/* Global common subexpression elimination/Partial redundancy elimination and global constant/copy propagation for GNU compiler. Copyright (C) 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc. This file is part of GNU CC. GNU CC is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. GNU CC is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with GNU CC; see the file COPYING. If not, write to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ /* TODO - reordering of memory allocation and freeing to be more space efficient - do rough calc of how many regs are needed in each block, and a rough calc of how many regs are available in each class and use that to throttle back the code in cases where RTX_COST is minimal. - dead store elimination - a store to the same address as a load does not kill the load if the source of the store is also the destination of the load. Handling this allows more load motion, particularly out of loops. - ability to realloc sbitmap vectors would allow one initial computation of reg_set_in_block with only subsequent additions, rather than recomputing it for each pass */ /* References searched while implementing this. Compilers Principles, Techniques and Tools Aho, Sethi, Ullman Addison-Wesley, 1988 Global Optimization by Suppression of Partial Redundancies E. Morel, C. Renvoise communications of the acm, Vol. 22, Num. 2, Feb. 1979 A Portable Machine-Independent Global Optimizer - Design and Measurements Frederick Chow Stanford Ph.D. thesis, Dec. 1983 A Fast Algorithm for Code Movement Optimization D.M. Dhamdhere SIGPLAN Notices, Vol. 23, Num. 10, Oct. 1988 A Solution to a Problem with Morel and Renvoise's Global Optimization by Suppression of Partial Redundancies K-H Drechsler, M.P. Stadel ACM TOPLAS, Vol. 10, Num. 4, Oct. 1988 Practical Adaptation of the Global Optimization Algorithm of Morel and Renvoise D.M. Dhamdhere ACM TOPLAS, Vol. 13, Num. 2. Apr. 1991 Efficiently Computing Static Single Assignment Form and the Control Dependence Graph R. Cytron, J. Ferrante, B.K. Rosen, M.N. Wegman, and F.K. Zadeck ACM TOPLAS, Vol. 13, Num. 4, Oct. 1991 Lazy Code Motion J. Knoop, O. Ruthing, B. Steffen ACM SIGPLAN Notices Vol. 27, Num. 7, Jul. 1992, '92 Conference on PLDI What's In a Region? Or Computing Control Dependence Regions in Near-Linear Time for Reducible Flow Control Thomas Ball ACM Letters on Programming Languages and Systems, Vol. 2, Num. 1-4, Mar-Dec 1993 An Efficient Representation for Sparse Sets Preston Briggs, Linda Torczon ACM Letters on Programming Languages and Systems, Vol. 2, Num. 1-4, Mar-Dec 1993 A Variation of Knoop, Ruthing, and Steffen's Lazy Code Motion K-H Drechsler, M.P. Stadel ACM SIGPLAN Notices, Vol. 28, Num. 5, May 1993 Partial Dead Code Elimination J. Knoop, O. Ruthing, B. Steffen ACM SIGPLAN Notices, Vol. 29, Num. 6, Jun. 1994 Effective Partial Redundancy Elimination P. Briggs, K.D. Cooper ACM SIGPLAN Notices, Vol. 29, Num. 6, Jun. 1994 The Program Structure Tree: Computing Control Regions in Linear Time R. Johnson, D. Pearson, K. Pingali ACM SIGPLAN Notices, Vol. 29, Num. 6, Jun. 1994 Optimal Code Motion: Theory and Practice J. Knoop, O. Ruthing, B. Steffen ACM TOPLAS, Vol. 16, Num. 4, Jul. 1994 The power of assignment motion J. Knoop, O. Ruthing, B. Steffen ACM SIGPLAN Notices Vol. 30, Num. 6, Jun. 1995, '95 Conference on PLDI Global code motion / global value numbering C. Click ACM SIGPLAN Notices Vol. 30, Num. 6, Jun. 1995, '95 Conference on PLDI Value Driven Redundancy Elimination L.T. Simpson Rice University Ph.D. thesis, Apr. 1996 Value Numbering L.T. Simpson Massively Scalar Compiler Project, Rice University, Sep. 1996 High Performance Compilers for Parallel Computing Michael Wolfe Addison-Wesley, 1996 Advanced Compiler Design and Implementation Steven Muchnick Morgan Kaufmann, 1997 Building an Optimizing Compiler Robert Morgan Digital Press, 1998 People wishing to speed up the code here should read: Elimination Algorithms for Data Flow Analysis B.G. Ryder, M.C. Paull ACM Computing Surveys, Vol. 18, Num. 3, Sep. 1986 How to Analyze Large Programs Efficiently and Informatively D.M. Dhamdhere, B.K. Rosen, F.K. Zadeck ACM SIGPLAN Notices Vol. 27, Num. 7, Jul. 1992, '92 Conference on PLDI People wishing to do something different can find various possibilities in the above papers and elsewhere. */ #include "config.h" #include "system.h" #include "toplev.h" #include "rtl.h" #include "tm_p.h" #include "regs.h" #include "hard-reg-set.h" #include "flags.h" #include "real.h" #include "insn-config.h" #include "recog.h" #include "basic-block.h" #include "output.h" #include "function.h" #include "expr.h" #include "ggc.h" #include "obstack.h" #define obstack_chunk_alloc gmalloc #define obstack_chunk_free free /* Maximum number of passes to perform. */ #define MAX_PASSES 1 /* Propagate flow information through back edges and thus enable PRE's moving loop invariant calculations out of loops. Originally this tended to create worse overall code, but several improvements during the development of PRE seem to have made following back edges generally a win. Note much of the loop invariant code motion done here would normally be done by loop.c, which has more heuristics for when to move invariants out of loops. At some point we might need to move some of those heuristics into gcse.c. */ #define FOLLOW_BACK_EDGES 1 /* We support GCSE via Partial Redundancy Elimination. PRE optimizations are a superset of those done by GCSE. We perform the following steps: 1) Compute basic block information. 2) Compute table of places where registers are set. 3) Perform copy/constant propagation. 4) Perform global cse. 5) Perform another pass of copy/constant propagation. Two passes of copy/constant propagation are done because the first one enables more GCSE and the second one helps to clean up the copies that GCSE creates. This is needed more for PRE than for Classic because Classic GCSE will try to use an existing register containing the common subexpression rather than create a new one. This is harder to do for PRE because of the code motion (which Classic GCSE doesn't do). Expressions we are interested in GCSE-ing are of the form (set (pseudo-reg) (expression)). Function want_to_gcse_p says what these are. PRE handles moving invariant expressions out of loops (by treating them as partially redundant). Eventually it would be nice to replace cse.c/gcse.c with SSA (static single assignment) based GVN (global value numbering). L. T. Simpson's paper (Rice University) on value numbering is a useful reference for this. ********************** We used to support multiple passes but there are diminishing returns in doing so. The first pass usually makes 90% of the changes that are doable. A second pass can make a few more changes made possible by the first pass. Experiments show any further passes don't make enough changes to justify the expense. A study of spec92 using an unlimited number of passes: [1 pass] = 1208 substitutions, [2] = 577, [3] = 202, [4] = 192, [5] = 83, [6] = 34, [7] = 17, [8] = 9, [9] = 4, [10] = 4, [11] = 2, [12] = 2, [13] = 1, [15] = 1, [16] = 2, [41] = 1 It was found doing copy propagation between each pass enables further substitutions. PRE is quite expensive in complicated functions because the DFA can take awhile to converge. Hence we only perform one pass. Macro MAX_PASSES can be modified if one wants to experiment. ********************** The steps for PRE are: 1) Build the hash table of expressions we wish to GCSE (expr_hash_table). 2) Perform the data flow analysis for PRE. 3) Delete the redundant instructions 4) Insert the required copies [if any] that make the partially redundant instructions fully redundant. 5) For other reaching expressions, insert an instruction to copy the value to a newly created pseudo that will reach the redundant instruction. The deletion is done first so that when we do insertions we know which pseudo reg to use. Various papers have argued that PRE DFA is expensive (O(n^2)) and others argue it is not. The number of iterations for the algorithm to converge is typically 2-4 so I don't view it as that expensive (relatively speaking). PRE GCSE depends heavily on the second CSE pass to clean up the copies we create. To make an expression reach the place where it's redundant, the result of the expression is copied to a new register, and the redundant expression is deleted by replacing it with this new register. Classic GCSE doesn't have this problem as much as it computes the reaching defs of each register in each block and thus can try to use an existing register. ********************** A fair bit of simplicity is created by creating small functions for simple tasks, even when the function is only called in one place. This may measurably slow things down [or may not] by creating more function call overhead than is necessary. The source is laid out so that it's trivial to make the affected functions inline so that one can measure what speed up, if any, can be achieved, and maybe later when things settle things can be rearranged. Help stamp out big monolithic functions! */ /* GCSE global vars. */ /* -dG dump file. */ static FILE *gcse_file; /* Note whether or not we should run jump optimization after gcse. We want to do this for two cases. * If we changed any jumps via cprop. * If we added any labels via edge splitting. */ static int run_jump_opt_after_gcse; /* Bitmaps are normally not included in debugging dumps. However it's useful to be able to print them from GDB. We could create special functions for this, but it's simpler to just allow passing stderr to the dump_foo fns. Since stderr can be a macro, we store a copy here. */ static FILE *debug_stderr; /* An obstack for our working variables. */ static struct obstack gcse_obstack; /* Non-zero for each mode that supports (set (reg) (reg)). This is trivially true for integer and floating point values. It may or may not be true for condition codes. */ static char can_copy_p[(int) NUM_MACHINE_MODES]; /* Non-zero if can_copy_p has been initialized. */ static int can_copy_init_p; struct reg_use {rtx reg_rtx; }; /* Hash table of expressions. */ struct expr { /* The expression (SET_SRC for expressions, PATTERN for assignments). */ rtx expr; /* Index in the available expression bitmaps. */ int bitmap_index; /* Next entry with the same hash. */ struct expr *next_same_hash; /* List of anticipatable occurrences in basic blocks in the function. An "anticipatable occurrence" is one that is the first occurrence in the basic block, the operands are not modified in the basic block prior to the occurrence and the output is not used between the start of the block and the occurrence. */ struct occr *antic_occr; /* List of available occurrence in basic blocks in the function. An "available occurrence" is one that is the last occurrence in the basic block and the operands are not modified by following statements in the basic block [including this insn]. */ struct occr *avail_occr; /* Non-null if the computation is PRE redundant. The value is the newly created pseudo-reg to record a copy of the expression in all the places that reach the redundant copy. */ rtx reaching_reg; }; /* Occurrence of an expression. There is one per basic block. If a pattern appears more than once the last appearance is used [or first for anticipatable expressions]. */ struct occr { /* Next occurrence of this expression. */ struct occr *next; /* The insn that computes the expression. */ rtx insn; /* Non-zero if this [anticipatable] occurrence has been deleted. */ char deleted_p; /* Non-zero if this [available] occurrence has been copied to reaching_reg. */ /* ??? This is mutually exclusive with deleted_p, so they could share the same byte. */ char copied_p; }; /* Expression and copy propagation hash tables. Each hash table is an array of buckets. ??? It is known that if it were an array of entries, structure elements `next_same_hash' and `bitmap_index' wouldn't be necessary. However, it is not clear whether in the final analysis a sufficient amount of memory would be saved as the size of the available expression bitmaps would be larger [one could build a mapping table without holes afterwards though]. Someday I'll perform the computation and figure it out. */ /* Total size of the expression hash table, in elements. */ static unsigned int expr_hash_table_size; /* The table itself. This is an array of `expr_hash_table_size' elements. */ static struct expr **expr_hash_table; /* Total size of the copy propagation hash table, in elements. */ static unsigned int set_hash_table_size; /* The table itself. This is an array of `set_hash_table_size' elements. */ static struct expr **set_hash_table; /* Mapping of uids to cuids. Only real insns get cuids. */ static int *uid_cuid; /* Highest UID in UID_CUID. */ static int max_uid; /* Get the cuid of an insn. */ #ifdef ENABLE_CHECKING #define INSN_CUID(INSN) (INSN_UID (INSN) > max_uid ? (abort (), 0) : uid_cuid[INSN_UID (INSN)]) #else #define INSN_CUID(INSN) (uid_cuid[INSN_UID (INSN)]) #endif /* Number of cuids. */ static int max_cuid; /* Mapping of cuids to insns. */ static rtx *cuid_insn; /* Get insn from cuid. */ #define CUID_INSN(CUID) (cuid_insn[CUID]) /* Maximum register number in function prior to doing gcse + 1. Registers created during this pass have regno >= max_gcse_regno. This is named with "gcse" to not collide with global of same name. */ static unsigned int max_gcse_regno; /* Maximum number of cse-able expressions found. */ static int n_exprs; /* Maximum number of assignments for copy propagation found. */ static int n_sets; /* Table of registers that are modified. For each register, each element is a list of places where the pseudo-reg is set. For simplicity, GCSE is done on sets of pseudo-regs only. PRE GCSE only requires knowledge of which blocks kill which regs [and thus could use a bitmap instead of the lists `reg_set_table' uses]. `reg_set_table' and could be turned into an array of bitmaps (num-bbs x num-regs) [however perhaps it may be useful to keep the data as is]. One advantage of recording things this way is that `reg_set_table' is fairly sparse with respect to pseudo regs but for hard regs could be fairly dense [relatively speaking]. And recording sets of pseudo-regs in lists speeds up functions like compute_transp since in the case of pseudo-regs we only need to iterate over the number of times a pseudo-reg is set, not over the number of basic blocks [clearly there is a bit of a slow down in the cases where a pseudo is set more than once in a block, however it is believed that the net effect is to speed things up]. This isn't done for hard-regs because recording call-clobbered hard-regs in `reg_set_table' at each function call can consume a fair bit of memory, and iterating over hard-regs stored this way in compute_transp will be more expensive. */ typedef struct reg_set { /* The next setting of this register. */ struct reg_set *next; /* The insn where it was set. */ rtx insn; } reg_set; static reg_set **reg_set_table; /* Size of `reg_set_table'. The table starts out at max_gcse_regno + slop, and is enlarged as necessary. */ static int reg_set_table_size; /* Amount to grow `reg_set_table' by when it's full. */ #define REG_SET_TABLE_SLOP 100 /* Bitmap containing one bit for each register in the program. Used when performing GCSE to track which registers have been set since the start of the basic block. */ static sbitmap reg_set_bitmap; /* For each block, a bitmap of registers set in the block. This is used by expr_killed_p and compute_transp. It is computed during hash table computation and not by compute_sets as it includes registers added since the last pass (or between cprop and gcse) and it's currently not easy to realloc sbitmap vectors. */ static sbitmap *reg_set_in_block; /* For each block, non-zero if memory is set in that block. This is computed during hash table computation and is used by expr_killed_p and compute_transp. ??? Handling of memory is very simple, we don't make any attempt to optimize things (later). ??? This can be computed by compute_sets since the information doesn't change. */ static char *mem_set_in_block; /* Various variables for statistics gathering. */ /* Memory used in a pass. This isn't intended to be absolutely precise. Its intent is only to keep an eye on memory usage. */ static int bytes_used; /* GCSE substitutions made. */ static int gcse_subst_count; /* Number of copy instructions created. */ static int gcse_create_count; /* Number of constants propagated. */ static int const_prop_count; /* Number of copys propagated. */ static int copy_prop_count; /* These variables are used by classic GCSE. Normally they'd be defined a bit later, but `rd_gen' needs to be declared sooner. */ /* Each block has a bitmap of each type. The length of each blocks bitmap is: max_cuid - for reaching definitions n_exprs - for available expressions Thus we view the bitmaps as 2 dimensional arrays. i.e. rd_kill[block_num][cuid_num] ae_kill[block_num][expr_num] */ /* For reaching defs */ static sbitmap *rd_kill, *rd_gen, *reaching_defs, *rd_out; /* for available exprs */ static sbitmap *ae_kill, *ae_gen, *ae_in, *ae_out; /* Objects of this type are passed around by the null-pointer check removal routines. */ struct null_pointer_info { /* The basic block being processed. */ int current_block; /* The first register to be handled in this pass. */ unsigned int min_reg; /* One greater than the last register to be handled in this pass. */ unsigned int max_reg; sbitmap *nonnull_local; sbitmap *nonnull_killed; }; static void compute_can_copy PARAMS ((void)); static char *gmalloc PARAMS ((unsigned int)); static char *grealloc PARAMS ((char *, unsigned int)); static char *gcse_alloc PARAMS ((unsigned long)); static void alloc_gcse_mem PARAMS ((rtx)); static void free_gcse_mem PARAMS ((void)); static void alloc_reg_set_mem PARAMS ((int)); static void free_reg_set_mem PARAMS ((void)); static int get_bitmap_width PARAMS ((int, int, int)); static void record_one_set PARAMS ((int, rtx)); static void record_set_info PARAMS ((rtx, rtx, void *)); static void compute_sets PARAMS ((rtx)); static void hash_scan_insn PARAMS ((rtx, int, int)); static void hash_scan_set PARAMS ((rtx, rtx, int)); static void hash_scan_clobber PARAMS ((rtx, rtx)); static void hash_scan_call PARAMS ((rtx, rtx)); static int want_to_gcse_p PARAMS ((rtx)); static int oprs_unchanged_p PARAMS ((rtx, rtx, int)); static int oprs_anticipatable_p PARAMS ((rtx, rtx)); static int oprs_available_p PARAMS ((rtx, rtx)); static void insert_expr_in_table PARAMS ((rtx, enum machine_mode, rtx, int, int)); static void insert_set_in_table PARAMS ((rtx, rtx)); static unsigned int hash_expr PARAMS ((rtx, enum machine_mode, int *, int)); static unsigned int hash_expr_1 PARAMS ((rtx, enum machine_mode, int *)); static unsigned int hash_string_1 PARAMS ((const char *)); static unsigned int hash_set PARAMS ((int, int)); static int expr_equiv_p PARAMS ((rtx, rtx)); static void record_last_reg_set_info PARAMS ((rtx, int)); static void record_last_mem_set_info PARAMS ((rtx)); static void record_last_set_info PARAMS ((rtx, rtx, void *)); static void compute_hash_table PARAMS ((int)); static void alloc_set_hash_table PARAMS ((int)); static void free_set_hash_table PARAMS ((void)); static void compute_set_hash_table PARAMS ((void)); static void alloc_expr_hash_table PARAMS ((unsigned int)); static void free_expr_hash_table PARAMS ((void)); static void compute_expr_hash_table PARAMS ((void)); static void dump_hash_table PARAMS ((FILE *, const char *, struct expr **, int, int)); static struct expr *lookup_expr PARAMS ((rtx)); static struct expr *lookup_set PARAMS ((unsigned int, rtx)); static struct expr *next_set PARAMS ((unsigned int, struct expr *)); static void reset_opr_set_tables PARAMS ((void)); static int oprs_not_set_p PARAMS ((rtx, rtx)); static void mark_call PARAMS ((rtx)); static void mark_set PARAMS ((rtx, rtx)); static void mark_clobber PARAMS ((rtx, rtx)); static void mark_oprs_set PARAMS ((rtx)); static void alloc_cprop_mem PARAMS ((int, int)); static void free_cprop_mem PARAMS ((void)); static void compute_transp PARAMS ((rtx, int, sbitmap *, int)); static void compute_transpout PARAMS ((void)); static void compute_local_properties PARAMS ((sbitmap *, sbitmap *, sbitmap *, int)); static void compute_cprop_data PARAMS ((void)); static void find_used_regs PARAMS ((rtx)); static int try_replace_reg PARAMS ((rtx, rtx, rtx)); static struct expr *find_avail_set PARAMS ((int, rtx)); static int cprop_jump PARAMS ((rtx, rtx, rtx)); #ifdef HAVE_cc0 static int cprop_cc0_jump PARAMS ((rtx, struct reg_use *, rtx)); #endif static int cprop_insn PARAMS ((rtx, int)); static int cprop PARAMS ((int)); static int one_cprop_pass PARAMS ((int, int)); static void alloc_pre_mem PARAMS ((int, int)); static void free_pre_mem PARAMS ((void)); static void compute_pre_data PARAMS ((void)); static int pre_expr_reaches_here_p PARAMS ((int, struct expr *, int)); static void insert_insn_end_bb PARAMS ((struct expr *, int, int)); static void pre_insert_copy_insn PARAMS ((struct expr *, rtx)); static void pre_insert_copies PARAMS ((void)); static int pre_delete PARAMS ((void)); static int pre_gcse PARAMS ((void)); static int one_pre_gcse_pass PARAMS ((int)); static void add_label_notes PARAMS ((rtx, rtx)); static void alloc_code_hoist_mem PARAMS ((int, int)); static void free_code_hoist_mem PARAMS ((void)); static void compute_code_hoist_vbeinout PARAMS ((void)); static void compute_code_hoist_data PARAMS ((void)); static int hoist_expr_reaches_here_p PARAMS ((int, int, int, char *)); static void hoist_code PARAMS ((void)); static int one_code_hoisting_pass PARAMS ((void)); static void alloc_rd_mem PARAMS ((int, int)); static void free_rd_mem PARAMS ((void)); static void handle_rd_kill_set PARAMS ((rtx, int, int)); static void compute_kill_rd PARAMS ((void)); static void compute_rd PARAMS ((void)); static void alloc_avail_expr_mem PARAMS ((int, int)); static void free_avail_expr_mem PARAMS ((void)); static void compute_ae_gen PARAMS ((void)); static int expr_killed_p PARAMS ((rtx, int)); static void compute_ae_kill PARAMS ((sbitmap *, sbitmap *)); static int expr_reaches_here_p PARAMS ((struct occr *, struct expr *, int, int)); static rtx computing_insn PARAMS ((struct expr *, rtx)); static int def_reaches_here_p PARAMS ((rtx, rtx)); static int can_disregard_other_sets PARAMS ((struct reg_set **, rtx, int)); static int handle_avail_expr PARAMS ((rtx, struct expr *)); static int classic_gcse PARAMS ((void)); static int one_classic_gcse_pass PARAMS ((int)); static void invalidate_nonnull_info PARAMS ((rtx, rtx, void *)); static void delete_null_pointer_checks_1 PARAMS ((unsigned int *, sbitmap *, sbitmap *, struct null_pointer_info *)); static rtx process_insert_insn PARAMS ((struct expr *)); static int pre_edge_insert PARAMS ((struct edge_list *, struct expr **)); static int expr_reaches_here_p_work PARAMS ((struct occr *, struct expr *, int, int, char *)); static int pre_expr_reaches_here_p_work PARAMS ((int, struct expr *, int, char *)); /* Entry point for global common subexpression elimination. F is the first instruction in the function. */ int gcse_main (f, file) rtx f; FILE *file; { int changed, pass; /* Bytes used at start of pass. */ int initial_bytes_used; /* Maximum number of bytes used by a pass. */ int max_pass_bytes; /* Point to release obstack data from for each pass. */ char *gcse_obstack_bottom; /* We do not construct an accurate cfg in functions which call setjmp, so just punt to be safe. */ if (current_function_calls_setjmp) return 0; /* Assume that we do not need to run jump optimizations after gcse. */ run_jump_opt_after_gcse = 0; /* For calling dump_foo fns from gdb. */ debug_stderr = stderr; gcse_file = file; /* Identify the basic block information for this function, including successors and predecessors. */ max_gcse_regno = max_reg_num (); if (file) dump_flow_info (file); /* Return if there's nothing to do. */ if (n_basic_blocks <= 1) return 0; /* Trying to perform global optimizations on flow graphs which have a high connectivity will take a long time and is unlikely to be particularly useful. In normal circumstances a cfg should have about twice has many edges as blocks. But we do not want to punish small functions which have a couple switch statements. So we require a relatively large number of basic blocks and the ratio of edges to blocks to be high. */ if (n_basic_blocks > 1000 && n_edges / n_basic_blocks >= 20) { if (warn_disabled_optimization) warning ("GCSE disabled: %d > 1000 basic blocks and %d >= 20 edges/basic block", n_basic_blocks, n_edges / n_basic_blocks); return 0; } /* See what modes support reg/reg copy operations. */ if (! can_copy_init_p) { compute_can_copy (); can_copy_init_p = 1; } gcc_obstack_init (&gcse_obstack); bytes_used = 0; /* Record where pseudo-registers are set. This data is kept accurate during each pass. ??? We could also record hard-reg information here [since it's unchanging], however it is currently done during hash table computation. It may be tempting to compute MEM set information here too, but MEM sets will be subject to code motion one day and thus we need to compute information about memory sets when we build the hash tables. */ alloc_reg_set_mem (max_gcse_regno); compute_sets (f); pass = 0; initial_bytes_used = bytes_used; max_pass_bytes = 0; gcse_obstack_bottom = gcse_alloc (1); changed = 1; while (changed && pass < MAX_PASSES) { changed = 0; if (file) fprintf (file, "GCSE pass %d\n\n", pass + 1); /* Initialize bytes_used to the space for the pred/succ lists, and the reg_set_table data. */ bytes_used = initial_bytes_used; /* Each pass may create new registers, so recalculate each time. */ max_gcse_regno = max_reg_num (); alloc_gcse_mem (f); /* Don't allow constant propagation to modify jumps during this pass. */ changed = one_cprop_pass (pass + 1, 0); if (optimize_size) changed |= one_classic_gcse_pass (pass + 1); else { changed |= one_pre_gcse_pass (pass + 1); free_reg_set_mem (); alloc_reg_set_mem (max_reg_num ()); compute_sets (f); run_jump_opt_after_gcse = 1; } if (max_pass_bytes < bytes_used) max_pass_bytes = bytes_used; /* Free up memory, then reallocate for code hoisting. We can not re-use the existing allocated memory because the tables will not have info for the insns or registers created by partial redundancy elimination. */ free_gcse_mem (); /* It does not make sense to run code hoisting unless we optimizing for code size -- it rarely makes programs faster, and can make them bigger if we did partial redundancy elimination (when optimizing for space, we use a classic gcse algorithm instead of partial redundancy algorithms). */ if (optimize_size) { max_gcse_regno = max_reg_num (); alloc_gcse_mem (f); changed |= one_code_hoisting_pass (); free_gcse_mem (); if (max_pass_bytes < bytes_used) max_pass_bytes = bytes_used; } if (file) { fprintf (file, "\n"); fflush (file); } obstack_free (&gcse_obstack, gcse_obstack_bottom); pass++; } /* Do one last pass of copy propagation, including cprop into conditional jumps. */ max_gcse_regno = max_reg_num (); alloc_gcse_mem (f); /* This time, go ahead and allow cprop to alter jumps. */ one_cprop_pass (pass + 1, 1); free_gcse_mem (); if (file) { fprintf (file, "GCSE of %s: %d basic blocks, ", current_function_name, n_basic_blocks); fprintf (file, "%d pass%s, %d bytes\n\n", pass, pass > 1 ? "es" : "", max_pass_bytes); } obstack_free (&gcse_obstack, NULL_PTR); free_reg_set_mem (); return run_jump_opt_after_gcse; } /* Misc. utilities. */ /* Compute which modes support reg/reg copy operations. */ static void compute_can_copy () { int i; #ifndef AVOID_CCMODE_COPIES rtx reg,insn; #endif memset (can_copy_p, 0, NUM_MACHINE_MODES); start_sequence (); for (i = 0; i < NUM_MACHINE_MODES; i++) if (GET_MODE_CLASS (i) == MODE_CC) { #ifdef AVOID_CCMODE_COPIES can_copy_p[i] = 0; #else reg = gen_rtx_REG ((enum machine_mode) i, LAST_VIRTUAL_REGISTER + 1); insn = emit_insn (gen_rtx_SET (VOIDmode, reg, reg)); if (recog (PATTERN (insn), insn, NULL_PTR) >= 0) can_copy_p[i] = 1; #endif } else can_copy_p[i] = 1; end_sequence (); } /* Cover function to xmalloc to record bytes allocated. */ static char * gmalloc (size) unsigned int size; { bytes_used += size; return xmalloc (size); } /* Cover function to xrealloc. We don't record the additional size since we don't know it. It won't affect memory usage stats much anyway. */ static char * grealloc (ptr, size) char *ptr; unsigned int size; { return xrealloc (ptr, size); } /* Cover function to obstack_alloc. We don't need to record the bytes allocated here since obstack_chunk_alloc is set to gmalloc. */ static char * gcse_alloc (size) unsigned long size; { return (char *) obstack_alloc (&gcse_obstack, size); } /* Allocate memory for the cuid mapping array, and reg/memory set tracking tables. This is called at the start of each pass. */ static void alloc_gcse_mem (f) rtx f; { int i,n; rtx insn; /* Find the largest UID and create a mapping from UIDs to CUIDs. CUIDs are like UIDs except they increase monotonically, have no gaps, and only apply to real insns. */ max_uid = get_max_uid (); n = (max_uid + 1) * sizeof (int); uid_cuid = (int *) gmalloc (n); memset ((char *) uid_cuid, 0, n); for (insn = f, i = 0; insn; insn = NEXT_INSN (insn)) { if (INSN_P (insn)) uid_cuid[INSN_UID (insn)] = i++; else uid_cuid[INSN_UID (insn)] = i; } /* Create a table mapping cuids to insns. */ max_cuid = i; n = (max_cuid + 1) * sizeof (rtx); cuid_insn = (rtx *) gmalloc (n); memset ((char *) cuid_insn, 0, n); for (insn = f, i = 0; insn; insn = NEXT_INSN (insn)) if (INSN_P (insn)) CUID_INSN (i++) = insn; /* Allocate vars to track sets of regs. */ reg_set_bitmap = (sbitmap) sbitmap_alloc (max_gcse_regno); /* Allocate vars to track sets of regs, memory per block. */ reg_set_in_block = (sbitmap *) sbitmap_vector_alloc (n_basic_blocks, max_gcse_regno); mem_set_in_block = (char *) gmalloc (n_basic_blocks); } /* Free memory allocated by alloc_gcse_mem. */ static void free_gcse_mem () { free (uid_cuid); free (cuid_insn); free (reg_set_bitmap); free (reg_set_in_block); free (mem_set_in_block); } /* Many of the global optimization algorithms work by solving dataflow equations for various expressions. Initially, some local value is computed for each expression in each block. Then, the values across the various blocks are combined (by following flow graph edges) to arrive at global values. Conceptually, each set of equations is independent. We may therefore solve all the equations in parallel, solve them one at a time, or pick any intermediate approach. When you're going to need N two-dimensional bitmaps, each X (say, the number of blocks) by Y (say, the number of expressions), call this function. It's not important what X and Y represent; only that Y correspond to the things that can be done in parallel. This function will return an appropriate chunking factor C; you should solve C sets of equations in parallel. By going through this function, we can easily trade space against time; by solving fewer equations in parallel we use less space. */ static int get_bitmap_width (n, x, y) int n; int x; int y; { /* It's not really worth figuring out *exactly* how much memory will be used by a particular choice. The important thing is to get something approximately right. */ size_t max_bitmap_memory = 10 * 1024 * 1024; /* The number of bytes we'd use for a single column of minimum width. */ size_t column_size = n * x * sizeof (SBITMAP_ELT_TYPE); /* Often, it's reasonable just to solve all the equations in parallel. */ if (column_size * SBITMAP_SET_SIZE (y) <= max_bitmap_memory) return y; /* Otherwise, pick the largest width we can, without going over the limit. */ return SBITMAP_ELT_BITS * ((max_bitmap_memory + column_size - 1) / column_size); } /* Compute the local properties of each recorded expression. Local properties are those that are defined by the block, irrespective of other blocks. An expression is transparent in a block if its operands are not modified in the block. An expression is computed (locally available) in a block if it is computed at least once and expression would contain the same value if the computation was moved to the end of the block. An expression is locally anticipatable in a block if it is computed at least once and expression would contain the same value if the computation was moved to the beginning of the block. We call this routine for cprop, pre and code hoisting. They all compute basically the same information and thus can easily share this code. TRANSP, COMP, and ANTLOC are destination sbitmaps for recording local properties. If NULL, then it is not necessary to compute or record that particular property. SETP controls which hash table to look at. If zero, this routine looks at the expr hash table; if nonzero this routine looks at the set hash table. Additionally, TRANSP is computed as ~TRANSP, since this is really cprop's ABSALTERED. */ static void compute_local_properties (transp, comp, antloc, setp) sbitmap *transp; sbitmap *comp; sbitmap *antloc; int setp; { unsigned int i, hash_table_size; struct expr **hash_table; /* Initialize any bitmaps that were passed in. */ if (transp) { if (setp) sbitmap_vector_zero (transp, n_basic_blocks); else sbitmap_vector_ones (transp, n_basic_blocks); } if (comp) sbitmap_vector_zero (comp, n_basic_blocks); if (antloc) sbitmap_vector_zero (antloc, n_basic_blocks); /* We use the same code for cprop, pre and hoisting. For cprop we care about the set hash table, for pre and hoisting we care about the expr hash table. */ hash_table_size = setp ? set_hash_table_size : expr_hash_table_size; hash_table = setp ? set_hash_table : expr_hash_table; for (i = 0; i < hash_table_size; i++) { struct expr *expr; for (expr = hash_table[i]; expr != NULL; expr = expr->next_same_hash) { int indx = expr->bitmap_index; struct occr *occr; /* The expression is transparent in this block if it is not killed. We start by assuming all are transparent [none are killed], and then reset the bits for those that are. */ if (transp) compute_transp (expr->expr, indx, transp, setp); /* The occurrences recorded in antic_occr are exactly those that we want to set to non-zero in ANTLOC. */ if (antloc) for (occr = expr->antic_occr; occr != NULL; occr = occr->next) { SET_BIT (antloc[BLOCK_NUM (occr->insn)], indx); /* While we're scanning the table, this is a good place to initialize this. */ occr->deleted_p = 0; } /* The occurrences recorded in avail_occr are exactly those that we want to set to non-zero in COMP. */ if (comp) for (occr = expr->avail_occr; occr != NULL; occr = occr->next) { SET_BIT (comp[BLOCK_NUM (occr->insn)], indx); /* While we're scanning the table, this is a good place to initialize this. */ occr->copied_p = 0; } /* While we're scanning the table, this is a good place to initialize this. */ expr->reaching_reg = 0; } } } /* Register set information. `reg_set_table' records where each register is set or otherwise modified. */ static struct obstack reg_set_obstack; static void alloc_reg_set_mem (n_regs) int n_regs; { unsigned int n; reg_set_table_size = n_regs + REG_SET_TABLE_SLOP; n = reg_set_table_size * sizeof (struct reg_set *); reg_set_table = (struct reg_set **) gmalloc (n); memset ((char *) reg_set_table, 0, n); gcc_obstack_init (®_set_obstack); } static void free_reg_set_mem () { free (reg_set_table); obstack_free (®_set_obstack, NULL_PTR); } /* Record REGNO in the reg_set table. */ static void record_one_set (regno, insn) int regno; rtx insn; { /* Allocate a new reg_set element and link it onto the list. */ struct reg_set *new_reg_info; /* If the table isn't big enough, enlarge it. */ if (regno >= reg_set_table_size) { int new_size = regno + REG_SET_TABLE_SLOP; reg_set_table = (struct reg_set **) grealloc ((char *) reg_set_table, new_size * sizeof (struct reg_set *)); memset ((char *) (reg_set_table + reg_set_table_size), 0, (new_size - reg_set_table_size) * sizeof (struct reg_set *)); reg_set_table_size = new_size; } new_reg_info = (struct reg_set *) obstack_alloc (®_set_obstack, sizeof (struct reg_set)); bytes_used += sizeof (struct reg_set); new_reg_info->insn = insn; new_reg_info->next = reg_set_table[regno]; reg_set_table[regno] = new_reg_info; } /* Called from compute_sets via note_stores to handle one SET or CLOBBER in an insn. The DATA is really the instruction in which the SET is occurring. */ static void record_set_info (dest, setter, data) rtx dest, setter ATTRIBUTE_UNUSED; void *data; { rtx record_set_insn = (rtx) data; if (GET_CODE (dest) == REG && REGNO (dest) >= FIRST_PSEUDO_REGISTER) record_one_set (REGNO (dest), record_set_insn); } /* Scan the function and record each set of each pseudo-register. This is called once, at the start of the gcse pass. See the comments for `reg_set_table' for further documenation. */ static void compute_sets (f) rtx f; { rtx insn; for (insn = f; insn != 0; insn = NEXT_INSN (insn)) if (INSN_P (insn)) note_stores (PATTERN (insn), record_set_info, insn); } /* Hash table support. */ /* For each register, the cuid of the first/last insn in the block to set it, or -1 if not set. */ #define NEVER_SET -1 static int *reg_first_set; static int *reg_last_set; /* While computing "first/last set" info, this is the CUID of first/last insn to set memory or -1 if not set. `mem_last_set' is also used when performing GCSE to record whether memory has been set since the beginning of the block. Note that handling of memory is very simple, we don't make any attempt to optimize things (later). */ static int mem_first_set; static int mem_last_set; /* See whether X, the source of a set, is something we want to consider for GCSE. */ static int want_to_gcse_p (x) rtx x; { static rtx test_insn = 0; int num_clobbers = 0; int icode; switch (GET_CODE (x)) { case REG: case SUBREG: case CONST_INT: case CONST_DOUBLE: case CALL: return 0; default: break; } /* If this is a valid operand, we are OK. If it's VOIDmode, we aren't. */ if (general_operand (x, GET_MODE (x))) return 1; else if (GET_MODE (x) == VOIDmode) return 0; /* Otherwise, check if we can make a valid insn from it. First initialize our test insn if we haven't already. */ if (test_insn == 0) { test_insn = make_insn_raw (gen_rtx_SET (VOIDmode, gen_rtx_REG (word_mode, FIRST_PSEUDO_REGISTER * 2), const0_rtx)); NEXT_INSN (test_insn) = PREV_INSN (test_insn) = 0; ggc_add_rtx_root (&test_insn, 1); } /* Now make an insn like the one we would make when GCSE'ing and see if valid. */ PUT_MODE (SET_DEST (PATTERN (test_insn)), GET_MODE (x)); SET_SRC (PATTERN (test_insn)) = x; return ((icode = recog (PATTERN (test_insn), test_insn, &num_clobbers)) >= 0 && (num_clobbers == 0 || ! added_clobbers_hard_reg_p (icode))); } /* Return non-zero if the operands of expression X are unchanged from the start of INSN's basic block up to but not including INSN (if AVAIL_P == 0), or from INSN to the end of INSN's basic block (if AVAIL_P != 0). */ static int oprs_unchanged_p (x, insn, avail_p) rtx x, insn; int avail_p; { int i, j; enum rtx_code code; const char *fmt; if (x == 0) return 1; code = GET_CODE (x); switch (code) { case REG: if (avail_p) return (reg_last_set[REGNO (x)] == NEVER_SET || reg_last_set[REGNO (x)] < INSN_CUID (insn)); else return (reg_first_set[REGNO (x)] == NEVER_SET || reg_first_set[REGNO (x)] >= INSN_CUID (insn)); case MEM: if (avail_p && mem_last_set != NEVER_SET && mem_last_set >= INSN_CUID (insn)) return 0; else if (! avail_p && mem_first_set != NEVER_SET && mem_first_set < INSN_CUID (insn)) return 0; else return oprs_unchanged_p (XEXP (x, 0), insn, avail_p); case PRE_DEC: case PRE_INC: case POST_DEC: case POST_INC: case PRE_MODIFY: case POST_MODIFY: return 0; case PC: case CC0: /*FIXME*/ case CONST: case CONST_INT: case CONST_DOUBLE: case SYMBOL_REF: case LABEL_REF: case ADDR_VEC: case ADDR_DIFF_VEC: return 1; default: break; } for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) return oprs_unchanged_p (XEXP (x, i), insn, avail_p); else if (! oprs_unchanged_p (XEXP (x, i), insn, avail_p)) return 0; } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) if (! oprs_unchanged_p (XVECEXP (x, i, j), insn, avail_p)) return 0; } return 1; } /* Return non-zero if the operands of expression X are unchanged from the start of INSN's basic block up to but not including INSN. */ static int oprs_anticipatable_p (x, insn) rtx x, insn; { return oprs_unchanged_p (x, insn, 0); } /* Return non-zero if the operands of expression X are unchanged from INSN to the end of INSN's basic block. */ static int oprs_available_p (x, insn) rtx x, insn; { return oprs_unchanged_p (x, insn, 1); } /* Hash expression X. MODE is only used if X is a CONST_INT. DO_NOT_RECORD_P is a boolean indicating if a volatile operand is found or if the expression contains something we don't want to insert in the table. ??? One might want to merge this with canon_hash. Later. */ static unsigned int hash_expr (x, mode, do_not_record_p, hash_table_size) rtx x; enum machine_mode mode; int *do_not_record_p; int hash_table_size; { unsigned int hash; *do_not_record_p = 0; hash = hash_expr_1 (x, mode, do_not_record_p); return hash % hash_table_size; } /* Hash a string. Just add its bytes up. */ static inline unsigned hash_string_1 (ps) const char *ps; { unsigned hash = 0; const unsigned char *p = (const unsigned char *)ps; if (p) while (*p) hash += *p++; return hash; } /* Subroutine of hash_expr to do the actual work. */ static unsigned int hash_expr_1 (x, mode, do_not_record_p) rtx x; enum machine_mode mode; int *do_not_record_p; { int i, j; unsigned hash = 0; enum rtx_code code; const char *fmt; /* Used to turn recursion into iteration. We can't rely on GCC's tail-recursion eliminatio since we need to keep accumulating values in HASH. */ if (x == 0) return hash; repeat: code = GET_CODE (x); switch (code) { case REG: hash += ((unsigned int) REG << 7) + REGNO (x); return hash; case CONST_INT: hash += (((unsigned int) CONST_INT << 7) + (unsigned int) mode + (unsigned int) INTVAL (x)); return hash; case CONST_DOUBLE: /* This is like the general case, except that it only counts the integers representing the constant. */ hash += (unsigned int) code + (unsigned int) GET_MODE (x); if (GET_MODE (x) != VOIDmode) for (i = 2; i < GET_RTX_LENGTH (CONST_DOUBLE); i++) hash += (unsigned int) XWINT (x, i); else hash += ((unsigned int) CONST_DOUBLE_LOW (x) + (unsigned int) CONST_DOUBLE_HIGH (x)); return hash; /* Assume there is only one rtx object for any given label. */ case LABEL_REF: /* We don't hash on the address of the CODE_LABEL to avoid bootstrap differences and differences between each stage's debugging dumps. */ hash += (((unsigned int) LABEL_REF << 7) + CODE_LABEL_NUMBER (XEXP (x, 0))); return hash; case SYMBOL_REF: { /* Don't hash on the symbol's address to avoid bootstrap differences. Different hash values may cause expressions to be recorded in different orders and thus different registers to be used in the final assembler. This also avoids differences in the dump files between various stages. */ unsigned int h = 0; const unsigned char *p = (const unsigned char *) XSTR (x, 0); while (*p) h += (h << 7) + *p++; /* ??? revisit */ hash += ((unsigned int) SYMBOL_REF << 7) + h; return hash; } case MEM: if (MEM_VOLATILE_P (x)) { *do_not_record_p = 1; return 0; } hash += (unsigned int) MEM; hash += MEM_ALIAS_SET (x); x = XEXP (x, 0); goto repeat; case PRE_DEC: case PRE_INC: case POST_DEC: case POST_INC: case PC: case CC0: case CALL: case UNSPEC_VOLATILE: *do_not_record_p = 1; return 0; case ASM_OPERANDS: if (MEM_VOLATILE_P (x)) { *do_not_record_p = 1; return 0; } else { /* We don't want to take the filename and line into account. */ hash += (unsigned) code + (unsigned) GET_MODE (x) + hash_string_1 (ASM_OPERANDS_TEMPLATE (x)) + hash_string_1 (ASM_OPERANDS_OUTPUT_CONSTRAINT (x)) + (unsigned) ASM_OPERANDS_OUTPUT_IDX (x); if (ASM_OPERANDS_INPUT_LENGTH (x)) { for (i = 1; i < ASM_OPERANDS_INPUT_LENGTH (x); i++) { hash += (hash_expr_1 (ASM_OPERANDS_INPUT (x, i), GET_MODE (ASM_OPERANDS_INPUT (x, i)), do_not_record_p) + hash_string_1 (ASM_OPERANDS_INPUT_CONSTRAINT (x, i))); } hash += hash_string_1 (ASM_OPERANDS_INPUT_CONSTRAINT (x, 0)); x = ASM_OPERANDS_INPUT (x, 0); mode = GET_MODE (x); goto repeat; } return hash; } default: break; } hash += (unsigned) code + (unsigned) GET_MODE (x); for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) { x = XEXP (x, i); goto repeat; } hash += hash_expr_1 (XEXP (x, i), 0, do_not_record_p); if (*do_not_record_p) return 0; } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) { hash += hash_expr_1 (XVECEXP (x, i, j), 0, do_not_record_p); if (*do_not_record_p) return 0; } else if (fmt[i] == 's') hash += hash_string_1 (XSTR (x, i)); else if (fmt[i] == 'i') hash += (unsigned int) XINT (x, i); else abort (); } return hash; } /* Hash a set of register REGNO. Sets are hashed on the register that is set. This simplifies the PRE copy propagation code. ??? May need to make things more elaborate. Later, as necessary. */ static unsigned int hash_set (regno, hash_table_size) int regno; int hash_table_size; { unsigned int hash; hash = regno; return hash % hash_table_size; } /* Return non-zero if exp1 is equivalent to exp2. ??? Borrowed from cse.c. Might want to remerge with cse.c. Later. */ static int expr_equiv_p (x, y) rtx x, y; { register int i, j; register enum rtx_code code; register const char *fmt; if (x == y) return 1; if (x == 0 || y == 0) return x == y; code = GET_CODE (x); if (code != GET_CODE (y)) return 0; /* (MULT:SI x y) and (MULT:HI x y) are NOT equivalent. */ if (GET_MODE (x) != GET_MODE (y)) return 0; switch (code) { case PC: case CC0: return x == y; case CONST_INT: return INTVAL (x) == INTVAL (y); case LABEL_REF: return XEXP (x, 0) == XEXP (y, 0); case SYMBOL_REF: return XSTR (x, 0) == XSTR (y, 0); case REG: return REGNO (x) == REGNO (y); case MEM: /* Can't merge two expressions in different alias sets, since we can decide that the expression is transparent in a block when it isn't, due to it being set with the different alias set. */ if (MEM_ALIAS_SET (x) != MEM_ALIAS_SET (y)) return 0; break; /* For commutative operations, check both orders. */ case PLUS: case MULT: case AND: case IOR: case XOR: case NE: case EQ: return ((expr_equiv_p (XEXP (x, 0), XEXP (y, 0)) && expr_equiv_p (XEXP (x, 1), XEXP (y, 1))) || (expr_equiv_p (XEXP (x, 0), XEXP (y, 1)) && expr_equiv_p (XEXP (x, 1), XEXP (y, 0)))); case ASM_OPERANDS: /* We don't use the generic code below because we want to disregard filename and line numbers. */ /* A volatile asm isn't equivalent to any other. */ if (MEM_VOLATILE_P (x) || MEM_VOLATILE_P (y)) return 0; if (GET_MODE (x) != GET_MODE (y) || strcmp (ASM_OPERANDS_TEMPLATE (x), ASM_OPERANDS_TEMPLATE (y)) || strcmp (ASM_OPERANDS_OUTPUT_CONSTRAINT (x), ASM_OPERANDS_OUTPUT_CONSTRAINT (y)) || ASM_OPERANDS_OUTPUT_IDX (x) != ASM_OPERANDS_OUTPUT_IDX (y) || ASM_OPERANDS_INPUT_LENGTH (x) != ASM_OPERANDS_INPUT_LENGTH (y)) return 0; if (ASM_OPERANDS_INPUT_LENGTH (x)) { for (i = ASM_OPERANDS_INPUT_LENGTH (x) - 1; i >= 0; i--) if (! expr_equiv_p (ASM_OPERANDS_INPUT (x, i), ASM_OPERANDS_INPUT (y, i)) || strcmp (ASM_OPERANDS_INPUT_CONSTRAINT (x, i), ASM_OPERANDS_INPUT_CONSTRAINT (y, i))) return 0; } return 1; default: break; } /* Compare the elements. If any pair of corresponding elements fail to match, return 0 for the whole thing. */ fmt = GET_RTX_FORMAT (code); for (i = GET_RTX_LENGTH (code) - 1; i >= 0; i--) { switch (fmt[i]) { case 'e': if (! expr_equiv_p (XEXP (x, i), XEXP (y, i))) return 0; break; case 'E': if (XVECLEN (x, i) != XVECLEN (y, i)) return 0; for (j = 0; j < XVECLEN (x, i); j++) if (! expr_equiv_p (XVECEXP (x, i, j), XVECEXP (y, i, j))) return 0; break; case 's': if (strcmp (XSTR (x, i), XSTR (y, i))) return 0; break; case 'i': if (XINT (x, i) != XINT (y, i)) return 0; break; case 'w': if (XWINT (x, i) != XWINT (y, i)) return 0; break; case '0': break; default: abort (); } } return 1; } /* Insert expression X in INSN in the hash table. If it is already present, record it as the last occurrence in INSN's basic block. MODE is the mode of the value X is being stored into. It is only used if X is a CONST_INT. ANTIC_P is non-zero if X is an anticipatable expression. AVAIL_P is non-zero if X is an available expression. */ static void insert_expr_in_table (x, mode, insn, antic_p, avail_p) rtx x; enum machine_mode mode; rtx insn; int antic_p, avail_p; { int found, do_not_record_p; unsigned int hash; struct expr *cur_expr, *last_expr = NULL; struct occr *antic_occr, *avail_occr; struct occr *last_occr = NULL; hash = hash_expr (x, mode, &do_not_record_p, expr_hash_table_size); /* Do not insert expression in table if it contains volatile operands, or if hash_expr determines the expression is something we don't want to or can't handle. */ if (do_not_record_p) return; cur_expr = expr_hash_table[hash]; found = 0; while (cur_expr && 0 == (found = expr_equiv_p (cur_expr->expr, x))) { /* If the expression isn't found, save a pointer to the end of the list. */ last_expr = cur_expr; cur_expr = cur_expr->next_same_hash; } if (! found) { cur_expr = (struct expr *) gcse_alloc (sizeof (struct expr)); bytes_used += sizeof (struct expr); if (expr_hash_table[hash] == NULL) /* This is the first pattern that hashed to this index. */ expr_hash_table[hash] = cur_expr; else /* Add EXPR to end of this hash chain. */ last_expr->next_same_hash = cur_expr; /* Set the fields of the expr element. */ cur_expr->expr = x; cur_expr->bitmap_index = n_exprs++; cur_expr->next_same_hash = NULL; cur_expr->antic_occr = NULL; cur_expr->avail_occr = NULL; } /* Now record the occurrence(s). */ if (antic_p) { antic_occr = cur_expr->antic_occr; /* Search for another occurrence in the same basic block. */ while (antic_occr && BLOCK_NUM (antic_occr->insn) != BLOCK_NUM (insn)) { /* If an occurrence isn't found, save a pointer to the end of the list. */ last_occr = antic_occr; antic_occr = antic_occr->next; } if (antic_occr) /* Found another instance of the expression in the same basic block. Prefer the currently recorded one. We want the first one in the block and the block is scanned from start to end. */ ; /* nothing to do */ else { /* First occurrence of this expression in this basic block. */ antic_occr = (struct occr *) gcse_alloc (sizeof (struct occr)); bytes_used += sizeof (struct occr); /* First occurrence of this expression in any block? */ if (cur_expr->antic_occr == NULL) cur_expr->antic_occr = antic_occr; else last_occr->next = antic_occr; antic_occr->insn = insn; antic_occr->next = NULL; } } if (avail_p) { avail_occr = cur_expr->avail_occr; /* Search for another occurrence in the same basic block. */ while (avail_occr && BLOCK_NUM (avail_occr->insn) != BLOCK_NUM (insn)) { /* If an occurrence isn't found, save a pointer to the end of the list. */ last_occr = avail_occr; avail_occr = avail_occr->next; } if (avail_occr) /* Found another instance of the expression in the same basic block. Prefer this occurrence to the currently recorded one. We want the last one in the block and the block is scanned from start to end. */ avail_occr->insn = insn; else { /* First occurrence of this expression in this basic block. */ avail_occr = (struct occr *) gcse_alloc (sizeof (struct occr)); bytes_used += sizeof (struct occr); /* First occurrence of this expression in any block? */ if (cur_expr->avail_occr == NULL) cur_expr->avail_occr = avail_occr; else last_occr->next = avail_occr; avail_occr->insn = insn; avail_occr->next = NULL; } } } /* Insert pattern X in INSN in the hash table. X is a SET of a reg to either another reg or a constant. If it is already present, record it as the last occurrence in INSN's basic block. */ static void insert_set_in_table (x, insn) rtx x; rtx insn; { int found; unsigned int hash; struct expr *cur_expr, *last_expr = NULL; struct occr *cur_occr, *last_occr = NULL; if (GET_CODE (x) != SET || GET_CODE (SET_DEST (x)) != REG) abort (); hash = hash_set (REGNO (SET_DEST (x)), set_hash_table_size); cur_expr = set_hash_table[hash]; found = 0; while (cur_expr && 0 == (found = expr_equiv_p (cur_expr->expr, x))) { /* If the expression isn't found, save a pointer to the end of the list. */ last_expr = cur_expr; cur_expr = cur_expr->next_same_hash; } if (! found) { cur_expr = (struct expr *) gcse_alloc (sizeof (struct expr)); bytes_used += sizeof (struct expr); if (set_hash_table[hash] == NULL) /* This is the first pattern that hashed to this index. */ set_hash_table[hash] = cur_expr; else /* Add EXPR to end of this hash chain. */ last_expr->next_same_hash = cur_expr; /* Set the fields of the expr element. We must copy X because it can be modified when copy propagation is performed on its operands. */ cur_expr->expr = copy_rtx (x); cur_expr->bitmap_index = n_sets++; cur_expr->next_same_hash = NULL; cur_expr->antic_occr = NULL; cur_expr->avail_occr = NULL; } /* Now record the occurrence. */ cur_occr = cur_expr->avail_occr; /* Search for another occurrence in the same basic block. */ while (cur_occr && BLOCK_NUM (cur_occr->insn) != BLOCK_NUM (insn)) { /* If an occurrence isn't found, save a pointer to the end of the list. */ last_occr = cur_occr; cur_occr = cur_occr->next; } if (cur_occr) /* Found another instance of the expression in the same basic block. Prefer this occurrence to the currently recorded one. We want the last one in the block and the block is scanned from start to end. */ cur_occr->insn = insn; else { /* First occurrence of this expression in this basic block. */ cur_occr = (struct occr *) gcse_alloc (sizeof (struct occr)); bytes_used += sizeof (struct occr); /* First occurrence of this expression in any block? */ if (cur_expr->avail_occr == NULL) cur_expr->avail_occr = cur_occr; else last_occr->next = cur_occr; cur_occr->insn = insn; cur_occr->next = NULL; } } /* Scan pattern PAT of INSN and add an entry to the hash table. If SET_P is non-zero, this is for the assignment hash table, otherwise it is for the expression hash table. */ static void hash_scan_set (pat, insn, set_p) rtx pat, insn; int set_p; { rtx src = SET_SRC (pat); rtx dest = SET_DEST (pat); rtx note; if (GET_CODE (src) == CALL) hash_scan_call (src, insn); else if (GET_CODE (dest) == REG) { unsigned int regno = REGNO (dest); rtx tmp; /* If this is a single set and we are doing constant propagation, see if a REG_NOTE shows this equivalent to a constant. */ if (set_p && (note = find_reg_equal_equiv_note (insn)) != 0 && CONSTANT_P (XEXP (note, 0))) src = XEXP (note, 0), pat = gen_rtx_SET (VOIDmode, dest, src); /* Only record sets of pseudo-regs in the hash table. */ if (! set_p && regno >= FIRST_PSEUDO_REGISTER /* Don't GCSE something if we can't do a reg/reg copy. */ && can_copy_p [GET_MODE (dest)] /* Is SET_SRC something we want to gcse? */ && want_to_gcse_p (src) /* Don't CSE a nop. */ && src != dest) { /* An expression is not anticipatable if its operands are modified before this insn. */ int antic_p = oprs_anticipatable_p (src, insn); /* An expression is not available if its operands are subsequently modified, including this insn. */ int avail_p = oprs_available_p (src, insn); insert_expr_in_table (src, GET_MODE (dest), insn, antic_p, avail_p); } /* Record sets for constant/copy propagation. */ else if (set_p && regno >= FIRST_PSEUDO_REGISTER && ((GET_CODE (src) == REG && REGNO (src) >= FIRST_PSEUDO_REGISTER && can_copy_p [GET_MODE (dest)] && REGNO (src) != regno) || GET_CODE (src) == CONST_INT || GET_CODE (src) == SYMBOL_REF || GET_CODE (src) == CONST_DOUBLE) /* A copy is not available if its src or dest is subsequently modified. Here we want to search from INSN+1 on, but oprs_available_p searches from INSN on. */ && (insn == BLOCK_END (BLOCK_NUM (insn)) || ((tmp = next_nonnote_insn (insn)) != NULL_RTX && oprs_available_p (pat, tmp)))) insert_set_in_table (pat, insn); } } static void hash_scan_clobber (x, insn) rtx x ATTRIBUTE_UNUSED, insn ATTRIBUTE_UNUSED; { /* Currently nothing to do. */ } static void hash_scan_call (x, insn) rtx x ATTRIBUTE_UNUSED, insn ATTRIBUTE_UNUSED; { /* Currently nothing to do. */ } /* Process INSN and add hash table entries as appropriate. Only available expressions that set a single pseudo-reg are recorded. Single sets in a PARALLEL could be handled, but it's an extra complication that isn't dealt with right now. The trick is handling the CLOBBERs that are also in the PARALLEL. Later. If SET_P is non-zero, this is for the assignment hash table, otherwise it is for the expression hash table. If IN_LIBCALL_BLOCK nonzero, we are in a libcall block, and should not record any expressions. */ static void hash_scan_insn (insn, set_p, in_libcall_block) rtx insn; int set_p; int in_libcall_block; { rtx pat = PATTERN (insn); int i; if (in_libcall_block) return; /* Pick out the sets of INSN and for other forms of instructions record what's been modified. */ if (GET_CODE (pat) == SET) hash_scan_set (pat, insn, set_p); else if (GET_CODE (pat) == PARALLEL) for (i = 0; i < XVECLEN (pat, 0); i++) { rtx x = XVECEXP (pat, 0, i); if (GET_CODE (x) == SET) hash_scan_set (x, insn, set_p); else if (GET_CODE (x) == CLOBBER) hash_scan_clobber (x, insn); else if (GET_CODE (x) == CALL) hash_scan_call (x, insn); } else if (GET_CODE (pat) == CLOBBER) hash_scan_clobber (pat, insn); else if (GET_CODE (pat) == CALL) hash_scan_call (pat, insn); } static void dump_hash_table (file, name, table, table_size, total_size) FILE *file; const char *name; struct expr **table; int table_size, total_size; { int i; /* Flattened out table, so it's printed in proper order. */ struct expr **flat_table; unsigned int *hash_val; struct expr *expr; flat_table = (struct expr **) xcalloc (total_size, sizeof (struct expr *)); hash_val = (unsigned int *) xmalloc (total_size * sizeof (unsigned int)); for (i = 0; i < table_size; i++) for (expr = table[i]; expr != NULL; expr = expr->next_same_hash) { flat_table[expr->bitmap_index] = expr; hash_val[expr->bitmap_index] = i; } fprintf (file, "%s hash table (%d buckets, %d entries)\n", name, table_size, total_size); for (i = 0; i < total_size; i++) if (flat_table[i] != 0) { expr = flat_table[i]; fprintf (file, "Index %d (hash value %d)\n ", expr->bitmap_index, hash_val[i]); print_rtl (file, expr->expr); fprintf (file, "\n"); } fprintf (file, "\n"); free (flat_table); free (hash_val); } /* Record register first/last/block set information for REGNO in INSN. reg_first_set records the first place in the block where the register is set and is used to compute "anticipatability". reg_last_set records the last place in the block where the register is set and is used to compute "availability". reg_set_in_block records whether the register is set in the block and is used to compute "transparency". */ static void record_last_reg_set_info (insn, regno) rtx insn; int regno; { if (reg_first_set[regno] == NEVER_SET) reg_first_set[regno] = INSN_CUID (insn); reg_last_set[regno] = INSN_CUID (insn); SET_BIT (reg_set_in_block[BLOCK_NUM (insn)], regno); } /* Record memory first/last/block set information for INSN. */ static void record_last_mem_set_info (insn) rtx insn; { if (mem_first_set == NEVER_SET) mem_first_set = INSN_CUID (insn); mem_last_set = INSN_CUID (insn); mem_set_in_block[BLOCK_NUM (insn)] = 1; } /* Called from compute_hash_table via note_stores to handle one SET or CLOBBER in an insn. DATA is really the instruction in which the SET is taking place. */ static void record_last_set_info (dest, setter, data) rtx dest, setter ATTRIBUTE_UNUSED; void *data; { rtx last_set_insn = (rtx) data; if (GET_CODE (dest) == SUBREG) dest = SUBREG_REG (dest); if (GET_CODE (dest) == REG) record_last_reg_set_info (last_set_insn, REGNO (dest)); else if (GET_CODE (dest) == MEM /* Ignore pushes, they clobber nothing. */ && ! push_operand (dest, GET_MODE (dest))) record_last_mem_set_info (last_set_insn); } /* Top level function to create an expression or assignment hash table. Expression entries are placed in the hash table if - they are of the form (set (pseudo-reg) src), - src is something we want to perform GCSE on, - none of the operands are subsequently modified in the block Assignment entries are placed in the hash table if - they are of the form (set (pseudo-reg) src), - src is something we want to perform const/copy propagation on, - none of the operands or target are subsequently modified in the block Currently src must be a pseudo-reg or a const_int. F is the first insn. SET_P is non-zero for computing the assignment hash table. */ static void compute_hash_table (set_p) int set_p; { int bb; /* While we compute the hash table we also compute a bit array of which registers are set in which blocks. We also compute which blocks set memory, in the absence of aliasing support [which is TODO]. ??? This isn't needed during const/copy propagation, but it's cheap to compute. Later. */ sbitmap_vector_zero (reg_set_in_block, n_basic_blocks); memset ((char *) mem_set_in_block, 0, n_basic_blocks); /* Some working arrays used to track first and last set in each block. */ /* ??? One could use alloca here, but at some size a threshold is crossed beyond which one should use malloc. Are we at that threshold here? */ reg_first_set = (int *) gmalloc (max_gcse_regno * sizeof (int)); reg_last_set = (int *) gmalloc (max_gcse_regno * sizeof (int)); for (bb = 0; bb < n_basic_blocks; bb++) { rtx insn; unsigned int regno; int in_libcall_block; unsigned int i; /* First pass over the instructions records information used to determine when registers and memory are first and last set. ??? The mem_set_in_block and hard-reg reg_set_in_block computation could be moved to compute_sets since they currently don't change. */ for (i = 0; i < max_gcse_regno; i++) reg_first_set[i] = reg_last_set[i] = NEVER_SET; mem_first_set = NEVER_SET; mem_last_set = NEVER_SET; for (insn = BLOCK_HEAD (bb); insn && insn != NEXT_INSN (BLOCK_END (bb)); insn = NEXT_INSN (insn)) { #ifdef NON_SAVING_SETJMP if (NON_SAVING_SETJMP && GET_CODE (insn) == NOTE && NOTE_LINE_NUMBER (insn) == NOTE_INSN_SETJMP) { for (regno = 0; regno < FIRST_PSEUDO_REGISTER; regno++) record_last_reg_set_info (insn, regno); continue; } #endif if (! INSN_P (insn)) continue; if (GET_CODE (insn) == CALL_INSN) { for (regno = 0; regno < FIRST_PSEUDO_REGISTER; regno++) if ((call_used_regs[regno] && regno != STACK_POINTER_REGNUM #if HARD_FRAME_POINTER_REGNUM != FRAME_POINTER_REGNUM && regno != HARD_FRAME_POINTER_REGNUM #endif #if ARG_POINTER_REGNUM != FRAME_POINTER_REGNUM && ! (regno == ARG_POINTER_REGNUM && fixed_regs[regno]) #endif #if !defined (PIC_OFFSET_TABLE_REG_CALL_CLOBBERED) && ! (regno == PIC_OFFSET_TABLE_REGNUM && flag_pic) #endif && regno != FRAME_POINTER_REGNUM) || global_regs[regno]) record_last_reg_set_info (insn, regno); if (! CONST_CALL_P (insn)) record_last_mem_set_info (insn); } note_stores (PATTERN (insn), record_last_set_info, insn); } /* The next pass builds the hash table. */ for (insn = BLOCK_HEAD (bb), in_libcall_block = 0; insn && insn != NEXT_INSN (BLOCK_END (bb)); insn = NEXT_INSN (insn)) if (INSN_P (insn)) { if (find_reg_note (insn, REG_LIBCALL, NULL_RTX)) in_libcall_block = 1; else if (find_reg_note (insn, REG_RETVAL, NULL_RTX)) in_libcall_block = 0; hash_scan_insn (insn, set_p, in_libcall_block); } } free (reg_first_set); free (reg_last_set); /* Catch bugs early. */ reg_first_set = reg_last_set = 0; } /* Allocate space for the set hash table. N_INSNS is the number of instructions in the function. It is used to determine the number of buckets to use. */ static void alloc_set_hash_table (n_insns) int n_insns; { int n; set_hash_table_size = n_insns / 4; if (set_hash_table_size < 11) set_hash_table_size = 11; /* Attempt to maintain efficient use of hash table. Making it an odd number is simplest for now. ??? Later take some measurements. */ set_hash_table_size |= 1; n = set_hash_table_size * sizeof (struct expr *); set_hash_table = (struct expr **) gmalloc (n); } /* Free things allocated by alloc_set_hash_table. */ static void free_set_hash_table () { free (set_hash_table); } /* Compute the hash table for doing copy/const propagation. */ static void compute_set_hash_table () { /* Initialize count of number of entries in hash table. */ n_sets = 0; memset ((char *) set_hash_table, 0, set_hash_table_size * sizeof (struct expr *)); compute_hash_table (1); } /* Allocate space for the expression hash table. N_INSNS is the number of instructions in the function. It is used to determine the number of buckets to use. */ static void alloc_expr_hash_table (n_insns) unsigned int n_insns; { int n; expr_hash_table_size = n_insns / 2; /* Make sure the amount is usable. */ if (expr_hash_table_size < 11) expr_hash_table_size = 11; /* Attempt to maintain efficient use of hash table. Making it an odd number is simplest for now. ??? Later take some measurements. */ expr_hash_table_size |= 1; n = expr_hash_table_size * sizeof (struct expr *); expr_hash_table = (struct expr **) gmalloc (n); } /* Free things allocated by alloc_expr_hash_table. */ static void free_expr_hash_table () { free (expr_hash_table); } /* Compute the hash table for doing GCSE. */ static void compute_expr_hash_table () { /* Initialize count of number of entries in hash table. */ n_exprs = 0; memset ((char *) expr_hash_table, 0, expr_hash_table_size * sizeof (struct expr *)); compute_hash_table (0); } /* Expression tracking support. */ /* Lookup pattern PAT in the expression table. The result is a pointer to the table entry, or NULL if not found. */ static struct expr * lookup_expr (pat) rtx pat; { int do_not_record_p; unsigned int hash = hash_expr (pat, GET_MODE (pat), &do_not_record_p, expr_hash_table_size); struct expr *expr; if (do_not_record_p) return NULL; expr = expr_hash_table[hash]; while (expr && ! expr_equiv_p (expr->expr, pat)) expr = expr->next_same_hash; return expr; } /* Lookup REGNO in the set table. If PAT is non-NULL look for the entry that matches it, otherwise return the first entry for REGNO. The result is a pointer to the table entry, or NULL if not found. */ static struct expr * lookup_set (regno, pat) unsigned int regno; rtx pat; { unsigned int hash = hash_set (regno, set_hash_table_size); struct expr *expr; expr = set_hash_table[hash]; if (pat) { while (expr && ! expr_equiv_p (expr->expr, pat)) expr = expr->next_same_hash; } else { while (expr && REGNO (SET_DEST (expr->expr)) != regno) expr = expr->next_same_hash; } return expr; } /* Return the next entry for REGNO in list EXPR. */ static struct expr * next_set (regno, expr) unsigned int regno; struct expr *expr; { do expr = expr->next_same_hash; while (expr && REGNO (SET_DEST (expr->expr)) != regno); return expr; } /* Reset tables used to keep track of what's still available [since the start of the block]. */ static void reset_opr_set_tables () { /* Maintain a bitmap of which regs have been set since beginning of the block. */ sbitmap_zero (reg_set_bitmap); /* Also keep a record of the last instruction to modify memory. For now this is very trivial, we only record whether any memory location has been modified. */ mem_last_set = 0; } /* Return non-zero if the operands of X are not set before INSN in INSN's basic block. */ static int oprs_not_set_p (x, insn) rtx x, insn; { int i, j; enum rtx_code code; const char *fmt; if (x == 0) return 1; code = GET_CODE (x); switch (code) { case PC: case CC0: case CONST: case CONST_INT: case CONST_DOUBLE: case SYMBOL_REF: case LABEL_REF: case ADDR_VEC: case ADDR_DIFF_VEC: return 1; case MEM: if (mem_last_set != 0) return 0; else return oprs_not_set_p (XEXP (x, 0), insn); case REG: return ! TEST_BIT (reg_set_bitmap, REGNO (x)); default: break; } for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) return oprs_not_set_p (XEXP (x, i), insn); if (! oprs_not_set_p (XEXP (x, i), insn)) return 0; } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) if (! oprs_not_set_p (XVECEXP (x, i, j), insn)) return 0; } return 1; } /* Mark things set by a CALL. */ static void mark_call (insn) rtx insn; { mem_last_set = INSN_CUID (insn); } /* Mark things set by a SET. */ static void mark_set (pat, insn) rtx pat, insn; { rtx dest = SET_DEST (pat); while (GET_CODE (dest) == SUBREG || GET_CODE (dest) == ZERO_EXTRACT || GET_CODE (dest) == SIGN_EXTRACT || GET_CODE (dest) == STRICT_LOW_PART) dest = XEXP (dest, 0); if (GET_CODE (dest) == REG) SET_BIT (reg_set_bitmap, REGNO (dest)); else if (GET_CODE (dest) == MEM) mem_last_set = INSN_CUID (insn); if (GET_CODE (SET_SRC (pat)) == CALL) mark_call (insn); } /* Record things set by a CLOBBER. */ static void mark_clobber (pat, insn) rtx pat, insn; { rtx clob = XEXP (pat, 0); while (GET_CODE (clob) == SUBREG || GET_CODE (clob) == STRICT_LOW_PART) clob = XEXP (clob, 0); if (GET_CODE (clob) == REG) SET_BIT (reg_set_bitmap, REGNO (clob)); else mem_last_set = INSN_CUID (insn); } /* Record things set by INSN. This data is used by oprs_not_set_p. */ static void mark_oprs_set (insn) rtx insn; { rtx pat = PATTERN (insn); int i; if (GET_CODE (pat) == SET) mark_set (pat, insn); else if (GET_CODE (pat) == PARALLEL) for (i = 0; i < XVECLEN (pat, 0); i++) { rtx x = XVECEXP (pat, 0, i); if (GET_CODE (x) == SET) mark_set (x, insn); else if (GET_CODE (x) == CLOBBER) mark_clobber (x, insn); else if (GET_CODE (x) == CALL) mark_call (insn); } else if (GET_CODE (pat) == CLOBBER) mark_clobber (pat, insn); else if (GET_CODE (pat) == CALL) mark_call (insn); } /* Classic GCSE reaching definition support. */ /* Allocate reaching def variables. */ static void alloc_rd_mem (n_blocks, n_insns) int n_blocks, n_insns; { rd_kill = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_insns); sbitmap_vector_zero (rd_kill, n_basic_blocks); rd_gen = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_insns); sbitmap_vector_zero (rd_gen, n_basic_blocks); reaching_defs = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_insns); sbitmap_vector_zero (reaching_defs, n_basic_blocks); rd_out = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_insns); sbitmap_vector_zero (rd_out, n_basic_blocks); } /* Free reaching def variables. */ static void free_rd_mem () { free (rd_kill); free (rd_gen); free (reaching_defs); free (rd_out); } /* Add INSN to the kills of BB. REGNO, set in BB, is killed by INSN. */ static void handle_rd_kill_set (insn, regno, bb) rtx insn; int regno, bb; { struct reg_set *this_reg; for (this_reg = reg_set_table[regno]; this_reg; this_reg = this_reg ->next) if (BLOCK_NUM (this_reg->insn) != BLOCK_NUM (insn)) SET_BIT (rd_kill[bb], INSN_CUID (this_reg->insn)); } /* Compute the set of kill's for reaching definitions. */ static void compute_kill_rd () { int bb, cuid; unsigned int regno; int i; /* For each block For each set bit in `gen' of the block (i.e each insn which generates a definition in the block) Call the reg set by the insn corresponding to that bit regx Look at the linked list starting at reg_set_table[regx] For each setting of regx in the linked list, which is not in this block Set the bit in `kill' corresponding to that insn. */ for (bb = 0; bb < n_basic_blocks; bb++) for (cuid = 0; cuid < max_cuid; cuid++) if (TEST_BIT (rd_gen[bb], cuid)) { rtx insn = CUID_INSN (cuid); rtx pat = PATTERN (insn); if (GET_CODE (insn) == CALL_INSN) { for (regno = 0; regno < FIRST_PSEUDO_REGISTER; regno++) { if ((call_used_regs[regno] && regno != STACK_POINTER_REGNUM #if HARD_FRAME_POINTER_REGNUM != FRAME_POINTER_REGNUM && regno != HARD_FRAME_POINTER_REGNUM #endif #if ARG_POINTER_REGNUM != FRAME_POINTER_REGNUM && ! (regno == ARG_POINTER_REGNUM && fixed_regs[regno]) #endif #if !defined (PIC_OFFSET_TABLE_REG_CALL_CLOBBERED) && ! (regno == PIC_OFFSET_TABLE_REGNUM && flag_pic) #endif && regno != FRAME_POINTER_REGNUM) || global_regs[regno]) handle_rd_kill_set (insn, regno, bb); } } if (GET_CODE (pat) == PARALLEL) { for (i = XVECLEN (pat, 0) - 1; i >= 0; i--) { enum rtx_code code = GET_CODE (XVECEXP (pat, 0, i)); if ((code == SET || code == CLOBBER) && GET_CODE (XEXP (XVECEXP (pat, 0, i), 0)) == REG) handle_rd_kill_set (insn, REGNO (XEXP (XVECEXP (pat, 0, i), 0)), bb); } } else if (GET_CODE (pat) == SET && GET_CODE (SET_DEST (pat)) == REG) /* Each setting of this register outside of this block must be marked in the set of kills in this block. */ handle_rd_kill_set (insn, REGNO (SET_DEST (pat)), bb); } } /* Compute the reaching definitions as in Compilers Principles, Techniques, and Tools. Aho, Sethi, Ullman, Chapter 10. It is the same algorithm as used for computing available expressions but applied to the gens and kills of reaching definitions. */ static void compute_rd () { int bb, changed, passes; for (bb = 0; bb < n_basic_blocks; bb++) sbitmap_copy (rd_out[bb] /*dst*/, rd_gen[bb] /*src*/); passes = 0; changed = 1; while (changed) { changed = 0; for (bb = 0; bb < n_basic_blocks; bb++) { sbitmap_union_of_preds (reaching_defs[bb], rd_out, bb); changed |= sbitmap_union_of_diff (rd_out[bb], rd_gen[bb], reaching_defs[bb], rd_kill[bb]); } passes++; } if (gcse_file) fprintf (gcse_file, "reaching def computation: %d passes\n", passes); } /* Classic GCSE available expression support. */ /* Allocate memory for available expression computation. */ static void alloc_avail_expr_mem (n_blocks, n_exprs) int n_blocks, n_exprs; { ae_kill = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_exprs); sbitmap_vector_zero (ae_kill, n_basic_blocks); ae_gen = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_exprs); sbitmap_vector_zero (ae_gen, n_basic_blocks); ae_in = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_exprs); sbitmap_vector_zero (ae_in, n_basic_blocks); ae_out = (sbitmap *) sbitmap_vector_alloc (n_blocks, n_exprs); sbitmap_vector_zero (ae_out, n_basic_blocks); } static void free_avail_expr_mem () { free (ae_kill); free (ae_gen); free (ae_in); free (ae_out); } /* Compute the set of available expressions generated in each basic block. */ static void compute_ae_gen () { unsigned int i; struct expr *expr; struct occr *occr; /* For each recorded occurrence of each expression, set ae_gen[bb][expr]. This is all we have to do because an expression is not recorded if it is not available, and the only expressions we want to work with are the ones that are recorded. */ for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr != 0; expr = expr->next_same_hash) for (occr = expr->avail_occr; occr != 0; occr = occr->next) SET_BIT (ae_gen[BLOCK_NUM (occr->insn)], expr->bitmap_index); } /* Return non-zero if expression X is killed in BB. */ static int expr_killed_p (x, bb) rtx x; int bb; { int i, j; enum rtx_code code; const char *fmt; if (x == 0) return 1; code = GET_CODE (x); switch (code) { case REG: return TEST_BIT (reg_set_in_block[bb], REGNO (x)); case MEM: if (mem_set_in_block[bb]) return 1; else return expr_killed_p (XEXP (x, 0), bb); case PC: case CC0: /*FIXME*/ case CONST: case CONST_INT: case CONST_DOUBLE: case SYMBOL_REF: case LABEL_REF: case ADDR_VEC: case ADDR_DIFF_VEC: return 0; default: break; } for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) return expr_killed_p (XEXP (x, i), bb); else if (expr_killed_p (XEXP (x, i), bb)) return 1; } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) if (expr_killed_p (XVECEXP (x, i, j), bb)) return 1; } return 0; } /* Compute the set of available expressions killed in each basic block. */ static void compute_ae_kill (ae_gen, ae_kill) sbitmap *ae_gen, *ae_kill; { int bb; unsigned int i; struct expr *expr; for (bb = 0; bb < n_basic_blocks; bb++) for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr; expr = expr->next_same_hash) { /* Skip EXPR if generated in this block. */ if (TEST_BIT (ae_gen[bb], expr->bitmap_index)) continue; if (expr_killed_p (expr->expr, bb)) SET_BIT (ae_kill[bb], expr->bitmap_index); } } /* Actually perform the Classic GCSE optimizations. */ /* Return non-zero if occurrence OCCR of expression EXPR reaches block BB. CHECK_SELF_LOOP is non-zero if we should consider a block reaching itself as a positive reach. We want to do this when there are two computations of the expression in the block. VISITED is a pointer to a working buffer for tracking which BB's have been visited. It is NULL for the top-level call. We treat reaching expressions that go through blocks containing the same reaching expression as "not reaching". E.g. if EXPR is generated in blocks 2 and 3, INSN is in block 4, and 2->3->4, we treat the expression in block 2 as not reaching. The intent is to improve the probability of finding only one reaching expression and to reduce register lifetimes by picking the closest such expression. */ static int expr_reaches_here_p_work (occr, expr, bb, check_self_loop, visited) struct occr *occr; struct expr *expr; int bb; int check_self_loop; char *visited; { edge pred; for (pred = BASIC_BLOCK(bb)->pred; pred != NULL; pred = pred->pred_next) { int pred_bb = pred->src->index; if (visited[pred_bb]) /* This predecessor has already been visited. Nothing to do. */ ; else if (pred_bb == bb) { /* BB loops on itself. */ if (check_self_loop && TEST_BIT (ae_gen[pred_bb], expr->bitmap_index) && BLOCK_NUM (occr->insn) == pred_bb) return 1; visited[pred_bb] = 1; } /* Ignore this predecessor if it kills the expression. */ else if (TEST_BIT (ae_kill[pred_bb], expr->bitmap_index)) visited[pred_bb] = 1; /* Does this predecessor generate this expression? */ else if (TEST_BIT (ae_gen[pred_bb], expr->bitmap_index)) { /* Is this the occurrence we're looking for? Note that there's only one generating occurrence per block so we just need to check the block number. */ if (BLOCK_NUM (occr->insn) == pred_bb) return 1; visited[pred_bb] = 1; } /* Neither gen nor kill. */ else { visited[pred_bb] = 1; if (expr_reaches_here_p_work (occr, expr, pred_bb, check_self_loop, visited)) return 1; } } /* All paths have been checked. */ return 0; } /* This wrapper for expr_reaches_here_p_work() is to ensure that any memory allocated for that function is returned. */ static int expr_reaches_here_p (occr, expr, bb, check_self_loop) struct occr *occr; struct expr *expr; int bb; int check_self_loop; { int rval; char *visited = (char *) xcalloc (n_basic_blocks, 1); rval = expr_reaches_here_p_work (occr, expr, bb, check_self_loop, visited); free (visited); return rval; } /* Return the instruction that computes EXPR that reaches INSN's basic block. If there is more than one such instruction, return NULL. Called only by handle_avail_expr. */ static rtx computing_insn (expr, insn) struct expr *expr; rtx insn; { int bb = BLOCK_NUM (insn); if (expr->avail_occr->next == NULL) { if (BLOCK_NUM (expr->avail_occr->insn) == bb) /* The available expression is actually itself (i.e. a loop in the flow graph) so do nothing. */ return NULL; /* (FIXME) Case that we found a pattern that was created by a substitution that took place. */ return expr->avail_occr->insn; } else { /* Pattern is computed more than once. Search backwards from this insn to see how many of these computations actually reach this insn. */ struct occr *occr; rtx insn_computes_expr = NULL; int can_reach = 0; for (occr = expr->avail_occr; occr != NULL; occr = occr->next) { if (BLOCK_NUM (occr->insn) == bb) { /* The expression is generated in this block. The only time we care about this is when the expression is generated later in the block [and thus there's a loop]. We let the normal cse pass handle the other cases. */ if (INSN_CUID (insn) < INSN_CUID (occr->insn) && expr_reaches_here_p (occr, expr, bb, 1)) { can_reach++; if (can_reach > 1) return NULL; insn_computes_expr = occr->insn; } } else if (expr_reaches_here_p (occr, expr, bb, 0)) { can_reach++; if (can_reach > 1) return NULL; insn_computes_expr = occr->insn; } } if (insn_computes_expr == NULL) abort (); return insn_computes_expr; } } /* Return non-zero if the definition in DEF_INSN can reach INSN. Only called by can_disregard_other_sets. */ static int def_reaches_here_p (insn, def_insn) rtx insn, def_insn; { rtx reg; if (TEST_BIT (reaching_defs[BLOCK_NUM (insn)], INSN_CUID (def_insn))) return 1; if (BLOCK_NUM (insn) == BLOCK_NUM (def_insn)) { if (INSN_CUID (def_insn) < INSN_CUID (insn)) { if (GET_CODE (PATTERN (def_insn)) == PARALLEL) return 1; else if (GET_CODE (PATTERN (def_insn)) == CLOBBER) reg = XEXP (PATTERN (def_insn), 0); else if (GET_CODE (PATTERN (def_insn)) == SET) reg = SET_DEST (PATTERN (def_insn)); else abort (); return ! reg_set_between_p (reg, NEXT_INSN (def_insn), insn); } else return 0; } return 0; } /* Return non-zero if *ADDR_THIS_REG can only have one value at INSN. The value returned is the number of definitions that reach INSN. Returning a value of zero means that [maybe] more than one definition reaches INSN and the caller can't perform whatever optimization it is trying. i.e. it is always safe to return zero. */ static int can_disregard_other_sets (addr_this_reg, insn, for_combine) struct reg_set **addr_this_reg; rtx insn; int for_combine; { int number_of_reaching_defs = 0; struct reg_set *this_reg; for (this_reg = *addr_this_reg; this_reg != 0; this_reg = this_reg->next) if (def_reaches_here_p (insn, this_reg->insn)) { number_of_reaching_defs++; /* Ignore parallels for now. */ if (GET_CODE (PATTERN (this_reg->insn)) == PARALLEL) return 0; if (!for_combine && (GET_CODE (PATTERN (this_reg->insn)) == CLOBBER || ! rtx_equal_p (SET_SRC (PATTERN (this_reg->insn)), SET_SRC (PATTERN (insn))))) /* A setting of the reg to a different value reaches INSN. */ return 0; if (number_of_reaching_defs > 1) { /* If in this setting the value the register is being set to is equal to the previous value the register was set to and this setting reaches the insn we are trying to do the substitution on then we are ok. */ if (GET_CODE (PATTERN (this_reg->insn)) == CLOBBER) return 0; else if (! rtx_equal_p (SET_SRC (PATTERN (this_reg->insn)), SET_SRC (PATTERN (insn)))) return 0; } *addr_this_reg = this_reg; } return number_of_reaching_defs; } /* Expression computed by insn is available and the substitution is legal, so try to perform the substitution. The result is non-zero if any changes were made. */ static int handle_avail_expr (insn, expr) rtx insn; struct expr *expr; { rtx pat, insn_computes_expr; rtx to; struct reg_set *this_reg; int found_setting, use_src; int changed = 0; /* We only handle the case where one computation of the expression reaches this instruction. */ insn_computes_expr = computing_insn (expr, insn); if (insn_computes_expr == NULL) return 0; found_setting = 0; use_src = 0; /* At this point we know only one computation of EXPR outside of this block reaches this insn. Now try to find a register that the expression is computed into. */ if (GET_CODE (SET_SRC (PATTERN (insn_computes_expr))) == REG) { /* This is the case when the available expression that reaches here has already been handled as an available expression. */ unsigned int regnum_for_replacing = REGNO (SET_SRC (PATTERN (insn_computes_expr))); /* If the register was created by GCSE we can't use `reg_set_table', however we know it's set only once. */ if (regnum_for_replacing >= max_gcse_regno /* If the register the expression is computed into is set only once, or only one set reaches this insn, we can use it. */ || (((this_reg = reg_set_table[regnum_for_replacing]), this_reg->next == NULL) || can_disregard_other_sets (&this_reg, insn, 0))) { use_src = 1; found_setting = 1; } } if (!found_setting) { unsigned int regnum_for_replacing = REGNO (SET_DEST (PATTERN (insn_computes_expr))); /* This shouldn't happen. */ if (regnum_for_replacing >= max_gcse_regno) abort (); this_reg = reg_set_table[regnum_for_replacing]; /* If the register the expression is computed into is set only once, or only one set reaches this insn, use it. */ if (this_reg->next == NULL || can_disregard_other_sets (&this_reg, insn, 0)) found_setting = 1; } if (found_setting) { pat = PATTERN (insn); if (use_src) to = SET_SRC (PATTERN (insn_computes_expr)); else to = SET_DEST (PATTERN (insn_computes_expr)); changed = validate_change (insn, &SET_SRC (pat), to, 0); /* We should be able to ignore the return code from validate_change but to play it safe we check. */ if (changed) { gcse_subst_count++; if (gcse_file != NULL) { fprintf (gcse_file, "GCSE: Replacing the source in insn %d with", INSN_UID (insn)); fprintf (gcse_file, " reg %d %s insn %d\n", REGNO (to), use_src ? "from" : "set in", INSN_UID (insn_computes_expr)); } } } /* The register that the expr is computed into is set more than once. */ else if (1 /*expensive_op(this_pattrn->op) && do_expensive_gcse)*/) { /* Insert an insn after insnx that copies the reg set in insnx into a new pseudo register call this new register REGN. From insnb until end of basic block or until REGB is set replace all uses of REGB with REGN. */ rtx new_insn; to = gen_reg_rtx (GET_MODE (SET_DEST (PATTERN (insn_computes_expr)))); /* Generate the new insn. */ /* ??? If the change fails, we return 0, even though we created an insn. I think this is ok. */ new_insn = emit_insn_after (gen_rtx_SET (VOIDmode, to, SET_DEST (PATTERN (insn_computes_expr))), insn_computes_expr); /* Keep block number table up to date. */ set_block_num (new_insn, BLOCK_NUM (insn_computes_expr)); /* Keep register set table up to date. */ record_one_set (REGNO (to), new_insn); gcse_create_count++; if (gcse_file != NULL) { fprintf (gcse_file, "GCSE: Creating insn %d to copy value of reg %d", INSN_UID (NEXT_INSN (insn_computes_expr)), REGNO (SET_SRC (PATTERN (NEXT_INSN (insn_computes_expr))))); fprintf (gcse_file, ", computed in insn %d,\n", INSN_UID (insn_computes_expr)); fprintf (gcse_file, " into newly allocated reg %d\n", REGNO (to)); } pat = PATTERN (insn); /* Do register replacement for INSN. */ changed = validate_change (insn, &SET_SRC (pat), SET_DEST (PATTERN (NEXT_INSN (insn_computes_expr))), 0); /* We should be able to ignore the return code from validate_change but to play it safe we check. */ if (changed) { gcse_subst_count++; if (gcse_file != NULL) { fprintf (gcse_file, "GCSE: Replacing the source in insn %d with reg %d ", INSN_UID (insn), REGNO (SET_DEST (PATTERN (NEXT_INSN (insn_computes_expr))))); fprintf (gcse_file, "set in insn %d\n", INSN_UID (insn_computes_expr)); } } } return changed; } /* Perform classic GCSE. This is called by one_classic_gcse_pass after all the dataflow analysis has been done. The result is non-zero if a change was made. */ static int classic_gcse () { int bb, changed; rtx insn; /* Note we start at block 1. */ changed = 0; for (bb = 1; bb < n_basic_blocks; bb++) { /* Reset tables used to keep track of what's still valid [since the start of the block]. */ reset_opr_set_tables (); for (insn = BLOCK_HEAD (bb); insn != NULL && insn != NEXT_INSN (BLOCK_END (bb)); insn = NEXT_INSN (insn)) { /* Is insn of form (set (pseudo-reg) ...)? */ if (GET_CODE (insn) == INSN && GET_CODE (PATTERN (insn)) == SET && GET_CODE (SET_DEST (PATTERN (insn))) == REG && REGNO (SET_DEST (PATTERN (insn))) >= FIRST_PSEUDO_REGISTER) { rtx pat = PATTERN (insn); rtx src = SET_SRC (pat); struct expr *expr; if (want_to_gcse_p (src) /* Is the expression recorded? */ && ((expr = lookup_expr (src)) != NULL) /* Is the expression available [at the start of the block]? */ && TEST_BIT (ae_in[bb], expr->bitmap_index) /* Are the operands unchanged since the start of the block? */ && oprs_not_set_p (src, insn)) changed |= handle_avail_expr (insn, expr); } /* Keep track of everything modified by this insn. */ /* ??? Need to be careful w.r.t. mods done to INSN. */ if (INSN_P (insn)) mark_oprs_set (insn); } } return changed; } /* Top level routine to perform one classic GCSE pass. Return non-zero if a change was made. */ static int one_classic_gcse_pass (pass) int pass; { int changed = 0; gcse_subst_count = 0; gcse_create_count = 0; alloc_expr_hash_table (max_cuid); alloc_rd_mem (n_basic_blocks, max_cuid); compute_expr_hash_table (); if (gcse_file) dump_hash_table (gcse_file, "Expression", expr_hash_table, expr_hash_table_size, n_exprs); if (n_exprs > 0) { compute_kill_rd (); compute_rd (); alloc_avail_expr_mem (n_basic_blocks, n_exprs); compute_ae_gen (); compute_ae_kill (ae_gen, ae_kill); compute_available (ae_gen, ae_kill, ae_out, ae_in); changed = classic_gcse (); free_avail_expr_mem (); } free_rd_mem (); free_expr_hash_table (); if (gcse_file) { fprintf (gcse_file, "\n"); fprintf (gcse_file, "GCSE of %s, pass %d: %d bytes needed, %d substs,", current_function_name, pass, bytes_used, gcse_subst_count); fprintf (gcse_file, "%d insns created\n", gcse_create_count); } return changed; } /* Compute copy/constant propagation working variables. */ /* Local properties of assignments. */ static sbitmap *cprop_pavloc; static sbitmap *cprop_absaltered; /* Global properties of assignments (computed from the local properties). */ static sbitmap *cprop_avin; static sbitmap *cprop_avout; /* Allocate vars used for copy/const propagation. N_BLOCKS is the number of basic blocks. N_SETS is the number of sets. */ static void alloc_cprop_mem (n_blocks, n_sets) int n_blocks, n_sets; { cprop_pavloc = sbitmap_vector_alloc (n_blocks, n_sets); cprop_absaltered = sbitmap_vector_alloc (n_blocks, n_sets); cprop_avin = sbitmap_vector_alloc (n_blocks, n_sets); cprop_avout = sbitmap_vector_alloc (n_blocks, n_sets); } /* Free vars used by copy/const propagation. */ static void free_cprop_mem () { free (cprop_pavloc); free (cprop_absaltered); free (cprop_avin); free (cprop_avout); } /* For each block, compute whether X is transparent. X is either an expression or an assignment [though we don't care which, for this context an assignment is treated as an expression]. For each block where an element of X is modified, set (SET_P == 1) or reset (SET_P == 0) the INDX bit in BMAP. */ static void compute_transp (x, indx, bmap, set_p) rtx x; int indx; sbitmap *bmap; int set_p; { int bb, i, j; enum rtx_code code; reg_set *r; const char *fmt; /* repeat is used to turn tail-recursion into iteration since GCC can't do it when there's no return value. */ repeat: if (x == 0) return; code = GET_CODE (x); switch (code) { case REG: if (set_p) { if (REGNO (x) < FIRST_PSEUDO_REGISTER) { for (bb = 0; bb < n_basic_blocks; bb++) if (TEST_BIT (reg_set_in_block[bb], REGNO (x))) SET_BIT (bmap[bb], indx); } else { for (r = reg_set_table[REGNO (x)]; r != NULL; r = r->next) SET_BIT (bmap[BLOCK_NUM (r->insn)], indx); } } else { if (REGNO (x) < FIRST_PSEUDO_REGISTER) { for (bb = 0; bb < n_basic_blocks; bb++) if (TEST_BIT (reg_set_in_block[bb], REGNO (x))) RESET_BIT (bmap[bb], indx); } else { for (r = reg_set_table[REGNO (x)]; r != NULL; r = r->next) RESET_BIT (bmap[BLOCK_NUM (r->insn)], indx); } } return; case MEM: if (set_p) { for (bb = 0; bb < n_basic_blocks; bb++) if (mem_set_in_block[bb]) SET_BIT (bmap[bb], indx); } else { for (bb = 0; bb < n_basic_blocks; bb++) if (mem_set_in_block[bb]) RESET_BIT (bmap[bb], indx); } x = XEXP (x, 0); goto repeat; case PC: case CC0: /*FIXME*/ case CONST: case CONST_INT: case CONST_DOUBLE: case SYMBOL_REF: case LABEL_REF: case ADDR_VEC: case ADDR_DIFF_VEC: return; default: break; } for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) { x = XEXP (x, i); goto repeat; } compute_transp (XEXP (x, i), indx, bmap, set_p); } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) compute_transp (XVECEXP (x, i, j), indx, bmap, set_p); } } /* Top level routine to do the dataflow analysis needed by copy/const propagation. */ static void compute_cprop_data () { compute_local_properties (cprop_absaltered, cprop_pavloc, NULL, 1); compute_available (cprop_pavloc, cprop_absaltered, cprop_avout, cprop_avin); } /* Copy/constant propagation. */ /* Maximum number of register uses in an insn that we handle. */ #define MAX_USES 8 /* Table of uses found in an insn. Allocated statically to avoid alloc/free complexity and overhead. */ static struct reg_use reg_use_table[MAX_USES]; /* Index into `reg_use_table' while building it. */ static int reg_use_count; /* Set up a list of register numbers used in INSN. The found uses are stored in `reg_use_table'. `reg_use_count' is initialized to zero before entry, and contains the number of uses in the table upon exit. ??? If a register appears multiple times we will record it multiple times. This doesn't hurt anything but it will slow things down. */ static void find_used_regs (x) rtx x; { int i, j; enum rtx_code code; const char *fmt; /* repeat is used to turn tail-recursion into iteration since GCC can't do it when there's no return value. */ repeat: if (x == 0) return; code = GET_CODE (x); switch (code) { case REG: if (reg_use_count == MAX_USES) return; reg_use_table[reg_use_count].reg_rtx = x; reg_use_count++; return; case MEM: x = XEXP (x, 0); goto repeat; case PC: case CC0: case CONST: case CONST_INT: case CONST_DOUBLE: case SYMBOL_REF: case LABEL_REF: case CLOBBER: case ADDR_VEC: case ADDR_DIFF_VEC: case ASM_INPUT: /*FIXME*/ return; case SET: if (GET_CODE (SET_DEST (x)) == MEM) find_used_regs (SET_DEST (x)); x = SET_SRC (x); goto repeat; default: break; } /* Recursively scan the operands of this expression. */ for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') { /* If we are about to do the last recursive call needed at this level, change it into iteration. This function is called enough to be worth it. */ if (i == 0) { x = XEXP (x, 0); goto repeat; } find_used_regs (XEXP (x, i)); } else if (fmt[i] == 'E') for (j = 0; j < XVECLEN (x, i); j++) find_used_regs (XVECEXP (x, i, j)); } } /* Try to replace all non-SET_DEST occurrences of FROM in INSN with TO. Returns non-zero is successful. */ static int try_replace_reg (from, to, insn) rtx from, to, insn; { rtx note = find_reg_equal_equiv_note (insn); rtx src = 0; int success = 0; rtx set = single_set (insn); /* If this is a single set, try to simplify the source of the set given our substitution. We could perhaps try this for multiple SETs, but it probably won't buy us anything. */ if (set != 0) { src = simplify_replace_rtx (SET_SRC (set), from, to); /* Try this two ways: first just replace SET_SRC. If that doesn't work and this is a PARALLEL, try to replace the whole pattern with a new SET. */ if (validate_change (insn, &SET_SRC (set), src, 0)) success = 1; else if (GET_CODE (PATTERN (insn)) == PARALLEL && validate_change (insn, &PATTERN (insn), gen_rtx_SET (VOIDmode, SET_DEST (set), src), 0)) success = 1; } /* Otherwise, try to do a global replacement within the insn. */ if (!success) success = validate_replace_src (from, to, insn); /* If we've failed to do replacement, have a single SET, and don't already have a note, add a REG_EQUAL note to not lose information. */ if (!success && note == 0 && set != 0) note = REG_NOTES (insn) = gen_rtx_EXPR_LIST (REG_EQUAL, src, REG_NOTES (insn)); /* If there is already a NOTE, update the expression in it with our replacement. */ else if (note != 0) XEXP (note, 0) = simplify_replace_rtx (XEXP (note, 0), from, to); /* REG_EQUAL may get simplified into register. We don't allow that. Remove that note. This code ought not to hapen, because previous code ought to syntetize reg-reg move, but be on the safe side. */ if (note && REG_P (XEXP (note, 0))) remove_note (insn, note); return success; } /* Find a set of REGNOs that are available on entry to INSN's block. Returns NULL no such set is found. */ static struct expr * find_avail_set (regno, insn) int regno; rtx insn; { /* SET1 contains the last set found that can be returned to the caller for use in a substitution. */ struct expr *set1 = 0; /* Loops are not possible here. To get a loop we would need two sets available at the start of the block containing INSN. ie we would need two sets like this available at the start of the block: (set (reg X) (reg Y)) (set (reg Y) (reg X)) This can not happen since the set of (reg Y) would have killed the set of (reg X) making it unavailable at the start of this block. */ while (1) { rtx src; struct expr *set = lookup_set (regno, NULL_RTX); /* Find a set that is available at the start of the block which contains INSN. */ while (set) { if (TEST_BIT (cprop_avin[BLOCK_NUM (insn)], set->bitmap_index)) break; set = next_set (regno, set); } /* If no available set was found we've reached the end of the (possibly empty) copy chain. */ if (set == 0) break; if (GET_CODE (set->expr) != SET) abort (); src = SET_SRC (set->expr); /* We know the set is available. Now check that SRC is ANTLOC (i.e. none of the source operands have changed since the start of the block). If the source operand changed, we may still use it for the next iteration of this loop, but we may not use it for substitutions. */ if (CONSTANT_P (src) || oprs_not_set_p (src, insn)) set1 = set; /* If the source of the set is anything except a register, then we have reached the end of the copy chain. */ if (GET_CODE (src) != REG) break; /* Follow the copy chain, ie start another iteration of the loop and see if we have an available copy into SRC. */ regno = REGNO (src); } /* SET1 holds the last set that was available and anticipatable at INSN. */ return set1; } /* Subroutine of cprop_insn that tries to propagate constants into JUMP_INSNS. INSN must be a conditional jump. FROM is what we will try to replace, SRC is the constant we will try to substitute for it. Returns nonzero if a change was made. We know INSN has just a SET. */ static int cprop_jump (insn, from, src) rtx insn; rtx from; rtx src; { rtx set = PATTERN (insn); rtx new = simplify_replace_rtx (SET_SRC (set), from, src); /* If no simplification can be made, then try the next register. */ if (rtx_equal_p (new, SET_SRC (set))) return 0; /* If this is now a no-op leave it that way, but update LABEL_NUSED if necessary. */ if (new == pc_rtx) { SET_SRC (set) = new; if (JUMP_LABEL (insn) != 0) --LABEL_NUSES (JUMP_LABEL (insn)); } /* Otherwise, this must be a valid instruction. */ else if (! validate_change (insn, &SET_SRC (set), new, 0)) return 0; /* If this has turned into an unconditional jump, then put a barrier after it so that the unreachable code will be deleted. */ if (GET_CODE (SET_SRC (set)) == LABEL_REF) emit_barrier_after (insn); run_jump_opt_after_gcse = 1; const_prop_count++; if (gcse_file != NULL) { fprintf (gcse_file, "CONST-PROP: Replacing reg %d in insn %d with constant ", REGNO (from), INSN_UID (insn)); print_rtl (gcse_file, src); fprintf (gcse_file, "\n"); } return 1; } #ifdef HAVE_cc0 /* Subroutine of cprop_insn that tries to propagate constants into JUMP_INSNS for machines that have CC0. INSN is a single set that stores into CC0; the insn following it is a conditional jump. REG_USED is the use we will try to replace, SRC is the constant we will try to substitute for it. Returns nonzero if a change was made. */ static int cprop_cc0_jump (insn, reg_used, src) rtx insn; struct reg_use *reg_used; rtx src; { /* First substitute in the SET_SRC of INSN, then substitute that for CC0 in JUMP. */ rtx jump = NEXT_INSN (insn); rtx new_src = simplify_replace_rtx (SET_SRC (PATTERN (insn)), reg_used->reg_rtx, src); if (! cprop_jump (jump, cc0_rtx, new_src)) return 0; /* If we succeeded, delete the cc0 setter. */ PUT_CODE (insn, NOTE); NOTE_LINE_NUMBER (insn) = NOTE_INSN_DELETED; NOTE_SOURCE_FILE (insn) = 0; return 1; } #endif /* Perform constant and copy propagation on INSN. The result is non-zero if a change was made. */ static int cprop_insn (insn, alter_jumps) rtx insn; int alter_jumps; { struct reg_use *reg_used; int changed = 0; rtx note; /* Only propagate into SETs. Note that a conditional jump is a SET with pc_rtx as the destination. */ if (GET_CODE (insn) != INSN && GET_CODE (insn) != JUMP_INSN) return 0; reg_use_count = 0; find_used_regs (PATTERN (insn)); note = find_reg_equal_equiv_note (insn); /* We may win even when propagating constants into notes. */ if (note) find_used_regs (XEXP (note, 0)); for (reg_used = ®_use_table[0]; reg_use_count > 0; reg_used++, reg_use_count--) { unsigned int regno = REGNO (reg_used->reg_rtx); rtx pat, src; struct expr *set; /* Ignore registers created by GCSE. We do this because ... */ if (regno >= max_gcse_regno) continue; /* If the register has already been set in this block, there's nothing we can do. */ if (! oprs_not_set_p (reg_used->reg_rtx, insn)) continue; /* Find an assignment that sets reg_used and is available at the start of the block. */ set = find_avail_set (regno, insn); if (! set) continue; pat = set->expr; /* ??? We might be able to handle PARALLELs. Later. */ if (GET_CODE (pat) != SET) abort (); src = SET_SRC (pat); /* Constant propagation. */ if (GET_CODE (src) == CONST_INT || GET_CODE (src) == CONST_DOUBLE || GET_CODE (src) == SYMBOL_REF) { /* Handle normal insns first. */ if (GET_CODE (insn) == INSN && try_replace_reg (reg_used->reg_rtx, src, insn)) { changed = 1; const_prop_count++; if (gcse_file != NULL) { fprintf (gcse_file, "CONST-PROP: Replacing reg %d in ", regno); fprintf (gcse_file, "insn %d with constant ", INSN_UID (insn)); print_rtl (gcse_file, src); fprintf (gcse_file, "\n"); } /* The original insn setting reg_used may or may not now be deletable. We leave the deletion to flow. */ } /* Try to propagate a CONST_INT into a conditional jump. We're pretty specific about what we will handle in this code, we can extend this as necessary over time. Right now the insn in question must look like (set (pc) (if_then_else ...)) */ else if (alter_jumps && GET_CODE (insn) == JUMP_INSN && condjump_p (insn) && ! simplejump_p (insn)) changed |= cprop_jump (insn, reg_used->reg_rtx, src); #ifdef HAVE_cc0 /* Similar code for machines that use a pair of CC0 setter and conditional jump insn. */ else if (alter_jumps && GET_CODE (PATTERN (insn)) == SET && SET_DEST (PATTERN (insn)) == cc0_rtx && GET_CODE (NEXT_INSN (insn)) == JUMP_INSN && condjump_p (NEXT_INSN (insn)) && ! simplejump_p (NEXT_INSN (insn)) && cprop_cc0_jump (insn, reg_used, src)) { changed = 1; break; } #endif } else if (GET_CODE (src) == REG && REGNO (src) >= FIRST_PSEUDO_REGISTER && REGNO (src) != regno) { if (try_replace_reg (reg_used->reg_rtx, src, insn)) { changed = 1; copy_prop_count++; if (gcse_file != NULL) { fprintf (gcse_file, "COPY-PROP: Replacing reg %d in insn %d", regno, INSN_UID (insn)); fprintf (gcse_file, " with reg %d\n", REGNO (src)); } /* The original insn setting reg_used may or may not now be deletable. We leave the deletion to flow. */ /* FIXME: If it turns out that the insn isn't deletable, then we may have unnecessarily extended register lifetimes and made things worse. */ } } } return changed; } /* Forward propagate copies. This includes copies and constants. Return non-zero if a change was made. */ static int cprop (alter_jumps) int alter_jumps; { int bb, changed; rtx insn; /* Note we start at block 1. */ changed = 0; for (bb = 1; bb < n_basic_blocks; bb++) { /* Reset tables used to keep track of what's still valid [since the start of the block]. */ reset_opr_set_tables (); for (insn = BLOCK_HEAD (bb); insn != NULL && insn != NEXT_INSN (BLOCK_END (bb)); insn = NEXT_INSN (insn)) if (INSN_P (insn)) { changed |= cprop_insn (insn, alter_jumps); /* Keep track of everything modified by this insn. */ /* ??? Need to be careful w.r.t. mods done to INSN. Don't call mark_oprs_set if we turned the insn into a NOTE. */ if (GET_CODE (insn) != NOTE) mark_oprs_set (insn); } } if (gcse_file != NULL) fprintf (gcse_file, "\n"); return changed; } /* Perform one copy/constant propagation pass. F is the first insn in the function. PASS is the pass count. */ static int one_cprop_pass (pass, alter_jumps) int pass; int alter_jumps; { int changed = 0; const_prop_count = 0; copy_prop_count = 0; alloc_set_hash_table (max_cuid); compute_set_hash_table (); if (gcse_file) dump_hash_table (gcse_file, "SET", set_hash_table, set_hash_table_size, n_sets); if (n_sets > 0) { alloc_cprop_mem (n_basic_blocks, n_sets); compute_cprop_data (); changed = cprop (alter_jumps); free_cprop_mem (); } free_set_hash_table (); if (gcse_file) { fprintf (gcse_file, "CPROP of %s, pass %d: %d bytes needed, ", current_function_name, pass, bytes_used); fprintf (gcse_file, "%d const props, %d copy props\n\n", const_prop_count, copy_prop_count); } return changed; } /* Compute PRE+LCM working variables. */ /* Local properties of expressions. */ /* Nonzero for expressions that are transparent in the block. */ static sbitmap *transp; /* Nonzero for expressions that are transparent at the end of the block. This is only zero for expressions killed by abnormal critical edge created by a calls. */ static sbitmap *transpout; /* Nonzero for expressions that are computed (available) in the block. */ static sbitmap *comp; /* Nonzero for expressions that are locally anticipatable in the block. */ static sbitmap *antloc; /* Nonzero for expressions where this block is an optimal computation point. */ static sbitmap *pre_optimal; /* Nonzero for expressions which are redundant in a particular block. */ static sbitmap *pre_redundant; /* Nonzero for expressions which should be inserted on a specific edge. */ static sbitmap *pre_insert_map; /* Nonzero for expressions which should be deleted in a specific block. */ static sbitmap *pre_delete_map; /* Contains the edge_list returned by pre_edge_lcm. */ static struct edge_list *edge_list; /* Redundant insns. */ static sbitmap pre_redundant_insns; /* Allocate vars used for PRE analysis. */ static void alloc_pre_mem (n_blocks, n_exprs) int n_blocks, n_exprs; { transp = sbitmap_vector_alloc (n_blocks, n_exprs); comp = sbitmap_vector_alloc (n_blocks, n_exprs); antloc = sbitmap_vector_alloc (n_blocks, n_exprs); pre_optimal = NULL; pre_redundant = NULL; pre_insert_map = NULL; pre_delete_map = NULL; ae_in = NULL; ae_out = NULL; ae_kill = sbitmap_vector_alloc (n_blocks, n_exprs); /* pre_insert and pre_delete are allocated later. */ } /* Free vars used for PRE analysis. */ static void free_pre_mem () { free (transp); free (comp); /* ANTLOC and AE_KILL are freed just after pre_lcm finishes. */ if (pre_optimal) free (pre_optimal); if (pre_redundant) free (pre_redundant); if (pre_insert_map) free (pre_insert_map); if (pre_delete_map) free (pre_delete_map); if (ae_in) free (ae_in); if (ae_out) free (ae_out); transp = comp = NULL; pre_optimal = pre_redundant = pre_insert_map = pre_delete_map = NULL; ae_in = ae_out = NULL; } /* Top level routine to do the dataflow analysis needed by PRE. */ static void compute_pre_data () { sbitmap trapping_expr; int i; unsigned int ui; compute_local_properties (transp, comp, antloc, 0); sbitmap_vector_zero (ae_kill, n_basic_blocks); /* Collect expressions which might trap. */ trapping_expr = sbitmap_alloc (n_exprs); sbitmap_zero (trapping_expr); for (ui = 0; ui < expr_hash_table_size; ui++) { struct expr *e; for (e = expr_hash_table[ui]; e != NULL; e = e->next_same_hash) if (may_trap_p (e->expr)) SET_BIT (trapping_expr, e->bitmap_index); } /* Compute ae_kill for each basic block using: ~(TRANSP | COMP) This is significantly faster than compute_ae_kill. */ for (i = 0; i < n_basic_blocks; i++) { edge e; /* If the current block is the destination of an abnormal edge, we kill all trapping expressions because we won't be able to properly place the instruction on the edge. So make them neither anticipatable nor transparent. This is fairly conservative. */ for (e = BASIC_BLOCK (i)->pred; e ; e = e->pred_next) if (e->flags & EDGE_ABNORMAL) { sbitmap_difference (antloc[i], antloc[i], trapping_expr); sbitmap_difference (transp[i], transp[i], trapping_expr); break; } sbitmap_a_or_b (ae_kill[i], transp[i], comp[i]); sbitmap_not (ae_kill[i], ae_kill[i]); } edge_list = pre_edge_lcm (gcse_file, n_exprs, transp, comp, antloc, ae_kill, &pre_insert_map, &pre_delete_map); free (antloc); antloc = NULL; free (ae_kill); ae_kill = NULL; free (trapping_expr); } /* PRE utilities */ /* Return non-zero if an occurrence of expression EXPR in OCCR_BB would reach block BB. VISITED is a pointer to a working buffer for tracking which BB's have been visited. It is NULL for the top-level call. We treat reaching expressions that go through blocks containing the same reaching expression as "not reaching". E.g. if EXPR is generated in blocks 2 and 3, INSN is in block 4, and 2->3->4, we treat the expression in block 2 as not reaching. The intent is to improve the probability of finding only one reaching expression and to reduce register lifetimes by picking the closest such expression. */ static int pre_expr_reaches_here_p_work (occr_bb, expr, bb, visited) int occr_bb; struct expr *expr; int bb; char *visited; { edge pred; for (pred = BASIC_BLOCK (bb)->pred; pred != NULL; pred = pred->pred_next) { int pred_bb = pred->src->index; if (pred->src == ENTRY_BLOCK_PTR /* Has predecessor has already been visited? */ || visited[pred_bb]) ;/* Nothing to do. */ /* Does this predecessor generate this expression? */ else if (TEST_BIT (comp[pred_bb], expr->bitmap_index)) { /* Is this the occurrence we're looking for? Note that there's only one generating occurrence per block so we just need to check the block number. */ if (occr_bb == pred_bb) return 1; visited[pred_bb] = 1; } /* Ignore this predecessor if it kills the expression. */ else if (! TEST_BIT (transp[pred_bb], expr->bitmap_index)) visited[pred_bb] = 1; /* Neither gen nor kill. */ else { visited[pred_bb] = 1; if (pre_expr_reaches_here_p_work (occr_bb, expr, pred_bb, visited)) return 1; } } /* All paths have been checked. */ return 0; } /* The wrapper for pre_expr_reaches_here_work that ensures that any memory allocated for that function is returned. */ static int pre_expr_reaches_here_p (occr_bb, expr, bb) int occr_bb; struct expr *expr; int bb; { int rval; char *visited = (char *) xcalloc (n_basic_blocks, 1); rval = pre_expr_reaches_here_p_work(occr_bb, expr, bb, visited); free (visited); return rval; } /* Given an expr, generate RTL which we can insert at the end of a BB, or on an edge. Set the block number of any insns generated to the value of BB. */ static rtx process_insert_insn (expr) struct expr *expr; { rtx reg = expr->reaching_reg; rtx exp = copy_rtx (expr->expr); rtx pat; start_sequence (); /* If the expression is something that's an operand, like a constant, just copy it to a register. */ if (general_operand (exp, GET_MODE (reg))) emit_move_insn (reg, exp); /* Otherwise, make a new insn to compute this expression and make sure the insn will be recognized (this also adds any needed CLOBBERs). Copy the expression to make sure we don't have any sharing issues. */ if (insn_invalid_p (emit_insn (gen_rtx_SET (VOIDmode, reg, exp)))) abort (); pat = gen_sequence (); end_sequence (); return pat; } /* Add EXPR to the end of basic block BB. This is used by both the PRE and code hoisting. For PRE, we want to verify that the expr is either transparent or locally anticipatable in the target block. This check makes no sense for code hoisting. */ static void insert_insn_end_bb (expr, bb, pre) struct expr *expr; int bb; int pre; { rtx insn = BLOCK_END (bb); rtx new_insn; rtx reg = expr->reaching_reg; int regno = REGNO (reg); rtx pat; int i; pat = process_insert_insn (expr); /* If the last insn is a jump, insert EXPR in front [taking care to handle cc0, etc. properly]. */ if (GET_CODE (insn) == JUMP_INSN) { #ifdef HAVE_cc0 rtx note; #endif /* If this is a jump table, then we can't insert stuff here. Since we know the previous real insn must be the tablejump, we insert the new instruction just before the tablejump. */ if (GET_CODE (PATTERN (insn)) == ADDR_VEC || GET_CODE (PATTERN (insn)) == ADDR_DIFF_VEC) insn = prev_real_insn (insn); #ifdef HAVE_cc0 /* FIXME: 'twould be nice to call prev_cc0_setter here but it aborts if cc0 isn't set. */ note = find_reg_note (insn, REG_CC_SETTER, NULL_RTX); if (note) insn = XEXP (note, 0); else { rtx maybe_cc0_setter = prev_nonnote_insn (insn); if (maybe_cc0_setter && INSN_P (maybe_cc0_setter) && sets_cc0_p (PATTERN (maybe_cc0_setter))) insn = maybe_cc0_setter; } #endif /* FIXME: What if something in cc0/jump uses value set in new insn? */ new_insn = emit_block_insn_before (pat, insn, BASIC_BLOCK (bb)); } /* Likewise if the last insn is a call, as will happen in the presence of exception handling. */ else if (GET_CODE (insn) == CALL_INSN) { HARD_REG_SET parm_regs; int nparm_regs; rtx p; /* Keeping in mind SMALL_REGISTER_CLASSES and parameters in registers, we search backward and place the instructions before the first parameter is loaded. Do this for everyone for consistency and a presumtion that we'll get better code elsewhere as well. It should always be the case that we can put these instructions anywhere in the basic block with performing PRE optimizations. Check this. */ if (pre && !TEST_BIT (antloc[bb], expr->bitmap_index) && !TEST_BIT (transp[bb], expr->bitmap_index)) abort (); /* Since different machines initialize their parameter registers in different orders, assume nothing. Collect the set of all parameter registers. */ CLEAR_HARD_REG_SET (parm_regs); nparm_regs = 0; for (p = CALL_INSN_FUNCTION_USAGE (insn); p ; p = XEXP (p, 1)) if (GET_CODE (XEXP (p, 0)) == USE && GET_CODE (XEXP (XEXP (p, 0), 0)) == REG) { if (REGNO (XEXP (XEXP (p, 0), 0)) >= FIRST_PSEUDO_REGISTER) abort (); SET_HARD_REG_BIT (parm_regs, REGNO (XEXP (XEXP (p, 0), 0))); nparm_regs++; } /* Search backward for the first set of a register in this set. */ while (nparm_regs && BLOCK_HEAD (bb) != insn) { insn = PREV_INSN (insn); p = single_set (insn); if (p && GET_CODE (SET_DEST (p)) == REG && REGNO (SET_DEST (p)) < FIRST_PSEUDO_REGISTER && TEST_HARD_REG_BIT (parm_regs, REGNO (SET_DEST (p)))) { CLEAR_HARD_REG_BIT (parm_regs, REGNO (SET_DEST (p))); nparm_regs--; } } /* If we found all the parameter loads, then we want to insert before the first parameter load. If we did not find all the parameter loads, then we might have stopped on the head of the block, which could be a CODE_LABEL. If we inserted before the CODE_LABEL, then we would be putting the insn in the wrong basic block. In that case, put the insn after the CODE_LABEL. Also, respect NOTE_INSN_BASIC_BLOCK. */ while (GET_CODE (insn) == CODE_LABEL || NOTE_INSN_BASIC_BLOCK_P (insn)) insn = NEXT_INSN (insn); new_insn = emit_block_insn_before (pat, insn, BASIC_BLOCK (bb)); } else { new_insn = emit_insn_after (pat, insn); BLOCK_END (bb) = new_insn; } /* Keep block number table up to date. Note, PAT could be a multiple insn sequence, we have to make sure that each insn in the sequence is handled. */ if (GET_CODE (pat) == SEQUENCE) { for (i = 0; i < XVECLEN (pat, 0); i++) { rtx insn = XVECEXP (pat, 0, i); set_block_num (insn, bb); if (INSN_P (insn)) add_label_notes (PATTERN (insn), new_insn); note_stores (PATTERN (insn), record_set_info, insn); } } else { add_label_notes (SET_SRC (pat), new_insn); set_block_num (new_insn, bb); /* Keep register set table up to date. */ record_one_set (regno, new_insn); } gcse_create_count++; if (gcse_file) { fprintf (gcse_file, "PRE/HOIST: end of bb %d, insn %d, ", bb, INSN_UID (new_insn)); fprintf (gcse_file, "copying expression %d to reg %d\n", expr->bitmap_index, regno); } } /* Insert partially redundant expressions on edges in the CFG to make the expressions fully redundant. */ static int pre_edge_insert (edge_list, index_map) struct edge_list *edge_list; struct expr **index_map; { int e, i, j, num_edges, set_size, did_insert = 0; sbitmap *inserted; /* Where PRE_INSERT_MAP is nonzero, we add the expression on that edge if it reaches any of the deleted expressions. */ set_size = pre_insert_map[0]->size; num_edges = NUM_EDGES (edge_list); inserted = sbitmap_vector_alloc (num_edges, n_exprs); sbitmap_vector_zero (inserted, num_edges); for (e = 0; e < num_edges; e++) { int indx; basic_block pred = INDEX_EDGE_PRED_BB (edge_list, e); int bb = pred->index; for (i = indx = 0; i < set_size; i++, indx += SBITMAP_ELT_BITS) { SBITMAP_ELT_TYPE insert = pre_insert_map[e]->elms[i]; for (j = indx; insert && j < n_exprs; j++, insert >>= 1) if ((insert & 1) != 0 && index_map[j]->reaching_reg != NULL_RTX) { struct expr *expr = index_map[j]; struct occr *occr; /* Now look at each deleted occurence of this expression. */ for (occr = expr->antic_occr; occr != NULL; occr = occr->next) { if (! occr->deleted_p) continue; /* Insert this expression on this edge if if it would reach the deleted occurence in BB. */ if (!TEST_BIT (inserted[e], j)) { rtx insn; edge eg = INDEX_EDGE (edge_list, e); /* We can't insert anything on an abnormal and critical edge, so we insert the insn at the end of the previous block. There are several alternatives detailed in Morgans book P277 (sec 10.5) for handling this situation. This one is easiest for now. */ if ((eg->flags & EDGE_ABNORMAL) == EDGE_ABNORMAL) insert_insn_end_bb (index_map[j], bb, 0); else { insn = process_insert_insn (index_map[j]); insert_insn_on_edge (insn, eg); } if (gcse_file) { fprintf (gcse_file, "PRE/HOIST: edge (%d,%d), ", bb, INDEX_EDGE_SUCC_BB (edge_list, e)->index); fprintf (gcse_file, "copy expression %d\n", expr->bitmap_index); } SET_BIT (inserted[e], j); did_insert = 1; gcse_create_count++; } } } } } free (inserted); return did_insert; } /* Copy the result of INSN to REG. INDX is the expression number. */ static void pre_insert_copy_insn (expr, insn) struct expr *expr; rtx insn; { rtx reg = expr->reaching_reg; int regno = REGNO (reg); int indx = expr->bitmap_index; rtx set = single_set (insn); rtx new_insn; int bb = BLOCK_NUM (insn); if (!set) abort (); new_insn = emit_insn_after (gen_rtx_SET (VOIDmode, reg, SET_DEST (set)), insn); /* Keep block number table up to date. */ set_block_num (new_insn, bb); /* Keep register set table up to date. */ record_one_set (regno, new_insn); if (insn == BLOCK_END (bb)) BLOCK_END (bb) = new_insn; gcse_create_count++; if (gcse_file) fprintf (gcse_file, "PRE: bb %d, insn %d, copy expression %d in insn %d to reg %d\n", BLOCK_NUM (insn), INSN_UID (new_insn), indx, INSN_UID (insn), regno); } /* Copy available expressions that reach the redundant expression to `reaching_reg'. */ static void pre_insert_copies () { unsigned int i; struct expr *expr; struct occr *occr; struct occr *avail; /* For each available expression in the table, copy the result to `reaching_reg' if the expression reaches a deleted one. ??? The current algorithm is rather brute force. Need to do some profiling. */ for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr != NULL; expr = expr->next_same_hash) { /* If the basic block isn't reachable, PPOUT will be TRUE. However, we don't want to insert a copy here because the expression may not really be redundant. So only insert an insn if the expression was deleted. This test also avoids further processing if the expression wasn't deleted anywhere. */ if (expr->reaching_reg == NULL) continue; for (occr = expr->antic_occr; occr != NULL; occr = occr->next) { if (! occr->deleted_p) continue; for (avail = expr->avail_occr; avail != NULL; avail = avail->next) { rtx insn = avail->insn; /* No need to handle this one if handled already. */ if (avail->copied_p) continue; /* Don't handle this one if it's a redundant one. */ if (TEST_BIT (pre_redundant_insns, INSN_CUID (insn))) continue; /* Or if the expression doesn't reach the deleted one. */ if (! pre_expr_reaches_here_p (BLOCK_NUM (avail->insn), expr, BLOCK_NUM (occr->insn))) continue; /* Copy the result of avail to reaching_reg. */ pre_insert_copy_insn (expr, insn); avail->copied_p = 1; } } } } /* Delete redundant computations. Deletion is done by changing the insn to copy the `reaching_reg' of the expression into the result of the SET. It is left to later passes (cprop, cse2, flow, combine, regmove) to propagate the copy or eliminate it. Returns non-zero if a change is made. */ static int pre_delete () { unsigned int i; int changed; struct expr *expr; struct occr *occr; changed = 0; for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr != NULL; expr = expr->next_same_hash) { int indx = expr->bitmap_index; /* We only need to search antic_occr since we require ANTLOC != 0. */ for (occr = expr->antic_occr; occr != NULL; occr = occr->next) { rtx insn = occr->insn; rtx set; int bb = BLOCK_NUM (insn); if (TEST_BIT (pre_delete_map[bb], indx)) { set = single_set (insn); if (! set) abort (); /* Create a pseudo-reg to store the result of reaching expressions into. Get the mode for the new pseudo from the mode of the original destination pseudo. */ if (expr->reaching_reg == NULL) expr->reaching_reg = gen_reg_rtx (GET_MODE (SET_DEST (set))); /* In theory this should never fail since we're creating a reg->reg copy. However, on the x86 some of the movXX patterns actually contain clobbers of scratch regs. This may cause the insn created by validate_change to not match any pattern and thus cause validate_change to fail. */ if (validate_change (insn, &SET_SRC (set), expr->reaching_reg, 0)) { occr->deleted_p = 1; SET_BIT (pre_redundant_insns, INSN_CUID (insn)); changed = 1; gcse_subst_count++; } if (gcse_file) { fprintf (gcse_file, "PRE: redundant insn %d (expression %d) in ", INSN_UID (insn), indx); fprintf (gcse_file, "bb %d, reaching reg is %d\n", bb, REGNO (expr->reaching_reg)); } } } } return changed; } /* Perform GCSE optimizations using PRE. This is called by one_pre_gcse_pass after all the dataflow analysis has been done. This is based on the original Morel-Renvoise paper Fred Chow's thesis, and lazy code motion from Knoop, Ruthing and Steffen as described in Advanced Compiler Design and Implementation. ??? A new pseudo reg is created to hold the reaching expression. The nice thing about the classical approach is that it would try to use an existing reg. If the register can't be adequately optimized [i.e. we introduce reload problems], one could add a pass here to propagate the new register through the block. ??? We don't handle single sets in PARALLELs because we're [currently] not able to copy the rest of the parallel when we insert copies to create full redundancies from partial redundancies. However, there's no reason why we can't handle PARALLELs in the cases where there are no partial redundancies. */ static int pre_gcse () { unsigned int i; int did_insert, changed; struct expr **index_map; struct expr *expr; /* Compute a mapping from expression number (`bitmap_index') to hash table entry. */ index_map = (struct expr **) xcalloc (n_exprs, sizeof (struct expr *)); for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr != NULL; expr = expr->next_same_hash) index_map[expr->bitmap_index] = expr; /* Reset bitmap used to track which insns are redundant. */ pre_redundant_insns = sbitmap_alloc (max_cuid); sbitmap_zero (pre_redundant_insns); /* Delete the redundant insns first so that - we know what register to use for the new insns and for the other ones with reaching expressions - we know which insns are redundant when we go to create copies */ changed = pre_delete (); did_insert = pre_edge_insert (edge_list, index_map); /* In other places with reaching expressions, copy the expression to the specially allocated pseudo-reg that reaches the redundant expr. */ pre_insert_copies (); if (did_insert) { commit_edge_insertions (); changed = 1; } free (index_map); free (pre_redundant_insns); return changed; } /* Top level routine to perform one PRE GCSE pass. Return non-zero if a change was made. */ static int one_pre_gcse_pass (pass) int pass; { int changed = 0; gcse_subst_count = 0; gcse_create_count = 0; alloc_expr_hash_table (max_cuid); add_noreturn_fake_exit_edges (); compute_expr_hash_table (); if (gcse_file) dump_hash_table (gcse_file, "Expression", expr_hash_table, expr_hash_table_size, n_exprs); if (n_exprs > 0) { alloc_pre_mem (n_basic_blocks, n_exprs); compute_pre_data (); changed |= pre_gcse (); free_edge_list (edge_list); free_pre_mem (); } remove_fake_edges (); free_expr_hash_table (); if (gcse_file) { fprintf (gcse_file, "\nPRE GCSE of %s, pass %d: %d bytes needed, ", current_function_name, pass, bytes_used); fprintf (gcse_file, "%d substs, %d insns created\n", gcse_subst_count, gcse_create_count); } return changed; } /* If X contains any LABEL_REF's, add REG_LABEL notes for them to INSN. If notes are added to an insn which references a CODE_LABEL, the LABEL_NUSES count is incremented. We have to add REG_LABEL notes, because the following loop optimization pass requires them. */ /* ??? This is very similar to the loop.c add_label_notes function. We could probably share code here. */ /* ??? If there was a jump optimization pass after gcse and before loop, then we would not need to do this here, because jump would add the necessary REG_LABEL notes. */ static void add_label_notes (x, insn) rtx x; rtx insn; { enum rtx_code code = GET_CODE (x); int i, j; const char *fmt; if (code == LABEL_REF && !LABEL_REF_NONLOCAL_P (x)) { /* This code used to ignore labels that referred to dispatch tables to avoid flow generating (slighly) worse code. We no longer ignore such label references (see LABEL_REF handling in mark_jump_label for additional information). */ REG_NOTES (insn) = gen_rtx_EXPR_LIST (REG_LABEL, XEXP (x, 0), REG_NOTES (insn)); if (LABEL_P (XEXP (x, 0))) LABEL_NUSES (XEXP (x, 0))++; return; } for (i = GET_RTX_LENGTH (code) - 1, fmt = GET_RTX_FORMAT (code); i >= 0; i--) { if (fmt[i] == 'e') add_label_notes (XEXP (x, i), insn); else if (fmt[i] == 'E') for (j = XVECLEN (x, i) - 1; j >= 0; j--) add_label_notes (XVECEXP (x, i, j), insn); } } /* Compute transparent outgoing information for each block. An expression is transparent to an edge unless it is killed by the edge itself. This can only happen with abnormal control flow, when the edge is traversed through a call. This happens with non-local labels and exceptions. This would not be necessary if we split the edge. While this is normally impossible for abnormal critical edges, with some effort it should be possible with exception handling, since we still have control over which handler should be invoked. But due to increased EH table sizes, this may not be worthwhile. */ static void compute_transpout () { int bb; unsigned int i; struct expr *expr; sbitmap_vector_ones (transpout, n_basic_blocks); for (bb = 0; bb < n_basic_blocks; ++bb) { /* Note that flow inserted a nop a the end of basic blocks that end in call instructions for reasons other than abnormal control flow. */ if (GET_CODE (BLOCK_END (bb)) != CALL_INSN) continue; for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr ; expr = expr->next_same_hash) if (GET_CODE (expr->expr) == MEM) { if (GET_CODE (XEXP (expr->expr, 0)) == SYMBOL_REF && CONSTANT_POOL_ADDRESS_P (XEXP (expr->expr, 0))) continue; /* ??? Optimally, we would use interprocedural alias analysis to determine if this mem is actually killed by this call. */ RESET_BIT (transpout[bb], expr->bitmap_index); } } } /* Removal of useless null pointer checks */ /* Called via note_stores. X is set by SETTER. If X is a register we must invalidate nonnull_local and set nonnull_killed. DATA is really a `null_pointer_info *'. We ignore hard registers. */ static void invalidate_nonnull_info (x, setter, data) rtx x; rtx setter ATTRIBUTE_UNUSED; void *data; { unsigned int regno; struct null_pointer_info *npi = (struct null_pointer_info *) data; while (GET_CODE (x) == SUBREG) x = SUBREG_REG (x); /* Ignore anything that is not a register or is a hard register. */ if (GET_CODE (x) != REG || REGNO (x) < npi->min_reg || REGNO (x) >= npi->max_reg) return; regno = REGNO (x) - npi->min_reg; RESET_BIT (npi->nonnull_local[npi->current_block], regno); SET_BIT (npi->nonnull_killed[npi->current_block], regno); } /* Do null-pointer check elimination for the registers indicated in NPI. NONNULL_AVIN and NONNULL_AVOUT are pre-allocated sbitmaps; they are not our responsibility to free. */ static void delete_null_pointer_checks_1 (block_reg, nonnull_avin, nonnull_avout, npi) unsigned int *block_reg; sbitmap *nonnull_avin; sbitmap *nonnull_avout; struct null_pointer_info *npi; { int bb; int current_block; sbitmap *nonnull_local = npi->nonnull_local; sbitmap *nonnull_killed = npi->nonnull_killed; /* Compute local properties, nonnull and killed. A register will have the nonnull property if at the end of the current block its value is known to be nonnull. The killed property indicates that somewhere in the block any information we had about the register is killed. Note that a register can have both properties in a single block. That indicates that it's killed, then later in the block a new value is computed. */ sbitmap_vector_zero (nonnull_local, n_basic_blocks); sbitmap_vector_zero (nonnull_killed, n_basic_blocks); for (current_block = 0; current_block < n_basic_blocks; current_block++) { rtx insn, stop_insn; /* Set the current block for invalidate_nonnull_info. */ npi->current_block = current_block; /* Scan each insn in the basic block looking for memory references and register sets. */ stop_insn = NEXT_INSN (BLOCK_END (current_block)); for (insn = BLOCK_HEAD (current_block); insn != stop_insn; insn = NEXT_INSN (insn)) { rtx set; rtx reg; /* Ignore anything that is not a normal insn. */ if (! INSN_P (insn)) continue; /* Basically ignore anything that is not a simple SET. We do have to make sure to invalidate nonnull_local and set nonnull_killed for such insns though. */ set = single_set (insn); if (!set) { note_stores (PATTERN (insn), invalidate_nonnull_info, npi); continue; } /* See if we've got a useable memory load. We handle it first in case it uses its address register as a dest (which kills the nonnull property). */ if (GET_CODE (SET_SRC (set)) == MEM && GET_CODE ((reg = XEXP (SET_SRC (set), 0))) == REG && REGNO (reg) >= npi->min_reg && REGNO (reg) < npi->max_reg) SET_BIT (nonnull_local[current_block], REGNO (reg) - npi->min_reg); /* Now invalidate stuff clobbered by this insn. */ note_stores (PATTERN (insn), invalidate_nonnull_info, npi); /* And handle stores, we do these last since any sets in INSN can not kill the nonnull property if it is derived from a MEM appearing in a SET_DEST. */ if (GET_CODE (SET_DEST (set)) == MEM && GET_CODE ((reg = XEXP (SET_DEST (set), 0))) == REG && REGNO (reg) >= npi->min_reg && REGNO (reg) < npi->max_reg) SET_BIT (nonnull_local[current_block], REGNO (reg) - npi->min_reg); } } /* Now compute global properties based on the local properties. This is a classic global availablity algorithm. */ compute_available (nonnull_local, nonnull_killed, nonnull_avout, nonnull_avin); /* Now look at each bb and see if it ends with a compare of a value against zero. */ for (bb = 0; bb < n_basic_blocks; bb++) { rtx last_insn = BLOCK_END (bb); rtx condition, earliest; int compare_and_branch; /* Since MIN_REG is always at least FIRST_PSEUDO_REGISTER, and since BLOCK_REG[BB] is zero if this block did not end with a comparison against zero, this condition works. */ if (block_reg[bb] < npi->min_reg || block_reg[bb] >= npi->max_reg) continue; /* LAST_INSN is a conditional jump. Get its condition. */ condition = get_condition (last_insn, &earliest); /* If we can't determine the condition then skip. */ if (! condition) continue; /* Is the register known to have a nonzero value? */ if (!TEST_BIT (nonnull_avout[bb], block_reg[bb] - npi->min_reg)) continue; /* Try to compute whether the compare/branch at the loop end is one or two instructions. */ if (earliest == last_insn) compare_and_branch = 1; else if (earliest == prev_nonnote_insn (last_insn)) compare_and_branch = 2; else continue; /* We know the register in this comparison is nonnull at exit from this block. We can optimize this comparison. */ if (GET_CODE (condition) == NE) { rtx new_jump; new_jump = emit_jump_insn_before (gen_jump (JUMP_LABEL (last_insn)), last_insn); JUMP_LABEL (new_jump) = JUMP_LABEL (last_insn); LABEL_NUSES (JUMP_LABEL (new_jump))++; emit_barrier_after (new_jump); } delete_insn (last_insn); if (compare_and_branch == 2) delete_insn (earliest); /* Don't check this block again. (Note that BLOCK_END is invalid here; we deleted the last instruction in the block.) */ block_reg[bb] = 0; } } /* Find EQ/NE comparisons against zero which can be (indirectly) evaluated at compile time. This is conceptually similar to global constant/copy propagation and classic global CSE (it even uses the same dataflow equations as cprop). If a register is used as memory address with the form (mem (reg)), then we know that REG can not be zero at that point in the program. Any instruction which sets REG "kills" this property. So, if every path leading to a conditional branch has an available memory reference of that form, then we know the register can not have the value zero at the conditional branch. So we merely need to compute the local properies and propagate that data around the cfg, then optimize where possible. We run this pass two times. Once before CSE, then again after CSE. This has proven to be the most profitable approach. It is rare for new optimization opportunities of this nature to appear after the first CSE pass. This could probably be integrated with global cprop with a little work. */ void delete_null_pointer_checks (f) rtx f ATTRIBUTE_UNUSED; { sbitmap *nonnull_avin, *nonnull_avout; unsigned int *block_reg; int bb; int reg; int regs_per_pass; int max_reg; struct null_pointer_info npi; /* If we have only a single block, then there's nothing to do. */ if (n_basic_blocks <= 1) return; /* Trying to perform global optimizations on flow graphs which have a high connectivity will take a long time and is unlikely to be particularly useful. In normal circumstances a cfg should have about twice has many edges as blocks. But we do not want to punish small functions which have a couple switch statements. So we require a relatively large number of basic blocks and the ratio of edges to blocks to be high. */ if (n_basic_blocks > 1000 && n_edges / n_basic_blocks >= 20) return; /* We need four bitmaps, each with a bit for each register in each basic block. */ max_reg = max_reg_num (); regs_per_pass = get_bitmap_width (4, n_basic_blocks, max_reg); /* Allocate bitmaps to hold local and global properties. */ npi.nonnull_local = sbitmap_vector_alloc (n_basic_blocks, regs_per_pass); npi.nonnull_killed = sbitmap_vector_alloc (n_basic_blocks, regs_per_pass); nonnull_avin = sbitmap_vector_alloc (n_basic_blocks, regs_per_pass); nonnull_avout = sbitmap_vector_alloc (n_basic_blocks, regs_per_pass); /* Go through the basic blocks, seeing whether or not each block ends with a conditional branch whose condition is a comparison against zero. Record the register compared in BLOCK_REG. */ block_reg = (unsigned int *) xcalloc (n_basic_blocks, sizeof (int)); for (bb = 0; bb < n_basic_blocks; bb++) { rtx last_insn = BLOCK_END (bb); rtx condition, earliest, reg; /* We only want conditional branches. */ if (GET_CODE (last_insn) != JUMP_INSN || !any_condjump_p (last_insn) || !onlyjump_p (last_insn)) continue; /* LAST_INSN is a conditional jump. Get its condition. */ condition = get_condition (last_insn, &earliest); /* If we were unable to get the condition, or it is not a equality comparison against zero then there's nothing we can do. */ if (!condition || (GET_CODE (condition) != NE && GET_CODE (condition) != EQ) || GET_CODE (XEXP (condition, 1)) != CONST_INT || (XEXP (condition, 1) != CONST0_RTX (GET_MODE (XEXP (condition, 0))))) continue; /* We must be checking a register against zero. */ reg = XEXP (condition, 0); if (GET_CODE (reg) != REG) continue; block_reg[bb] = REGNO (reg); } /* Go through the algorithm for each block of registers. */ for (reg = FIRST_PSEUDO_REGISTER; reg < max_reg; reg += regs_per_pass) { npi.min_reg = reg; npi.max_reg = MIN (reg + regs_per_pass, max_reg); delete_null_pointer_checks_1 (block_reg, nonnull_avin, nonnull_avout, &npi); } /* Free the table of registers compared at the end of every block. */ free (block_reg); /* Free bitmaps. */ free (npi.nonnull_local); free (npi.nonnull_killed); free (nonnull_avin); free (nonnull_avout); } /* Code Hoisting variables and subroutines. */ /* Very busy expressions. */ static sbitmap *hoist_vbein; static sbitmap *hoist_vbeout; /* Hoistable expressions. */ static sbitmap *hoist_exprs; /* Dominator bitmaps. */ static sbitmap *dominators; /* ??? We could compute post dominators and run this algorithm in reverse to to perform tail merging, doing so would probably be more effective than the tail merging code in jump.c. It's unclear if tail merging could be run in parallel with code hoisting. It would be nice. */ /* Allocate vars used for code hoisting analysis. */ static void alloc_code_hoist_mem (n_blocks, n_exprs) int n_blocks, n_exprs; { antloc = sbitmap_vector_alloc (n_blocks, n_exprs); transp = sbitmap_vector_alloc (n_blocks, n_exprs); comp = sbitmap_vector_alloc (n_blocks, n_exprs); hoist_vbein = sbitmap_vector_alloc (n_blocks, n_exprs); hoist_vbeout = sbitmap_vector_alloc (n_blocks, n_exprs); hoist_exprs = sbitmap_vector_alloc (n_blocks, n_exprs); transpout = sbitmap_vector_alloc (n_blocks, n_exprs); dominators = sbitmap_vector_alloc (n_blocks, n_blocks); } /* Free vars used for code hoisting analysis. */ static void free_code_hoist_mem () { free (antloc); free (transp); free (comp); free (hoist_vbein); free (hoist_vbeout); free (hoist_exprs); free (transpout); free (dominators); } /* Compute the very busy expressions at entry/exit from each block. An expression is very busy if all paths from a given point compute the expression. */ static void compute_code_hoist_vbeinout () { int bb, changed, passes; sbitmap_vector_zero (hoist_vbeout, n_basic_blocks); sbitmap_vector_zero (hoist_vbein, n_basic_blocks); passes = 0; changed = 1; while (changed) { changed = 0; /* We scan the blocks in the reverse order to speed up the convergence. */ for (bb = n_basic_blocks - 1; bb >= 0; bb--) { changed |= sbitmap_a_or_b_and_c (hoist_vbein[bb], antloc[bb], hoist_vbeout[bb], transp[bb]); if (bb != n_basic_blocks - 1) sbitmap_intersection_of_succs (hoist_vbeout[bb], hoist_vbein, bb); } passes++; } if (gcse_file) fprintf (gcse_file, "hoisting vbeinout computation: %d passes\n", passes); } /* Top level routine to do the dataflow analysis needed by code hoisting. */ static void compute_code_hoist_data () { compute_local_properties (transp, comp, antloc, 0); compute_transpout (); compute_code_hoist_vbeinout (); calculate_dominance_info (NULL, dominators, CDI_DOMINATORS); if (gcse_file) fprintf (gcse_file, "\n"); } /* Determine if the expression identified by EXPR_INDEX would reach BB unimpared if it was placed at the end of EXPR_BB. It's unclear exactly what Muchnick meant by "unimpared". It seems to me that the expression must either be computed or transparent in *every* block in the path(s) from EXPR_BB to BB. Any other definition would allow the expression to be hoisted out of loops, even if the expression wasn't a loop invariant. Contrast this to reachability for PRE where an expression is considered reachable if *any* path reaches instead of *all* paths. */ static int hoist_expr_reaches_here_p (expr_bb, expr_index, bb, visited) int expr_bb; int expr_index; int bb; char *visited; { edge pred; int visited_allocated_locally = 0; if (visited == NULL) { visited_allocated_locally = 1; visited = xcalloc (n_basic_blocks, 1); } for (pred = BASIC_BLOCK (bb)->pred; pred != NULL; pred = pred->pred_next) { int pred_bb = pred->src->index; if (pred->src == ENTRY_BLOCK_PTR) break; else if (visited[pred_bb]) continue; /* Does this predecessor generate this expression? */ else if (TEST_BIT (comp[pred_bb], expr_index)) break; else if (! TEST_BIT (transp[pred_bb], expr_index)) break; /* Not killed. */ else { visited[pred_bb] = 1; if (! hoist_expr_reaches_here_p (expr_bb, expr_index, pred_bb, visited)) break; } } if (visited_allocated_locally) free (visited); return (pred == NULL); } /* Actually perform code hoisting. */ static void hoist_code () { int bb, dominated; unsigned int i; struct expr **index_map; struct expr *expr; sbitmap_vector_zero (hoist_exprs, n_basic_blocks); /* Compute a mapping from expression number (`bitmap_index') to hash table entry. */ index_map = (struct expr **) xcalloc (n_exprs, sizeof (struct expr *)); for (i = 0; i < expr_hash_table_size; i++) for (expr = expr_hash_table[i]; expr != NULL; expr = expr->next_same_hash) index_map[expr->bitmap_index] = expr; /* Walk over each basic block looking for potentially hoistable expressions, nothing gets hoisted from the entry block. */ for (bb = 0; bb < n_basic_blocks; bb++) { int found = 0; int insn_inserted_p; /* Examine each expression that is very busy at the exit of this block. These are the potentially hoistable expressions. */ for (i = 0; i < hoist_vbeout[bb]->n_bits; i++) { int hoistable = 0; if (TEST_BIT (hoist_vbeout[bb], i) && TEST_BIT (transpout[bb], i)) { /* We've found a potentially hoistable expression, now we look at every block BB dominates to see if it computes the expression. */ for (dominated = 0; dominated < n_basic_blocks; dominated++) { /* Ignore self dominance. */ if (bb == dominated || ! TEST_BIT (dominators[dominated], bb)) continue; /* We've found a dominated block, now see if it computes the busy expression and whether or not moving that expression to the "beginning" of that block is safe. */ if (!TEST_BIT (antloc[dominated], i)) continue; /* Note if the expression would reach the dominated block unimpared if it was placed at the end of BB. Keep track of how many times this expression is hoistable from a dominated block into BB. */ if (hoist_expr_reaches_here_p (bb, i, dominated, NULL)) hoistable++; } /* If we found more than one hoistable occurence of this expression, then note it in the bitmap of expressions to hoist. It makes no sense to hoist things which are computed in only one BB, and doing so tends to pessimize register allocation. One could increase this value to try harder to avoid any possible code expansion due to register allocation issues; however experiments have shown that the vast majority of hoistable expressions are only movable from two successors, so raising this threshhold is likely to nullify any benefit we get from code hoisting. */ if (hoistable > 1) { SET_BIT (hoist_exprs[bb], i); found = 1; } } } /* If we found nothing to hoist, then quit now. */ if (! found) continue; /* Loop over all the hoistable expressions. */ for (i = 0; i < hoist_exprs[bb]->n_bits; i++) { /* We want to insert the expression into BB only once, so note when we've inserted it. */ insn_inserted_p = 0; /* These tests should be the same as the tests above. */ if (TEST_BIT (hoist_vbeout[bb], i)) { /* We've found a potentially hoistable expression, now we look at every block BB dominates to see if it computes the expression. */ for (dominated = 0; dominated < n_basic_blocks; dominated++) { /* Ignore self dominance. */ if (bb == dominated || ! TEST_BIT (dominators[dominated], bb)) continue; /* We've found a dominated block, now see if it computes the busy expression and whether or not moving that expression to the "beginning" of that block is safe. */ if (!TEST_BIT (antloc[dominated], i)) continue; /* The expression is computed in the dominated block and it would be safe to compute it at the start of the dominated block. Now we have to determine if the expresion would reach the dominated block if it was placed at the end of BB. */ if (hoist_expr_reaches_here_p (bb, i, dominated, NULL)) { struct expr *expr = index_map[i]; struct occr *occr = expr->antic_occr; rtx insn; rtx set; /* Find the right occurence of this expression. */ while (BLOCK_NUM (occr->insn) != dominated && occr) occr = occr->next; /* Should never happen. */ if (!occr) abort (); insn = occr->insn; set = single_set (insn); if (! set) abort (); /* Create a pseudo-reg to store the result of reaching expressions into. Get the mode for the new pseudo from the mode of the original destination pseudo. */ if (expr->reaching_reg == NULL) expr->reaching_reg = gen_reg_rtx (GET_MODE (SET_DEST (set))); /* In theory this should never fail since we're creating a reg->reg copy. However, on the x86 some of the movXX patterns actually contain clobbers of scratch regs. This may cause the insn created by validate_change to not match any pattern and thus cause validate_change to fail. */ if (validate_change (insn, &SET_SRC (set), expr->reaching_reg, 0)) { occr->deleted_p = 1; if (!insn_inserted_p) { insert_insn_end_bb (index_map[i], bb, 0); insn_inserted_p = 1; } } } } } } } free (index_map); } /* Top level routine to perform one code hoisting (aka unification) pass Return non-zero if a change was made. */ static int one_code_hoisting_pass () { int changed = 0; alloc_expr_hash_table (max_cuid); compute_expr_hash_table (); if (gcse_file) dump_hash_table (gcse_file, "Code Hosting Expressions", expr_hash_table, expr_hash_table_size, n_exprs); if (n_exprs > 0) { alloc_code_hoist_mem (n_basic_blocks, n_exprs); compute_code_hoist_data (); hoist_code (); free_code_hoist_mem (); } free_expr_hash_table (); return changed; }