NUM_MACHINE_MODES; j++)
1524 CLEAR_HARD_REG_SET (ira_useful_class_mode_regs[cl][j]);
1525 for (k = ira_class_hard_regs_num[cl] - 1; k >= 0; k--)
1527 hard_regno = ira_class_hard_regs[cl][k];
1528 if (TEST_HARD_REG_BIT (ira_prohibited_class_mode_regs[cl][j], hard_regno))
1529 continue;
1530 nregs = hard_regno_nregs[hard_regno][j];
1531 if (hard_regno + nregs > FIRST_PSEUDO_REGISTER)
1533 SET_HARD_REG_BIT (ira_prohibited_class_mode_regs[cl][j],
1534 hard_regno);
1535 continue;
1537 pclass = ira_pressure_class_translate[REGNO_REG_CLASS (hard_regno)];
1538 for (nregs-- ;nregs >= 0; nregs--)
1539 if (((enum reg_class) pclass
1540 != ira_pressure_class_translate[REGNO_REG_CLASS
1541 (hard_regno + nregs)]))
1543 SET_HARD_REG_BIT (ira_prohibited_class_mode_regs[cl][j],
1544 hard_regno);
1545 break;
1547 if (!TEST_HARD_REG_BIT (ira_prohibited_class_mode_regs[cl][j],
1548 hard_regno))
1549 add_to_hard_reg_set (&ira_useful_class_mode_regs[cl][j],
1550 (enum machine_mode) j, hard_regno);
1555 /* Allocate and initialize IRA_REGISTER_MOVE_COST, IRA_MAY_MOVE_IN_COST
1556 and IRA_MAY_MOVE_OUT_COST for MODE. */
1557 void
1558 ira_init_register_move_cost (enum machine_mode mode)
1560 static unsigned short last_move_cost[N_REG_CLASSES][N_REG_CLASSES];
1561 bool all_match = true;
1562 unsigned int cl1, cl2;
1564 ira_assert (ira_register_move_cost[mode] == NULL
1565 && ira_may_move_in_cost[mode] == NULL
1566 && ira_may_move_out_cost[mode] == NULL);
1567 ira_assert (have_regs_of_mode[mode]);
1568 for (cl1 = 0; cl1 < N_REG_CLASSES; cl1++)
1569 if (contains_reg_of_mode[cl1][mode])
1570 for (cl2 = 0; cl2 < N_REG_CLASSES; cl2++)
1572 int cost;
1573 if (!contains_reg_of_mode[cl2][mode])
1574 cost = 65535;
1575 else
1577 cost = register_move_cost (mode, (enum reg_class) cl1,
1578 (enum reg_class) cl2);
1579 ira_assert (cost < 65535);
1581 all_match &= (last_move_cost[cl1][cl2] == cost);
1582 last_move_cost[cl1][cl2] = cost;
1584 if (all_match && last_mode_for_init_move_cost != -1)
1586 ira_register_move_cost[mode]
1587 = ira_register_move_cost[last_mode_for_init_move_cost];
1588 ira_may_move_in_cost[mode]
1589 = ira_may_move_in_cost[last_mode_for_init_move_cost];
1590 ira_may_move_out_cost[mode]
1591 = ira_may_move_out_cost[last_mode_for_init_move_cost];
1592 return;
1594 last_mode_for_init_move_cost = mode;
1595 ira_register_move_cost[mode] = XNEWVEC (move_table, N_REG_CLASSES);
1596 ira_may_move_in_cost[mode] = XNEWVEC (move_table, N_REG_CLASSES);
1597 ira_may_move_out_cost[mode] = XNEWVEC (move_table, N_REG_CLASSES);
1598 for (cl1 = 0; cl1 < N_REG_CLASSES; cl1++)
1599 if (contains_reg_of_mode[cl1][mode])
1600 for (cl2 = 0; cl2 < N_REG_CLASSES; cl2++)
1602 int cost;
1603 enum reg_class *p1, *p2;
1605 if (last_move_cost[cl1][cl2] == 65535)
1607 ira_register_move_cost[mode][cl1][cl2] = 65535;
1608 ira_may_move_in_cost[mode][cl1][cl2] = 65535;
1609 ira_may_move_out_cost[mode][cl1][cl2] = 65535;
1611 else
1613 cost = last_move_cost[cl1][cl2];
1615 for (p2 = &reg_class_subclasses[cl2][0];
1616 *p2 != LIM_REG_CLASSES; p2++)
1617 if (ira_class_hard_regs_num[*p2] > 0
1618 && (ira_reg_class_max_nregs[*p2][mode]
1619 <= ira_class_hard_regs_num[*p2]))
1620 cost = MAX (cost, ira_register_move_cost[mode][cl1][*p2]);
1622 for (p1 = &reg_class_subclasses[cl1][0];
1623 *p1 != LIM_REG_CLASSES; p1++)
1624 if (ira_class_hard_regs_num[*p1] > 0
1625 && (ira_reg_class_max_nregs[*p1][mode]
1626 <= ira_class_hard_regs_num[*p1]))
1627 cost = MAX (cost, ira_register_move_cost[mode][*p1][cl2]);
1629 ira_assert (cost <= 65535);
1630 ira_register_move_cost[mode][cl1][cl2] = cost;
1632 if (ira_class_subset_p[cl1][cl2])
1633 ira_may_move_in_cost[mode][cl1][cl2] = 0;
1634 else
1635 ira_may_move_in_cost[mode][cl1][cl2] = cost;
1637 if (ira_class_subset_p[cl2][cl1])
1638 ira_may_move_out_cost[mode][cl1][cl2] = 0;
1639 else
1640 ira_may_move_out_cost[mode][cl1][cl2] = cost;
1643 else
1644 for (cl2 = 0; cl2 < N_REG_CLASSES; cl2++)
1646 ira_register_move_cost[mode][cl1][cl2] = 65535;
1647 ira_may_move_in_cost[mode][cl1][cl2] = 65535;
1648 ira_may_move_out_cost[mode][cl1][cl2] = 65535;
1653 /* This is called once during compiler work. It sets up
1654 different arrays whose values don't depend on the compiled
1655 function. */
1656 void
1657 ira_init_once (void)
1659 ira_init_costs_once ();
1660 lra_init_once ();
1663 /* Free ira_max_register_move_cost, ira_may_move_in_cost and
1664 ira_may_move_out_cost for each mode. */
1665 static void
1666 free_register_move_costs (void)
1668 int mode, i;
1670 /* Reset move_cost and friends, making sure we only free shared
1671 table entries once. */
1672 for (mode = 0; mode < MAX_MACHINE_MODE; mode++)
1673 if (ira_register_move_cost[mode])
1675 for (i = 0;
1676 i < mode && (ira_register_move_cost[i]
1677 != ira_register_move_cost[mode]);
1678 i++)
1680 if (i == mode)
1682 free (ira_register_move_cost[mode]);
1683 free (ira_may_move_in_cost[mode]);
1684 free (ira_may_move_out_cost[mode]);
1687 memset (ira_register_move_cost, 0, sizeof ira_register_move_cost);
1688 memset (ira_may_move_in_cost, 0, sizeof ira_may_move_in_cost);
1689 memset (ira_may_move_out_cost, 0, sizeof ira_may_move_out_cost);
1690 last_mode_for_init_move_cost = -1;
1693 /* This is called every time when register related information is
1694 changed. */
1695 void
1696 ira_init (void)
1698 free_register_move_costs ();
1699 setup_reg_mode_hard_regset ();
1700 setup_alloc_regs (flag_omit_frame_pointer != 0);
1701 setup_class_subset_and_memory_move_costs ();
1702 setup_reg_class_nregs ();
1703 setup_prohibited_class_mode_regs ();
1704 find_reg_classes ();
1705 clarify_prohibited_class_mode_regs ();
1706 setup_hard_regno_aclass ();
1707 ira_init_costs ();
1708 lra_init ();
1711 /* Function called once at the end of compiler work. */
1712 void
1713 ira_finish_once (void)
1715 ira_finish_costs_once ();
1716 free_register_move_costs ();
1717 lra_finish_once ();
1721 #define ira_prohibited_mode_move_regs_initialized_p \
1722 (this_target_ira_int->x_ira_prohibited_mode_move_regs_initialized_p)
1724 /* Set up IRA_PROHIBITED_MODE_MOVE_REGS. */
1725 static void
1726 setup_prohibited_mode_move_regs (void)
1728 int i, j;
1729 rtx test_reg1, test_reg2, move_pat, move_insn;
1731 if (ira_prohibited_mode_move_regs_initialized_p)
1732 return;
1733 ira_prohibited_mode_move_regs_initialized_p = true;
1734 test_reg1 = gen_rtx_REG (VOIDmode, 0);
1735 test_reg2 = gen_rtx_REG (VOIDmode, 0);
1736 move_pat = gen_rtx_SET (VOIDmode, test_reg1, test_reg2);
1737 move_insn = gen_rtx_INSN (VOIDmode, 0, 0, 0, 0, move_pat, 0, -1, 0);
1738 for (i = 0; i < NUM_MACHINE_MODES; i++)
1740 SET_HARD_REG_SET (ira_prohibited_mode_move_regs[i]);
1741 for (j = 0; j < FIRST_PSEUDO_REGISTER; j++)
1743 if (! HARD_REGNO_MODE_OK (j, (enum machine_mode) i))
1744 continue;
1745 SET_REGNO_RAW (test_reg1, j);
1746 PUT_MODE (test_reg1, (enum machine_mode) i);
1747 SET_REGNO_RAW (test_reg2, j);
1748 PUT_MODE (test_reg2, (enum machine_mode) i);
1749 INSN_CODE (move_insn) = -1;
1750 recog_memoized (move_insn);
1751 if (INSN_CODE (move_insn) < 0)
1752 continue;
1753 extract_insn (move_insn);
1754 if (! constrain_operands (1))
1755 continue;
1756 CLEAR_HARD_REG_BIT (ira_prohibited_mode_move_regs[i], j);
1763 /* Return nonzero if REGNO is a particularly bad choice for reloading X. */
1764 static bool
1765 ira_bad_reload_regno_1 (int regno, rtx x)
1767 int x_regno, n, i;
1768 ira_allocno_t a;
1769 enum reg_class pref;
1771 /* We only deal with pseudo regs. */
1772 if (! x || GET_CODE (x) != REG)
1773 return false;
1775 x_regno = REGNO (x);
1776 if (x_regno < FIRST_PSEUDO_REGISTER)
1777 return false;
1779 /* If the pseudo prefers REGNO explicitly, then do not consider
1780 REGNO a bad spill choice. */
1781 pref = reg_preferred_class (x_regno);
1782 if (reg_class_size[pref] == 1)
1783 return !TEST_HARD_REG_BIT (reg_class_contents[pref], regno);
1785 /* If the pseudo conflicts with REGNO, then we consider REGNO a
1786 poor choice for a reload regno. */
1787 a = ira_regno_allocno_map[x_regno];
1788 n = ALLOCNO_NUM_OBJECTS (a);
1789 for (i = 0; i < n; i++)
1791 ira_object_t obj = ALLOCNO_OBJECT (a, i);
1792 if (TEST_HARD_REG_BIT (OBJECT_TOTAL_CONFLICT_HARD_REGS (obj), regno))
1793 return true;
1795 return false;
1798 /* Return nonzero if REGNO is a particularly bad choice for reloading
1799 IN or OUT. */
1800 bool
1801 ira_bad_reload_regno (int regno, rtx in, rtx out)
1803 return (ira_bad_reload_regno_1 (regno, in)
1804 || ira_bad_reload_regno_1 (regno, out));
1807 /* Return TRUE if *LOC contains an asm. */
1808 static int
1809 insn_contains_asm_1 (rtx *loc, void *data ATTRIBUTE_UNUSED)
1811 if ( !*loc)
1812 return FALSE;
1813 if (GET_CODE (*loc) == ASM_OPERANDS)
1814 return TRUE;
1815 return FALSE;
1819 /* Return TRUE if INSN contains an ASM. */
1820 static bool
1821 insn_contains_asm (rtx insn)
1823 return for_each_rtx (&insn, insn_contains_asm_1, NULL);
1826 /* Add register clobbers from asm statements. */
1827 static void
1828 compute_regs_asm_clobbered (void)
1830 basic_block bb;
1832 FOR_EACH_BB (bb)
1834 rtx insn;
1835 FOR_BB_INSNS_REVERSE (bb, insn)
1837 df_ref *def_rec;
1839 if (insn_contains_asm (insn))
1840 for (def_rec = DF_INSN_DEFS (insn); *def_rec; def_rec++)
1842 df_ref def = *def_rec;
1843 unsigned int dregno = DF_REF_REGNO (def);
1844 if (HARD_REGISTER_NUM_P (dregno))
1845 add_to_hard_reg_set (&crtl->asm_clobbers,
1846 GET_MODE (DF_REF_REAL_REG (def)),
1847 dregno);
1854 /* Set up ELIMINABLE_REGSET, IRA_NO_ALLOC_REGS, and REGS_EVER_LIVE.
1855 If the function is called from IRA (not from the insn scheduler or
1856 RTL loop invariant motion), FROM_IRA_P is true. */
1857 void
1858 ira_setup_eliminable_regset (bool from_ira_p)
1860 #ifdef ELIMINABLE_REGS
1861 int i;
1862 static const struct {const int from, to; } eliminables[] = ELIMINABLE_REGS;
1863 #endif
1864 /* FIXME: If EXIT_IGNORE_STACK is set, we will not save and restore
1865 sp for alloca. So we can't eliminate the frame pointer in that
1866 case. At some point, we should improve this by emitting the
1867 sp-adjusting insns for this case. */
1868 frame_pointer_needed
1869 = (! flag_omit_frame_pointer
1870 || (cfun->calls_alloca && EXIT_IGNORE_STACK)
1871 /* We need the frame pointer to catch stack overflow exceptions
1872 if the stack pointer is moving. */
1873 || (flag_stack_check && STACK_CHECK_MOVING_SP)
1874 || crtl->accesses_prior_frames
1875 || crtl->stack_realign_needed
1876 || targetm.frame_pointer_required ());
1878 if (from_ira_p && ira_use_lra_p)
1879 /* It can change FRAME_POINTER_NEEDED. We call it only from IRA
1880 because it is expensive. */
1881 lra_init_elimination ();
1883 if (frame_pointer_needed)
1884 df_set_regs_ever_live (HARD_FRAME_POINTER_REGNUM, true);
1886 COPY_HARD_REG_SET (ira_no_alloc_regs, no_unit_alloc_regs);
1887 CLEAR_HARD_REG_SET (eliminable_regset);
1889 compute_regs_asm_clobbered ();
1891 /* Build the regset of all eliminable registers and show we can't
1892 use those that we already know won't be eliminated. */
1893 #ifdef ELIMINABLE_REGS
1894 for (i = 0; i < (int) ARRAY_SIZE (eliminables); i++)
1896 bool cannot_elim
1897 = (! targetm.can_eliminate (eliminables[i].from, eliminables[i].to)
1898 || (eliminables[i].to == STACK_POINTER_REGNUM && frame_pointer_needed));
1900 if (!TEST_HARD_REG_BIT (crtl->asm_clobbers, eliminables[i].from))
1902 SET_HARD_REG_BIT (eliminable_regset, eliminables[i].from);
1904 if (cannot_elim)
1905 SET_HARD_REG_BIT (ira_no_alloc_regs, eliminables[i].from);
1907 else if (cannot_elim)
1908 error ("%s cannot be used in asm here",
1909 reg_names[eliminables[i].from]);
1910 else
1911 df_set_regs_ever_live (eliminables[i].from, true);
1913 #if !HARD_FRAME_POINTER_IS_FRAME_POINTER
1914 if (!TEST_HARD_REG_BIT (crtl->asm_clobbers, HARD_FRAME_POINTER_REGNUM))
1916 SET_HARD_REG_BIT (eliminable_regset, HARD_FRAME_POINTER_REGNUM);
1917 if (frame_pointer_needed)
1918 SET_HARD_REG_BIT (ira_no_alloc_regs, HARD_FRAME_POINTER_REGNUM);
1920 else if (frame_pointer_needed)
1921 error ("%s cannot be used in asm here",
1922 reg_names[HARD_FRAME_POINTER_REGNUM]);
1923 else
1924 df_set_regs_ever_live (HARD_FRAME_POINTER_REGNUM, true);
1925 #endif
1927 #else
1928 if (!TEST_HARD_REG_BIT (crtl->asm_clobbers, HARD_FRAME_POINTER_REGNUM))
1930 SET_HARD_REG_BIT (eliminable_regset, FRAME_POINTER_REGNUM);
1931 if (frame_pointer_needed)
1932 SET_HARD_REG_BIT (ira_no_alloc_regs, FRAME_POINTER_REGNUM);
1934 else if (frame_pointer_needed)
1935 error ("%s cannot be used in asm here", reg_names[FRAME_POINTER_REGNUM]);
1936 else
1937 df_set_regs_ever_live (FRAME_POINTER_REGNUM, true);
1938 #endif
1943 /* Vector of substitutions of register numbers,
1944 used to map pseudo regs into hardware regs.
1945 This is set up as a result of register allocation.
1946 Element N is the hard reg assigned to pseudo reg N,
1947 or is -1 if no hard reg was assigned.
1948 If N is a hard reg number, element N is N. */
1949 short *reg_renumber;
1951 /* Set up REG_RENUMBER and CALLER_SAVE_NEEDED (used by reload) from
1952 the allocation found by IRA. */
1953 static void
1954 setup_reg_renumber (void)
1956 int regno, hard_regno;
1957 ira_allocno_t a;
1958 ira_allocno_iterator ai;
1960 caller_save_needed = 0;
1961 FOR_EACH_ALLOCNO (a, ai)
1963 if (ira_use_lra_p && ALLOCNO_CAP_MEMBER (a) != NULL)
1964 continue;
1965 /* There are no caps at this point. */
1966 ira_assert (ALLOCNO_CAP_MEMBER (a) == NULL);
1967 if (! ALLOCNO_ASSIGNED_P (a))
1968 /* It can happen if A is not referenced but partially anticipated
1969 somewhere in a region. */
1970 ALLOCNO_ASSIGNED_P (a) = true;
1971 ira_free_allocno_updated_costs (a);
1972 hard_regno = ALLOCNO_HARD_REGNO (a);
1973 regno = ALLOCNO_REGNO (a);
1974 reg_renumber[regno] = (hard_regno < 0 ? -1 : hard_regno);
1975 if (hard_regno >= 0)
1977 int i, nwords;
1978 enum reg_class pclass;
1979 ira_object_t obj;
1981 pclass = ira_pressure_class_translate[REGNO_REG_CLASS (hard_regno)];
1982 nwords = ALLOCNO_NUM_OBJECTS (a);
1983 for (i = 0; i < nwords; i++)
1985 obj = ALLOCNO_OBJECT (a, i);
1986 IOR_COMPL_HARD_REG_SET (OBJECT_TOTAL_CONFLICT_HARD_REGS (obj),
1987 reg_class_contents[pclass]);
1989 if (ALLOCNO_CALLS_CROSSED_NUM (a) != 0
1990 && ira_hard_reg_set_intersection_p (hard_regno, ALLOCNO_MODE (a),
1991 call_used_reg_set))
1993 ira_assert (!optimize || flag_caller_saves
1994 || (ALLOCNO_CALLS_CROSSED_NUM (a)
1995 == ALLOCNO_CHEAP_CALLS_CROSSED_NUM (a))
1996 || regno >= ira_reg_equiv_len
1997 || ira_equiv_no_lvalue_p (regno));
1998 caller_save_needed = 1;
2004 /* Set up allocno assignment flags for further allocation
2005 improvements. */
2006 static void
2007 setup_allocno_assignment_flags (void)
2009 int hard_regno;
2010 ira_allocno_t a;
2011 ira_allocno_iterator ai;
2013 FOR_EACH_ALLOCNO (a, ai)
2015 if (! ALLOCNO_ASSIGNED_P (a))
2016 /* It can happen if A is not referenced but partially anticipated
2017 somewhere in a region. */
2018 ira_free_allocno_updated_costs (a);
2019 hard_regno = ALLOCNO_HARD_REGNO (a);
2020 /* Don't assign hard registers to allocnos which are destination
2021 of removed store at the end of loop. It has no sense to keep
2022 the same value in different hard registers. It is also
2023 impossible to assign hard registers correctly to such
2024 allocnos because the cost info and info about intersected
2025 calls are incorrect for them. */
2026 ALLOCNO_ASSIGNED_P (a) = (hard_regno >= 0
2027 || ALLOCNO_EMIT_DATA (a)->mem_optimized_dest_p
2028 || (ALLOCNO_MEMORY_COST (a)
2029 - ALLOCNO_CLASS_COST (a)) < 0);
2030 ira_assert
2031 (hard_regno < 0
2032 || ira_hard_reg_in_set_p (hard_regno, ALLOCNO_MODE (a),
2033 reg_class_contents[ALLOCNO_CLASS (a)]));
2037 /* Evaluate overall allocation cost and the costs for using hard
2038 registers and memory for allocnos. */
2039 static void
2040 calculate_allocation_cost (void)
2042 int hard_regno, cost;
2043 ira_allocno_t a;
2044 ira_allocno_iterator ai;
2046 ira_overall_cost = ira_reg_cost = ira_mem_cost = 0;
2047 FOR_EACH_ALLOCNO (a, ai)
2049 hard_regno = ALLOCNO_HARD_REGNO (a);
2050 ira_assert (hard_regno < 0
2051 || (ira_hard_reg_in_set_p
2052 (hard_regno, ALLOCNO_MODE (a),
2053 reg_class_contents[ALLOCNO_CLASS (a)])));
2054 if (hard_regno < 0)
2056 cost = ALLOCNO_MEMORY_COST (a);
2057 ira_mem_cost += cost;
2059 else if (ALLOCNO_HARD_REG_COSTS (a) != NULL)
2061 cost = (ALLOCNO_HARD_REG_COSTS (a)
2062 [ira_class_hard_reg_index
2063 [ALLOCNO_CLASS (a)][hard_regno]]);
2064 ira_reg_cost += cost;
2066 else
2068 cost = ALLOCNO_CLASS_COST (a);
2069 ira_reg_cost += cost;
2071 ira_overall_cost += cost;
2074 if (internal_flag_ira_verbose > 0 && ira_dump_file != NULL)
2076 fprintf (ira_dump_file,
2077 "+++Costs: overall %d, reg %d, mem %d, ld %d, st %d, move %d\n",
2078 ira_overall_cost, ira_reg_cost, ira_mem_cost,
2079 ira_load_cost, ira_store_cost, ira_shuffle_cost);
2080 fprintf (ira_dump_file, "+++ move loops %d, new jumps %d\n",
2081 ira_move_loops_num, ira_additional_jumps_num);
2086 #ifdef ENABLE_IRA_CHECKING
2087 /* Check the correctness of the allocation. We do need this because
2088 of complicated code to transform more one region internal
2089 representation into one region representation. */
2090 static void
2091 check_allocation (void)
2093 ira_allocno_t a;
2094 int hard_regno, nregs, conflict_nregs;
2095 ira_allocno_iterator ai;
2097 FOR_EACH_ALLOCNO (a, ai)
2099 int n = ALLOCNO_NUM_OBJECTS (a);
2100 int i;
2102 if (ALLOCNO_CAP_MEMBER (a) != NULL
2103 || (hard_regno = ALLOCNO_HARD_REGNO (a)) < 0)
2104 continue;
2105 nregs = hard_regno_nregs[hard_regno][ALLOCNO_MODE (a)];
2106 if (nregs == 1)
2107 /* We allocated a single hard register. */
2108 n = 1;
2109 else if (n > 1)
2110 /* We allocated multiple hard registers, and we will test
2111 conflicts in a granularity of single hard regs. */
2112 nregs = 1;
2114 for (i = 0; i < n; i++)
2116 ira_object_t obj = ALLOCNO_OBJECT (a, i);
2117 ira_object_t conflict_obj;
2118 ira_object_conflict_iterator oci;
2119 int this_regno = hard_regno;
2120 if (n > 1)
2122 if (REG_WORDS_BIG_ENDIAN)
2123 this_regno += n - i - 1;
2124 else
2125 this_regno += i;
2127 FOR_EACH_OBJECT_CONFLICT (obj, conflict_obj, oci)
2129 ira_allocno_t conflict_a = OBJECT_ALLOCNO (conflict_obj);
2130 int conflict_hard_regno = ALLOCNO_HARD_REGNO (conflict_a);
2131 if (conflict_hard_regno < 0)
2132 continue;
2134 conflict_nregs
2135 = (hard_regno_nregs
2136 [conflict_hard_regno][ALLOCNO_MODE (conflict_a)]);
2138 if (ALLOCNO_NUM_OBJECTS (conflict_a) > 1
2139 && conflict_nregs == ALLOCNO_NUM_OBJECTS (conflict_a))
2141 if (REG_WORDS_BIG_ENDIAN)
2142 conflict_hard_regno += (ALLOCNO_NUM_OBJECTS (conflict_a)
2143 - OBJECT_SUBWORD (conflict_obj) - 1);
2144 else
2145 conflict_hard_regno += OBJECT_SUBWORD (conflict_obj);
2146 conflict_nregs = 1;
2149 if ((conflict_hard_regno <= this_regno
2150 && this_regno < conflict_hard_regno + conflict_nregs)
2151 || (this_regno <= conflict_hard_regno
2152 && conflict_hard_regno < this_regno + nregs))
2154 fprintf (stderr, "bad allocation for %d and %d\n",
2155 ALLOCNO_REGNO (a), ALLOCNO_REGNO (conflict_a));
2156 gcc_unreachable ();
2162 #endif
2164 /* Allocate REG_EQUIV_INIT. Set up it from IRA_REG_EQUIV which should
2165 be already calculated. */
2166 static void
2167 setup_reg_equiv_init (void)
2169 int i;
2170 int max_regno = max_reg_num ();
2172 for (i = 0; i < max_regno; i++)
2173 reg_equiv_init (i) = ira_reg_equiv[i].init_insns;
2176 /* Update equiv regno from movement of FROM_REGNO to TO_REGNO. INSNS
2177 are insns which were generated for such movement. It is assumed
2178 that FROM_REGNO and TO_REGNO always have the same value at the
2179 point of any move containing such registers. This function is used
2180 to update equiv info for register shuffles on the region borders
2181 and for caller save/restore insns. */
2182 void
2183 ira_update_equiv_info_by_shuffle_insn (int to_regno, int from_regno, rtx insns)
2185 rtx insn, x, note;
2187 if (! ira_reg_equiv[from_regno].defined_p
2188 && (! ira_reg_equiv[to_regno].defined_p
2189 || ((x = ira_reg_equiv[to_regno].memory) != NULL_RTX
2190 && ! MEM_READONLY_P (x))))
2191 return;
2192 insn = insns;
2193 if (NEXT_INSN (insn) != NULL_RTX)
2195 if (! ira_reg_equiv[to_regno].defined_p)
2197 ira_assert (ira_reg_equiv[to_regno].init_insns == NULL_RTX);
2198 return;
2200 ira_reg_equiv[to_regno].defined_p = false;
2201 ira_reg_equiv[to_regno].memory
2202 = ira_reg_equiv[to_regno].constant
2203 = ira_reg_equiv[to_regno].invariant
2204 = ira_reg_equiv[to_regno].init_insns = NULL_RTX;
2205 if (internal_flag_ira_verbose > 3 && ira_dump_file != NULL)
2206 fprintf (ira_dump_file,
2207 " Invalidating equiv info for reg %d\n", to_regno);
2208 return;
2210 /* It is possible that FROM_REGNO still has no equivalence because
2211 in shuffles to_regno<-from_regno and from_regno<-to_regno the 2nd
2212 insn was not processed yet. */
2213 if (ira_reg_equiv[from_regno].defined_p)
2215 ira_reg_equiv[to_regno].defined_p = true;
2216 if ((x = ira_reg_equiv[from_regno].memory) != NULL_RTX)
2218 ira_assert (ira_reg_equiv[from_regno].invariant == NULL_RTX
2219 && ira_reg_equiv[from_regno].constant == NULL_RTX);
2220 ira_assert (ira_reg_equiv[to_regno].memory == NULL_RTX
2221 || rtx_equal_p (ira_reg_equiv[to_regno].memory, x));
2222 ira_reg_equiv[to_regno].memory = x;
2223 if (! MEM_READONLY_P (x))
2224 /* We don't add the insn to insn init list because memory
2225 equivalence is just to say what memory is better to use
2226 when the pseudo is spilled. */
2227 return;
2229 else if ((x = ira_reg_equiv[from_regno].constant) != NULL_RTX)
2231 ira_assert (ira_reg_equiv[from_regno].invariant == NULL_RTX);
2232 ira_assert (ira_reg_equiv[to_regno].constant == NULL_RTX
2233 || rtx_equal_p (ira_reg_equiv[to_regno].constant, x));
2234 ira_reg_equiv[to_regno].constant = x;
2236 else
2238 x = ira_reg_equiv[from_regno].invariant;
2239 ira_assert (x != NULL_RTX);
2240 ira_assert (ira_reg_equiv[to_regno].invariant == NULL_RTX
2241 || rtx_equal_p (ira_reg_equiv[to_regno].invariant, x));
2242 ira_reg_equiv[to_regno].invariant = x;
2244 if (find_reg_note (insn, REG_EQUIV, x) == NULL_RTX)
2246 note = set_unique_reg_note (insn, REG_EQUIV, x);
2247 gcc_assert (note != NULL_RTX);
2248 if (internal_flag_ira_verbose > 3 && ira_dump_file != NULL)
2250 fprintf (ira_dump_file,
2251 " Adding equiv note to insn %u for reg %d ",
2252 INSN_UID (insn), to_regno);
2253 dump_value_slim (ira_dump_file, x, 1);
2254 fprintf (ira_dump_file, "\n");
2258 ira_reg_equiv[to_regno].init_insns
2259 = gen_rtx_INSN_LIST (VOIDmode, insn,
2260 ira_reg_equiv[to_regno].init_insns);
2261 if (internal_flag_ira_verbose > 3 && ira_dump_file != NULL)
2262 fprintf (ira_dump_file,
2263 " Adding equiv init move insn %u to reg %d\n",
2264 INSN_UID (insn), to_regno);
2267 /* Fix values of array REG_EQUIV_INIT after live range splitting done
2268 by IRA. */
2269 static void
2270 fix_reg_equiv_init (void)
2272 int max_regno = max_reg_num ();
2273 int i, new_regno, max;
2274 rtx x, prev, next, insn, set;
2276 if (max_regno_before_ira < max_regno)
2278 max = vec_safe_length (reg_equivs);
2279 grow_reg_equivs ();
2280 for (i = FIRST_PSEUDO_REGISTER; i < max; i++)
2281 for (prev = NULL_RTX, x = reg_equiv_init (i);
2282 x != NULL_RTX;
2283 x = next)
2285 next = XEXP (x, 1);
2286 insn = XEXP (x, 0);
2287 set = single_set (insn);
2288 ira_assert (set != NULL_RTX
2289 && (REG_P (SET_DEST (set)) || REG_P (SET_SRC (set))));
2290 if (REG_P (SET_DEST (set))
2291 && ((int) REGNO (SET_DEST (set)) == i
2292 || (int) ORIGINAL_REGNO (SET_DEST (set)) == i))
2293 new_regno = REGNO (SET_DEST (set));
2294 else if (REG_P (SET_SRC (set))
2295 && ((int) REGNO (SET_SRC (set)) == i
2296 || (int) ORIGINAL_REGNO (SET_SRC (set)) == i))
2297 new_regno = REGNO (SET_SRC (set));
2298 else
2299 gcc_unreachable ();
2300 if (new_regno == i)
2301 prev = x;
2302 else
2304 /* Remove the wrong list element. */
2305 if (prev == NULL_RTX)
2306 reg_equiv_init (i) = next;
2307 else
2308 XEXP (prev, 1) = next;
2309 XEXP (x, 1) = reg_equiv_init (new_regno);
2310 reg_equiv_init (new_regno) = x;
2316 #ifdef ENABLE_IRA_CHECKING
2317 /* Print redundant memory-memory copies. */
2318 static void
2319 print_redundant_copies (void)
2321 int hard_regno;
2322 ira_allocno_t a;
2323 ira_copy_t cp, next_cp;
2324 ira_allocno_iterator ai;
2326 FOR_EACH_ALLOCNO (a, ai)
2328 if (ALLOCNO_CAP_MEMBER (a) != NULL)
2329 /* It is a cap. */
2330 continue;
2331 hard_regno = ALLOCNO_HARD_REGNO (a);
2332 if (hard_regno >= 0)
2333 continue;
2334 for (cp = ALLOCNO_COPIES (a); cp != NULL; cp = next_cp)
2335 if (cp->first == a)
2336 next_cp = cp->next_first_allocno_copy;
2337 else
2339 next_cp = cp->next_second_allocno_copy;
2340 if (internal_flag_ira_verbose > 4 && ira_dump_file != NULL
2341 && cp->insn != NULL_RTX
2342 && ALLOCNO_HARD_REGNO (cp->first) == hard_regno)
2343 fprintf (ira_dump_file,
2344 " Redundant move from %d(freq %d):%d\n",
2345 INSN_UID (cp->insn), cp->freq, hard_regno);
2349 #endif
2351 /* Setup preferred and alternative classes for new pseudo-registers
2352 created by IRA starting with START. */
2353 static void
2354 setup_preferred_alternate_classes_for_new_pseudos (int start)
2356 int i, old_regno;
2357 int max_regno = max_reg_num ();
2359 for (i = start; i < max_regno; i++)
2361 old_regno = ORIGINAL_REGNO (regno_reg_rtx[i]);
2362 ira_assert (i != old_regno);
2363 setup_reg_classes (i, reg_preferred_class (old_regno),
2364 reg_alternate_class (old_regno),
2365 reg_allocno_class (old_regno));
2366 if (internal_flag_ira_verbose > 2 && ira_dump_file != NULL)
2367 fprintf (ira_dump_file,
2368 " New r%d: setting preferred %s, alternative %s\n",
2369 i, reg_class_names[reg_preferred_class (old_regno)],
2370 reg_class_names[reg_alternate_class (old_regno)]);
2375 /* The number of entries allocated in teg_info. */
2376 static int allocated_reg_info_size;
2378 /* Regional allocation can create new pseudo-registers. This function
2379 expands some arrays for pseudo-registers. */
2380 static void
2381 expand_reg_info (void)
2383 int i;
2384 int size = max_reg_num ();
2386 resize_reg_info ();
2387 for (i = allocated_reg_info_size; i < size; i++)
2388 setup_reg_classes (i, GENERAL_REGS, ALL_REGS, GENERAL_REGS);
2389 setup_preferred_alternate_classes_for_new_pseudos (allocated_reg_info_size);
2390 allocated_reg_info_size = size;
2393 /* Return TRUE if there is too high register pressure in the function.
2394 It is used to decide when stack slot sharing is worth to do. */
2395 static bool
2396 too_high_register_pressure_p (void)
2398 int i;
2399 enum reg_class pclass;
2401 for (i = 0; i < ira_pressure_classes_num; i++)
2403 pclass = ira_pressure_classes[i];
2404 if (ira_loop_tree_root->reg_pressure[pclass] > 10000)
2405 return true;
2407 return false;
2412 /* Indicate that hard register number FROM was eliminated and replaced with
2413 an offset from hard register number TO. The status of hard registers live
2414 at the start of a basic block is updated by replacing a use of FROM with
2415 a use of TO. */
2417 void
2418 mark_elimination (int from, int to)
2420 basic_block bb;
2421 bitmap r;
2423 FOR_EACH_BB (bb)
2425 r = DF_LR_IN (bb);
2426 if (bitmap_bit_p (r, from))
2428 bitmap_clear_bit (r, from);
2429 bitmap_set_bit (r, to);
2431 if (! df_live)
2432 continue;
2433 r = DF_LIVE_IN (bb);
2434 if (bitmap_bit_p (r, from))
2436 bitmap_clear_bit (r, from);
2437 bitmap_set_bit (r, to);
2444 /* The length of the following array. */
2445 int ira_reg_equiv_len;
2447 /* Info about equiv. info for each register. */
2448 struct ira_reg_equiv *ira_reg_equiv;
2450 /* Expand ira_reg_equiv if necessary. */
2451 void
2452 ira_expand_reg_equiv (void)
2454 int old = ira_reg_equiv_len;
2456 if (ira_reg_equiv_len > max_reg_num ())
2457 return;
2458 ira_reg_equiv_len = max_reg_num () * 3 / 2 + 1;
2459 ira_reg_equiv
2460 = (struct ira_reg_equiv *) xrealloc (ira_reg_equiv,
2461 ira_reg_equiv_len
2462 * sizeof (struct ira_reg_equiv));
2463 gcc_assert (old < ira_reg_equiv_len);
2464 memset (ira_reg_equiv + old, 0,
2465 sizeof (struct ira_reg_equiv) * (ira_reg_equiv_len - old));
2468 static void
2469 init_reg_equiv (void)
2471 ira_reg_equiv_len = 0;
2472 ira_reg_equiv = NULL;
2473 ira_expand_reg_equiv ();
2476 static void
2477 finish_reg_equiv (void)
2479 free (ira_reg_equiv);
2484 struct equivalence
2486 /* Set when a REG_EQUIV note is found or created. Use to
2487 keep track of what memory accesses might be created later,
2488 e.g. by reload. */
2489 rtx replacement;
2490 rtx *src_p;
2491 /* The list of each instruction which initializes this register. */
2492 rtx init_insns;
2493 /* Loop depth is used to recognize equivalences which appear
2494 to be present within the same loop (or in an inner loop). */
2495 int loop_depth;
2496 /* Nonzero if this had a preexisting REG_EQUIV note. */
2497 int is_arg_equivalence;
2498 /* Set when an attempt should be made to replace a register
2499 with the associated src_p entry. */
2500 char replace;
2503 /* reg_equiv[N] (where N is a pseudo reg number) is the equivalence
2504 structure for that register. */
2505 static struct equivalence *reg_equiv;
2507 /* Used for communication between the following two functions: contains
2508 a MEM that we wish to ensure remains unchanged. */
2509 static rtx equiv_mem;
2511 /* Set nonzero if EQUIV_MEM is modified. */
2512 static int equiv_mem_modified;
2514 /* If EQUIV_MEM is modified by modifying DEST, indicate that it is modified.
2515 Called via note_stores. */
2516 static void
2517 validate_equiv_mem_from_store (rtx dest, const_rtx set ATTRIBUTE_UNUSED,
2518 void *data ATTRIBUTE_UNUSED)
2520 if ((REG_P (dest)
2521 && reg_overlap_mentioned_p (dest, equiv_mem))
2522 || (MEM_P (dest)
2523 && anti_dependence (equiv_mem, dest)))
2524 equiv_mem_modified = 1;
2527 /* Verify that no store between START and the death of REG invalidates
2528 MEMREF. MEMREF is invalidated by modifying a register used in MEMREF,
2529 by storing into an overlapping memory location, or with a non-const
2530 CALL_INSN.
2532 Return 1 if MEMREF remains valid. */
2533 static int
2534 validate_equiv_mem (rtx start, rtx reg, rtx memref)
2536 rtx insn;
2537 rtx note;
2539 equiv_mem = memref;
2540 equiv_mem_modified = 0;
2542 /* If the memory reference has side effects or is volatile, it isn't a
2543 valid equivalence. */
2544 if (side_effects_p (memref))
2545 return 0;
2547 for (insn = start; insn && ! equiv_mem_modified; insn = NEXT_INSN (insn))
2549 if (! INSN_P (insn))
2550 continue;
2552 if (find_reg_note (insn, REG_DEAD, reg))
2553 return 1;
2555 /* This used to ignore readonly memory and const/pure calls. The problem
2556 is the equivalent form may reference a pseudo which gets assigned a
2557 call clobbered hard reg. When we later replace REG with its
2558 equivalent form, the value in the call-clobbered reg has been
2559 changed and all hell breaks loose. */
2560 if (CALL_P (insn))
2561 return 0;
2563 note_stores (PATTERN (insn), validate_equiv_mem_from_store, NULL);
2565 /* If a register mentioned in MEMREF is modified via an
2566 auto-increment, we lose the equivalence. Do the same if one
2567 dies; although we could extend the life, it doesn't seem worth
2568 the trouble. */
2570 for (note = REG_NOTES (insn); note; note = XEXP (note, 1))
2571 if ((REG_NOTE_KIND (note) == REG_INC
2572 || REG_NOTE_KIND (note) == REG_DEAD)
2573 && REG_P (XEXP (note, 0))
2574 && reg_overlap_mentioned_p (XEXP (note, 0), memref))
2575 return 0;
2578 return 0;
2581 /* Returns zero if X is known to be invariant. */
2582 static int
2583 equiv_init_varies_p (rtx x)
2585 RTX_CODE code = GET_CODE (x);
2586 int i;
2587 const char *fmt;
2589 switch (code)
2591 case MEM:
2592 return !MEM_READONLY_P (x) || equiv_init_varies_p (XEXP (x, 0));
2594 case CONST:
2595 CASE_CONST_ANY:
2596 case SYMBOL_REF:
2597 case LABEL_REF:
2598 return 0;
2600 case REG:
2601 return reg_equiv[REGNO (x)].replace == 0 && rtx_varies_p (x, 0);
2603 case ASM_OPERANDS:
2604 if (MEM_VOLATILE_P (x))
2605 return 1;
2607 /* Fall through. */
2609 default:
2610 break;
2613 fmt = GET_RTX_FORMAT (code);
2614 for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--)
2615 if (fmt[i] == 'e')
2617 if (equiv_init_varies_p (XEXP (x, i)))
2618 return 1;
2620 else if (fmt[i] == 'E')
2622 int j;
2623 for (j = 0; j < XVECLEN (x, i); j++)
2624 if (equiv_init_varies_p (XVECEXP (x, i, j)))
2625 return 1;
2628 return 0;
2631 /* Returns nonzero if X (used to initialize register REGNO) is movable.
2632 X is only movable if the registers it uses have equivalent initializations
2633 which appear to be within the same loop (or in an inner loop) and movable
2634 or if they are not candidates for local_alloc and don't vary. */
2635 static int
2636 equiv_init_movable_p (rtx x, int regno)
2638 int i, j;
2639 const char *fmt;
2640 enum rtx_code code = GET_CODE (x);
2642 switch (code)
2644 case SET:
2645 return equiv_init_movable_p (SET_SRC (x), regno);
2647 case CC0:
2648 case CLOBBER:
2649 return 0;
2651 case PRE_INC:
2652 case PRE_DEC:
2653 case POST_INC:
2654 case POST_DEC:
2655 case PRE_MODIFY:
2656 case POST_MODIFY:
2657 return 0;
2659 case REG:
2660 return ((reg_equiv[REGNO (x)].loop_depth >= reg_equiv[regno].loop_depth
2661 && reg_equiv[REGNO (x)].replace)
2662 || (REG_BASIC_BLOCK (REGNO (x)) < NUM_FIXED_BLOCKS
2663 && ! rtx_varies_p (x, 0)));
2665 case UNSPEC_VOLATILE:
2666 return 0;
2668 case ASM_OPERANDS:
2669 if (MEM_VOLATILE_P (x))
2670 return 0;
2672 /* Fall through. */
2674 default:
2675 break;
2678 fmt = GET_RTX_FORMAT (code);
2679 for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--)
2680 switch (fmt[i])
2682 case 'e':
2683 if (! equiv_init_movable_p (XEXP (x, i), regno))
2684 return 0;
2685 break;
2686 case 'E':
2687 for (j = XVECLEN (x, i) - 1; j >= 0; j--)
2688 if (! equiv_init_movable_p (XVECEXP (x, i, j), regno))
2689 return 0;
2690 break;
2693 return 1;
2696 /* TRUE if X uses any registers for which reg_equiv[REGNO].replace is
2697 true. */
2698 static int
2699 contains_replace_regs (rtx x)
2701 int i, j;
2702 const char *fmt;
2703 enum rtx_code code = GET_CODE (x);
2705 switch (code)
2707 case CONST:
2708 case LABEL_REF:
2709 case SYMBOL_REF:
2710 CASE_CONST_ANY:
2711 case PC:
2712 case CC0:
2713 case HIGH:
2714 return 0;
2716 case REG:
2717 return reg_equiv[REGNO (x)].replace;
2719 default:
2720 break;
2723 fmt = GET_RTX_FORMAT (code);
2724 for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--)
2725 switch (fmt[i])
2727 case 'e':
2728 if (contains_replace_regs (XEXP (x, i)))
2729 return 1;
2730 break;
2731 case 'E':
2732 for (j = XVECLEN (x, i) - 1; j >= 0; j--)
2733 if (contains_replace_regs (XVECEXP (x, i, j)))
2734 return 1;
2735 break;
2738 return 0;
2741 /* TRUE if X references a memory location that would be affected by a store
2742 to MEMREF. */
2743 static int
2744 memref_referenced_p (rtx memref, rtx x)
2746 int i, j;
2747 const char *fmt;
2748 enum rtx_code code = GET_CODE (x);
2750 switch (code)
2752 case CONST:
2753 case LABEL_REF:
2754 case SYMBOL_REF:
2755 CASE_CONST_ANY:
2756 case PC:
2757 case CC0:
2758 case HIGH:
2759 case LO_SUM:
2760 return 0;
2762 case REG:
2763 return (reg_equiv[REGNO (x)].replacement
2764 && memref_referenced_p (memref,
2765 reg_equiv[REGNO (x)].replacement));
2767 case MEM:
2768 if (true_dependence (memref, VOIDmode, x))
2769 return 1;
2770 break;
2772 case SET:
2773 /* If we are setting a MEM, it doesn't count (its address does), but any
2774 other SET_DEST that has a MEM in it is referencing the MEM. */
2775 if (MEM_P (SET_DEST (x)))
2777 if (memref_referenced_p (memref, XEXP (SET_DEST (x), 0)))
2778 return 1;
2780 else if (memref_referenced_p (memref, SET_DEST (x)))
2781 return 1;
2783 return memref_referenced_p (memref, SET_SRC (x));
2785 default:
2786 break;
2789 fmt = GET_RTX_FORMAT (code);
2790 for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--)
2791 switch (fmt[i])
2793 case 'e':
2794 if (memref_referenced_p (memref, XEXP (x, i)))
2795 return 1;
2796 break;
2797 case 'E':
2798 for (j = XVECLEN (x, i) - 1; j >= 0; j--)
2799 if (memref_referenced_p (memref, XVECEXP (x, i, j)))
2800 return 1;
2801 break;
2804 return 0;
2807 /* TRUE if some insn in the range (START, END] references a memory location
2808 that would be affected by a store to MEMREF. */
2809 static int
2810 memref_used_between_p (rtx memref, rtx start, rtx end)
2812 rtx insn;
2814 for (insn = NEXT_INSN (start); insn != NEXT_INSN (end);
2815 insn = NEXT_INSN (insn))
2817 if (!NONDEBUG_INSN_P (insn))
2818 continue;
2820 if (memref_referenced_p (memref, PATTERN (insn)))
2821 return 1;
2823 /* Nonconst functions may access memory. */
2824 if (CALL_P (insn) && (! RTL_CONST_CALL_P (insn)))
2825 return 1;
2828 return 0;
2831 /* Mark REG as having no known equivalence.
2832 Some instructions might have been processed before and furnished
2833 with REG_EQUIV notes for this register; these notes will have to be
2834 removed.
2835 STORE is the piece of RTL that does the non-constant / conflicting
2836 assignment - a SET, CLOBBER or REG_INC note. It is currently not used,
2837 but needs to be there because this function is called from note_stores. */
2838 static void
2839 no_equiv (rtx reg, const_rtx store ATTRIBUTE_UNUSED,
2840 void *data ATTRIBUTE_UNUSED)
2842 int regno;
2843 rtx list;
2845 if (!REG_P (reg))
2846 return;
2847 regno = REGNO (reg);
2848 list = reg_equiv[regno].init_insns;
2849 if (list == const0_rtx)
2850 return;
2851 reg_equiv[regno].init_insns = const0_rtx;
2852 reg_equiv[regno].replacement = NULL_RTX;
2853 /* This doesn't matter for equivalences made for argument registers, we
2854 should keep their initialization insns. */
2855 if (reg_equiv[regno].is_arg_equivalence)
2856 return;
2857 ira_reg_equiv[regno].defined_p = false;
2858 ira_reg_equiv[regno].init_insns = NULL_RTX;
2859 for (; list; list = XEXP (list, 1))
2861 rtx insn = XEXP (list, 0);
2862 remove_note (insn, find_reg_note (insn, REG_EQUIV, NULL_RTX));
2866 /* Check whether the SUBREG is a paradoxical subreg and set the result
2867 in PDX_SUBREGS. */
2869 static int
2870 set_paradoxical_subreg (rtx *subreg, void *pdx_subregs)
2872 rtx reg;
2874 if ((*subreg) == NULL_RTX)
2875 return 1;
2876 if (GET_CODE (*subreg) != SUBREG)
2877 return 0;
2878 reg = SUBREG_REG (*subreg);
2879 if (!REG_P (reg))
2880 return 0;
2882 if (paradoxical_subreg_p (*subreg))
2883 ((bool *)pdx_subregs)[REGNO (reg)] = true;
2885 return 0;
2888 /* In DEBUG_INSN location adjust REGs from CLEARED_REGS bitmap to the
2889 equivalent replacement. */
2891 static rtx
2892 adjust_cleared_regs (rtx loc, const_rtx old_rtx ATTRIBUTE_UNUSED, void *data)
2894 if (REG_P (loc))
2896 bitmap cleared_regs = (bitmap) data;
2897 if (bitmap_bit_p (cleared_regs, REGNO (loc)))
2898 return simplify_replace_fn_rtx (*reg_equiv[REGNO (loc)].src_p,
2899 NULL_RTX, adjust_cleared_regs, data);
2901 return NULL_RTX;
2904 /* Nonzero if we recorded an equivalence for a LABEL_REF. */
2905 static int recorded_label_ref;
2907 /* Find registers that are equivalent to a single value throughout the
2908 compilation (either because they can be referenced in memory or are
2909 set once from a single constant). Lower their priority for a
2910 register.
2912 If such a register is only referenced once, try substituting its
2913 value into the using insn. If it succeeds, we can eliminate the
2914 register completely.
2916 Initialize init_insns in ira_reg_equiv array.
2918 Return non-zero if jump label rebuilding should be done. */
2919 static int
2920 update_equiv_regs (void)
2922 rtx insn;
2923 basic_block bb;
2924 int loop_depth;
2925 bitmap cleared_regs;
2926 bool *pdx_subregs;
2928 /* We need to keep track of whether or not we recorded a LABEL_REF so
2929 that we know if the jump optimizer needs to be rerun. */
2930 recorded_label_ref = 0;
2932 /* Use pdx_subregs to show whether a reg is used in a paradoxical
2933 subreg. */
2934 pdx_subregs = XCNEWVEC (bool, max_regno);
2936 reg_equiv = XCNEWVEC (struct equivalence, max_regno);
2937 grow_reg_equivs ();
2939 init_alias_analysis ();
2941 /* Scan insns and set pdx_subregs[regno] if the reg is used in a
2942 paradoxical subreg. Don't set such reg sequivalent to a mem,
2943 because lra will not substitute such equiv memory in order to
2944 prevent access beyond allocated memory for paradoxical memory subreg. */
2945 FOR_EACH_BB (bb)
2946 FOR_BB_INSNS (bb, insn)
2947 if (NONDEBUG_INSN_P (insn))
2948 for_each_rtx (&insn, set_paradoxical_subreg, (void *) pdx_subregs);
2950 /* Scan the insns and find which registers have equivalences. Do this
2951 in a separate scan of the insns because (due to -fcse-follow-jumps)
2952 a register can be set below its use. */
2953 FOR_EACH_BB (bb)
2955 loop_depth = bb_loop_depth (bb);
2957 for (insn = BB_HEAD (bb);
2958 insn != NEXT_INSN (BB_END (bb));
2959 insn = NEXT_INSN (insn))
2961 rtx note;
2962 rtx set;
2963 rtx dest, src;
2964 int regno;
2966 if (! INSN_P (insn))
2967 continue;
2969 for (note = REG_NOTES (insn); note; note = XEXP (note, 1))
2970 if (REG_NOTE_KIND (note) == REG_INC)
2971 no_equiv (XEXP (note, 0), note, NULL);
2973 set = single_set (insn);
2975 /* If this insn contains more (or less) than a single SET,
2976 only mark all destinations as having no known equivalence. */
2977 if (set == 0)
2979 note_stores (PATTERN (insn), no_equiv, NULL);
2980 continue;
2982 else if (GET_CODE (PATTERN (insn)) == PARALLEL)
2984 int i;
2986 for (i = XVECLEN (PATTERN (insn), 0) - 1; i >= 0; i--)
2988 rtx part = XVECEXP (PATTERN (insn), 0, i);
2989 if (part != set)
2990 note_stores (part, no_equiv, NULL);
2994 dest = SET_DEST (set);
2995 src = SET_SRC (set);
2997 /* See if this is setting up the equivalence between an argument
2998 register and its stack slot. */
2999 note = find_reg_note (insn, REG_EQUIV, NULL_RTX);
3000 if (note)
3002 gcc_assert (REG_P (dest));
3003 regno = REGNO (dest);
3005 /* Note that we don't want to clear init_insns in
3006 ira_reg_equiv even if there are multiple sets of this
3007 register. */
3008 reg_equiv[regno].is_arg_equivalence = 1;
3010 /* The insn result can have equivalence memory although
3011 the equivalence is not set up by the insn. We add
3012 this insn to init insns as it is a flag for now that
3013 regno has an equivalence. We will remove the insn
3014 from init insn list later. */
3015 if (rtx_equal_p (src, XEXP (note, 0)) || MEM_P (XEXP (note, 0)))
3016 ira_reg_equiv[regno].init_insns
3017 = gen_rtx_INSN_LIST (VOIDmode, insn,
3018 ira_reg_equiv[regno].init_insns);
3020 /* Continue normally in case this is a candidate for
3021 replacements. */
3024 if (!optimize)
3025 continue;
3027 /* We only handle the case of a pseudo register being set
3028 once, or always to the same value. */
3029 /* ??? The mn10200 port breaks if we add equivalences for
3030 values that need an ADDRESS_REGS register and set them equivalent
3031 to a MEM of a pseudo. The actual problem is in the over-conservative
3032 handling of INPADDR_ADDRESS / INPUT_ADDRESS / INPUT triples in
3033 calculate_needs, but we traditionally work around this problem
3034 here by rejecting equivalences when the destination is in a register
3035 that's likely spilled. This is fragile, of course, since the
3036 preferred class of a pseudo depends on all instructions that set
3037 or use it. */
3039 if (!REG_P (dest)
3040 || (regno = REGNO (dest)) < FIRST_PSEUDO_REGISTER
3041 || reg_equiv[regno].init_insns == const0_rtx
3042 || (targetm.class_likely_spilled_p (reg_preferred_class (regno))
3043 && MEM_P (src) && ! reg_equiv[regno].is_arg_equivalence))
3045 /* This might be setting a SUBREG of a pseudo, a pseudo that is
3046 also set somewhere else to a constant. */
3047 note_stores (set, no_equiv, NULL);
3048 continue;
3051 /* Don't set reg (if pdx_subregs[regno] == true) equivalent to a mem. */
3052 if (MEM_P (src) && pdx_subregs[regno])
3054 note_stores (set, no_equiv, NULL);
3055 continue;
3058 note = find_reg_note (insn, REG_EQUAL, NULL_RTX);
3060 /* cse sometimes generates function invariants, but doesn't put a
3061 REG_EQUAL note on the insn. Since this note would be redundant,
3062 there's no point creating it earlier than here. */
3063 if (! note && ! rtx_varies_p (src, 0))
3064 note = set_unique_reg_note (insn, REG_EQUAL, copy_rtx (src));
3066 /* Don't bother considering a REG_EQUAL note containing an EXPR_LIST
3067 since it represents a function call */
3068 if (note && GET_CODE (XEXP (note, 0)) == EXPR_LIST)
3069 note = NULL_RTX;
3071 if (DF_REG_DEF_COUNT (regno) != 1
3072 && (! note
3073 || rtx_varies_p (XEXP (note, 0), 0)
3074 || (reg_equiv[regno].replacement
3075 && ! rtx_equal_p (XEXP (note, 0),
3076 reg_equiv[regno].replacement))))
3078 no_equiv (dest, set, NULL);
3079 continue;
3081 /* Record this insn as initializing this register. */
3082 reg_equiv[regno].init_insns
3083 = gen_rtx_INSN_LIST (VOIDmode, insn, reg_equiv[regno].init_insns);
3085 /* If this register is known to be equal to a constant, record that
3086 it is always equivalent to the constant. */
3087 if (DF_REG_DEF_COUNT (regno) == 1
3088 && note && ! rtx_varies_p (XEXP (note, 0), 0))
3090 rtx note_value = XEXP (note, 0);
3091 remove_note (insn, note);
3092 set_unique_reg_note (insn, REG_EQUIV, note_value);
3095 /* If this insn introduces a "constant" register, decrease the priority
3096 of that register. Record this insn if the register is only used once
3097 more and the equivalence value is the same as our source.
3099 The latter condition is checked for two reasons: First, it is an
3100 indication that it may be more efficient to actually emit the insn
3101 as written (if no registers are available, reload will substitute
3102 the equivalence). Secondly, it avoids problems with any registers
3103 dying in this insn whose death notes would be missed.
3105 If we don't have a REG_EQUIV note, see if this insn is loading
3106 a register used only in one basic block from a MEM. If so, and the
3107 MEM remains unchanged for the life of the register, add a REG_EQUIV
3108 note. */
3110 note = find_reg_note (insn, REG_EQUIV, NULL_RTX);
3112 if (note == 0 && REG_BASIC_BLOCK (regno) >= NUM_FIXED_BLOCKS
3113 && MEM_P (SET_SRC (set))
3114 && validate_equiv_mem (insn, dest, SET_SRC (set)))
3115 note = set_unique_reg_note (insn, REG_EQUIV, copy_rtx (SET_SRC (set)));
3117 if (note)
3119 int regno = REGNO (dest);
3120 rtx x = XEXP (note, 0);
3122 /* If we haven't done so, record for reload that this is an
3123 equivalencing insn. */
3124 if (!reg_equiv[regno].is_arg_equivalence)
3125 ira_reg_equiv[regno].init_insns
3126 = gen_rtx_INSN_LIST (VOIDmode, insn,
3127 ira_reg_equiv[regno].init_insns);
3129 /* Record whether or not we created a REG_EQUIV note for a LABEL_REF.
3130 We might end up substituting the LABEL_REF for uses of the
3131 pseudo here or later. That kind of transformation may turn an
3132 indirect jump into a direct jump, in which case we must rerun the
3133 jump optimizer to ensure that the JUMP_LABEL fields are valid. */
3134 if (GET_CODE (x) == LABEL_REF
3135 || (GET_CODE (x) == CONST
3136 && GET_CODE (XEXP (x, 0)) == PLUS
3137 && (GET_CODE (XEXP (XEXP (x, 0), 0)) == LABEL_REF)))
3138 recorded_label_ref = 1;
3140 reg_equiv[regno].replacement = x;
3141 reg_equiv[regno].src_p = &SET_SRC (set);
3142 reg_equiv[regno].loop_depth = loop_depth;
3144 /* Don't mess with things live during setjmp. */
3145 if (REG_LIVE_LENGTH (regno) >= 0 && optimize)
3147 /* Note that the statement below does not affect the priority
3148 in local-alloc! */
3149 REG_LIVE_LENGTH (regno) *= 2;
3151 /* If the register is referenced exactly twice, meaning it is
3152 set once and used once, indicate that the reference may be
3153 replaced by the equivalence we computed above. Do this
3154 even if the register is only used in one block so that
3155 dependencies can be handled where the last register is
3156 used in a different block (i.e. HIGH / LO_SUM sequences)
3157 and to reduce the number of registers alive across
3158 calls. */
3160 if (REG_N_REFS (regno) == 2
3161 && (rtx_equal_p (x, src)
3162 || ! equiv_init_varies_p (src))
3163 && NONJUMP_INSN_P (insn)
3164 && equiv_init_movable_p (PATTERN (insn), regno))
3165 reg_equiv[regno].replace = 1;
3171 if (!optimize)
3172 goto out;
3174 /* A second pass, to gather additional equivalences with memory. This needs
3175 to be done after we know which registers we are going to replace. */
3177 for (insn = get_insns (); insn; insn = NEXT_INSN (insn))
3179 rtx set, src, dest;
3180 unsigned regno;
3182 if (! INSN_P (insn))
3183 continue;
3185 set = single_set (insn);
3186 if (! set)
3187 continue;
3189 dest = SET_DEST (set);
3190 src = SET_SRC (set);
3192 /* If this sets a MEM to the contents of a REG that is only used
3193 in a single basic block, see if the register is always equivalent
3194 to that memory location and if moving the store from INSN to the
3195 insn that set REG is safe. If so, put a REG_EQUIV note on the
3196 initializing insn.
3198 Don't add a REG_EQUIV note if the insn already has one. The existing
3199 REG_EQUIV is likely more useful than the one we are adding.
3201 If one of the regs in the address has reg_equiv[REGNO].replace set,
3202 then we can't add this REG_EQUIV note. The reg_equiv[REGNO].replace
3203 optimization may move the set of this register immediately before
3204 insn, which puts it after reg_equiv[REGNO].init_insns, and hence
3205 the mention in the REG_EQUIV note would be to an uninitialized
3206 pseudo. */
3208 if (MEM_P (dest) && REG_P (src)
3209 && (regno = REGNO (src)) >= FIRST_PSEUDO_REGISTER
3210 && REG_BASIC_BLOCK (regno) >= NUM_FIXED_BLOCKS
3211 && DF_REG_DEF_COUNT (regno) == 1
3212 && reg_equiv[regno].init_insns != 0
3213 && reg_equiv[regno].init_insns != const0_rtx
3214 && ! find_reg_note (XEXP (reg_equiv[regno].init_insns, 0),
3215 REG_EQUIV, NULL_RTX)
3216 && ! contains_replace_regs (XEXP (dest, 0))
3217 && ! pdx_subregs[regno])
3219 rtx init_insn = XEXP (reg_equiv[regno].init_insns, 0);
3220 if (validate_equiv_mem (init_insn, src, dest)
3221 && ! memref_used_between_p (dest, init_insn, insn)
3222 /* Attaching a REG_EQUIV note will fail if INIT_INSN has
3223 multiple sets. */
3224 && set_unique_reg_note (init_insn, REG_EQUIV, copy_rtx (dest)))
3226 /* This insn makes the equivalence, not the one initializing
3227 the register. */
3228 ira_reg_equiv[regno].init_insns
3229 = gen_rtx_INSN_LIST (VOIDmode, insn, NULL_RTX);
3230 df_notes_rescan (init_insn);
3235 cleared_regs = BITMAP_ALLOC (NULL);
3236 /* Now scan all regs killed in an insn to see if any of them are
3237 registers only used that once. If so, see if we can replace the
3238 reference with the equivalent form. If we can, delete the
3239 initializing reference and this register will go away. If we
3240 can't replace the reference, and the initializing reference is
3241 within the same loop (or in an inner loop), then move the register
3242 initialization just before the use, so that they are in the same
3243 basic block. */
3244 FOR_EACH_BB_REVERSE (bb)
3246 loop_depth = bb_loop_depth (bb);
3247 for (insn = BB_END (bb);
3248 insn != PREV_INSN (BB_HEAD (bb));
3249 insn = PREV_INSN (insn))
3251 rtx link;
3253 if (! INSN_P (insn))
3254 continue;
3256 /* Don't substitute into a non-local goto, this confuses CFG. */
3257 if (JUMP_P (insn)
3258 && find_reg_note (insn, REG_NON_LOCAL_GOTO, NULL_RTX))
3259 continue;
3261 for (link = REG_NOTES (insn); link; link = XEXP (link, 1))
3263 if (REG_NOTE_KIND (link) == REG_DEAD
3264 /* Make sure this insn still refers to the register. */
3265 && reg_mentioned_p (XEXP (link, 0), PATTERN (insn)))
3267 int regno = REGNO (XEXP (link, 0));
3268 rtx equiv_insn;
3270 if (! reg_equiv[regno].replace
3271 || reg_equiv[regno].loop_depth < loop_depth
3272 /* There is no sense to move insns if we did
3273 register pressure-sensitive scheduling was
3274 done because it will not improve allocation
3275 but worsen insn schedule with a big
3276 probability. */
3277 || (flag_sched_pressure && flag_schedule_insns))
3278 continue;
3280 /* reg_equiv[REGNO].replace gets set only when
3281 REG_N_REFS[REGNO] is 2, i.e. the register is set
3282 once and used once. (If it were only set, but
3283 not used, flow would have deleted the setting
3284 insns.) Hence there can only be one insn in
3285 reg_equiv[REGNO].init_insns. */
3286 gcc_assert (reg_equiv[regno].init_insns
3287 && !XEXP (reg_equiv[regno].init_insns, 1));
3288 equiv_insn = XEXP (reg_equiv[regno].init_insns, 0);
3290 /* We may not move instructions that can throw, since
3291 that changes basic block boundaries and we are not
3292 prepared to adjust the CFG to match. */
3293 if (can_throw_internal (equiv_insn))
3294 continue;
3296 if (asm_noperands (PATTERN (equiv_insn)) < 0
3297 && validate_replace_rtx (regno_reg_rtx[regno],
3298 *(reg_equiv[regno].src_p), insn))
3300 rtx equiv_link;
3301 rtx last_link;
3302 rtx note;
3304 /* Find the last note. */
3305 for (last_link = link; XEXP (last_link, 1);
3306 last_link = XEXP (last_link, 1))
3309 /* Append the REG_DEAD notes from equiv_insn. */
3310 equiv_link = REG_NOTES (equiv_insn);
3311 while (equiv_link)
3313 note = equiv_link;
3314 equiv_link = XEXP (equiv_link, 1);
3315 if (REG_NOTE_KIND (note) == REG_DEAD)
3317 remove_note (equiv_insn, note);
3318 XEXP (last_link, 1) = note;
3319 XEXP (note, 1) = NULL_RTX;
3320 last_link = note;
3324 remove_death (regno, insn);
3325 SET_REG_N_REFS (regno, 0);
3326 REG_FREQ (regno) = 0;
3327 delete_insn (equiv_insn);
3329 reg_equiv[regno].init_insns
3330 = XEXP (reg_equiv[regno].init_insns, 1);
3332 ira_reg_equiv[regno].init_insns = NULL_RTX;
3333 bitmap_set_bit (cleared_regs, regno);
3335 /* Move the initialization of the register to just before
3336 INSN. Update the flow information. */
3337 else if (prev_nondebug_insn (insn) != equiv_insn)
3339 rtx new_insn;
3341 new_insn = emit_insn_before (PATTERN (equiv_insn), insn);
3342 REG_NOTES (new_insn) = REG_NOTES (equiv_insn);
3343 REG_NOTES (equiv_insn) = 0;
3344 /* Rescan it to process the notes. */
3345 df_insn_rescan (new_insn);
3347 /* Make sure this insn is recognized before
3348 reload begins, otherwise
3349 eliminate_regs_in_insn will die. */
3350 INSN_CODE (new_insn) = INSN_CODE (equiv_insn);
3352 delete_insn (equiv_insn);
3354 XEXP (reg_equiv[regno].init_insns, 0) = new_insn;
3356 REG_BASIC_BLOCK (regno) = bb->index;
3357 REG_N_CALLS_CROSSED (regno) = 0;
3358 REG_FREQ_CALLS_CROSSED (regno) = 0;
3359 REG_N_THROWING_CALLS_CROSSED (regno) = 0;
3360 REG_LIVE_LENGTH (regno) = 2;
3362 if (insn == BB_HEAD (bb))
3363 BB_HEAD (bb) = PREV_INSN (insn);
3365 ira_reg_equiv[regno].init_insns
3366 = gen_rtx_INSN_LIST (VOIDmode, new_insn, NULL_RTX);
3367 bitmap_set_bit (cleared_regs, regno);
3374 if (!bitmap_empty_p (cleared_regs))
3376 FOR_EACH_BB (bb)
3378 bitmap_and_compl_into (DF_LR_IN (bb), cleared_regs);
3379 bitmap_and_compl_into (DF_LR_OUT (bb), cleared_regs);
3380 if (! df_live)
3381 continue;
3382 bitmap_and_compl_into (DF_LIVE_IN (bb), cleared_regs);
3383 bitmap_and_compl_into (DF_LIVE_OUT (bb), cleared_regs);
3386 /* Last pass - adjust debug insns referencing cleared regs. */
3387 if (MAY_HAVE_DEBUG_INSNS)
3388 for (insn = get_insns (); insn; insn = NEXT_INSN (insn))
3389 if (DEBUG_INSN_P (insn))
3391 rtx old_loc = INSN_VAR_LOCATION_LOC (insn);
3392 INSN_VAR_LOCATION_LOC (insn)
3393 = simplify_replace_fn_rtx (old_loc, NULL_RTX,
3394 adjust_cleared_regs,
3395 (void *) cleared_regs);
3396 if (old_loc != INSN_VAR_LOCATION_LOC (insn))
3397 df_insn_rescan (insn);
3401 BITMAP_FREE (cleared_regs);
3403 out:
3404 /* Clean up. */
3406 end_alias_analysis ();
3407 free (reg_equiv);
3408 free (pdx_subregs);
3409 return recorded_label_ref;
3414 /* Set up fields memory, constant, and invariant from init_insns in
3415 the structures of array ira_reg_equiv. */
3416 static void
3417 setup_reg_equiv (void)
3419 int i;
3420 rtx elem, prev_elem, next_elem, insn, set, x;
3422 for (i = FIRST_PSEUDO_REGISTER; i < ira_reg_equiv_len; i++)
3423 for (prev_elem = NULL, elem = ira_reg_equiv[i].init_insns;
3424 elem;
3425 prev_elem = elem, elem = next_elem)
3427 next_elem = XEXP (elem, 1);
3428 insn = XEXP (elem, 0);
3429 set = single_set (insn);
3431 /* Init insns can set up equivalence when the reg is a destination or
3432 a source (in this case the destination is memory). */
3433 if (set != 0 && (REG_P (SET_DEST (set)) || REG_P (SET_SRC (set))))
3435 if ((x = find_reg_note (insn, REG_EQUIV, NULL_RTX)) != NULL)
3437 x = XEXP (x, 0);
3438 if (REG_P (SET_DEST (set))
3439 && REGNO (SET_DEST (set)) == (unsigned int) i
3440 && ! rtx_equal_p (SET_SRC (set), x) && MEM_P (x))
3442 /* This insn reporting the equivalence but
3443 actually not setting it. Remove it from the
3444 list. */
3445 if (prev_elem == NULL)
3446 ira_reg_equiv[i].init_insns = next_elem;
3447 else
3448 XEXP (prev_elem, 1) = next_elem;
3449 elem = prev_elem;
3452 else if (REG_P (SET_DEST (set))
3453 && REGNO (SET_DEST (set)) == (unsigned int) i)
3454 x = SET_SRC (set);
3455 else
3457 gcc_assert (REG_P (SET_SRC (set))
3458 && REGNO (SET_SRC (set)) == (unsigned int) i);
3459 x = SET_DEST (set);
3461 if (! function_invariant_p (x)
3462 || ! flag_pic
3463 /* A function invariant is often CONSTANT_P but may
3464 include a register. We promise to only pass
3465 CONSTANT_P objects to LEGITIMATE_PIC_OPERAND_P. */
3466 || (CONSTANT_P (x) && LEGITIMATE_PIC_OPERAND_P (x)))
3468 /* It can happen that a REG_EQUIV note contains a MEM
3469 that is not a legitimate memory operand. As later
3470 stages of reload assume that all addresses found in
3471 the lra_regno_equiv_* arrays were originally
3472 legitimate, we ignore such REG_EQUIV notes. */
3473 if (memory_operand (x, VOIDmode))
3475 ira_reg_equiv[i].defined_p = true;
3476 ira_reg_equiv[i].memory = x;
3477 continue;
3479 else if (function_invariant_p (x))
3481 enum machine_mode mode;
3483 mode = GET_MODE (SET_DEST (set));
3484 if (GET_CODE (x) == PLUS
3485 || x == frame_pointer_rtx || x == arg_pointer_rtx)
3486 /* This is PLUS of frame pointer and a constant,
3487 or fp, or argp. */
3488 ira_reg_equiv[i].invariant = x;
3489 else if (targetm.legitimate_constant_p (mode, x))
3490 ira_reg_equiv[i].constant = x;
3491 else
3493 ira_reg_equiv[i].memory = force_const_mem (mode, x);
3494 if (ira_reg_equiv[i].memory == NULL_RTX)
3496 ira_reg_equiv[i].defined_p = false;
3497 ira_reg_equiv[i].init_insns = NULL_RTX;
3498 break;
3501 ira_reg_equiv[i].defined_p = true;
3502 continue;
3506 ira_reg_equiv[i].defined_p = false;
3507 ira_reg_equiv[i].init_insns = NULL_RTX;
3508 break;
3514 /* Print chain C to FILE. */
3515 static void
3516 print_insn_chain (FILE *file, struct insn_chain *c)
3518 fprintf (file, "insn=%d, ", INSN_UID (c->insn));
3519 bitmap_print (file, &c->live_throughout, "live_throughout: ", ", ");
3520 bitmap_print (file, &c->dead_or_set, "dead_or_set: ", "\n");
3524 /* Print all reload_insn_chains to FILE. */
3525 static void
3526 print_insn_chains (FILE *file)
3528 struct insn_chain *c;
3529 for (c = reload_insn_chain; c ; c = c->next)
3530 print_insn_chain (file, c);
3533 /* Return true if pseudo REGNO should be added to set live_throughout
3534 or dead_or_set of the insn chains for reload consideration. */
3535 static bool
3536 pseudo_for_reload_consideration_p (int regno)
3538 /* Consider spilled pseudos too for IRA because they still have a
3539 chance to get hard-registers in the reload when IRA is used. */
3540 return (reg_renumber[regno] >= 0 || ira_conflicts_p);
3543 /* Init LIVE_SUBREGS[ALLOCNUM] and LIVE_SUBREGS_USED[ALLOCNUM] using
3544 REG to the number of nregs, and INIT_VALUE to get the
3545 initialization. ALLOCNUM need not be the regno of REG. */
3546 static void
3547 init_live_subregs (bool init_value, sbitmap *live_subregs,
3548 bitmap live_subregs_used, int allocnum, rtx reg)
3550 unsigned int regno = REGNO (SUBREG_REG (reg));
3551 int size = GET_MODE_SIZE (GET_MODE (regno_reg_rtx[regno]));
3553 gcc_assert (size > 0);
3555 /* Been there, done that. */
3556 if (bitmap_bit_p (live_subregs_used, allocnum))
3557 return;
3559 /* Create a new one. */
3560 if (live_subregs[allocnum] == NULL)
3561 live_subregs[allocnum] = sbitmap_alloc (size);
3563 /* If the entire reg was live before blasting into subregs, we need
3564 to init all of the subregs to ones else init to 0. */
3565 if (init_value)
3566 bitmap_ones (live_subregs[allocnum]);
3567 else
3568 bitmap_clear (live_subregs[allocnum]);
3570 bitmap_set_bit (live_subregs_used, allocnum);
3573 /* Walk the insns of the current function and build reload_insn_chain,
3574 and record register life information. */
3575 static void
3576 build_insn_chain (void)
3578 unsigned int i;
3579 struct insn_chain **p = &reload_insn_chain;
3580 basic_block bb;
3581 struct insn_chain *c = NULL;
3582 struct insn_chain *next = NULL;
3583 bitmap live_relevant_regs = BITMAP_ALLOC (NULL);
3584 bitmap elim_regset = BITMAP_ALLOC (NULL);
3585 /* live_subregs is a vector used to keep accurate information about
3586 which hardregs are live in multiword pseudos. live_subregs and
3587 live_subregs_used are indexed by pseudo number. The live_subreg
3588 entry for a particular pseudo is only used if the corresponding
3589 element is non zero in live_subregs_used. The sbitmap size of
3590 live_subreg[allocno] is number of bytes that the pseudo can
3591 occupy. */
3592 sbitmap *live_subregs = XCNEWVEC (sbitmap, max_regno);
3593 bitmap live_subregs_used = BITMAP_ALLOC (NULL);
3595 for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
3596 if (TEST_HARD_REG_BIT (eliminable_regset, i))
3597 bitmap_set_bit (elim_regset, i);
3598 FOR_EACH_BB_REVERSE (bb)
3600 bitmap_iterator bi;
3601 rtx insn;
3603 CLEAR_REG_SET (live_relevant_regs);
3604 bitmap_clear (live_subregs_used);
3606 EXECUTE_IF_SET_IN_BITMAP (df_get_live_out (bb), 0, i, bi)
3608 if (i >= FIRST_PSEUDO_REGISTER)
3609 break;
3610 bitmap_set_bit (live_relevant_regs, i);
3613 EXECUTE_IF_SET_IN_BITMAP (df_get_live_out (bb),
3614 FIRST_PSEUDO_REGISTER, i, bi)
3616 if (pseudo_for_reload_consideration_p (i))
3617 bitmap_set_bit (live_relevant_regs, i);
3620 FOR_BB_INSNS_REVERSE (bb, insn)
3622 if (!NOTE_P (insn) && !BARRIER_P (insn))
3624 unsigned int uid = INSN_UID (insn);
3625 df_ref *def_rec;
3626 df_ref *use_rec;
3628 c = new_insn_chain ();
3629 c->next = next;
3630 next = c;
3631 *p = c;
3632 p = &c->prev;
3634 c->insn = insn;
3635 c->block = bb->index;
3637 if (NONDEBUG_INSN_P (insn))
3638 for (def_rec = DF_INSN_UID_DEFS (uid); *def_rec; def_rec++)
3640 df_ref def = *def_rec;
3641 unsigned int regno = DF_REF_REGNO (def);
3643 /* Ignore may clobbers because these are generated
3644 from calls. However, every other kind of def is
3645 added to dead_or_set. */
3646 if (!DF_REF_FLAGS_IS_SET (def, DF_REF_MAY_CLOBBER))
3648 if (regno < FIRST_PSEUDO_REGISTER)
3650 if (!fixed_regs[regno])
3651 bitmap_set_bit (&c->dead_or_set, regno);
3653 else if (pseudo_for_reload_consideration_p (regno))
3654 bitmap_set_bit (&c->dead_or_set, regno);
3657 if ((regno < FIRST_PSEUDO_REGISTER
3658 || reg_renumber[regno] >= 0
3659 || ira_conflicts_p)
3660 && (!DF_REF_FLAGS_IS_SET (def, DF_REF_CONDITIONAL)))
3662 rtx reg = DF_REF_REG (def);
3664 /* We can model subregs, but not if they are
3665 wrapped in ZERO_EXTRACTS. */
3666 if (GET_CODE (reg) == SUBREG
3667 && !DF_REF_FLAGS_IS_SET (def, DF_REF_ZERO_EXTRACT))
3669 unsigned int start = SUBREG_BYTE (reg);
3670 unsigned int last = start
3671 + GET_MODE_SIZE (GET_MODE (reg));
3673 init_live_subregs
3674 (bitmap_bit_p (live_relevant_regs, regno),
3675 live_subregs, live_subregs_used, regno, reg);
3677 if (!DF_REF_FLAGS_IS_SET
3678 (def, DF_REF_STRICT_LOW_PART))
3680 /* Expand the range to cover entire words.
3681 Bytes added here are "don't care". */
3682 start
3683 = start / UNITS_PER_WORD * UNITS_PER_WORD;
3684 last = ((last + UNITS_PER_WORD - 1)
3685 / UNITS_PER_WORD * UNITS_PER_WORD);
3688 /* Ignore the paradoxical bits. */
3689 if (last > SBITMAP_SIZE (live_subregs[regno]))
3690 last = SBITMAP_SIZE (live_subregs[regno]);
3692 while (start < last)
3694 bitmap_clear_bit (live_subregs[regno], start);
3695 start++;
3698 if (bitmap_empty_p (live_subregs[regno]))
3700 bitmap_clear_bit (live_subregs_used, regno);
3701 bitmap_clear_bit (live_relevant_regs, regno);
3703 else
3704 /* Set live_relevant_regs here because
3705 that bit has to be true to get us to
3706 look at the live_subregs fields. */
3707 bitmap_set_bit (live_relevant_regs, regno);
3709 else
3711 /* DF_REF_PARTIAL is generated for
3712 subregs, STRICT_LOW_PART, and
3713 ZERO_EXTRACT. We handle the subreg
3714 case above so here we have to keep from
3715 modeling the def as a killing def. */
3716 if (!DF_REF_FLAGS_IS_SET (def, DF_REF_PARTIAL))
3718 bitmap_clear_bit (live_subregs_used, regno);
3719 bitmap_clear_bit (live_relevant_regs, regno);
3725 bitmap_and_compl_into (live_relevant_regs, elim_regset);
3726 bitmap_copy (&c->live_throughout, live_relevant_regs);
3728 if (NONDEBUG_INSN_P (insn))
3729 for (use_rec = DF_INSN_UID_USES (uid); *use_rec; use_rec++)
3731 df_ref use = *use_rec;
3732 unsigned int regno = DF_REF_REGNO (use);
3733 rtx reg = DF_REF_REG (use);
3735 /* DF_REF_READ_WRITE on a use means that this use
3736 is fabricated from a def that is a partial set
3737 to a multiword reg. Here, we only model the
3738 subreg case that is not wrapped in ZERO_EXTRACT
3739 precisely so we do not need to look at the
3740 fabricated use. */
3741 if (DF_REF_FLAGS_IS_SET (use, DF_REF_READ_WRITE)
3742 && !DF_REF_FLAGS_IS_SET (use, DF_REF_ZERO_EXTRACT)
3743 && DF_REF_FLAGS_IS_SET (use, DF_REF_SUBREG))
3744 continue;
3746 /* Add the last use of each var to dead_or_set. */
3747 if (!bitmap_bit_p (live_relevant_regs, regno))
3749 if (regno < FIRST_PSEUDO_REGISTER)
3751 if (!fixed_regs[regno])
3752 bitmap_set_bit (&c->dead_or_set, regno);
3754 else if (pseudo_for_reload_consideration_p (regno))
3755 bitmap_set_bit (&c->dead_or_set, regno);
3758 if (regno < FIRST_PSEUDO_REGISTER
3759 || pseudo_for_reload_consideration_p (regno))
3761 if (GET_CODE (reg) == SUBREG
3762 && !DF_REF_FLAGS_IS_SET (use,
3763 DF_REF_SIGN_EXTRACT
3764 | DF_REF_ZERO_EXTRACT))
3766 unsigned int start = SUBREG_BYTE (reg);
3767 unsigned int last = start
3768 + GET_MODE_SIZE (GET_MODE (reg));
3770 init_live_subregs
3771 (bitmap_bit_p (live_relevant_regs, regno),
3772 live_subregs, live_subregs_used, regno, reg);
3774 /* Ignore the paradoxical bits. */
3775 if (last > SBITMAP_SIZE (live_subregs[regno]))
3776 last = SBITMAP_SIZE (live_subregs[regno]);
3778 while (start < last)
3780 bitmap_set_bit (live_subregs[regno], start);
3781 start++;
3784 else
3785 /* Resetting the live_subregs_used is
3786 effectively saying do not use the subregs
3787 because we are reading the whole
3788 pseudo. */
3789 bitmap_clear_bit (live_subregs_used, regno);
3790 bitmap_set_bit (live_relevant_regs, regno);
3796 /* FIXME!! The following code is a disaster. Reload needs to see the
3797 labels and jump tables that are just hanging out in between
3798 the basic blocks. See pr33676. */
3799 insn = BB_HEAD (bb);
3801 /* Skip over the barriers and cruft. */
3802 while (insn && (BARRIER_P (insn) || NOTE_P (insn)
3803 || BLOCK_FOR_INSN (insn) == bb))
3804 insn = PREV_INSN (insn);
3806 /* While we add anything except barriers and notes, the focus is
3807 to get the labels and jump tables into the
3808 reload_insn_chain. */
3809 while (insn)
3811 if (!NOTE_P (insn) && !BARRIER_P (insn))
3813 if (BLOCK_FOR_INSN (insn))
3814 break;
3816 c = new_insn_chain ();
3817 c->next = next;
3818 next = c;
3819 *p = c;
3820 p = &c->prev;
3822 /* The block makes no sense here, but it is what the old
3823 code did. */
3824 c->block = bb->index;
3825 c->insn = insn;
3826 bitmap_copy (&c->live_throughout, live_relevant_regs);
3828 insn = PREV_INSN (insn);
3832 reload_insn_chain = c;
3833 *p = NULL;
3835 for (i = 0; i < (unsigned int) max_regno; i++)
3836 if (live_subregs[i] != NULL)
3837 sbitmap_free (live_subregs[i]);
3838 free (live_subregs);
3839 BITMAP_FREE (live_subregs_used);
3840 BITMAP_FREE (live_relevant_regs);
3841 BITMAP_FREE (elim_regset);
3843 if (dump_file)
3844 print_insn_chains (dump_file);
3847 /* Examine the rtx found in *LOC, which is read or written to as determined
3848 by TYPE. Return false if we find a reason why an insn containing this
3849 rtx should not be moved (such as accesses to non-constant memory), true
3850 otherwise. */
3851 static bool
3852 rtx_moveable_p (rtx *loc, enum op_type type)
3854 const char *fmt;
3855 rtx x = *loc;
3856 enum rtx_code code = GET_CODE (x);
3857 int i, j;
3859 code = GET_CODE (x);
3860 switch (code)
3862 case CONST:
3863 CASE_CONST_ANY:
3864 case SYMBOL_REF:
3865 case LABEL_REF:
3866 return true;
3868 case PC:
3869 return type == OP_IN;
3871 case CC0:
3872 return false;
3874 case REG:
3875 if (x == frame_pointer_rtx)
3876 return true;
3877 if (HARD_REGISTER_P (x))
3878 return false;
3880 return true;
3882 case MEM:
3883 if (type == OP_IN && MEM_READONLY_P (x))
3884 return rtx_moveable_p (&XEXP (x, 0), OP_IN);
3885 return false;
3887 case SET:
3888 return (rtx_moveable_p (&SET_SRC (x), OP_IN)
3889 && rtx_moveable_p (&SET_DEST (x), OP_OUT));
3891 case STRICT_LOW_PART:
3892 return rtx_moveable_p (&XEXP (x, 0), OP_OUT);
3894 case ZERO_EXTRACT:
3895 case SIGN_EXTRACT:
3896 return (rtx_moveable_p (&XEXP (x, 0), type)
3897 && rtx_moveable_p (&XEXP (x, 1), OP_IN)
3898 && rtx_moveable_p (&XEXP (x, 2), OP_IN));
3900 case CLOBBER:
3901 return rtx_moveable_p (&SET_DEST (x), OP_OUT);
3903 default:
3904 break;
3907 fmt = GET_RTX_FORMAT (code);
3908 for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--)
3910 if (fmt[i] == 'e')
3912 if (!rtx_moveable_p (&XEXP (x, i), type))
3913 return false;
3915 else if (fmt[i] == 'E')
3916 for (j = XVECLEN (x, i) - 1; j >= 0; j--)
3918 if (!rtx_moveable_p (&XVECEXP (x, i, j), type))
3919 return false;
3922 return true;
3925 /* A wrapper around dominated_by_p, which uses the information in UID_LUID
3926 to give dominance relationships between two insns I1 and I2. */
3927 static bool
3928 insn_dominated_by_p (rtx i1, rtx i2, int *uid_luid)
3930 basic_block bb1 = BLOCK_FOR_INSN (i1);
3931 basic_block bb2 = BLOCK_FOR_INSN (i2);
3933 if (bb1 == bb2)
3934 return uid_luid[INSN_UID (i2)] < uid_luid[INSN_UID (i1)];
3935 return dominated_by_p (CDI_DOMINATORS, bb1, bb2);
3938 /* Record the range of register numbers added by find_moveable_pseudos. */
3939 int first_moveable_pseudo, last_moveable_pseudo;
3941 /* These two vectors hold data for every register added by
3942 find_movable_pseudos, with index 0 holding data for the
3943 first_moveable_pseudo. */
3944 /* The original home register. */
3945 static vec<rtx> pseudo_replaced_reg;
3947 /* Look for instances where we have an instruction that is known to increase
3948 register pressure, and whose result is not used immediately. If it is
3949 possible to move the instruction downwards to just before its first use,
3950 split its lifetime into two ranges. We create a new pseudo to compute the
3951 value, and emit a move instruction just before the first use. If, after
3952 register allocation, the new pseudo remains unallocated, the function
3953 move_unallocated_pseudos then deletes the move instruction and places
3954 the computation just before the first use.
3956 Such a move is safe and profitable if all the input registers remain live
3957 and unchanged between the original computation and its first use. In such
3958 a situation, the computation is known to increase register pressure, and
3959 moving it is known to at least not worsen it.
3961 We restrict moves to only those cases where a register remains unallocated,
3962 in order to avoid interfering too much with the instruction schedule. As
3963 an exception, we may move insns which only modify their input register
3964 (typically induction variables), as this increases the freedom for our
3965 intended transformation, and does not limit the second instruction
3966 scheduler pass. */
3968 static void
3969 find_moveable_pseudos (void)
3971 unsigned i;
3972 int max_regs = max_reg_num ();
3973 int max_uid = get_max_uid ();
3974 basic_block bb;
3975 int *uid_luid = XNEWVEC (int, max_uid);
3976 rtx *closest_uses = XNEWVEC (rtx, max_regs);
3977 /* A set of registers which are live but not modified throughout a block. */
3978 bitmap_head *bb_transp_live = XNEWVEC (bitmap_head, last_basic_block);
3979 /* A set of registers which only exist in a given basic block. */
3980 bitmap_head *bb_local = XNEWVEC (bitmap_head, last_basic_block);
3981 /* A set of registers which are set once, in an instruction that can be
3982 moved freely downwards, but are otherwise transparent to a block. */
3983 bitmap_head *bb_moveable_reg_sets = XNEWVEC (bitmap_head, last_basic_block);
3984 bitmap_head live, used, set, interesting, unusable_as_input;
3985 bitmap_iterator bi;
3986 bitmap_initialize (&interesting, 0);
3988 first_moveable_pseudo = max_regs;
3989 pseudo_replaced_reg.release ();
3990 pseudo_replaced_reg.safe_grow_cleared (max_regs);
3992 df_analyze ();
3993 calculate_dominance_info (CDI_DOMINATORS);
3995 i = 0;
3996 bitmap_initialize (&live, 0);
3997 bitmap_initialize (&used, 0);
3998 bitmap_initialize (&set, 0);
3999 bitmap_initialize (&unusable_as_input, 0);
4000 FOR_EACH_BB (bb)
4002 rtx insn;
4003 bitmap transp = bb_transp_live + bb->index;
4004 bitmap moveable = bb_moveable_reg_sets + bb->index;
4005 bitmap local = bb_local + bb->index;
4007 bitmap_initialize (local, 0);
4008 bitmap_initialize (transp, 0);
4009 bitmap_initialize (moveable, 0);
4010 bitmap_copy (&live, df_get_live_out (bb));
4011 bitmap_and_into (&live, df_get_live_in (bb));
4012 bitmap_copy (transp, &live);
4013 bitmap_clear (moveable);
4014 bitmap_clear (&live);
4015 bitmap_clear (&used);
4016 bitmap_clear (&set);
4017 FOR_BB_INSNS (bb, insn)
4018 if (NONDEBUG_INSN_P (insn))
4020 df_ref *u_rec, *d_rec;
4022 uid_luid[INSN_UID (insn)] = i++;
4024 u_rec = DF_INSN_USES (insn);
4025 d_rec = DF_INSN_DEFS (insn);
4026 if (d_rec[0] != NULL && d_rec[1] == NULL
4027 && u_rec[0] != NULL && u_rec[1] == NULL
4028 && DF_REF_REGNO (*u_rec) == DF_REF_REGNO (*d_rec)
4029 && !bitmap_bit_p (&set, DF_REF_REGNO (*u_rec))
4030 && rtx_moveable_p (&PATTERN (insn), OP_IN))
4032 unsigned regno = DF_REF_REGNO (*u_rec);
4033 bitmap_set_bit (moveable, regno);
4034 bitmap_set_bit (&set, regno);
4035 bitmap_set_bit (&used, regno);
4036 bitmap_clear_bit (transp, regno);
4037 continue;
4039 while (*u_rec)
4041 unsigned regno = DF_REF_REGNO (*u_rec);
4042 bitmap_set_bit (&used, regno);
4043 if (bitmap_clear_bit (moveable, regno))
4044 bitmap_clear_bit (transp, regno);
4045 u_rec++;
4048 while (*d_rec)
4050 unsigned regno = DF_REF_REGNO (*d_rec);
4051 bitmap_set_bit (&set, regno);
4052 bitmap_clear_bit (transp, regno);
4053 bitmap_clear_bit (moveable, regno);
4054 d_rec++;
4059 bitmap_clear (&live);
4060 bitmap_clear (&used);
4061 bitmap_clear (&set);
4063 FOR_EACH_BB (bb)
4065 bitmap local = bb_local + bb->index;
4066 rtx insn;
4068 FOR_BB_INSNS (bb, insn)
4069 if (NONDEBUG_INSN_P (insn))
4071 rtx def_insn, closest_use, note;
4072 df_ref *def_rec, def, use;
4073 unsigned regno;
4074 bool all_dominated, all_local;
4075 enum machine_mode mode;
4077 def_rec = DF_INSN_DEFS (insn);
4078 /* There must be exactly one def in this insn. */
4079 def = *def_rec;
4080 if (!def || def_rec[1] || !single_set (insn))
4081 continue;
4082 /* This must be the only definition of the reg. We also limit
4083 which modes we deal with so that we can assume we can generate
4084 move instructions. */
4085 regno = DF_REF_REGNO (def);
4086 mode = GET_MODE (DF_REF_REG (def));
4087 if (DF_REG_DEF_COUNT (regno) != 1
4088 || !DF_REF_INSN_INFO (def)
4089 || HARD_REGISTER_NUM_P (regno)
4090 || DF_REG_EQ_USE_COUNT (regno) > 0
4091 || (!INTEGRAL_MODE_P (mode) && !FLOAT_MODE_P (mode)))
4092 continue;
4093 def_insn = DF_REF_INSN (def);
4095 for (note = REG_NOTES (def_insn); note; note = XEXP (note, 1))
4096 if (REG_NOTE_KIND (note) == REG_EQUIV && MEM_P (XEXP (note, 0)))
4097 break;
4099 if (note)
4101 if (dump_file)
4102 fprintf (dump_file, "Ignoring reg %d, has equiv memory\n",
4103 regno);
4104 bitmap_set_bit (&unusable_as_input, regno);
4105 continue;
4108 use = DF_REG_USE_CHAIN (regno);
4109 all_dominated = true;
4110 all_local = true;
4111 closest_use = NULL_RTX;
4112 for (; use; use = DF_REF_NEXT_REG (use))
4114 rtx insn;
4115 if (!DF_REF_INSN_INFO (use))
4117 all_dominated = false;
4118 all_local = false;
4119 break;
4121 insn = DF_REF_INSN (use);
4122 if (DEBUG_INSN_P (insn))
4123 continue;
4124 if (BLOCK_FOR_INSN (insn) != BLOCK_FOR_INSN (def_insn))
4125 all_local = false;
4126 if (!insn_dominated_by_p (insn, def_insn, uid_luid))
4127 all_dominated = false;
4128 if (closest_use != insn && closest_use != const0_rtx)
4130 if (closest_use == NULL_RTX)
4131 closest_use = insn;
4132 else if (insn_dominated_by_p (closest_use, insn, uid_luid))
4133 closest_use = insn;
4134 else if (!insn_dominated_by_p (insn, closest_use, uid_luid))
4135 closest_use = const0_rtx;
4138 if (!all_dominated)
4140 if (dump_file)
4141 fprintf (dump_file, "Reg %d not all uses dominated by set\n",
4142 regno);
4143 continue;
4145 if (all_local)
4146 bitmap_set_bit (local, regno);
4147 if (closest_use == const0_rtx || closest_use == NULL
4148 || next_nonnote_nondebug_insn (def_insn) == closest_use)
4150 if (dump_file)
4151 fprintf (dump_file, "Reg %d uninteresting%s\n", regno,
4152 closest_use == const0_rtx || closest_use == NULL
4153 ? " (no unique first use)" : "");
4154 continue;
4156 #ifdef HAVE_cc0
4157 if (reg_referenced_p (cc0_rtx, PATTERN (closest_use)))
4159 if (dump_file)
4160 fprintf (dump_file, "Reg %d: closest user uses cc0\n",
4161 regno);
4162 continue;
4164 #endif
4165 bitmap_set_bit (&interesting, regno);
4166 closest_uses[regno] = closest_use;
4168 if (dump_file && (all_local || all_dominated))
4170 fprintf (dump_file, "Reg %u:", regno);
4171 if (all_local)
4172 fprintf (dump_file, " local to bb %d", bb->index);
4173 if (all_dominated)
4174 fprintf (dump_file, " def dominates all uses");
4175 if (closest_use != const0_rtx)
4176 fprintf (dump_file, " has unique first use");
4177 fputs ("\n", dump_file);
4182 EXECUTE_IF_SET_IN_BITMAP (&interesting, 0, i, bi)
4184 df_ref def = DF_REG_DEF_CHAIN (i);
4185 rtx def_insn = DF_REF_INSN (def);
4186 basic_block def_block = BLOCK_FOR_INSN (def_insn);
4187 bitmap def_bb_local = bb_local + def_block->index;
4188 bitmap def_bb_moveable = bb_moveable_reg_sets + def_block->index;
4189 bitmap def_bb_transp = bb_transp_live + def_block->index;
4190 bool local_to_bb_p = bitmap_bit_p (def_bb_local, i);
4191 rtx use_insn = closest_uses[i];
4192 df_ref *def_insn_use_rec = DF_INSN_USES (def_insn);
4193 bool all_ok = true;
4194 bool all_transp = true;
4196 if (!REG_P (DF_REF_REG (def)))
4197 continue;
4199 if (!local_to_bb_p)
4201 if (dump_file)
4202 fprintf (dump_file, "Reg %u not local to one basic block\n",
4204 continue;
4206 if (reg_equiv_init (i) != NULL_RTX)
4208 if (dump_file)
4209 fprintf (dump_file, "Ignoring reg %u with equiv init insn\n",
4211 continue;
4213 if (!rtx_moveable_p (&PATTERN (def_insn), OP_IN))
4215 if (dump_file)
4216 fprintf (dump_file, "Found def insn %d for %d to be not moveable\n",
4217 INSN_UID (def_insn), i);
4218 continue;
4220 if (dump_file)
4221 fprintf (dump_file, "Examining insn %d, def for %d\n",
4222 INSN_UID (def_insn), i);
4223 while (*def_insn_use_rec != NULL)
4225 df_ref use = *def_insn_use_rec;
4226 unsigned regno = DF_REF_REGNO (use);
4227 if (bitmap_bit_p (&unusable_as_input, regno))
4229 all_ok = false;
4230 if (dump_file)
4231 fprintf (dump_file, " found unusable input reg %u.\n", regno);
4232 break;
4234 if (!bitmap_bit_p (def_bb_transp, regno))
4236 if (bitmap_bit_p (def_bb_moveable, regno)
4237 && !control_flow_insn_p (use_insn)
4238 #ifdef HAVE_cc0
4239 && !sets_cc0_p (use_insn)
4240 #endif
4243 if (modified_between_p (DF_REF_REG (use), def_insn, use_insn))
4245 rtx x = NEXT_INSN (def_insn);
4246 while (!modified_in_p (DF_REF_REG (use), x))
4248 gcc_assert (x != use_insn);
4249 x = NEXT_INSN (x);
4251 if (dump_file)
4252 fprintf (dump_file, " input reg %u modified but insn %d moveable\n",
4253 regno, INSN_UID (x));
4254 emit_insn_after (PATTERN (x), use_insn);
4255 set_insn_deleted (x);
4257 else
4259 if (dump_file)
4260 fprintf (dump_file, " input reg %u modified between def and use\n",
4261 regno);
4262 all_transp = false;
4265 else
4266 all_transp = false;
4269 def_insn_use_rec++;
4271 if (!all_ok)
4272 continue;
4273 if (!dbg_cnt (ira_move))
4274 break;
4275 if (dump_file)
4276 fprintf (dump_file, " all ok%s\n", all_transp ? " and transp" : "");
4278 if (all_transp)
4280 rtx def_reg = DF_REF_REG (def);
4281 rtx newreg = ira_create_new_reg (def_reg);
4282 if (validate_change (def_insn, DF_REF_LOC (def), newreg, 0))
4284 unsigned nregno = REGNO (newreg);
4285 emit_insn_before (gen_move_insn (def_reg, newreg), use_insn);
4286 nregno -= max_regs;
4287 pseudo_replaced_reg[nregno] = def_reg;
4292 FOR_EACH_BB (bb)
4294 bitmap_clear (bb_local + bb->index);
4295 bitmap_clear (bb_transp_live + bb->index);
4296 bitmap_clear (bb_moveable_reg_sets + bb->index);
4298 bitmap_clear (&interesting);
4299 bitmap_clear (&unusable_as_input);
4300 free (uid_luid);
4301 free (closest_uses);
4302 free (bb_local);
4303 free (bb_transp_live);
4304 free (bb_moveable_reg_sets);
4306 last_moveable_pseudo = max_reg_num ();
4308 fix_reg_equiv_init ();
4309 expand_reg_info ();
4310 regstat_free_n_sets_and_refs ();
4311 regstat_free_ri ();
4312 regstat_init_n_sets_and_refs ();
4313 regstat_compute_ri ();
4314 free_dominance_info (CDI_DOMINATORS);
4317 /* Perform the second half of the transformation started in
4318 find_moveable_pseudos. We look for instances where the newly introduced
4319 pseudo remains unallocated, and remove it by moving the definition to
4320 just before its use, replacing the move instruction generated by
4321 find_moveable_pseudos. */
4322 static void
4323 move_unallocated_pseudos (void)
4325 int i;
4326 for (i = first_moveable_pseudo; i < last_moveable_pseudo; i++)
4327 if (reg_renumber[i] < 0)
4329 int idx = i - first_moveable_pseudo;
4330 rtx other_reg = pseudo_replaced_reg[idx];
4331 rtx def_insn = DF_REF_INSN (DF_REG_DEF_CHAIN (i));
4332 /* The use must follow all definitions of OTHER_REG, so we can
4333 insert the new definition immediately after any of them. */
4334 df_ref other_def = DF_REG_DEF_CHAIN (REGNO (other_reg));
4335 rtx move_insn = DF_REF_INSN (other_def);
4336 rtx newinsn = emit_insn_after (PATTERN (def_insn), move_insn);
4337 rtx set;
4338 int success;
4340 if (dump_file)
4341 fprintf (dump_file, "moving def of %d (insn %d now) ",
4342 REGNO (other_reg), INSN_UID (def_insn));
4344 delete_insn (move_insn);
4345 while ((other_def = DF_REG_DEF_CHAIN (REGNO (other_reg))))
4346 delete_insn (DF_REF_INSN (other_def));
4347 delete_insn (def_insn);
4349 set = single_set (newinsn);
4350 success = validate_change (newinsn, &SET_DEST (set), other_reg, 0);
4351 gcc_assert (success);
4352 if (dump_file)
4353 fprintf (dump_file, " %d) rather than keep unallocated replacement %d\n",
4354 INSN_UID (newinsn), i);
4355 SET_REG_N_REFS (i, 0);
4359 /* If the backend knows where to allocate pseudos for hard
4360 register initial values, register these allocations now. */
4361 static void
4362 allocate_initial_values (void)
4364 if (targetm.allocate_initial_value)
4366 rtx hreg, preg, x;
4367 int i, regno;
4369 for (i = 0; HARD_REGISTER_NUM_P (i); i++)
4371 if (! initial_value_entry (i, &hreg, &preg))
4372 break;
4374 x = targetm.allocate_initial_value (hreg);
4375 regno = REGNO (preg);
4376 if (x && REG_N_SETS (regno) <= 1)
4378 if (MEM_P (x))
4379 reg_equiv_memory_loc (regno) = x;
4380 else
4382 basic_block bb;
4383 int new_regno;
4385 gcc_assert (REG_P (x));
4386 new_regno = REGNO (x);
4387 reg_renumber[regno] = new_regno;
4388 /* Poke the regno right into regno_reg_rtx so that even
4389 fixed regs are accepted. */
4390 SET_REGNO (preg, new_regno);
4391 /* Update global register liveness information. */
4392 FOR_EACH_BB (bb)
4394 if (REGNO_REG_SET_P (df_get_live_in (bb), regno))
4395 SET_REGNO_REG_SET (df_get_live_in (bb), new_regno);
4396 if (REGNO_REG_SET_P (df_get_live_out (bb), regno))
4397 SET_REGNO_REG_SET (df_get_live_out (bb), new_regno);
4403 gcc_checking_assert (! initial_value_entry (FIRST_PSEUDO_REGISTER,
4404 &hreg, &preg));
4409 /* True when we use LRA instead of reload pass for the current
4410 function. */
4411 bool ira_use_lra_p;
4413 /* True if we have allocno conflicts. It is false for non-optimized
4414 mode or when the conflict table is too big. */
4415 bool ira_conflicts_p;
4417 /* Saved between IRA and reload. */
4418 static int saved_flag_ira_share_spill_slots;
4420 /* This is the main entry of IRA. */
4421 static void
4422 ira (FILE *f)
4424 bool loops_p;
4425 int ira_max_point_before_emit;
4426 int rebuild_p;
4427 bool saved_flag_caller_saves = flag_caller_saves;
4428 enum ira_region saved_flag_ira_region = flag_ira_region;
4430 ira_conflicts_p = optimize > 0;
4432 ira_use_lra_p = targetm.lra_p ();
4433 /* If there are too many pseudos and/or basic blocks (e.g. 10K
4434 pseudos and 10K blocks or 100K pseudos and 1K blocks), we will
4435 use simplified and faster algorithms in LRA. */
4436 lra_simple_p
4437 = (ira_use_lra_p && max_reg_num () >= (1 << 26) / last_basic_block);
4438 if (lra_simple_p)
4440 /* It permits to skip live range splitting in LRA. */
4441 flag_caller_saves = false;
4442 /* There is no sense to do regional allocation when we use
4443 simplified LRA. */
4444 flag_ira_region = IRA_REGION_ONE;
4445 ira_conflicts_p = false;
4448 #ifndef IRA_NO_OBSTACK
4449 gcc_obstack_init (&ira_obstack);
4450 #endif
4451 bitmap_obstack_initialize (&ira_bitmap_obstack);
4453 if (flag_caller_saves)
4454 init_caller_save ();
4456 if (flag_ira_verbose < 10)
4458 internal_flag_ira_verbose = flag_ira_verbose;
4459 ira_dump_file = f;
4461 else
4463 internal_flag_ira_verbose = flag_ira_verbose - 10;
4464 ira_dump_file = stderr;
4467 setup_prohibited_mode_move_regs ();
4469 df_note_add_problem ();
4471 /* DF_LIVE can't be used in the register allocator, too many other
4472 parts of the compiler depend on using the "classic" liveness
4473 interpretation of the DF_LR problem. See PR38711.
4474 Remove the problem, so that we don't spend time updating it in
4475 any of the df_analyze() calls during IRA/LRA. */
4476 if (optimize > 1)
4477 df_remove_problem (df_live);
4478 gcc_checking_assert (df_live == NULL);
4480 #ifdef ENABLE_CHECKING
4481 df->changeable_flags |= DF_VERIFY_SCHEDULED;
4482 #endif
4483 df_analyze ();
4484 df_clear_flags (DF_NO_INSN_RESCAN);
4485 regstat_init_n_sets_and_refs ();
4486 regstat_compute_ri ();
4488 /* If we are not optimizing, then this is the only place before
4489 register allocation where dataflow is done. And that is needed
4490 to generate these warnings. */
4491 if (warn_clobbered)
4492 generate_setjmp_warnings ();
4494 /* Determine if the current function is a leaf before running IRA
4495 since this can impact optimizations done by the prologue and
4496 epilogue thus changing register elimination offsets. */
4497 crtl->is_leaf = leaf_function_p ();
4499 if (resize_reg_info () && flag_ira_loop_pressure)
4500 ira_set_pseudo_classes (true, ira_dump_file);
4502 init_reg_equiv ();
4503 rebuild_p = update_equiv_regs ();
4504 setup_reg_equiv ();
4505 setup_reg_equiv_init ();
4507 if (optimize && rebuild_p)
4509 timevar_push (TV_JUMP);
4510 rebuild_jump_labels (get_insns ());
4511 if (purge_all_dead_edges ())
4512 delete_unreachable_blocks ();
4513 timevar_pop (TV_JUMP);
4516 allocated_reg_info_size = max_reg_num ();
4518 if (delete_trivially_dead_insns (get_insns (), max_reg_num ()))
4519 df_analyze ();
4521 /* It is not worth to do such improvement when we use a simple
4522 allocation because of -O0 usage or because the function is too
4523 big. */
4524 if (ira_conflicts_p)
4525 find_moveable_pseudos ();
4527 max_regno_before_ira = max_reg_num ();
4528 ira_setup_eliminable_regset (true);
4530 ira_overall_cost = ira_reg_cost = ira_mem_cost = 0;
4531 ira_load_cost = ira_store_cost = ira_shuffle_cost = 0;
4532 ira_move_loops_num = ira_additional_jumps_num = 0;
4534 ira_assert (current_loops == NULL);
4535 if (flag_ira_region == IRA_REGION_ALL || flag_ira_region == IRA_REGION_MIXED)
4536 loop_optimizer_init (AVOID_CFG_MODIFICATIONS | LOOPS_HAVE_RECORDED_EXITS);
4538 if (internal_flag_ira_verbose > 0 && ira_dump_file != NULL)
4539 fprintf (ira_dump_file, "Building IRA IR\n");
4540 loops_p = ira_build ();
4542 ira_assert (ira_conflicts_p || !loops_p);
4544 saved_flag_ira_share_spill_slots = flag_ira_share_spill_slots;
4545 if (too_high_register_pressure_p () || cfun->calls_setjmp)
4546 /* It is just wasting compiler's time to pack spilled pseudos into
4547 stack slots in this case -- prohibit it. We also do this if
4548 there is setjmp call because a variable not modified between
4549 setjmp and longjmp the compiler is required to preserve its
4550 value and sharing slots does not guarantee it. */
4551 flag_ira_share_spill_slots = FALSE;
4553 ira_color ();
4555 ira_max_point_before_emit = ira_max_point;
4557 ira_initiate_emit_data ();
4559 ira_emit (loops_p);
4561 max_regno = max_reg_num ();
4562 if (ira_conflicts_p)
4564 if (! loops_p)
4566 if (! ira_use_lra_p)
4567 ira_initiate_assign ();
4569 else
4571 expand_reg_info ();
4573 if (ira_use_lra_p)
4575 ira_allocno_t a;
4576 ira_allocno_iterator ai;
4578 FOR_EACH_ALLOCNO (a, ai)
4579 ALLOCNO_REGNO (a) = REGNO (ALLOCNO_EMIT_DATA (a)->reg);
4581 else
4583 if (internal_flag_ira_verbose > 0 && ira_dump_file != NULL)
4584 fprintf (ira_dump_file, "Flattening IR\n");
4585 ira_flattening (max_regno_before_ira, ira_max_point_before_emit);
4587 /* New insns were generated: add notes and recalculate live
4588 info. */
4589 df_analyze ();
4591 /* ??? Rebuild the loop tree, but why? Does the loop tree
4592 change if new insns were generated? Can that be handled
4593 by updating the loop tree incrementally? */
4594 loop_optimizer_finalize ();
4595 free_dominance_info (CDI_DOMINATORS);
4596 loop_optimizer_init (AVOID_CFG_MODIFICATIONS
4597 | LOOPS_HAVE_RECORDED_EXITS);
4599 if (! ira_use_lra_p)
4601 setup_allocno_assignment_flags ();
4602 ira_initiate_assign ();
4603 ira_reassign_conflict_allocnos (max_regno);
4608 ira_finish_emit_data ();
4610 setup_reg_renumber ();
4612 calculate_allocation_cost ();
4614 #ifdef ENABLE_IRA_CHECKING
4615 if (ira_conflicts_p)
4616 check_allocation ();
4617 #endif
4619 if (max_regno != max_regno_before_ira)
4621 regstat_free_n_sets_and_refs ();
4622 regstat_free_ri ();
4623 regstat_init_n_sets_and_refs ();
4624 regstat_compute_ri ();
4627 overall_cost_before = ira_overall_cost;
4628 if (! ira_conflicts_p)
4629 grow_reg_equivs ();
4630 else
4632 fix_reg_equiv_init ();
4634 #ifdef ENABLE_IRA_CHECKING
4635 print_redundant_copies ();
4636 #endif
4638 ira_spilled_reg_stack_slots_num = 0;
4639 ira_spilled_reg_stack_slots
4640 = ((struct ira_spilled_reg_stack_slot *)
4641 ira_allocate (max_regno
4642 * sizeof (struct ira_spilled_reg_stack_slot)));
4643 memset (ira_spilled_reg_stack_slots, 0,
4644 max_regno * sizeof (struct ira_spilled_reg_stack_slot));
4646 allocate_initial_values ();
4648 /* See comment for find_moveable_pseudos call. */
4649 if (ira_conflicts_p)
4650 move_unallocated_pseudos ();
4652 /* Restore original values. */
4653 if (lra_simple_p)
4655 flag_caller_saves = saved_flag_caller_saves;
4656 flag_ira_region = saved_flag_ira_region;
4660 static void
4661 do_reload (void)
4663 basic_block bb;
4664 bool need_dce;
4666 if (flag_ira_verbose < 10)
4667 ira_dump_file = dump_file;
4669 timevar_push (TV_RELOAD);
4670 if (ira_use_lra_p)
4672 if (current_loops != NULL)
4674 loop_optimizer_finalize ();
4675 free_dominance_info (CDI_DOMINATORS);
4677 FOR_ALL_BB (bb)
4678 bb->loop_father = NULL;
4679 current_loops = NULL;
4681 if (ira_conflicts_p)
4682 ira_free (ira_spilled_reg_stack_slots);
4684 ira_destroy ();
4686 lra (ira_dump_file);
4687 /* ???!!! Move it before lra () when we use ira_reg_equiv in
4688 LRA. */
4689 vec_free (reg_equivs);
4690 reg_equivs = NULL;
4691 need_dce = false;
4693 else
4695 df_set_flags (DF_NO_INSN_RESCAN);
4696 build_insn_chain ();
4698 need_dce = reload (get_insns (), ira_conflicts_p);
4702 timevar_pop (TV_RELOAD);
4704 timevar_push (TV_IRA);
4706 if (ira_conflicts_p && ! ira_use_lra_p)
4708 ira_free (ira_spilled_reg_stack_slots);
4709 ira_finish_assign ();
4712 if (internal_flag_ira_verbose > 0 && ira_dump_file != NULL
4713 && overall_cost_before != ira_overall_cost)
4714 fprintf (ira_dump_file, "+++Overall after reload %d\n", ira_overall_cost);
4716 flag_ira_share_spill_slots = saved_flag_ira_share_spill_slots;
4718 if (! ira_use_lra_p)
4720 ira_destroy ();
4721 if (current_loops != NULL)
4723 loop_optimizer_finalize ();
4724 free_dominance_info (CDI_DOMINATORS);
4726 FOR_ALL_BB (bb)
4727 bb->loop_father = NULL;
4728 current_loops = NULL;
4730 regstat_free_ri ();
4731 regstat_free_n_sets_and_refs ();
4734 if (optimize)
4735 cleanup_cfg (CLEANUP_EXPENSIVE);
4737 finish_reg_equiv ();
4739 bitmap_obstack_release (&ira_bitmap_obstack);
4740 #ifndef IRA_NO_OBSTACK
4741 obstack_free (&ira_obstack, NULL);
4742 #endif
4744 /* The code after the reload has changed so much that at this point
4745 we might as well just rescan everything. Note that
4746 df_rescan_all_insns is not going to help here because it does not
4747 touch the artificial uses and defs. */
4748 df_finish_pass (true);
4749 df_scan_alloc (NULL);
4750 df_scan_blocks ();
4752 if (optimize > 1)
4754 df_live_add_problem ();
4755 df_live_set_all_dirty ();
4758 if (optimize)
4759 df_analyze ();
4761 if (need_dce && optimize)
4762 run_fast_dce ();
4764 timevar_pop (TV_IRA);
4767 /* Run the integrated register allocator. */
4768 static unsigned int
4769 rest_of_handle_ira (void)
4771 ira (dump_file);
4772 return 0;
4775 namespace {
4777 const pass_data pass_data_ira =
4779 RTL_PASS, /* type */
4780 "ira", /* name */
4781 OPTGROUP_NONE, /* optinfo_flags */
4782 false, /* has_gate */
4783 true, /* has_execute */
4784 TV_IRA, /* tv_id */
4785 0, /* properties_required */
4786 0, /* properties_provided */
4787 0, /* properties_destroyed */
4788 0, /* todo_flags_start */
4789 TODO_do_not_ggc_collect, /* todo_flags_finish */
4792 class pass_ira : public rtl_opt_pass
4794 public:
4795 pass_ira (gcc::context *ctxt)
4796 : rtl_opt_pass (pass_data_ira, ctxt)
4799 /* opt_pass methods: */
4800 unsigned int execute () { return rest_of_handle_ira (); }
4802 }; // class pass_ira
4804 } // anon namespace
4806 rtl_opt_pass *
4807 make_pass_ira (gcc::context *ctxt)
4809 return new pass_ira (ctxt);
4812 static unsigned int
4813 rest_of_handle_reload (void)
4815 do_reload ();
4816 return 0;
4819 namespace {
4821 const pass_data pass_data_reload =
4823 RTL_PASS, /* type */
4824 "reload", /* name */
4825 OPTGROUP_NONE, /* optinfo_flags */
4826 false, /* has_gate */
4827 true, /* has_execute */
4828 TV_RELOAD, /* tv_id */
4829 0, /* properties_required */
4830 0, /* properties_provided */
4831 0, /* properties_destroyed */
4832 0, /* todo_flags_start */
4833 0, /* todo_flags_finish */
4836 class pass_reload : public rtl_opt_pass
4838 public:
4839 pass_reload (gcc::context *ctxt)
4840 : rtl_opt_pass (pass_data_reload, ctxt)
4843 /* opt_pass methods: */
4844 unsigned int execute () { return rest_of_handle_reload (); }
4846 }; // class pass_reload
4848 } // anon namespace
4850 rtl_opt_pass *
4851 make_pass_reload (gcc::context *ctxt)
4853 return new pass_reload (ctxt);