/* Interprocedural constant propagation Copyright (C) 2005-2023 Free Software Foundation, Inc. Contributed by Razya Ladelsky and Martin Jambor This file is part of GCC. GCC 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 3, or (at your option) any later version. GCC 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 GCC; see the file COPYING3. If not see . */ /* Interprocedural constant propagation (IPA-CP). The goal of this transformation is to 1) discover functions which are always invoked with some arguments with the same known constant values and modify the functions so that the subsequent optimizations can take advantage of the knowledge, and 2) partial specialization - create specialized versions of functions transformed in this way if some parameters are known constants only in certain contexts but the estimated tradeoff between speedup and cost size is deemed good. The algorithm also propagates types and attempts to perform type based devirtualization. Types are propagated much like constants. The algorithm basically consists of three stages. In the first, functions are analyzed one at a time and jump functions are constructed for all known call-sites. In the second phase, the pass propagates information from the jump functions across the call to reveal what values are available at what call sites, performs estimations of effects of known values on functions and their callees, and finally decides what specialized extra versions should be created. In the third, the special versions materialize and appropriate calls are redirected. The algorithm used is to a certain extent based on "Interprocedural Constant Propagation", by David Callahan, Keith D Cooper, Ken Kennedy, Linda Torczon, Comp86, pg 152-161 and "A Methodology for Procedure Cloning" by Keith D Cooper, Mary W. Hall, and Ken Kennedy. First stage - intraprocedural analysis ======================================= This phase computes jump_function and modification flags. A jump function for a call-site represents the values passed as an actual arguments of a given call-site. In principle, there are three types of values: Pass through - the caller's formal parameter is passed as an actual argument, plus an operation on it can be performed. Constant - a constant is passed as an actual argument. Unknown - neither of the above. All jump function types are described in detail in ipa-prop.h, together with the data structures that represent them and methods of accessing them. ipcp_generate_summary() is the main function of the first stage. Second stage - interprocedural analysis ======================================== This stage is itself divided into two phases. In the first, we propagate known values over the call graph, in the second, we make cloning decisions. It uses a different algorithm than the original Callahan's paper. First, we traverse the functions topologically from callers to callees and, for each strongly connected component (SCC), we propagate constants according to previously computed jump functions. We also record what known values depend on other known values and estimate local effects. Finally, we propagate cumulative information about these effects from dependent values to those on which they depend. Second, we again traverse the call graph in the same topological order and make clones for functions which we know are called with the same values in all contexts and decide about extra specialized clones of functions just for some contexts - these decisions are based on both local estimates and cumulative estimates propagated from callees. ipcp_propagate_stage() and ipcp_decision_stage() together constitute the third stage. Third phase - materialization of clones, call statement updates. ============================================ This stage is currently performed by call graph code (mainly in cgraphunit.cc and tree-inline.cc) according to instructions inserted to the call graph by the second stage. */ #define INCLUDE_ALGORITHM #include "config.h" #include "system.h" #include "coretypes.h" #include "backend.h" #include "tree.h" #include "gimple-expr.h" #include "gimple.h" #include "predict.h" #include "alloc-pool.h" #include "tree-pass.h" #include "cgraph.h" #include "diagnostic.h" #include "fold-const.h" #include "gimple-iterator.h" #include "gimple-fold.h" #include "symbol-summary.h" #include "tree-vrp.h" #include "ipa-prop.h" #include "tree-pretty-print.h" #include "tree-inline.h" #include "ipa-fnsummary.h" #include "ipa-utils.h" #include "tree-ssa-ccp.h" #include "stringpool.h" #include "attribs.h" #include "dbgcnt.h" #include "symtab-clones.h" #include "gimple-range.h" template class ipcp_value; /* Describes a particular source for an IPA-CP value. */ template struct ipcp_value_source { public: /* Aggregate offset of the source, negative if the source is scalar value of the argument itself. */ HOST_WIDE_INT offset; /* The incoming edge that brought the value. */ cgraph_edge *cs; /* If the jump function that resulted into his value was a pass-through or an ancestor, this is the ipcp_value of the caller from which the described value has been derived. Otherwise it is NULL. */ ipcp_value *val; /* Next pointer in a linked list of sources of a value. */ ipcp_value_source *next; /* If the jump function that resulted into his value was a pass-through or an ancestor, this is the index of the parameter of the caller the jump function references. */ int index; }; /* Common ancestor for all ipcp_value instantiations. */ class ipcp_value_base { public: /* Time benefit and that specializing the function for this value would bring about in this function alone. */ sreal local_time_benefit; /* Time benefit that specializing the function for this value can bring about in it's callees. */ sreal prop_time_benefit; /* Size cost that specializing the function for this value would bring about in this function alone. */ int local_size_cost; /* Size cost that specializing the function for this value can bring about in it's callees. */ int prop_size_cost; ipcp_value_base () : local_time_benefit (0), prop_time_benefit (0), local_size_cost (0), prop_size_cost (0) {} }; /* Describes one particular value stored in struct ipcp_lattice. */ template class ipcp_value : public ipcp_value_base { public: /* The actual value for the given parameter. */ valtype value; /* The list of sources from which this value originates. */ ipcp_value_source *sources = nullptr; /* Next pointers in a linked list of all values in a lattice. */ ipcp_value *next = nullptr; /* Next pointers in a linked list of values in a strongly connected component of values. */ ipcp_value *scc_next = nullptr; /* Next pointers in a linked list of SCCs of values sorted topologically according their sources. */ ipcp_value *topo_next = nullptr; /* A specialized node created for this value, NULL if none has been (so far) created. */ cgraph_node *spec_node = nullptr; /* Depth first search number and low link for topological sorting of values. */ int dfs = 0; int low_link = 0; /* SCC number to identify values which recursively feed into each other. Values in the same SCC have the same SCC number. */ int scc_no = 0; /* Non zero if the value is generated from another value in the same lattice for a self-recursive call, the actual number is how many times the operation has been performed. In the unlikely event of the value being present in two chains fo self-recursive value generation chains, it is the maximum. */ unsigned self_recursion_generated_level = 0; /* True if this value is currently on the topo-sort stack. */ bool on_stack = false; void add_source (cgraph_edge *cs, ipcp_value *src_val, int src_idx, HOST_WIDE_INT offset); /* Return true if both THIS value and O feed into each other. */ bool same_scc (const ipcp_value *o) { return o->scc_no == scc_no; } /* Return true, if a this value has been generated for a self-recursive call as a result of an arithmetic pass-through jump-function acting on a value in the same lattice function. */ bool self_recursion_generated_p () { return self_recursion_generated_level > 0; } }; /* Lattice describing potential values of a formal parameter of a function, or a part of an aggregate. TOP is represented by a lattice with zero values and with contains_variable and bottom flags cleared. BOTTOM is represented by a lattice with the bottom flag set. In that case, values and contains_variable flag should be disregarded. */ template struct ipcp_lattice { public: /* The list of known values and types in this lattice. Note that values are not deallocated if a lattice is set to bottom because there may be value sources referencing them. */ ipcp_value *values; /* Number of known values and types in this lattice. */ int values_count; /* The lattice contains a variable component (in addition to values). */ bool contains_variable; /* The value of the lattice is bottom (i.e. variable and unusable for any propagation). */ bool bottom; inline bool is_single_const (); inline bool set_to_bottom (); inline bool set_contains_variable (); bool add_value (valtype newval, cgraph_edge *cs, ipcp_value *src_val = NULL, int src_idx = 0, HOST_WIDE_INT offset = -1, ipcp_value **val_p = NULL, unsigned same_lat_gen_level = 0); void print (FILE * f, bool dump_sources, bool dump_benefits); }; /* Lattice of tree values with an offset to describe a part of an aggregate. */ struct ipcp_agg_lattice : public ipcp_lattice { public: /* Offset that is being described by this lattice. */ HOST_WIDE_INT offset; /* Size so that we don't have to re-compute it every time we traverse the list. Must correspond to TYPE_SIZE of all lat values. */ HOST_WIDE_INT size; /* Next element of the linked list. */ struct ipcp_agg_lattice *next; }; /* Lattice of known bits, only capable of holding one value. Bitwise constant propagation propagates which bits of a value are constant. For eg: int f(int x) { return some_op (x); } int f1(int y) { if (cond) return f (y & 0xff); else return f (y & 0xf); } In the above case, the param 'x' will always have all the bits (except the bits in lsb) set to 0. Hence the mask of 'x' would be 0xff. The mask reflects that the bits in lsb are unknown. The actual propagated value is given by m_value & ~m_mask. */ class ipcp_bits_lattice { public: bool bottom_p () const { return m_lattice_val == IPA_BITS_VARYING; } bool top_p () const { return m_lattice_val == IPA_BITS_UNDEFINED; } bool constant_p () const { return m_lattice_val == IPA_BITS_CONSTANT; } bool set_to_bottom (); bool set_to_constant (widest_int, widest_int); bool known_nonzero_p () const; widest_int get_value () const { return m_value; } widest_int get_mask () const { return m_mask; } bool meet_with (ipcp_bits_lattice& other, unsigned, signop, enum tree_code, tree, bool); bool meet_with (widest_int, widest_int, unsigned); void print (FILE *); private: enum { IPA_BITS_UNDEFINED, IPA_BITS_CONSTANT, IPA_BITS_VARYING } m_lattice_val; /* Similar to ccp_lattice_t, mask represents which bits of value are constant. If a bit in mask is set to 0, then the corresponding bit in value is known to be constant. */ widest_int m_value, m_mask; bool meet_with_1 (widest_int, widest_int, unsigned, bool); void get_value_and_mask (tree, widest_int *, widest_int *); }; /* Lattice of value ranges. */ class ipcp_vr_lattice { public: Value_Range m_vr; inline bool bottom_p () const; inline bool top_p () const; inline bool set_to_bottom (); bool meet_with (const vrange &p_vr); bool meet_with (const ipcp_vr_lattice &other); void init (tree type); void print (FILE * f); private: bool meet_with_1 (const vrange &other_vr); }; inline void ipcp_vr_lattice::init (tree type) { if (type) m_vr.set_type (type); // Otherwise m_vr will default to unsupported_range. } /* Structure containing lattices for a parameter itself and for pieces of aggregates that are passed in the parameter or by a reference in a parameter plus some other useful flags. */ class ipcp_param_lattices { public: /* Lattice describing the value of the parameter itself. */ ipcp_lattice itself; /* Lattice describing the polymorphic contexts of a parameter. */ ipcp_lattice ctxlat; /* Lattices describing aggregate parts. */ ipcp_agg_lattice *aggs; /* Lattice describing known bits. */ ipcp_bits_lattice bits_lattice; /* Lattice describing value range. */ ipcp_vr_lattice m_value_range; /* Number of aggregate lattices */ int aggs_count; /* True if aggregate data were passed by reference (as opposed to by value). */ bool aggs_by_ref; /* All aggregate lattices contain a variable component (in addition to values). */ bool aggs_contain_variable; /* The value of all aggregate lattices is bottom (i.e. variable and unusable for any propagation). */ bool aggs_bottom; /* There is a virtual call based on this parameter. */ bool virt_call; }; /* Allocation pools for values and their sources in ipa-cp. */ object_allocator > ipcp_cst_values_pool ("IPA-CP constant values"); object_allocator > ipcp_poly_ctx_values_pool ("IPA-CP polymorphic contexts"); object_allocator > ipcp_sources_pool ("IPA-CP value sources"); object_allocator ipcp_agg_lattice_pool ("IPA_CP aggregate lattices"); /* Base count to use in heuristics when using profile feedback. */ static profile_count base_count; /* Original overall size of the program. */ static long overall_size, orig_overall_size; /* Node name to unique clone suffix number map. */ static hash_map *clone_num_suffixes; /* Return the param lattices structure corresponding to the Ith formal parameter of the function described by INFO. */ static inline class ipcp_param_lattices * ipa_get_parm_lattices (class ipa_node_params *info, int i) { gcc_assert (i >= 0 && i < ipa_get_param_count (info)); gcc_checking_assert (!info->ipcp_orig_node); gcc_checking_assert (info->lattices); return &(info->lattices[i]); } /* Return the lattice corresponding to the scalar value of the Ith formal parameter of the function described by INFO. */ static inline ipcp_lattice * ipa_get_scalar_lat (class ipa_node_params *info, int i) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); return &plats->itself; } /* Return the lattice corresponding to the scalar value of the Ith formal parameter of the function described by INFO. */ static inline ipcp_lattice * ipa_get_poly_ctx_lat (class ipa_node_params *info, int i) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); return &plats->ctxlat; } /* Return whether LAT is a lattice with a single constant and without an undefined value. */ template inline bool ipcp_lattice::is_single_const () { if (bottom || contains_variable || values_count != 1) return false; else return true; } /* Return true iff X and Y should be considered equal values by IPA-CP. */ static bool values_equal_for_ipcp_p (tree x, tree y) { gcc_checking_assert (x != NULL_TREE && y != NULL_TREE); if (x == y) return true; if (TREE_CODE (x) == ADDR_EXPR && TREE_CODE (y) == ADDR_EXPR && TREE_CODE (TREE_OPERAND (x, 0)) == CONST_DECL && TREE_CODE (TREE_OPERAND (y, 0)) == CONST_DECL) return operand_equal_p (DECL_INITIAL (TREE_OPERAND (x, 0)), DECL_INITIAL (TREE_OPERAND (y, 0)), 0); else return operand_equal_p (x, y, 0); } /* Print V which is extracted from a value in a lattice to F. */ static void print_ipcp_constant_value (FILE * f, tree v) { if (TREE_CODE (v) == ADDR_EXPR && TREE_CODE (TREE_OPERAND (v, 0)) == CONST_DECL) { fprintf (f, "& "); print_generic_expr (f, DECL_INITIAL (TREE_OPERAND (v, 0))); } else print_generic_expr (f, v); } /* Print V which is extracted from a value in a lattice to F. */ static void print_ipcp_constant_value (FILE * f, ipa_polymorphic_call_context v) { v.dump(f, false); } /* Print a lattice LAT to F. */ template void ipcp_lattice::print (FILE * f, bool dump_sources, bool dump_benefits) { ipcp_value *val; bool prev = false; if (bottom) { fprintf (f, "BOTTOM\n"); return; } if (!values_count && !contains_variable) { fprintf (f, "TOP\n"); return; } if (contains_variable) { fprintf (f, "VARIABLE"); prev = true; if (dump_benefits) fprintf (f, "\n"); } for (val = values; val; val = val->next) { if (dump_benefits && prev) fprintf (f, " "); else if (!dump_benefits && prev) fprintf (f, ", "); else prev = true; print_ipcp_constant_value (f, val->value); if (dump_sources) { ipcp_value_source *s; if (val->self_recursion_generated_p ()) fprintf (f, " [self_gen(%i), from:", val->self_recursion_generated_level); else fprintf (f, " [scc: %i, from:", val->scc_no); for (s = val->sources; s; s = s->next) fprintf (f, " %i(%f)", s->cs->caller->order, s->cs->sreal_frequency ().to_double ()); fprintf (f, "]"); } if (dump_benefits) fprintf (f, " [loc_time: %g, loc_size: %i, " "prop_time: %g, prop_size: %i]\n", val->local_time_benefit.to_double (), val->local_size_cost, val->prop_time_benefit.to_double (), val->prop_size_cost); } if (!dump_benefits) fprintf (f, "\n"); } void ipcp_bits_lattice::print (FILE *f) { if (top_p ()) fprintf (f, " Bits unknown (TOP)\n"); else if (bottom_p ()) fprintf (f, " Bits unusable (BOTTOM)\n"); else { fprintf (f, " Bits: value = "); print_hex (get_value (), f); fprintf (f, ", mask = "); print_hex (get_mask (), f); fprintf (f, "\n"); } } /* Print value range lattice to F. */ void ipcp_vr_lattice::print (FILE * f) { m_vr.dump (f); } /* Print all ipcp_lattices of all functions to F. */ static void print_all_lattices (FILE * f, bool dump_sources, bool dump_benefits) { struct cgraph_node *node; int i, count; fprintf (f, "\nLattices:\n"); FOR_EACH_FUNCTION_WITH_GIMPLE_BODY (node) { class ipa_node_params *info; info = ipa_node_params_sum->get (node); /* Skip unoptimized functions and constprop clones since we don't make lattices for them. */ if (!info || info->ipcp_orig_node) continue; fprintf (f, " Node: %s:\n", node->dump_name ()); count = ipa_get_param_count (info); for (i = 0; i < count; i++) { struct ipcp_agg_lattice *aglat; class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); fprintf (f, " param [%d]: ", i); plats->itself.print (f, dump_sources, dump_benefits); fprintf (f, " ctxs: "); plats->ctxlat.print (f, dump_sources, dump_benefits); plats->bits_lattice.print (f); fprintf (f, " "); plats->m_value_range.print (f); fprintf (f, "\n"); if (plats->virt_call) fprintf (f, " virt_call flag set\n"); if (plats->aggs_bottom) { fprintf (f, " AGGS BOTTOM\n"); continue; } if (plats->aggs_contain_variable) fprintf (f, " AGGS VARIABLE\n"); for (aglat = plats->aggs; aglat; aglat = aglat->next) { fprintf (f, " %soffset " HOST_WIDE_INT_PRINT_DEC ": ", plats->aggs_by_ref ? "ref " : "", aglat->offset); aglat->print (f, dump_sources, dump_benefits); } } } } /* Determine whether it is at all technically possible to create clones of NODE and store this information in the ipa_node_params structure associated with NODE. */ static void determine_versionability (struct cgraph_node *node, class ipa_node_params *info) { const char *reason = NULL; /* There are a number of generic reasons functions cannot be versioned. We also cannot remove parameters if there are type attributes such as fnspec present. */ if (node->alias || node->thunk) reason = "alias or thunk"; else if (!node->versionable) reason = "not a tree_versionable_function"; else if (node->get_availability () <= AVAIL_INTERPOSABLE) reason = "insufficient body availability"; else if (!opt_for_fn (node->decl, optimize) || !opt_for_fn (node->decl, flag_ipa_cp)) reason = "non-optimized function"; else if (lookup_attribute ("omp declare simd", DECL_ATTRIBUTES (node->decl))) { /* Ideally we should clone the SIMD clones themselves and create vector copies of them, so IPA-cp and SIMD clones can happily coexist, but that may not be worth the effort. */ reason = "function has SIMD clones"; } else if (lookup_attribute ("target_clones", DECL_ATTRIBUTES (node->decl))) { /* Ideally we should clone the target clones themselves and create copies of them, so IPA-cp and target clones can happily coexist, but that may not be worth the effort. */ reason = "function target_clones attribute"; } /* Don't clone decls local to a comdat group; it breaks and for C++ decloned constructors, inlining is always better anyway. */ else if (node->comdat_local_p ()) reason = "comdat-local function"; else if (node->calls_comdat_local) { /* TODO: call is versionable if we make sure that all callers are inside of a comdat group. */ reason = "calls comdat-local function"; } /* Functions calling BUILT_IN_VA_ARG_PACK and BUILT_IN_VA_ARG_PACK_LEN work only when inlined. Cloning them may still lead to better code because ipa-cp will not give up on cloning further. If the function is external this however leads to wrong code because we may end up producing offline copy of the function. */ if (DECL_EXTERNAL (node->decl)) for (cgraph_edge *edge = node->callees; !reason && edge; edge = edge->next_callee) if (fndecl_built_in_p (edge->callee->decl, BUILT_IN_NORMAL)) { if (DECL_FUNCTION_CODE (edge->callee->decl) == BUILT_IN_VA_ARG_PACK) reason = "external function which calls va_arg_pack"; if (DECL_FUNCTION_CODE (edge->callee->decl) == BUILT_IN_VA_ARG_PACK_LEN) reason = "external function which calls va_arg_pack_len"; } if (reason && dump_file && !node->alias && !node->thunk) fprintf (dump_file, "Function %s is not versionable, reason: %s.\n", node->dump_name (), reason); info->versionable = (reason == NULL); } /* Return true if it is at all technically possible to create clones of a NODE. */ static bool ipcp_versionable_function_p (struct cgraph_node *node) { ipa_node_params *info = ipa_node_params_sum->get (node); return info && info->versionable; } /* Structure holding accumulated information about callers of a node. */ struct caller_statistics { /* If requested (see below), self-recursive call counts are summed into this field. */ profile_count rec_count_sum; /* The sum of all ipa counts of all the other (non-recursive) calls. */ profile_count count_sum; /* Sum of all frequencies for all calls. */ sreal freq_sum; /* Number of calls and hot calls respectively. */ int n_calls, n_hot_calls; /* If itself is set up, also count the number of non-self-recursive calls. */ int n_nonrec_calls; /* If non-NULL, this is the node itself and calls from it should have their counts included in rec_count_sum and not count_sum. */ cgraph_node *itself; }; /* Initialize fields of STAT to zeroes and optionally set it up so that edges from IGNORED_CALLER are not counted. */ static inline void init_caller_stats (caller_statistics *stats, cgraph_node *itself = NULL) { stats->rec_count_sum = profile_count::zero (); stats->count_sum = profile_count::zero (); stats->n_calls = 0; stats->n_hot_calls = 0; stats->n_nonrec_calls = 0; stats->freq_sum = 0; stats->itself = itself; } /* Worker callback of cgraph_for_node_and_aliases accumulating statistics of non-thunk incoming edges to NODE. */ static bool gather_caller_stats (struct cgraph_node *node, void *data) { struct caller_statistics *stats = (struct caller_statistics *) data; struct cgraph_edge *cs; for (cs = node->callers; cs; cs = cs->next_caller) if (!cs->caller->thunk) { ipa_node_params *info = ipa_node_params_sum->get (cs->caller); if (info && info->node_dead) continue; if (cs->count.ipa ().initialized_p ()) { if (stats->itself && stats->itself == cs->caller) stats->rec_count_sum += cs->count.ipa (); else stats->count_sum += cs->count.ipa (); } stats->freq_sum += cs->sreal_frequency (); stats->n_calls++; if (stats->itself && stats->itself != cs->caller) stats->n_nonrec_calls++; if (cs->maybe_hot_p ()) stats->n_hot_calls ++; } return false; } /* Return true if this NODE is viable candidate for cloning. */ static bool ipcp_cloning_candidate_p (struct cgraph_node *node) { struct caller_statistics stats; gcc_checking_assert (node->has_gimple_body_p ()); if (!opt_for_fn (node->decl, flag_ipa_cp_clone)) { if (dump_file) fprintf (dump_file, "Not considering %s for cloning; " "-fipa-cp-clone disabled.\n", node->dump_name ()); return false; } if (node->optimize_for_size_p ()) { if (dump_file) fprintf (dump_file, "Not considering %s for cloning; " "optimizing it for size.\n", node->dump_name ()); return false; } init_caller_stats (&stats); node->call_for_symbol_thunks_and_aliases (gather_caller_stats, &stats, false); if (ipa_size_summaries->get (node)->self_size < stats.n_calls) { if (dump_file) fprintf (dump_file, "Considering %s for cloning; code might shrink.\n", node->dump_name ()); return true; } /* When profile is available and function is hot, propagate into it even if calls seems cold; constant propagation can improve function's speed significantly. */ if (stats.count_sum > profile_count::zero () && node->count.ipa ().initialized_p ()) { if (stats.count_sum > node->count.ipa ().apply_scale (90, 100)) { if (dump_file) fprintf (dump_file, "Considering %s for cloning; " "usually called directly.\n", node->dump_name ()); return true; } } if (!stats.n_hot_calls) { if (dump_file) fprintf (dump_file, "Not considering %s for cloning; no hot calls.\n", node->dump_name ()); return false; } if (dump_file) fprintf (dump_file, "Considering %s for cloning.\n", node->dump_name ()); return true; } template class value_topo_info { public: /* Head of the linked list of topologically sorted values. */ ipcp_value *values_topo; /* Stack for creating SCCs, represented by a linked list too. */ ipcp_value *stack; /* Counter driving the algorithm in add_val_to_toposort. */ int dfs_counter; value_topo_info () : values_topo (NULL), stack (NULL), dfs_counter (0) {} void add_val (ipcp_value *cur_val); void propagate_effects (); }; /* Arrays representing a topological ordering of call graph nodes and a stack of nodes used during constant propagation and also data required to perform topological sort of values and propagation of benefits in the determined order. */ class ipa_topo_info { public: /* Array with obtained topological order of cgraph nodes. */ struct cgraph_node **order; /* Stack of cgraph nodes used during propagation within SCC until all values in the SCC stabilize. */ struct cgraph_node **stack; int nnodes, stack_top; value_topo_info constants; value_topo_info contexts; ipa_topo_info () : order(NULL), stack(NULL), nnodes(0), stack_top(0), constants () {} }; /* Skip edges from and to nodes without ipa_cp enabled. Ignore not available symbols. */ static bool ignore_edge_p (cgraph_edge *e) { enum availability avail; cgraph_node *ultimate_target = e->callee->function_or_virtual_thunk_symbol (&avail, e->caller); return (avail <= AVAIL_INTERPOSABLE || !opt_for_fn (ultimate_target->decl, optimize) || !opt_for_fn (ultimate_target->decl, flag_ipa_cp)); } /* Allocate the arrays in TOPO and topologically sort the nodes into order. */ static void build_toporder_info (class ipa_topo_info *topo) { topo->order = XCNEWVEC (struct cgraph_node *, symtab->cgraph_count); topo->stack = XCNEWVEC (struct cgraph_node *, symtab->cgraph_count); gcc_checking_assert (topo->stack_top == 0); topo->nnodes = ipa_reduced_postorder (topo->order, true, ignore_edge_p); } /* Free information about strongly connected components and the arrays in TOPO. */ static void free_toporder_info (class ipa_topo_info *topo) { ipa_free_postorder_info (); free (topo->order); free (topo->stack); } /* Add NODE to the stack in TOPO, unless it is already there. */ static inline void push_node_to_stack (class ipa_topo_info *topo, struct cgraph_node *node) { ipa_node_params *info = ipa_node_params_sum->get (node); if (info->node_enqueued) return; info->node_enqueued = 1; topo->stack[topo->stack_top++] = node; } /* Pop a node from the stack in TOPO and return it or return NULL if the stack is empty. */ static struct cgraph_node * pop_node_from_stack (class ipa_topo_info *topo) { if (topo->stack_top) { struct cgraph_node *node; topo->stack_top--; node = topo->stack[topo->stack_top]; ipa_node_params_sum->get (node)->node_enqueued = 0; return node; } else return NULL; } /* Set lattice LAT to bottom and return true if it previously was not set as such. */ template inline bool ipcp_lattice::set_to_bottom () { bool ret = !bottom; bottom = true; return ret; } /* Mark lattice as containing an unknown value and return true if it previously was not marked as such. */ template inline bool ipcp_lattice::set_contains_variable () { bool ret = !contains_variable; contains_variable = true; return ret; } /* Set all aggregate lattices in PLATS to bottom and return true if they were not previously set as such. */ static inline bool set_agg_lats_to_bottom (class ipcp_param_lattices *plats) { bool ret = !plats->aggs_bottom; plats->aggs_bottom = true; return ret; } /* Mark all aggregate lattices in PLATS as containing an unknown value and return true if they were not previously marked as such. */ static inline bool set_agg_lats_contain_variable (class ipcp_param_lattices *plats) { bool ret = !plats->aggs_contain_variable; plats->aggs_contain_variable = true; return ret; } bool ipcp_vr_lattice::meet_with (const ipcp_vr_lattice &other) { return meet_with_1 (other.m_vr); } /* Meet the current value of the lattice with the range described by P_VR. */ bool ipcp_vr_lattice::meet_with (const vrange &p_vr) { return meet_with_1 (p_vr); } /* Meet the current value of the lattice with the range described by OTHER_VR. Return TRUE if anything changed. */ bool ipcp_vr_lattice::meet_with_1 (const vrange &other_vr) { if (bottom_p ()) return false; if (other_vr.varying_p ()) return set_to_bottom (); bool res; if (flag_checking) { Value_Range save (m_vr); res = m_vr.union_ (other_vr); gcc_assert (res == (m_vr != save)); } else res = m_vr.union_ (other_vr); return res; } /* Return true if value range information in the lattice is yet unknown. */ bool ipcp_vr_lattice::top_p () const { return m_vr.undefined_p (); } /* Return true if value range information in the lattice is known to be unusable. */ bool ipcp_vr_lattice::bottom_p () const { return m_vr.varying_p (); } /* Set value range information in the lattice to bottom. Return true if it previously was in a different state. */ bool ipcp_vr_lattice::set_to_bottom () { if (m_vr.varying_p ()) return false; /* Setting an unsupported type here forces the temporary to default to unsupported_range, which can handle VARYING/DEFINED ranges, but nothing else (union, intersect, etc). This allows us to set bottoms on any ranges, and is safe as all users of the lattice check for bottom first. */ m_vr.set_type (void_type_node); m_vr.set_varying (void_type_node); return true; } /* Set lattice value to bottom, if it already isn't the case. */ bool ipcp_bits_lattice::set_to_bottom () { if (bottom_p ()) return false; m_lattice_val = IPA_BITS_VARYING; m_value = 0; m_mask = -1; return true; } /* Set to constant if it isn't already. Only meant to be called when switching state from TOP. */ bool ipcp_bits_lattice::set_to_constant (widest_int value, widest_int mask) { gcc_assert (top_p ()); m_lattice_val = IPA_BITS_CONSTANT; m_value = wi::bit_and (wi::bit_not (mask), value); m_mask = mask; return true; } /* Return true if any of the known bits are non-zero. */ bool ipcp_bits_lattice::known_nonzero_p () const { if (!constant_p ()) return false; return wi::ne_p (wi::bit_and (wi::bit_not (m_mask), m_value), 0); } /* Convert operand to value, mask form. */ void ipcp_bits_lattice::get_value_and_mask (tree operand, widest_int *valuep, widest_int *maskp) { wide_int get_nonzero_bits (const_tree); if (TREE_CODE (operand) == INTEGER_CST) { *valuep = wi::to_widest (operand); *maskp = 0; } else { *valuep = 0; *maskp = -1; } } /* Meet operation, similar to ccp_lattice_meet, we xor values if this->value, value have different values at same bit positions, we want to drop that bit to varying. Return true if mask is changed. This function assumes that the lattice value is in CONSTANT state. If DROP_ALL_ONES, mask out any known bits with value one afterwards. */ bool ipcp_bits_lattice::meet_with_1 (widest_int value, widest_int mask, unsigned precision, bool drop_all_ones) { gcc_assert (constant_p ()); widest_int old_mask = m_mask; m_mask = (m_mask | mask) | (m_value ^ value); if (drop_all_ones) m_mask |= m_value; m_value &= ~m_mask; if (wi::sext (m_mask, precision) == -1) return set_to_bottom (); return m_mask != old_mask; } /* Meet the bits lattice with operand described by val.index) return false; if (elt.unit_offset < val.unit_offset) return true; return false; }); if (res == m_elts.end () || res->index != index || res->unit_offset != unit_offset) res = nullptr; /* TODO: perhaps remove the check (that the underlying array is indeed sorted) if it turns out it can be too slow? */ if (!flag_checking) return res; const ipa_argagg_value *slow_res = NULL; int prev_index = -1; unsigned prev_unit_offset = 0; for (const ipa_argagg_value &av : m_elts) { gcc_assert (prev_index < 0 || prev_index < av.index || prev_unit_offset < av.unit_offset); prev_index = av.index; prev_unit_offset = av.unit_offset; if (av.index == index && av.unit_offset == unit_offset) slow_res = &av; } gcc_assert (res == slow_res); return res; } /* Return the first item describing a constant stored for parameter with INDEX, regardless of offset or reference, or NULL if there is no such constant. */ const ipa_argagg_value * ipa_argagg_value_list::get_elt_for_index (int index) const { const ipa_argagg_value *res = std::lower_bound (m_elts.begin (), m_elts.end (), index, [] (const ipa_argagg_value &elt, unsigned idx) { return elt.index < idx; }); if (res == m_elts.end () || res->index != index) res = nullptr; return res; } /* Return the aggregate constant stored for INDEX at UNIT_OFFSET, not performing any check of whether value is passed by reference, or NULL_TREE if there is no such constant. */ tree ipa_argagg_value_list::get_value (int index, unsigned unit_offset) const { const ipa_argagg_value *av = get_elt (index, unit_offset); return av ? av->value : NULL_TREE; } /* Return the aggregate constant stored for INDEX at UNIT_OFFSET, if it is passed by reference or not according to BY_REF, or NULL_TREE if there is no such constant. */ tree ipa_argagg_value_list::get_value (int index, unsigned unit_offset, bool by_ref) const { const ipa_argagg_value *av = get_elt (index, unit_offset); if (av && av->by_ref == by_ref) return av->value; return NULL_TREE; } /* Return true if all elements present in OTHER are also present in this list. */ bool ipa_argagg_value_list::superset_of_p (const ipa_argagg_value_list &other) const { unsigned j = 0; for (unsigned i = 0; i < other.m_elts.size (); i++) { unsigned other_index = other.m_elts[i].index; unsigned other_offset = other.m_elts[i].unit_offset; while (j < m_elts.size () && (m_elts[j].index < other_index || (m_elts[j].index == other_index && m_elts[j].unit_offset < other_offset))) j++; if (j >= m_elts.size () || m_elts[j].index != other_index || m_elts[j].unit_offset != other_offset || m_elts[j].by_ref != other.m_elts[i].by_ref || !m_elts[j].value || !values_equal_for_ipcp_p (m_elts[j].value, other.m_elts[i].value)) return false; } return true; } /* Push all items in this list that describe parameter SRC_INDEX into RES as ones describing DST_INDEX while subtracting UNIT_DELTA from their unit offsets but skip those which would end up with a negative offset. */ void ipa_argagg_value_list::push_adjusted_values (unsigned src_index, unsigned dest_index, unsigned unit_delta, vec *res) const { const ipa_argagg_value *av = get_elt_for_index (src_index); if (!av) return; unsigned prev_unit_offset = 0; bool first = true; for (; av < m_elts.end (); ++av) { if (av->index > src_index) return; if (av->index == src_index && (av->unit_offset >= unit_delta) && av->value) { ipa_argagg_value new_av; gcc_checking_assert (av->value); new_av.value = av->value; new_av.unit_offset = av->unit_offset - unit_delta; new_av.index = dest_index; new_av.by_ref = av->by_ref; /* Quick check that the offsets we push are indeed increasing. */ gcc_assert (first || new_av.unit_offset > prev_unit_offset); prev_unit_offset = new_av.unit_offset; first = false; res->safe_push (new_av); } } } /* Push to RES information about single lattices describing aggregate values in PLATS as those describing parameter DEST_INDEX and the original offset minus UNIT_DELTA. Return true if any item has been pushed to RES. */ static bool push_agg_values_from_plats (ipcp_param_lattices *plats, int dest_index, unsigned unit_delta, vec *res) { if (plats->aggs_contain_variable) return false; bool pushed_sth = false; bool first = true; unsigned prev_unit_offset = 0; for (struct ipcp_agg_lattice *aglat = plats->aggs; aglat; aglat = aglat->next) if (aglat->is_single_const () && (aglat->offset / BITS_PER_UNIT - unit_delta) >= 0) { ipa_argagg_value iav; iav.value = aglat->values->value; iav.unit_offset = aglat->offset / BITS_PER_UNIT - unit_delta; iav.index = dest_index; iav.by_ref = plats->aggs_by_ref; gcc_assert (first || iav.unit_offset > prev_unit_offset); prev_unit_offset = iav.unit_offset; first = false; pushed_sth = true; res->safe_push (iav); } return pushed_sth; } /* Turn all values in LIST that are not present in OTHER into NULL_TREEs. Return the number of remaining valid entries. */ static unsigned intersect_argaggs_with (vec &elts, const vec &other) { unsigned valid_entries = 0; unsigned j = 0; for (unsigned i = 0; i < elts.length (); i++) { if (!elts[i].value) continue; unsigned this_index = elts[i].index; unsigned this_offset = elts[i].unit_offset; while (j < other.length () && (other[j].index < this_index || (other[j].index == this_index && other[j].unit_offset < this_offset))) j++; if (j >= other.length ()) { elts[i].value = NULL_TREE; continue; } if (other[j].index == this_index && other[j].unit_offset == this_offset && other[j].by_ref == elts[i].by_ref && other[j].value && values_equal_for_ipcp_p (other[j].value, elts[i].value)) valid_entries++; else elts[i].value = NULL_TREE; } return valid_entries; } /* Mark bot aggregate and scalar lattices as containing an unknown variable, return true is any of them has not been marked as such so far. */ static inline bool set_all_contains_variable (class ipcp_param_lattices *plats) { bool ret; ret = plats->itself.set_contains_variable (); ret |= plats->ctxlat.set_contains_variable (); ret |= set_agg_lats_contain_variable (plats); ret |= plats->bits_lattice.set_to_bottom (); ret |= plats->m_value_range.set_to_bottom (); return ret; } /* Worker of call_for_symbol_thunks_and_aliases, increment the integer DATA points to by the number of callers to NODE. */ static bool count_callers (cgraph_node *node, void *data) { int *caller_count = (int *) data; for (cgraph_edge *cs = node->callers; cs; cs = cs->next_caller) /* Local thunks can be handled transparently, but if the thunk cannot be optimized out, count it as a real use. */ if (!cs->caller->thunk || !cs->caller->local) ++*caller_count; return false; } /* Worker of call_for_symbol_thunks_and_aliases, it is supposed to be called on the one caller of some other node. Set the caller's corresponding flag. */ static bool set_single_call_flag (cgraph_node *node, void *) { cgraph_edge *cs = node->callers; /* Local thunks can be handled transparently, skip them. */ while (cs && cs->caller->thunk && cs->caller->local) cs = cs->next_caller; if (cs) if (ipa_node_params* info = ipa_node_params_sum->get (cs->caller)) { info->node_calling_single_call = true; return true; } return false; } /* Initialize ipcp_lattices. */ static void initialize_node_lattices (struct cgraph_node *node) { ipa_node_params *info = ipa_node_params_sum->get (node); struct cgraph_edge *ie; bool disable = false, variable = false; int i; gcc_checking_assert (node->has_gimple_body_p ()); if (!ipa_get_param_count (info)) disable = true; else if (node->local) { int caller_count = 0; node->call_for_symbol_thunks_and_aliases (count_callers, &caller_count, true); gcc_checking_assert (caller_count > 0); if (caller_count == 1) node->call_for_symbol_thunks_and_aliases (set_single_call_flag, NULL, true); } else { /* When cloning is allowed, we can assume that externally visible functions are not called. We will compensate this by cloning later. */ if (ipcp_versionable_function_p (node) && ipcp_cloning_candidate_p (node)) variable = true; else disable = true; } if (dump_file && (dump_flags & TDF_DETAILS) && !node->alias && !node->thunk) { fprintf (dump_file, "Initializing lattices of %s\n", node->dump_name ()); if (disable || variable) fprintf (dump_file, " Marking all lattices as %s\n", disable ? "BOTTOM" : "VARIABLE"); } auto_vec surviving_params; bool pre_modified = false; clone_info *cinfo = clone_info::get (node); if (!disable && cinfo && cinfo->param_adjustments) { /* At the moment all IPA optimizations should use the number of parameters of the prevailing decl as the m_always_copy_start. Handling any other value would complicate the code below, so for the time bing let's only assert it is so. */ gcc_assert ((cinfo->param_adjustments->m_always_copy_start == ipa_get_param_count (info)) || cinfo->param_adjustments->m_always_copy_start < 0); pre_modified = true; cinfo->param_adjustments->get_surviving_params (&surviving_params); if (dump_file && (dump_flags & TDF_DETAILS) && !node->alias && !node->thunk) { bool first = true; for (int j = 0; j < ipa_get_param_count (info); j++) { if (j < (int) surviving_params.length () && surviving_params[j]) continue; if (first) { fprintf (dump_file, " The following parameters are dead on arrival:"); first = false; } fprintf (dump_file, " %u", j); } if (!first) fprintf (dump_file, "\n"); } } for (i = 0; i < ipa_get_param_count (info); i++) { ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); tree type = ipa_get_type (info, i); if (disable || !ipa_get_type (info, i) || (pre_modified && (surviving_params.length () <= (unsigned) i || !surviving_params[i]))) { plats->itself.set_to_bottom (); plats->ctxlat.set_to_bottom (); set_agg_lats_to_bottom (plats); plats->bits_lattice.set_to_bottom (); plats->m_value_range.init (type); plats->m_value_range.set_to_bottom (); } else { plats->m_value_range.init (type); if (variable) set_all_contains_variable (plats); } } for (ie = node->indirect_calls; ie; ie = ie->next_callee) if (ie->indirect_info->polymorphic && ie->indirect_info->param_index >= 0) { gcc_checking_assert (ie->indirect_info->param_index >= 0); ipa_get_parm_lattices (info, ie->indirect_info->param_index)->virt_call = 1; } } /* Return true if VALUE can be safely IPA-CP propagated to a parameter of type PARAM_TYPE. */ static bool ipacp_value_safe_for_type (tree param_type, tree value) { tree val_type = TREE_TYPE (value); if (param_type == val_type || useless_type_conversion_p (param_type, val_type) || fold_convertible_p (param_type, value)) return true; else return false; } /* Return the result of a (possibly arithmetic) operation on the constant value INPUT. OPERAND is 2nd operand for binary operation. RES_TYPE is the type of the parameter to which the result is passed. Return NULL_TREE if that cannot be determined or be considered an interprocedural invariant. */ static tree ipa_get_jf_arith_result (enum tree_code opcode, tree input, tree operand, tree res_type) { tree res; if (opcode == NOP_EXPR) return input; if (!is_gimple_ip_invariant (input)) return NULL_TREE; if (opcode == ASSERT_EXPR) { if (values_equal_for_ipcp_p (input, operand)) return input; else return NULL_TREE; } if (!res_type) { if (TREE_CODE_CLASS (opcode) == tcc_comparison) res_type = boolean_type_node; else if (expr_type_first_operand_type_p (opcode)) res_type = TREE_TYPE (input); else return NULL_TREE; } if (TREE_CODE_CLASS (opcode) == tcc_unary) res = fold_unary (opcode, res_type, input); else res = fold_binary (opcode, res_type, input, operand); if (res && !is_gimple_ip_invariant (res)) return NULL_TREE; return res; } /* Return the result of a (possibly arithmetic) pass through jump function JFUNC on the constant value INPUT. RES_TYPE is the type of the parameter to which the result is passed. Return NULL_TREE if that cannot be determined or be considered an interprocedural invariant. */ static tree ipa_get_jf_pass_through_result (struct ipa_jump_func *jfunc, tree input, tree res_type) { return ipa_get_jf_arith_result (ipa_get_jf_pass_through_operation (jfunc), input, ipa_get_jf_pass_through_operand (jfunc), res_type); } /* Return the result of an ancestor jump function JFUNC on the constant value INPUT. Return NULL_TREE if that cannot be determined. */ static tree ipa_get_jf_ancestor_result (struct ipa_jump_func *jfunc, tree input) { gcc_checking_assert (TREE_CODE (input) != TREE_BINFO); if (TREE_CODE (input) == ADDR_EXPR) { gcc_checking_assert (is_gimple_ip_invariant_address (input)); poly_int64 off = ipa_get_jf_ancestor_offset (jfunc); if (known_eq (off, 0)) return input; poly_int64 byte_offset = exact_div (off, BITS_PER_UNIT); return build1 (ADDR_EXPR, TREE_TYPE (input), fold_build2 (MEM_REF, TREE_TYPE (TREE_TYPE (input)), input, build_int_cst (ptr_type_node, byte_offset))); } else if (ipa_get_jf_ancestor_keep_null (jfunc) && zerop (input)) return input; else return NULL_TREE; } /* Determine whether JFUNC evaluates to a single known constant value and if so, return it. Otherwise return NULL. INFO describes the caller node or the one it is inlined to, so that pass-through jump functions can be evaluated. PARM_TYPE is the type of the parameter to which the result is passed. */ tree ipa_value_from_jfunc (class ipa_node_params *info, struct ipa_jump_func *jfunc, tree parm_type) { if (jfunc->type == IPA_JF_CONST) return ipa_get_jf_constant (jfunc); else if (jfunc->type == IPA_JF_PASS_THROUGH || jfunc->type == IPA_JF_ANCESTOR) { tree input; int idx; if (jfunc->type == IPA_JF_PASS_THROUGH) idx = ipa_get_jf_pass_through_formal_id (jfunc); else idx = ipa_get_jf_ancestor_formal_id (jfunc); if (info->ipcp_orig_node) input = info->known_csts[idx]; else { ipcp_lattice *lat; if (!info->lattices || idx >= ipa_get_param_count (info)) return NULL_TREE; lat = ipa_get_scalar_lat (info, idx); if (!lat->is_single_const ()) return NULL_TREE; input = lat->values->value; } if (!input) return NULL_TREE; if (jfunc->type == IPA_JF_PASS_THROUGH) return ipa_get_jf_pass_through_result (jfunc, input, parm_type); else return ipa_get_jf_ancestor_result (jfunc, input); } else return NULL_TREE; } /* Determine whether JFUNC evaluates to single known polymorphic context, given that INFO describes the caller node or the one it is inlined to, CS is the call graph edge corresponding to JFUNC and CSIDX index of the described parameter. */ ipa_polymorphic_call_context ipa_context_from_jfunc (ipa_node_params *info, cgraph_edge *cs, int csidx, ipa_jump_func *jfunc) { ipa_edge_args *args = ipa_edge_args_sum->get (cs); ipa_polymorphic_call_context ctx; ipa_polymorphic_call_context *edge_ctx = cs ? ipa_get_ith_polymorhic_call_context (args, csidx) : NULL; if (edge_ctx && !edge_ctx->useless_p ()) ctx = *edge_ctx; if (jfunc->type == IPA_JF_PASS_THROUGH || jfunc->type == IPA_JF_ANCESTOR) { ipa_polymorphic_call_context srcctx; int srcidx; bool type_preserved = true; if (jfunc->type == IPA_JF_PASS_THROUGH) { if (ipa_get_jf_pass_through_operation (jfunc) != NOP_EXPR) return ctx; type_preserved = ipa_get_jf_pass_through_type_preserved (jfunc); srcidx = ipa_get_jf_pass_through_formal_id (jfunc); } else { type_preserved = ipa_get_jf_ancestor_type_preserved (jfunc); srcidx = ipa_get_jf_ancestor_formal_id (jfunc); } if (info->ipcp_orig_node) { if (info->known_contexts.exists ()) srcctx = info->known_contexts[srcidx]; } else { if (!info->lattices || srcidx >= ipa_get_param_count (info)) return ctx; ipcp_lattice *lat; lat = ipa_get_poly_ctx_lat (info, srcidx); if (!lat->is_single_const ()) return ctx; srcctx = lat->values->value; } if (srcctx.useless_p ()) return ctx; if (jfunc->type == IPA_JF_ANCESTOR) srcctx.offset_by (ipa_get_jf_ancestor_offset (jfunc)); if (!type_preserved) srcctx.possible_dynamic_type_change (cs->in_polymorphic_cdtor); srcctx.combine_with (ctx); return srcctx; } return ctx; } /* Emulate effects of unary OPERATION and/or conversion from SRC_TYPE to DST_TYPE on value range in SRC_VR and store it to DST_VR. Return true if the result is a range that is not VARYING nor UNDEFINED. */ static bool ipa_vr_operation_and_type_effects (vrange &dst_vr, const vrange &src_vr, enum tree_code operation, tree dst_type, tree src_type) { if (!irange::supports_p (dst_type) || !irange::supports_p (src_type)) return false; range_op_handler handler (operation); if (!handler) return false; Value_Range varying (dst_type); varying.set_varying (dst_type); return (handler.fold_range (dst_vr, dst_type, src_vr, varying) && !dst_vr.varying_p () && !dst_vr.undefined_p ()); } /* Same as above, but the SRC_VR argument is an IPA_VR which must first be extracted onto a vrange. */ static bool ipa_vr_operation_and_type_effects (vrange &dst_vr, const ipa_vr &src_vr, enum tree_code operation, tree dst_type, tree src_type) { Value_Range tmp; src_vr.get_vrange (tmp); return ipa_vr_operation_and_type_effects (dst_vr, tmp, operation, dst_type, src_type); } /* Determine range of JFUNC given that INFO describes the caller node or the one it is inlined to, CS is the call graph edge corresponding to JFUNC and PARM_TYPE of the parameter. */ void ipa_value_range_from_jfunc (vrange &vr, ipa_node_params *info, cgraph_edge *cs, ipa_jump_func *jfunc, tree parm_type) { vr.set_undefined (); if (jfunc->m_vr) ipa_vr_operation_and_type_effects (vr, *jfunc->m_vr, NOP_EXPR, parm_type, jfunc->m_vr->type ()); if (vr.singleton_p ()) return; if (jfunc->type == IPA_JF_PASS_THROUGH) { int idx; ipcp_transformation *sum = ipcp_get_transformation_summary (cs->caller->inlined_to ? cs->caller->inlined_to : cs->caller); if (!sum || !sum->m_vr) return; idx = ipa_get_jf_pass_through_formal_id (jfunc); if (!(*sum->m_vr)[idx].known_p ()) return; tree vr_type = ipa_get_type (info, idx); Value_Range srcvr; (*sum->m_vr)[idx].get_vrange (srcvr); enum tree_code operation = ipa_get_jf_pass_through_operation (jfunc); if (TREE_CODE_CLASS (operation) == tcc_unary) { Value_Range res (vr_type); if (ipa_vr_operation_and_type_effects (res, srcvr, operation, parm_type, vr_type)) vr.intersect (res); } else { Value_Range op_res (vr_type); Value_Range res (vr_type); tree op = ipa_get_jf_pass_through_operand (jfunc); Value_Range op_vr (vr_type); range_op_handler handler (operation); ipa_range_set_and_normalize (op_vr, op); if (!handler || !op_res.supports_type_p (vr_type) || !handler.fold_range (op_res, vr_type, srcvr, op_vr)) op_res.set_varying (vr_type); if (ipa_vr_operation_and_type_effects (res, op_res, NOP_EXPR, parm_type, vr_type)) vr.intersect (res); } } } /* Determine whether ITEM, jump function for an aggregate part, evaluates to a single known constant value and if so, return it. Otherwise return NULL. NODE and INFO describes the caller node or the one it is inlined to, and its related info. */ tree ipa_agg_value_from_jfunc (ipa_node_params *info, cgraph_node *node, const ipa_agg_jf_item *item) { tree value = NULL_TREE; int src_idx; if (item->offset < 0 || item->jftype == IPA_JF_UNKNOWN || item->offset >= (HOST_WIDE_INT) UINT_MAX * BITS_PER_UNIT) return NULL_TREE; if (item->jftype == IPA_JF_CONST) return item->value.constant; gcc_checking_assert (item->jftype == IPA_JF_PASS_THROUGH || item->jftype == IPA_JF_LOAD_AGG); src_idx = item->value.pass_through.formal_id; if (info->ipcp_orig_node) { if (item->jftype == IPA_JF_PASS_THROUGH) value = info->known_csts[src_idx]; else if (ipcp_transformation *ts = ipcp_get_transformation_summary (node)) { ipa_argagg_value_list avl (ts); value = avl.get_value (src_idx, item->value.load_agg.offset / BITS_PER_UNIT, item->value.load_agg.by_ref); } } else if (info->lattices) { class ipcp_param_lattices *src_plats = ipa_get_parm_lattices (info, src_idx); if (item->jftype == IPA_JF_PASS_THROUGH) { struct ipcp_lattice *lat = &src_plats->itself; if (!lat->is_single_const ()) return NULL_TREE; value = lat->values->value; } else if (src_plats->aggs && !src_plats->aggs_bottom && !src_plats->aggs_contain_variable && src_plats->aggs_by_ref == item->value.load_agg.by_ref) { struct ipcp_agg_lattice *aglat; for (aglat = src_plats->aggs; aglat; aglat = aglat->next) { if (aglat->offset > item->value.load_agg.offset) break; if (aglat->offset == item->value.load_agg.offset) { if (aglat->is_single_const ()) value = aglat->values->value; break; } } } } if (!value) return NULL_TREE; if (item->jftype == IPA_JF_LOAD_AGG) { tree load_type = item->value.load_agg.type; tree value_type = TREE_TYPE (value); /* Ensure value type is compatible with load type. */ if (!useless_type_conversion_p (load_type, value_type)) return NULL_TREE; } return ipa_get_jf_arith_result (item->value.pass_through.operation, value, item->value.pass_through.operand, item->type); } /* Process all items in AGG_JFUNC relative to caller (or the node the original caller is inlined to) NODE which described by INFO and push the results to RES as describing values passed in parameter DST_INDEX. */ void ipa_push_agg_values_from_jfunc (ipa_node_params *info, cgraph_node *node, ipa_agg_jump_function *agg_jfunc, unsigned dst_index, vec *res) { unsigned prev_unit_offset = 0; bool first = true; for (const ipa_agg_jf_item &item : agg_jfunc->items) { tree value = ipa_agg_value_from_jfunc (info, node, &item); if (!value) continue; ipa_argagg_value iav; iav.value = value; iav.unit_offset = item.offset / BITS_PER_UNIT; iav.index = dst_index; iav.by_ref = agg_jfunc->by_ref; gcc_assert (first || iav.unit_offset > prev_unit_offset); prev_unit_offset = iav.unit_offset; first = false; res->safe_push (iav); } } /* If checking is enabled, verify that no lattice is in the TOP state, i.e. not bottom, not containing a variable component and without any known value at the same time. */ DEBUG_FUNCTION void ipcp_verify_propagated_values (void) { struct cgraph_node *node; FOR_EACH_FUNCTION_WITH_GIMPLE_BODY (node) { ipa_node_params *info = ipa_node_params_sum->get (node); if (!opt_for_fn (node->decl, flag_ipa_cp) || !opt_for_fn (node->decl, optimize)) continue; int i, count = ipa_get_param_count (info); for (i = 0; i < count; i++) { ipcp_lattice *lat = ipa_get_scalar_lat (info, i); if (!lat->bottom && !lat->contains_variable && lat->values_count == 0) { if (dump_file) { symtab->dump (dump_file); fprintf (dump_file, "\nIPA lattices after constant " "propagation, before gcc_unreachable:\n"); print_all_lattices (dump_file, true, false); } gcc_unreachable (); } } } } /* Return true iff X and Y should be considered equal contexts by IPA-CP. */ static bool values_equal_for_ipcp_p (ipa_polymorphic_call_context x, ipa_polymorphic_call_context y) { return x.equal_to (y); } /* Add a new value source to the value represented by THIS, marking that a value comes from edge CS and (if the underlying jump function is a pass-through or an ancestor one) from a caller value SRC_VAL of a caller parameter described by SRC_INDEX. OFFSET is negative if the source was the scalar value of the parameter itself or the offset within an aggregate. */ template void ipcp_value::add_source (cgraph_edge *cs, ipcp_value *src_val, int src_idx, HOST_WIDE_INT offset) { ipcp_value_source *src; src = new (ipcp_sources_pool.allocate ()) ipcp_value_source; src->offset = offset; src->cs = cs; src->val = src_val; src->index = src_idx; src->next = sources; sources = src; } /* Allocate a new ipcp_value holding a tree constant, initialize its value to SOURCE and clear all other fields. */ static ipcp_value * allocate_and_init_ipcp_value (tree cst, unsigned same_lat_gen_level) { ipcp_value *val; val = new (ipcp_cst_values_pool.allocate ()) ipcp_value(); val->value = cst; val->self_recursion_generated_level = same_lat_gen_level; return val; } /* Allocate a new ipcp_value holding a polymorphic context, initialize its value to SOURCE and clear all other fields. */ static ipcp_value * allocate_and_init_ipcp_value (ipa_polymorphic_call_context ctx, unsigned same_lat_gen_level) { ipcp_value *val; val = new (ipcp_poly_ctx_values_pool.allocate ()) ipcp_value(); val->value = ctx; val->self_recursion_generated_level = same_lat_gen_level; return val; } /* Try to add NEWVAL to LAT, potentially creating a new ipcp_value for it. CS, SRC_VAL SRC_INDEX and OFFSET are meant for add_source and have the same meaning. OFFSET -1 means the source is scalar and not a part of an aggregate. If non-NULL, VAL_P records address of existing or newly added ipcp_value. If the value is generated for a self-recursive call as a result of an arithmetic pass-through jump-function acting on a value in the same lattice, SAME_LAT_GEN_LEVEL must be the length of such chain, otherwise it must be zero. If it is non-zero, PARAM_IPA_CP_VALUE_LIST_SIZE limit is ignored. */ template bool ipcp_lattice::add_value (valtype newval, cgraph_edge *cs, ipcp_value *src_val, int src_idx, HOST_WIDE_INT offset, ipcp_value **val_p, unsigned same_lat_gen_level) { ipcp_value *val, *last_val = NULL; if (val_p) *val_p = NULL; if (bottom) return false; for (val = values; val; last_val = val, val = val->next) if (values_equal_for_ipcp_p (val->value, newval)) { if (val_p) *val_p = val; if (val->self_recursion_generated_level < same_lat_gen_level) val->self_recursion_generated_level = same_lat_gen_level; if (ipa_edge_within_scc (cs)) { ipcp_value_source *s; for (s = val->sources; s; s = s->next) if (s->cs == cs && s->val == src_val) break; if (s) return false; } val->add_source (cs, src_val, src_idx, offset); return false; } if (!same_lat_gen_level && values_count == opt_for_fn (cs->caller->decl, param_ipa_cp_value_list_size)) { /* We can only free sources, not the values themselves, because sources of other values in this SCC might point to them. */ for (val = values; val; val = val->next) { while (val->sources) { ipcp_value_source *src = val->sources; val->sources = src->next; ipcp_sources_pool.remove ((ipcp_value_source*)src); } } values = NULL; return set_to_bottom (); } values_count++; val = allocate_and_init_ipcp_value (newval, same_lat_gen_level); val->add_source (cs, src_val, src_idx, offset); val->next = NULL; /* Add the new value to end of value list, which can reduce iterations of propagation stage for recursive function. */ if (last_val) last_val->next = val; else values = val; if (val_p) *val_p = val; return true; } /* A helper function that returns result of operation specified by OPCODE on the value of SRC_VAL. If non-NULL, OPND1_TYPE is expected type for the value of SRC_VAL. If the operation is binary, OPND2 is a constant value acting as its second operand. If non-NULL, RES_TYPE is expected type of the result. */ static tree get_val_across_arith_op (enum tree_code opcode, tree opnd1_type, tree opnd2, ipcp_value *src_val, tree res_type) { tree opnd1 = src_val->value; /* Skip source values that is incompatible with specified type. */ if (opnd1_type && !useless_type_conversion_p (opnd1_type, TREE_TYPE (opnd1))) return NULL_TREE; return ipa_get_jf_arith_result (opcode, opnd1, opnd2, res_type); } /* Propagate values through an arithmetic transformation described by a jump function associated with edge CS, taking values from SRC_LAT and putting them into DEST_LAT. OPND1_TYPE is expected type for the values in SRC_LAT. OPND2 is a constant value if transformation is a binary operation. SRC_OFFSET specifies offset in an aggregate if SRC_LAT describes lattice of a part of the aggregate. SRC_IDX is the index of the source parameter. RES_TYPE is the value type of result being propagated into. Return true if DEST_LAT changed. */ static bool propagate_vals_across_arith_jfunc (cgraph_edge *cs, enum tree_code opcode, tree opnd1_type, tree opnd2, ipcp_lattice *src_lat, ipcp_lattice *dest_lat, HOST_WIDE_INT src_offset, int src_idx, tree res_type) { ipcp_value *src_val; bool ret = false; /* Due to circular dependencies, propagating within an SCC through arithmetic transformation would create infinite number of values. But for self-feeding recursive function, we could allow propagation in a limited count, and this can enable a simple kind of recursive function versioning. For other scenario, we would just make lattices bottom. */ if (opcode != NOP_EXPR && ipa_edge_within_scc (cs)) { int i; int max_recursive_depth = opt_for_fn(cs->caller->decl, param_ipa_cp_max_recursive_depth); if (src_lat != dest_lat || max_recursive_depth < 1) return dest_lat->set_contains_variable (); /* No benefit if recursive execution is in low probability. */ if (cs->sreal_frequency () * 100 <= ((sreal) 1) * opt_for_fn (cs->caller->decl, param_ipa_cp_min_recursive_probability)) return dest_lat->set_contains_variable (); auto_vec *, 8> val_seeds; for (src_val = src_lat->values; src_val; src_val = src_val->next) { /* Now we do not use self-recursively generated value as propagation source, this is absolutely conservative, but could avoid explosion of lattice's value space, especially when one recursive function calls another recursive. */ if (src_val->self_recursion_generated_p ()) { ipcp_value_source *s; /* If the lattice has already been propagated for the call site, no need to do that again. */ for (s = src_val->sources; s; s = s->next) if (s->cs == cs) return dest_lat->set_contains_variable (); } else val_seeds.safe_push (src_val); } gcc_assert ((int) val_seeds.length () <= param_ipa_cp_value_list_size); /* Recursively generate lattice values with a limited count. */ FOR_EACH_VEC_ELT (val_seeds, i, src_val) { for (int j = 1; j < max_recursive_depth; j++) { tree cstval = get_val_across_arith_op (opcode, opnd1_type, opnd2, src_val, res_type); if (!cstval || !ipacp_value_safe_for_type (res_type, cstval)) break; ret |= dest_lat->add_value (cstval, cs, src_val, src_idx, src_offset, &src_val, j); gcc_checking_assert (src_val); } } ret |= dest_lat->set_contains_variable (); } else for (src_val = src_lat->values; src_val; src_val = src_val->next) { /* Now we do not use self-recursively generated value as propagation source, otherwise it is easy to make value space of normal lattice overflow. */ if (src_val->self_recursion_generated_p ()) { ret |= dest_lat->set_contains_variable (); continue; } tree cstval = get_val_across_arith_op (opcode, opnd1_type, opnd2, src_val, res_type); if (cstval && ipacp_value_safe_for_type (res_type, cstval)) ret |= dest_lat->add_value (cstval, cs, src_val, src_idx, src_offset); else ret |= dest_lat->set_contains_variable (); } return ret; } /* Propagate values through a pass-through jump function JFUNC associated with edge CS, taking values from SRC_LAT and putting them into DEST_LAT. SRC_IDX is the index of the source parameter. PARM_TYPE is the type of the parameter to which the result is passed. */ static bool propagate_vals_across_pass_through (cgraph_edge *cs, ipa_jump_func *jfunc, ipcp_lattice *src_lat, ipcp_lattice *dest_lat, int src_idx, tree parm_type) { return propagate_vals_across_arith_jfunc (cs, ipa_get_jf_pass_through_operation (jfunc), NULL_TREE, ipa_get_jf_pass_through_operand (jfunc), src_lat, dest_lat, -1, src_idx, parm_type); } /* Propagate values through an ancestor jump function JFUNC associated with edge CS, taking values from SRC_LAT and putting them into DEST_LAT. SRC_IDX is the index of the source parameter. */ static bool propagate_vals_across_ancestor (struct cgraph_edge *cs, struct ipa_jump_func *jfunc, ipcp_lattice *src_lat, ipcp_lattice *dest_lat, int src_idx, tree param_type) { ipcp_value *src_val; bool ret = false; if (ipa_edge_within_scc (cs)) return dest_lat->set_contains_variable (); for (src_val = src_lat->values; src_val; src_val = src_val->next) { tree t = ipa_get_jf_ancestor_result (jfunc, src_val->value); if (t && ipacp_value_safe_for_type (param_type, t)) ret |= dest_lat->add_value (t, cs, src_val, src_idx); else ret |= dest_lat->set_contains_variable (); } return ret; } /* Propagate scalar values across jump function JFUNC that is associated with edge CS and put the values into DEST_LAT. PARM_TYPE is the type of the parameter to which the result is passed. */ static bool propagate_scalar_across_jump_function (struct cgraph_edge *cs, struct ipa_jump_func *jfunc, ipcp_lattice *dest_lat, tree param_type) { if (dest_lat->bottom) return false; if (jfunc->type == IPA_JF_CONST) { tree val = ipa_get_jf_constant (jfunc); if (ipacp_value_safe_for_type (param_type, val)) return dest_lat->add_value (val, cs, NULL, 0); else return dest_lat->set_contains_variable (); } else if (jfunc->type == IPA_JF_PASS_THROUGH || jfunc->type == IPA_JF_ANCESTOR) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); ipcp_lattice *src_lat; int src_idx; bool ret; if (jfunc->type == IPA_JF_PASS_THROUGH) src_idx = ipa_get_jf_pass_through_formal_id (jfunc); else src_idx = ipa_get_jf_ancestor_formal_id (jfunc); src_lat = ipa_get_scalar_lat (caller_info, src_idx); if (src_lat->bottom) return dest_lat->set_contains_variable (); /* If we would need to clone the caller and cannot, do not propagate. */ if (!ipcp_versionable_function_p (cs->caller) && (src_lat->contains_variable || (src_lat->values_count > 1))) return dest_lat->set_contains_variable (); if (jfunc->type == IPA_JF_PASS_THROUGH) ret = propagate_vals_across_pass_through (cs, jfunc, src_lat, dest_lat, src_idx, param_type); else ret = propagate_vals_across_ancestor (cs, jfunc, src_lat, dest_lat, src_idx, param_type); if (src_lat->contains_variable) ret |= dest_lat->set_contains_variable (); return ret; } /* TODO: We currently do not handle member method pointers in IPA-CP (we only use it for indirect inlining), we should propagate them too. */ return dest_lat->set_contains_variable (); } /* Propagate scalar values across jump function JFUNC that is associated with edge CS and describes argument IDX and put the values into DEST_LAT. */ static bool propagate_context_across_jump_function (cgraph_edge *cs, ipa_jump_func *jfunc, int idx, ipcp_lattice *dest_lat) { if (dest_lat->bottom) return false; ipa_edge_args *args = ipa_edge_args_sum->get (cs); bool ret = false; bool added_sth = false; bool type_preserved = true; ipa_polymorphic_call_context edge_ctx, *edge_ctx_ptr = ipa_get_ith_polymorhic_call_context (args, idx); if (edge_ctx_ptr) edge_ctx = *edge_ctx_ptr; if (jfunc->type == IPA_JF_PASS_THROUGH || jfunc->type == IPA_JF_ANCESTOR) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); int src_idx; ipcp_lattice *src_lat; /* TODO: Once we figure out how to propagate speculations, it will probably be a good idea to switch to speculation if type_preserved is not set instead of punting. */ if (jfunc->type == IPA_JF_PASS_THROUGH) { if (ipa_get_jf_pass_through_operation (jfunc) != NOP_EXPR) goto prop_fail; type_preserved = ipa_get_jf_pass_through_type_preserved (jfunc); src_idx = ipa_get_jf_pass_through_formal_id (jfunc); } else { type_preserved = ipa_get_jf_ancestor_type_preserved (jfunc); src_idx = ipa_get_jf_ancestor_formal_id (jfunc); } src_lat = ipa_get_poly_ctx_lat (caller_info, src_idx); /* If we would need to clone the caller and cannot, do not propagate. */ if (!ipcp_versionable_function_p (cs->caller) && (src_lat->contains_variable || (src_lat->values_count > 1))) goto prop_fail; ipcp_value *src_val; for (src_val = src_lat->values; src_val; src_val = src_val->next) { ipa_polymorphic_call_context cur = src_val->value; if (!type_preserved) cur.possible_dynamic_type_change (cs->in_polymorphic_cdtor); if (jfunc->type == IPA_JF_ANCESTOR) cur.offset_by (ipa_get_jf_ancestor_offset (jfunc)); /* TODO: In cases we know how the context is going to be used, we can improve the result by passing proper OTR_TYPE. */ cur.combine_with (edge_ctx); if (!cur.useless_p ()) { if (src_lat->contains_variable && !edge_ctx.equal_to (cur)) ret |= dest_lat->set_contains_variable (); ret |= dest_lat->add_value (cur, cs, src_val, src_idx); added_sth = true; } } } prop_fail: if (!added_sth) { if (!edge_ctx.useless_p ()) ret |= dest_lat->add_value (edge_ctx, cs); else ret |= dest_lat->set_contains_variable (); } return ret; } /* Propagate bits across jfunc that is associated with edge cs and update dest_lattice accordingly. */ bool propagate_bits_across_jump_function (cgraph_edge *cs, int idx, ipa_jump_func *jfunc, ipcp_bits_lattice *dest_lattice) { if (dest_lattice->bottom_p ()) return false; enum availability availability; cgraph_node *callee = cs->callee->function_symbol (&availability); ipa_node_params *callee_info = ipa_node_params_sum->get (callee); tree parm_type = ipa_get_type (callee_info, idx); /* For K&R C programs, ipa_get_type() could return NULL_TREE. Avoid the transform for these cases. Similarly, we can have bad type mismatches with LTO, avoid doing anything with those too. */ if (!parm_type || (!INTEGRAL_TYPE_P (parm_type) && !POINTER_TYPE_P (parm_type))) { if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, "Setting dest_lattice to bottom, because type of " "param %i of %s is NULL or unsuitable for bits propagation\n", idx, cs->callee->dump_name ()); return dest_lattice->set_to_bottom (); } unsigned precision = TYPE_PRECISION (parm_type); signop sgn = TYPE_SIGN (parm_type); if (jfunc->type == IPA_JF_PASS_THROUGH || jfunc->type == IPA_JF_ANCESTOR) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); tree operand = NULL_TREE; enum tree_code code; unsigned src_idx; bool keep_null = false; if (jfunc->type == IPA_JF_PASS_THROUGH) { code = ipa_get_jf_pass_through_operation (jfunc); src_idx = ipa_get_jf_pass_through_formal_id (jfunc); if (code != NOP_EXPR) operand = ipa_get_jf_pass_through_operand (jfunc); } else { code = POINTER_PLUS_EXPR; src_idx = ipa_get_jf_ancestor_formal_id (jfunc); unsigned HOST_WIDE_INT offset = ipa_get_jf_ancestor_offset (jfunc) / BITS_PER_UNIT; keep_null = (ipa_get_jf_ancestor_keep_null (jfunc) || !offset); operand = build_int_cstu (size_type_node, offset); } class ipcp_param_lattices *src_lats = ipa_get_parm_lattices (caller_info, src_idx); /* Try to propagate bits if src_lattice is bottom, but jfunc is known. for eg consider: int f(int x) { g (x & 0xff); } Assume lattice for x is bottom, however we can still propagate result of x & 0xff == 0xff, which gets computed during ccp1 pass and we store it in jump function during analysis stage. */ if (!src_lats->bits_lattice.bottom_p ()) { bool drop_all_ones = keep_null && !src_lats->bits_lattice.known_nonzero_p (); return dest_lattice->meet_with (src_lats->bits_lattice, precision, sgn, code, operand, drop_all_ones); } } Value_Range vr (parm_type); if (jfunc->m_vr) { jfunc->m_vr->get_vrange (vr); if (!vr.undefined_p () && !vr.varying_p ()) { irange &r = as_a (vr); irange_bitmask bm = r.get_bitmask (); widest_int mask = widest_int::from (bm.mask (), TYPE_SIGN (parm_type)); widest_int value = widest_int::from (bm.value (), TYPE_SIGN (parm_type)); return dest_lattice->meet_with (value, mask, precision); } } return dest_lattice->set_to_bottom (); } /* Propagate value range across jump function JFUNC that is associated with edge CS with param of callee of PARAM_TYPE and update DEST_PLATS accordingly. */ static bool propagate_vr_across_jump_function (cgraph_edge *cs, ipa_jump_func *jfunc, class ipcp_param_lattices *dest_plats, tree param_type) { ipcp_vr_lattice *dest_lat = &dest_plats->m_value_range; if (dest_lat->bottom_p ()) return false; if (!param_type || (!INTEGRAL_TYPE_P (param_type) && !POINTER_TYPE_P (param_type))) return dest_lat->set_to_bottom (); if (jfunc->type == IPA_JF_PASS_THROUGH) { enum tree_code operation = ipa_get_jf_pass_through_operation (jfunc); ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); int src_idx = ipa_get_jf_pass_through_formal_id (jfunc); class ipcp_param_lattices *src_lats = ipa_get_parm_lattices (caller_info, src_idx); tree operand_type = ipa_get_type (caller_info, src_idx); if (src_lats->m_value_range.bottom_p ()) return dest_lat->set_to_bottom (); Value_Range vr (operand_type); if (TREE_CODE_CLASS (operation) == tcc_unary) ipa_vr_operation_and_type_effects (vr, src_lats->m_value_range.m_vr, operation, param_type, operand_type); /* A crude way to prevent unbounded number of value range updates in SCC components. We should allow limited number of updates within SCC, too. */ else if (!ipa_edge_within_scc (cs)) { tree op = ipa_get_jf_pass_through_operand (jfunc); Value_Range op_vr (TREE_TYPE (op)); Value_Range op_res (operand_type); range_op_handler handler (operation); ipa_range_set_and_normalize (op_vr, op); if (!handler || !op_res.supports_type_p (operand_type) || !handler.fold_range (op_res, operand_type, src_lats->m_value_range.m_vr, op_vr)) op_res.set_varying (operand_type); ipa_vr_operation_and_type_effects (vr, op_res, NOP_EXPR, param_type, operand_type); } if (!vr.undefined_p () && !vr.varying_p ()) { if (jfunc->m_vr) { Value_Range jvr (param_type); if (ipa_vr_operation_and_type_effects (jvr, *jfunc->m_vr, NOP_EXPR, param_type, jfunc->m_vr->type ())) vr.intersect (jvr); } return dest_lat->meet_with (vr); } } else if (jfunc->type == IPA_JF_CONST) { tree val = ipa_get_jf_constant (jfunc); if (TREE_CODE (val) == INTEGER_CST) { val = fold_convert (param_type, val); if (TREE_OVERFLOW_P (val)) val = drop_tree_overflow (val); Value_Range tmpvr (val, val); return dest_lat->meet_with (tmpvr); } } Value_Range vr (param_type); if (jfunc->m_vr && ipa_vr_operation_and_type_effects (vr, *jfunc->m_vr, NOP_EXPR, param_type, jfunc->m_vr->type ())) return dest_lat->meet_with (vr); else return dest_lat->set_to_bottom (); } /* If DEST_PLATS already has aggregate items, check that aggs_by_ref matches NEW_AGGS_BY_REF and if not, mark all aggs as bottoms and return true (in all other cases, return false). If there are no aggregate items, set aggs_by_ref to NEW_AGGS_BY_REF. */ static bool set_check_aggs_by_ref (class ipcp_param_lattices *dest_plats, bool new_aggs_by_ref) { if (dest_plats->aggs) { if (dest_plats->aggs_by_ref != new_aggs_by_ref) { set_agg_lats_to_bottom (dest_plats); return true; } } else dest_plats->aggs_by_ref = new_aggs_by_ref; return false; } /* Walk aggregate lattices in DEST_PLATS from ***AGLAT on, until ***aglat is an already existing lattice for the given OFFSET and SIZE, marking all skipped lattices as containing variable and checking for overlaps. If there is no already existing lattice for the OFFSET and VAL_SIZE, create one, initialize it with offset, size and contains_variable to PRE_EXISTING, and return true, unless there are too many already. If there are two many, return false. If there are overlaps turn whole DEST_PLATS to bottom and return false. If any skipped lattices were newly marked as containing variable, set *CHANGE to true. MAX_AGG_ITEMS is the maximum number of lattices. */ static bool merge_agg_lats_step (class ipcp_param_lattices *dest_plats, HOST_WIDE_INT offset, HOST_WIDE_INT val_size, struct ipcp_agg_lattice ***aglat, bool pre_existing, bool *change, int max_agg_items) { gcc_checking_assert (offset >= 0); while (**aglat && (**aglat)->offset < offset) { if ((**aglat)->offset + (**aglat)->size > offset) { set_agg_lats_to_bottom (dest_plats); return false; } *change |= (**aglat)->set_contains_variable (); *aglat = &(**aglat)->next; } if (**aglat && (**aglat)->offset == offset) { if ((**aglat)->size != val_size) { set_agg_lats_to_bottom (dest_plats); return false; } gcc_assert (!(**aglat)->next || (**aglat)->next->offset >= offset + val_size); return true; } else { struct ipcp_agg_lattice *new_al; if (**aglat && (**aglat)->offset < offset + val_size) { set_agg_lats_to_bottom (dest_plats); return false; } if (dest_plats->aggs_count == max_agg_items) return false; dest_plats->aggs_count++; new_al = ipcp_agg_lattice_pool.allocate (); memset (new_al, 0, sizeof (*new_al)); new_al->offset = offset; new_al->size = val_size; new_al->contains_variable = pre_existing; new_al->next = **aglat; **aglat = new_al; return true; } } /* Set all AGLAT and all other aggregate lattices reachable by next pointers as containing an unknown value. */ static bool set_chain_of_aglats_contains_variable (struct ipcp_agg_lattice *aglat) { bool ret = false; while (aglat) { ret |= aglat->set_contains_variable (); aglat = aglat->next; } return ret; } /* Merge existing aggregate lattices in SRC_PLATS to DEST_PLATS, subtracting DELTA_OFFSET. CS is the call graph edge and SRC_IDX the index of the source parameter used for lattice value sources. Return true if DEST_PLATS changed in any way. */ static bool merge_aggregate_lattices (struct cgraph_edge *cs, class ipcp_param_lattices *dest_plats, class ipcp_param_lattices *src_plats, int src_idx, HOST_WIDE_INT offset_delta) { bool pre_existing = dest_plats->aggs != NULL; struct ipcp_agg_lattice **dst_aglat; bool ret = false; if (set_check_aggs_by_ref (dest_plats, src_plats->aggs_by_ref)) return true; if (src_plats->aggs_bottom) return set_agg_lats_contain_variable (dest_plats); if (src_plats->aggs_contain_variable) ret |= set_agg_lats_contain_variable (dest_plats); dst_aglat = &dest_plats->aggs; int max_agg_items = opt_for_fn (cs->callee->function_symbol ()->decl, param_ipa_max_agg_items); for (struct ipcp_agg_lattice *src_aglat = src_plats->aggs; src_aglat; src_aglat = src_aglat->next) { HOST_WIDE_INT new_offset = src_aglat->offset - offset_delta; if (new_offset < 0) continue; if (merge_agg_lats_step (dest_plats, new_offset, src_aglat->size, &dst_aglat, pre_existing, &ret, max_agg_items)) { struct ipcp_agg_lattice *new_al = *dst_aglat; dst_aglat = &(*dst_aglat)->next; if (src_aglat->bottom) { ret |= new_al->set_contains_variable (); continue; } if (src_aglat->contains_variable) ret |= new_al->set_contains_variable (); for (ipcp_value *val = src_aglat->values; val; val = val->next) ret |= new_al->add_value (val->value, cs, val, src_idx, src_aglat->offset); } else if (dest_plats->aggs_bottom) return true; } ret |= set_chain_of_aglats_contains_variable (*dst_aglat); return ret; } /* Determine whether there is anything to propagate FROM SRC_PLATS through a pass-through JFUNC and if so, whether it has conform and conforms to the rules about propagating values passed by reference. */ static bool agg_pass_through_permissible_p (class ipcp_param_lattices *src_plats, struct ipa_jump_func *jfunc) { return src_plats->aggs && (!src_plats->aggs_by_ref || ipa_get_jf_pass_through_agg_preserved (jfunc)); } /* Propagate values through ITEM, jump function for a part of an aggregate, into corresponding aggregate lattice AGLAT. CS is the call graph edge associated with the jump function. Return true if AGLAT changed in any way. */ static bool propagate_aggregate_lattice (struct cgraph_edge *cs, struct ipa_agg_jf_item *item, struct ipcp_agg_lattice *aglat) { class ipa_node_params *caller_info; class ipcp_param_lattices *src_plats; struct ipcp_lattice *src_lat; HOST_WIDE_INT src_offset; int src_idx; tree load_type; bool ret; if (item->jftype == IPA_JF_CONST) { tree value = item->value.constant; gcc_checking_assert (is_gimple_ip_invariant (value)); return aglat->add_value (value, cs, NULL, 0); } gcc_checking_assert (item->jftype == IPA_JF_PASS_THROUGH || item->jftype == IPA_JF_LOAD_AGG); caller_info = ipa_node_params_sum->get (cs->caller); src_idx = item->value.pass_through.formal_id; src_plats = ipa_get_parm_lattices (caller_info, src_idx); if (item->jftype == IPA_JF_PASS_THROUGH) { load_type = NULL_TREE; src_lat = &src_plats->itself; src_offset = -1; } else { HOST_WIDE_INT load_offset = item->value.load_agg.offset; struct ipcp_agg_lattice *src_aglat; for (src_aglat = src_plats->aggs; src_aglat; src_aglat = src_aglat->next) if (src_aglat->offset >= load_offset) break; load_type = item->value.load_agg.type; if (!src_aglat || src_aglat->offset > load_offset || src_aglat->size != tree_to_shwi (TYPE_SIZE (load_type)) || src_plats->aggs_by_ref != item->value.load_agg.by_ref) return aglat->set_contains_variable (); src_lat = src_aglat; src_offset = load_offset; } if (src_lat->bottom || (!ipcp_versionable_function_p (cs->caller) && !src_lat->is_single_const ())) return aglat->set_contains_variable (); ret = propagate_vals_across_arith_jfunc (cs, item->value.pass_through.operation, load_type, item->value.pass_through.operand, src_lat, aglat, src_offset, src_idx, item->type); if (src_lat->contains_variable) ret |= aglat->set_contains_variable (); return ret; } /* Propagate scalar values across jump function JFUNC that is associated with edge CS and put the values into DEST_LAT. */ static bool propagate_aggs_across_jump_function (struct cgraph_edge *cs, struct ipa_jump_func *jfunc, class ipcp_param_lattices *dest_plats) { bool ret = false; if (dest_plats->aggs_bottom) return false; if (jfunc->type == IPA_JF_PASS_THROUGH && ipa_get_jf_pass_through_operation (jfunc) == NOP_EXPR) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); int src_idx = ipa_get_jf_pass_through_formal_id (jfunc); class ipcp_param_lattices *src_plats; src_plats = ipa_get_parm_lattices (caller_info, src_idx); if (agg_pass_through_permissible_p (src_plats, jfunc)) { /* Currently we do not produce clobber aggregate jump functions, replace with merging when we do. */ gcc_assert (!jfunc->agg.items); ret |= merge_aggregate_lattices (cs, dest_plats, src_plats, src_idx, 0); return ret; } } else if (jfunc->type == IPA_JF_ANCESTOR && ipa_get_jf_ancestor_agg_preserved (jfunc)) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); int src_idx = ipa_get_jf_ancestor_formal_id (jfunc); class ipcp_param_lattices *src_plats; src_plats = ipa_get_parm_lattices (caller_info, src_idx); if (src_plats->aggs && src_plats->aggs_by_ref) { /* Currently we do not produce clobber aggregate jump functions, replace with merging when we do. */ gcc_assert (!jfunc->agg.items); ret |= merge_aggregate_lattices (cs, dest_plats, src_plats, src_idx, ipa_get_jf_ancestor_offset (jfunc)); } else if (!src_plats->aggs_by_ref) ret |= set_agg_lats_to_bottom (dest_plats); else ret |= set_agg_lats_contain_variable (dest_plats); return ret; } if (jfunc->agg.items) { bool pre_existing = dest_plats->aggs != NULL; struct ipcp_agg_lattice **aglat = &dest_plats->aggs; struct ipa_agg_jf_item *item; int i; if (set_check_aggs_by_ref (dest_plats, jfunc->agg.by_ref)) return true; int max_agg_items = opt_for_fn (cs->callee->function_symbol ()->decl, param_ipa_max_agg_items); FOR_EACH_VEC_ELT (*jfunc->agg.items, i, item) { HOST_WIDE_INT val_size; if (item->offset < 0 || item->jftype == IPA_JF_UNKNOWN) continue; val_size = tree_to_shwi (TYPE_SIZE (item->type)); if (merge_agg_lats_step (dest_plats, item->offset, val_size, &aglat, pre_existing, &ret, max_agg_items)) { ret |= propagate_aggregate_lattice (cs, item, *aglat); aglat = &(*aglat)->next; } else if (dest_plats->aggs_bottom) return true; } ret |= set_chain_of_aglats_contains_variable (*aglat); } else ret |= set_agg_lats_contain_variable (dest_plats); return ret; } /* Return true if on the way cfrom CS->caller to the final (non-alias and non-thunk) destination, the call passes through a thunk. */ static bool call_passes_through_thunk (cgraph_edge *cs) { cgraph_node *alias_or_thunk = cs->callee; while (alias_or_thunk->alias) alias_or_thunk = alias_or_thunk->get_alias_target (); return alias_or_thunk->thunk; } /* Propagate constants from the caller to the callee of CS. INFO describes the caller. */ static bool propagate_constants_across_call (struct cgraph_edge *cs) { class ipa_node_params *callee_info; enum availability availability; cgraph_node *callee; class ipa_edge_args *args; bool ret = false; int i, args_count, parms_count; callee = cs->callee->function_symbol (&availability); if (!callee->definition) return false; gcc_checking_assert (callee->has_gimple_body_p ()); callee_info = ipa_node_params_sum->get (callee); if (!callee_info) return false; args = ipa_edge_args_sum->get (cs); parms_count = ipa_get_param_count (callee_info); if (parms_count == 0) return false; if (!args || !opt_for_fn (cs->caller->decl, flag_ipa_cp) || !opt_for_fn (cs->caller->decl, optimize)) { for (i = 0; i < parms_count; i++) ret |= set_all_contains_variable (ipa_get_parm_lattices (callee_info, i)); return ret; } args_count = ipa_get_cs_argument_count (args); /* If this call goes through a thunk we must not propagate to the first (0th) parameter. However, we might need to uncover a thunk from below a series of aliases first. */ if (call_passes_through_thunk (cs)) { ret |= set_all_contains_variable (ipa_get_parm_lattices (callee_info, 0)); i = 1; } else i = 0; for (; (i < args_count) && (i < parms_count); i++) { struct ipa_jump_func *jump_func = ipa_get_ith_jump_func (args, i); class ipcp_param_lattices *dest_plats; tree param_type = ipa_get_type (callee_info, i); dest_plats = ipa_get_parm_lattices (callee_info, i); if (availability == AVAIL_INTERPOSABLE) ret |= set_all_contains_variable (dest_plats); else { ret |= propagate_scalar_across_jump_function (cs, jump_func, &dest_plats->itself, param_type); ret |= propagate_context_across_jump_function (cs, jump_func, i, &dest_plats->ctxlat); ret |= propagate_bits_across_jump_function (cs, i, jump_func, &dest_plats->bits_lattice); ret |= propagate_aggs_across_jump_function (cs, jump_func, dest_plats); if (opt_for_fn (callee->decl, flag_ipa_vrp)) ret |= propagate_vr_across_jump_function (cs, jump_func, dest_plats, param_type); else ret |= dest_plats->m_value_range.set_to_bottom (); } } for (; i < parms_count; i++) ret |= set_all_contains_variable (ipa_get_parm_lattices (callee_info, i)); return ret; } /* If an indirect edge IE can be turned into a direct one based on KNOWN_VALS KNOWN_CONTEXTS, and known aggregates either in AVS or KNOWN_AGGS return the destination. The latter three can be NULL. If AGG_REPS is not NULL, KNOWN_AGGS is ignored. */ static tree ipa_get_indirect_edge_target_1 (struct cgraph_edge *ie, const vec &known_csts, const vec &known_contexts, const ipa_argagg_value_list &avs, bool *speculative) { int param_index = ie->indirect_info->param_index; HOST_WIDE_INT anc_offset; tree t = NULL; tree target = NULL; *speculative = false; if (param_index == -1) return NULL_TREE; if (!ie->indirect_info->polymorphic) { tree t = NULL; if (ie->indirect_info->agg_contents) { t = NULL; if ((unsigned) param_index < known_csts.length () && known_csts[param_index]) t = ipa_find_agg_cst_from_init (known_csts[param_index], ie->indirect_info->offset, ie->indirect_info->by_ref); if (!t && ie->indirect_info->guaranteed_unmodified) t = avs.get_value (param_index, ie->indirect_info->offset / BITS_PER_UNIT, ie->indirect_info->by_ref); } else if ((unsigned) param_index < known_csts.length ()) t = known_csts[param_index]; if (t && TREE_CODE (t) == ADDR_EXPR && TREE_CODE (TREE_OPERAND (t, 0)) == FUNCTION_DECL) return TREE_OPERAND (t, 0); else return NULL_TREE; } if (!opt_for_fn (ie->caller->decl, flag_devirtualize)) return NULL_TREE; gcc_assert (!ie->indirect_info->agg_contents); gcc_assert (!ie->indirect_info->by_ref); anc_offset = ie->indirect_info->offset; t = NULL; if ((unsigned) param_index < known_csts.length () && known_csts[param_index]) t = ipa_find_agg_cst_from_init (known_csts[param_index], ie->indirect_info->offset, true); /* Try to work out value of virtual table pointer value in replacements. */ /* or known aggregate values. */ if (!t) t = avs.get_value (param_index, ie->indirect_info->offset / BITS_PER_UNIT, true); /* If we found the virtual table pointer, lookup the target. */ if (t) { tree vtable; unsigned HOST_WIDE_INT offset; if (vtable_pointer_value_to_vtable (t, &vtable, &offset)) { bool can_refer; target = gimple_get_virt_method_for_vtable (ie->indirect_info->otr_token, vtable, offset, &can_refer); if (can_refer) { if (!target || fndecl_built_in_p (target, BUILT_IN_UNREACHABLE) || !possible_polymorphic_call_target_p (ie, cgraph_node::get (target))) { /* Do not speculate builtin_unreachable, it is stupid! */ if (ie->indirect_info->vptr_changed) return NULL; target = ipa_impossible_devirt_target (ie, target); } *speculative = ie->indirect_info->vptr_changed; if (!*speculative) return target; } } } /* Do we know the constant value of pointer? */ if (!t && (unsigned) param_index < known_csts.length ()) t = known_csts[param_index]; gcc_checking_assert (!t || TREE_CODE (t) != TREE_BINFO); ipa_polymorphic_call_context context; if (known_contexts.length () > (unsigned int) param_index) { context = known_contexts[param_index]; context.offset_by (anc_offset); if (ie->indirect_info->vptr_changed) context.possible_dynamic_type_change (ie->in_polymorphic_cdtor, ie->indirect_info->otr_type); if (t) { ipa_polymorphic_call_context ctx2 = ipa_polymorphic_call_context (t, ie->indirect_info->otr_type, anc_offset); if (!ctx2.useless_p ()) context.combine_with (ctx2, ie->indirect_info->otr_type); } } else if (t) { context = ipa_polymorphic_call_context (t, ie->indirect_info->otr_type, anc_offset); if (ie->indirect_info->vptr_changed) context.possible_dynamic_type_change (ie->in_polymorphic_cdtor, ie->indirect_info->otr_type); } else return NULL_TREE; vec targets; bool final; targets = possible_polymorphic_call_targets (ie->indirect_info->otr_type, ie->indirect_info->otr_token, context, &final); if (!final || targets.length () > 1) { struct cgraph_node *node; if (*speculative) return target; if (!opt_for_fn (ie->caller->decl, flag_devirtualize_speculatively) || ie->speculative || !ie->maybe_hot_p ()) return NULL; node = try_speculative_devirtualization (ie->indirect_info->otr_type, ie->indirect_info->otr_token, context); if (node) { *speculative = true; target = node->decl; } else return NULL; } else { *speculative = false; if (targets.length () == 1) target = targets[0]->decl; else target = ipa_impossible_devirt_target (ie, NULL_TREE); } if (target && !possible_polymorphic_call_target_p (ie, cgraph_node::get (target))) { if (*speculative) return NULL; target = ipa_impossible_devirt_target (ie, target); } return target; } /* If an indirect edge IE can be turned into a direct one based on data in AVALS, return the destination. Store into *SPECULATIVE a boolean determinig whether the discovered target is only speculative guess. */ tree ipa_get_indirect_edge_target (struct cgraph_edge *ie, ipa_call_arg_values *avals, bool *speculative) { ipa_argagg_value_list avl (avals); return ipa_get_indirect_edge_target_1 (ie, avals->m_known_vals, avals->m_known_contexts, avl, speculative); } /* Calculate devirtualization time bonus for NODE, assuming we know information about arguments stored in AVALS. */ static int devirtualization_time_bonus (struct cgraph_node *node, ipa_auto_call_arg_values *avals) { struct cgraph_edge *ie; int res = 0; for (ie = node->indirect_calls; ie; ie = ie->next_callee) { struct cgraph_node *callee; class ipa_fn_summary *isummary; enum availability avail; tree target; bool speculative; ipa_argagg_value_list avl (avals); target = ipa_get_indirect_edge_target_1 (ie, avals->m_known_vals, avals->m_known_contexts, avl, &speculative); if (!target) continue; /* Only bare minimum benefit for clearly un-inlineable targets. */ res += 1; callee = cgraph_node::get (target); if (!callee || !callee->definition) continue; callee = callee->function_symbol (&avail); if (avail < AVAIL_AVAILABLE) continue; isummary = ipa_fn_summaries->get (callee); if (!isummary || !isummary->inlinable) continue; int size = ipa_size_summaries->get (callee)->size; /* FIXME: The values below need re-considering and perhaps also integrating into the cost metrics, at lest in some very basic way. */ int max_inline_insns_auto = opt_for_fn (callee->decl, param_max_inline_insns_auto); if (size <= max_inline_insns_auto / 4) res += 31 / ((int)speculative + 1); else if (size <= max_inline_insns_auto / 2) res += 15 / ((int)speculative + 1); else if (size <= max_inline_insns_auto || DECL_DECLARED_INLINE_P (callee->decl)) res += 7 / ((int)speculative + 1); } return res; } /* Return time bonus incurred because of hints stored in ESTIMATES. */ static int hint_time_bonus (cgraph_node *node, const ipa_call_estimates &estimates) { int result = 0; ipa_hints hints = estimates.hints; if (hints & (INLINE_HINT_loop_iterations | INLINE_HINT_loop_stride)) result += opt_for_fn (node->decl, param_ipa_cp_loop_hint_bonus); sreal bonus_for_one = opt_for_fn (node->decl, param_ipa_cp_loop_hint_bonus); if (hints & INLINE_HINT_loop_iterations) result += (estimates.loops_with_known_iterations * bonus_for_one).to_int (); if (hints & INLINE_HINT_loop_stride) result += (estimates.loops_with_known_strides * bonus_for_one).to_int (); return result; } /* If there is a reason to penalize the function described by INFO in the cloning goodness evaluation, do so. */ static inline sreal incorporate_penalties (cgraph_node *node, ipa_node_params *info, sreal evaluation) { if (info->node_within_scc && !info->node_is_self_scc) evaluation = (evaluation * (100 - opt_for_fn (node->decl, param_ipa_cp_recursion_penalty))) / 100; if (info->node_calling_single_call) evaluation = (evaluation * (100 - opt_for_fn (node->decl, param_ipa_cp_single_call_penalty))) / 100; return evaluation; } /* Return true if cloning NODE is a good idea, given the estimated TIME_BENEFIT and SIZE_COST and with the sum of frequencies of incoming edges to the potential new clone in FREQUENCIES. */ static bool good_cloning_opportunity_p (struct cgraph_node *node, sreal time_benefit, sreal freq_sum, profile_count count_sum, int size_cost) { if (time_benefit == 0 || !opt_for_fn (node->decl, flag_ipa_cp_clone) || node->optimize_for_size_p ()) return false; gcc_assert (size_cost > 0); ipa_node_params *info = ipa_node_params_sum->get (node); int eval_threshold = opt_for_fn (node->decl, param_ipa_cp_eval_threshold); if (count_sum.nonzero_p ()) { gcc_assert (base_count.nonzero_p ()); sreal factor = count_sum.probability_in (base_count).to_sreal (); sreal evaluation = (time_benefit * factor) / size_cost; evaluation = incorporate_penalties (node, info, evaluation); evaluation *= 1000; if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " good_cloning_opportunity_p (time: %g, " "size: %i, count_sum: ", time_benefit.to_double (), size_cost); count_sum.dump (dump_file); fprintf (dump_file, "%s%s) -> evaluation: %.2f, threshold: %i\n", info->node_within_scc ? (info->node_is_self_scc ? ", self_scc" : ", scc") : "", info->node_calling_single_call ? ", single_call" : "", evaluation.to_double (), eval_threshold); } return evaluation.to_int () >= eval_threshold; } else { sreal evaluation = (time_benefit * freq_sum) / size_cost; evaluation = incorporate_penalties (node, info, evaluation); evaluation *= 1000; if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " good_cloning_opportunity_p (time: %g, " "size: %i, freq_sum: %g%s%s) -> evaluation: %.2f, " "threshold: %i\n", time_benefit.to_double (), size_cost, freq_sum.to_double (), info->node_within_scc ? (info->node_is_self_scc ? ", self_scc" : ", scc") : "", info->node_calling_single_call ? ", single_call" : "", evaluation.to_double (), eval_threshold); return evaluation.to_int () >= eval_threshold; } } /* Grow vectors in AVALS and fill them with information about values of parameters that are known to be independent of the context. Only calculate m_known_aggs if CALCULATE_AGGS is true. INFO describes the function. If REMOVABLE_PARAMS_COST is non-NULL, the movement cost of all removable parameters will be stored in it. TODO: Also grow context independent value range vectors. */ static bool gather_context_independent_values (class ipa_node_params *info, ipa_auto_call_arg_values *avals, bool calculate_aggs, int *removable_params_cost) { int i, count = ipa_get_param_count (info); bool ret = false; avals->m_known_vals.safe_grow_cleared (count, true); avals->m_known_contexts.safe_grow_cleared (count, true); if (removable_params_cost) *removable_params_cost = 0; for (i = 0; i < count; i++) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); ipcp_lattice *lat = &plats->itself; if (lat->is_single_const ()) { ipcp_value *val = lat->values; gcc_checking_assert (TREE_CODE (val->value) != TREE_BINFO); avals->m_known_vals[i] = val->value; if (removable_params_cost) *removable_params_cost += estimate_move_cost (TREE_TYPE (val->value), false); ret = true; } else if (removable_params_cost && !ipa_is_param_used (info, i)) *removable_params_cost += ipa_get_param_move_cost (info, i); if (!ipa_is_param_used (info, i)) continue; ipcp_lattice *ctxlat = &plats->ctxlat; /* Do not account known context as reason for cloning. We can see if it permits devirtualization. */ if (ctxlat->is_single_const ()) avals->m_known_contexts[i] = ctxlat->values->value; if (calculate_aggs) ret |= push_agg_values_from_plats (plats, i, 0, &avals->m_known_aggs); } return ret; } /* Perform time and size measurement of NODE with the context given in AVALS, calculate the benefit compared to the node without specialization and store it into VAL. Take into account REMOVABLE_PARAMS_COST of all context-independent or unused removable parameters and EST_MOVE_COST, the estimated movement of the considered parameter. */ static void perform_estimation_of_a_value (cgraph_node *node, ipa_auto_call_arg_values *avals, int removable_params_cost, int est_move_cost, ipcp_value_base *val) { sreal time_benefit; ipa_call_estimates estimates; estimate_ipcp_clone_size_and_time (node, avals, &estimates); /* Extern inline functions have no cloning local time benefits because they will be inlined anyway. The only reason to clone them is if it enables optimization in any of the functions they call. */ if (DECL_EXTERNAL (node->decl) && DECL_DECLARED_INLINE_P (node->decl)) time_benefit = 0; else time_benefit = (estimates.nonspecialized_time - estimates.time) + (devirtualization_time_bonus (node, avals) + hint_time_bonus (node, estimates) + removable_params_cost + est_move_cost); int size = estimates.size; gcc_checking_assert (size >=0); /* The inliner-heuristics based estimates may think that in certain contexts some functions do not have any size at all but we want all specializations to have at least a tiny cost, not least not to divide by zero. */ if (size == 0) size = 1; val->local_time_benefit = time_benefit; val->local_size_cost = size; } /* Get the overall limit oof growth based on parameters extracted from growth. it does not really make sense to mix functions with different overall growth limits but it is possible and if it happens, we do not want to select one limit at random. */ static long get_max_overall_size (cgraph_node *node) { long max_new_size = orig_overall_size; long large_unit = opt_for_fn (node->decl, param_ipa_cp_large_unit_insns); if (max_new_size < large_unit) max_new_size = large_unit; int unit_growth = opt_for_fn (node->decl, param_ipa_cp_unit_growth); max_new_size += max_new_size * unit_growth / 100 + 1; return max_new_size; } /* Return true if NODE should be cloned just for a parameter removal, possibly dumping a reason if not. */ static bool clone_for_param_removal_p (cgraph_node *node) { if (!node->can_change_signature) { if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Not considering cloning to remove parameters, " "function cannot change signature.\n"); return false; } if (node->can_be_local_p ()) { if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Not considering cloning to remove parameters, " "IPA-SRA can do it potentially better.\n"); return false; } return true; } /* Iterate over known values of parameters of NODE and estimate the local effects in terms of time and size they have. */ static void estimate_local_effects (struct cgraph_node *node) { ipa_node_params *info = ipa_node_params_sum->get (node); int count = ipa_get_param_count (info); bool always_const; int removable_params_cost; if (!count || !ipcp_versionable_function_p (node)) return; if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, "\nEstimating effects for %s.\n", node->dump_name ()); ipa_auto_call_arg_values avals; always_const = gather_context_independent_values (info, &avals, true, &removable_params_cost); int devirt_bonus = devirtualization_time_bonus (node, &avals); if (always_const || devirt_bonus || (removable_params_cost && clone_for_param_removal_p (node))) { struct caller_statistics stats; ipa_call_estimates estimates; init_caller_stats (&stats); node->call_for_symbol_thunks_and_aliases (gather_caller_stats, &stats, false); estimate_ipcp_clone_size_and_time (node, &avals, &estimates); sreal time = estimates.nonspecialized_time - estimates.time; time += devirt_bonus; time += hint_time_bonus (node, estimates); time += removable_params_cost; int size = estimates.size - stats.n_calls * removable_params_cost; if (dump_file) fprintf (dump_file, " - context independent values, size: %i, " "time_benefit: %f\n", size, (time).to_double ()); if (size <= 0 || node->local) { info->do_clone_for_all_contexts = true; if (dump_file) fprintf (dump_file, " Decided to specialize for all " "known contexts, code not going to grow.\n"); } else if (good_cloning_opportunity_p (node, time, stats.freq_sum, stats.count_sum, size)) { if (size + overall_size <= get_max_overall_size (node)) { info->do_clone_for_all_contexts = true; overall_size += size; if (dump_file) fprintf (dump_file, " Decided to specialize for all " "known contexts, growth (to %li) deemed " "beneficial.\n", overall_size); } else if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Not cloning for all contexts because " "maximum unit size would be reached with %li.\n", size + overall_size); } else if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Not cloning for all contexts because " "!good_cloning_opportunity_p.\n"); } for (int i = 0; i < count; i++) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); ipcp_lattice *lat = &plats->itself; ipcp_value *val; if (lat->bottom || !lat->values || avals.m_known_vals[i]) continue; for (val = lat->values; val; val = val->next) { gcc_checking_assert (TREE_CODE (val->value) != TREE_BINFO); avals.m_known_vals[i] = val->value; int emc = estimate_move_cost (TREE_TYPE (val->value), true); perform_estimation_of_a_value (node, &avals, removable_params_cost, emc, val); if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " - estimates for value "); print_ipcp_constant_value (dump_file, val->value); fprintf (dump_file, " for "); ipa_dump_param (dump_file, info, i); fprintf (dump_file, ": time_benefit: %g, size: %i\n", val->local_time_benefit.to_double (), val->local_size_cost); } } avals.m_known_vals[i] = NULL_TREE; } for (int i = 0; i < count; i++) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); if (!plats->virt_call) continue; ipcp_lattice *ctxlat = &plats->ctxlat; ipcp_value *val; if (ctxlat->bottom || !ctxlat->values || !avals.m_known_contexts[i].useless_p ()) continue; for (val = ctxlat->values; val; val = val->next) { avals.m_known_contexts[i] = val->value; perform_estimation_of_a_value (node, &avals, removable_params_cost, 0, val); if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " - estimates for polymorphic context "); print_ipcp_constant_value (dump_file, val->value); fprintf (dump_file, " for "); ipa_dump_param (dump_file, info, i); fprintf (dump_file, ": time_benefit: %g, size: %i\n", val->local_time_benefit.to_double (), val->local_size_cost); } } avals.m_known_contexts[i] = ipa_polymorphic_call_context (); } unsigned all_ctx_len = avals.m_known_aggs.length (); auto_vec all_ctx; all_ctx.reserve_exact (all_ctx_len); all_ctx.splice (avals.m_known_aggs); avals.m_known_aggs.safe_grow_cleared (all_ctx_len + 1); unsigned j = 0; for (int index = 0; index < count; index++) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, index); if (plats->aggs_bottom || !plats->aggs) continue; for (ipcp_agg_lattice *aglat = plats->aggs; aglat; aglat = aglat->next) { ipcp_value *val; if (aglat->bottom || !aglat->values /* If the following is true, the one value is already part of all context estimations. */ || (!plats->aggs_contain_variable && aglat->is_single_const ())) continue; unsigned unit_offset = aglat->offset / BITS_PER_UNIT; while (j < all_ctx_len && (all_ctx[j].index < index || (all_ctx[j].index == index && all_ctx[j].unit_offset < unit_offset))) { avals.m_known_aggs[j] = all_ctx[j]; j++; } for (unsigned k = j; k < all_ctx_len; k++) avals.m_known_aggs[k+1] = all_ctx[k]; for (val = aglat->values; val; val = val->next) { avals.m_known_aggs[j].value = val->value; avals.m_known_aggs[j].unit_offset = unit_offset; avals.m_known_aggs[j].index = index; avals.m_known_aggs[j].by_ref = plats->aggs_by_ref; perform_estimation_of_a_value (node, &avals, removable_params_cost, 0, val); if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " - estimates for value "); print_ipcp_constant_value (dump_file, val->value); fprintf (dump_file, " for "); ipa_dump_param (dump_file, info, index); fprintf (dump_file, "[%soffset: " HOST_WIDE_INT_PRINT_DEC "]: time_benefit: %g, size: %i\n", plats->aggs_by_ref ? "ref " : "", aglat->offset, val->local_time_benefit.to_double (), val->local_size_cost); } } } } } /* Add value CUR_VAL and all yet-unsorted values it is dependent on to the topological sort of values. */ template void value_topo_info::add_val (ipcp_value *cur_val) { ipcp_value_source *src; if (cur_val->dfs) return; dfs_counter++; cur_val->dfs = dfs_counter; cur_val->low_link = dfs_counter; cur_val->topo_next = stack; stack = cur_val; cur_val->on_stack = true; for (src = cur_val->sources; src; src = src->next) if (src->val) { if (src->val->dfs == 0) { add_val (src->val); if (src->val->low_link < cur_val->low_link) cur_val->low_link = src->val->low_link; } else if (src->val->on_stack && src->val->dfs < cur_val->low_link) cur_val->low_link = src->val->dfs; } if (cur_val->dfs == cur_val->low_link) { ipcp_value *v, *scc_list = NULL; do { v = stack; stack = v->topo_next; v->on_stack = false; v->scc_no = cur_val->dfs; v->scc_next = scc_list; scc_list = v; } while (v != cur_val); cur_val->topo_next = values_topo; values_topo = cur_val; } } /* Add all values in lattices associated with NODE to the topological sort if they are not there yet. */ static void add_all_node_vals_to_toposort (cgraph_node *node, ipa_topo_info *topo) { ipa_node_params *info = ipa_node_params_sum->get (node); int i, count = ipa_get_param_count (info); for (i = 0; i < count; i++) { class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); ipcp_lattice *lat = &plats->itself; struct ipcp_agg_lattice *aglat; if (!lat->bottom) { ipcp_value *val; for (val = lat->values; val; val = val->next) topo->constants.add_val (val); } if (!plats->aggs_bottom) for (aglat = plats->aggs; aglat; aglat = aglat->next) if (!aglat->bottom) { ipcp_value *val; for (val = aglat->values; val; val = val->next) topo->constants.add_val (val); } ipcp_lattice *ctxlat = &plats->ctxlat; if (!ctxlat->bottom) { ipcp_value *ctxval; for (ctxval = ctxlat->values; ctxval; ctxval = ctxval->next) topo->contexts.add_val (ctxval); } } } /* One pass of constants propagation along the call graph edges, from callers to callees (requires topological ordering in TOPO), iterate over strongly connected components. */ static void propagate_constants_topo (class ipa_topo_info *topo) { int i; for (i = topo->nnodes - 1; i >= 0; i--) { unsigned j; struct cgraph_node *v, *node = topo->order[i]; vec cycle_nodes = ipa_get_nodes_in_cycle (node); /* First, iteratively propagate within the strongly connected component until all lattices stabilize. */ FOR_EACH_VEC_ELT (cycle_nodes, j, v) if (v->has_gimple_body_p ()) { if (opt_for_fn (v->decl, flag_ipa_cp) && opt_for_fn (v->decl, optimize)) push_node_to_stack (topo, v); /* When V is not optimized, we can not push it to stack, but still we need to set all its callees lattices to bottom. */ else { for (cgraph_edge *cs = v->callees; cs; cs = cs->next_callee) propagate_constants_across_call (cs); } } v = pop_node_from_stack (topo); while (v) { struct cgraph_edge *cs; class ipa_node_params *info = NULL; bool self_scc = true; for (cs = v->callees; cs; cs = cs->next_callee) if (ipa_edge_within_scc (cs)) { cgraph_node *callee = cs->callee->function_symbol (); if (v != callee) self_scc = false; if (!info) { info = ipa_node_params_sum->get (v); info->node_within_scc = true; } if (propagate_constants_across_call (cs)) push_node_to_stack (topo, callee); } if (info) info->node_is_self_scc = self_scc; v = pop_node_from_stack (topo); } /* Afterwards, propagate along edges leading out of the SCC, calculates the local effects of the discovered constants and all valid values to their topological sort. */ FOR_EACH_VEC_ELT (cycle_nodes, j, v) if (v->has_gimple_body_p () && opt_for_fn (v->decl, flag_ipa_cp) && opt_for_fn (v->decl, optimize)) { struct cgraph_edge *cs; estimate_local_effects (v); add_all_node_vals_to_toposort (v, topo); for (cs = v->callees; cs; cs = cs->next_callee) if (!ipa_edge_within_scc (cs)) propagate_constants_across_call (cs); } cycle_nodes.release (); } } /* Propagate the estimated effects of individual values along the topological from the dependent values to those they depend on. */ template void value_topo_info::propagate_effects () { ipcp_value *base; hash_set *> processed_srcvals; for (base = values_topo; base; base = base->topo_next) { ipcp_value_source *src; ipcp_value *val; sreal time = 0; HOST_WIDE_INT size = 0; for (val = base; val; val = val->scc_next) { time = time + val->local_time_benefit + val->prop_time_benefit; size = size + val->local_size_cost + val->prop_size_cost; } for (val = base; val; val = val->scc_next) { processed_srcvals.empty (); for (src = val->sources; src; src = src->next) if (src->val && src->cs->maybe_hot_p ()) { if (!processed_srcvals.add (src->val)) { HOST_WIDE_INT prop_size = size + src->val->prop_size_cost; if (prop_size < INT_MAX) src->val->prop_size_cost = prop_size; else continue; } int special_factor = 1; if (val->same_scc (src->val)) special_factor = opt_for_fn(src->cs->caller->decl, param_ipa_cp_recursive_freq_factor); else if (val->self_recursion_generated_p () && (src->cs->callee->function_symbol () == src->cs->caller)) { int max_recur_gen_depth = opt_for_fn(src->cs->caller->decl, param_ipa_cp_max_recursive_depth); special_factor = max_recur_gen_depth - val->self_recursion_generated_level + 1; } src->val->prop_time_benefit += time * special_factor * src->cs->sreal_frequency (); } if (size < INT_MAX) { val->prop_time_benefit = time; val->prop_size_cost = size; } else { val->prop_time_benefit = 0; val->prop_size_cost = 0; } } } } /* Callback for qsort to sort counts of all edges. */ static int compare_edge_profile_counts (const void *a, const void *b) { const profile_count *cnt1 = (const profile_count *) a; const profile_count *cnt2 = (const profile_count *) b; if (*cnt1 < *cnt2) return 1; if (*cnt1 > *cnt2) return -1; return 0; } /* Propagate constants, polymorphic contexts and their effects from the summaries interprocedurally. */ static void ipcp_propagate_stage (class ipa_topo_info *topo) { struct cgraph_node *node; if (dump_file) fprintf (dump_file, "\n Propagating constants:\n\n"); base_count = profile_count::uninitialized (); bool compute_count_base = false; unsigned base_count_pos_percent = 0; FOR_EACH_DEFINED_FUNCTION (node) { if (node->has_gimple_body_p () && opt_for_fn (node->decl, flag_ipa_cp) && opt_for_fn (node->decl, optimize)) { ipa_node_params *info = ipa_node_params_sum->get (node); determine_versionability (node, info); unsigned nlattices = ipa_get_param_count (info); void *chunk = XCNEWVEC (class ipcp_param_lattices, nlattices); info->lattices = new (chunk) ipcp_param_lattices[nlattices]; initialize_node_lattices (node); } ipa_size_summary *s = ipa_size_summaries->get (node); if (node->definition && !node->alias && s != NULL) overall_size += s->self_size; if (node->count.ipa ().initialized_p ()) { compute_count_base = true; unsigned pos_percent = opt_for_fn (node->decl, param_ipa_cp_profile_count_base); base_count_pos_percent = MAX (base_count_pos_percent, pos_percent); } } if (compute_count_base) { auto_vec all_edge_counts; all_edge_counts.reserve_exact (symtab->edges_count); FOR_EACH_DEFINED_FUNCTION (node) for (cgraph_edge *cs = node->callees; cs; cs = cs->next_callee) { profile_count count = cs->count.ipa (); if (!count.nonzero_p ()) continue; enum availability avail; cgraph_node *tgt = cs->callee->function_or_virtual_thunk_symbol (&avail); ipa_node_params *info = ipa_node_params_sum->get (tgt); if (info && info->versionable) all_edge_counts.quick_push (count); } if (!all_edge_counts.is_empty ()) { gcc_assert (base_count_pos_percent <= 100); all_edge_counts.qsort (compare_edge_profile_counts); unsigned base_count_pos = ((all_edge_counts.length () * (base_count_pos_percent)) / 100); base_count = all_edge_counts[base_count_pos]; if (dump_file) { fprintf (dump_file, "\nSelected base_count from %u edges at " "position %u, arriving at: ", all_edge_counts.length (), base_count_pos); base_count.dump (dump_file); fprintf (dump_file, "\n"); } } else if (dump_file) fprintf (dump_file, "\nNo candidates with non-zero call count found, " "continuing as if without profile feedback.\n"); } orig_overall_size = overall_size; if (dump_file) fprintf (dump_file, "\noverall_size: %li\n", overall_size); propagate_constants_topo (topo); if (flag_checking) ipcp_verify_propagated_values (); topo->constants.propagate_effects (); topo->contexts.propagate_effects (); if (dump_file) { fprintf (dump_file, "\nIPA lattices after all propagation:\n"); print_all_lattices (dump_file, (dump_flags & TDF_DETAILS), true); } } /* Discover newly direct outgoing edges from NODE which is a new clone with known KNOWN_CSTS and make them direct. */ static void ipcp_discover_new_direct_edges (struct cgraph_node *node, vec known_csts, vec known_contexts, vec *aggvals) { struct cgraph_edge *ie, *next_ie; bool found = false; for (ie = node->indirect_calls; ie; ie = next_ie) { tree target; bool speculative; next_ie = ie->next_callee; ipa_argagg_value_list avs (aggvals); target = ipa_get_indirect_edge_target_1 (ie, known_csts, known_contexts, avs, &speculative); if (target) { bool agg_contents = ie->indirect_info->agg_contents; bool polymorphic = ie->indirect_info->polymorphic; int param_index = ie->indirect_info->param_index; struct cgraph_edge *cs = ipa_make_edge_direct_to_target (ie, target, speculative); found = true; if (cs && !agg_contents && !polymorphic) { ipa_node_params *info = ipa_node_params_sum->get (node); int c = ipa_get_controlled_uses (info, param_index); if (c != IPA_UNDESCRIBED_USE && !ipa_get_param_load_dereferenced (info, param_index)) { struct ipa_ref *to_del; c--; ipa_set_controlled_uses (info, param_index, c); if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " controlled uses count of param " "%i bumped down to %i\n", param_index, c); if (c == 0 && (to_del = node->find_reference (cs->callee, NULL, 0, IPA_REF_ADDR))) { if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " and even removing its " "cloning-created reference\n"); to_del->remove_reference (); } } } } } /* Turning calls to direct calls will improve overall summary. */ if (found) ipa_update_overall_fn_summary (node); } class edge_clone_summary; static call_summary *edge_clone_summaries = NULL; /* Edge clone summary. */ class edge_clone_summary { public: /* Default constructor. */ edge_clone_summary (): prev_clone (NULL), next_clone (NULL) {} /* Default destructor. */ ~edge_clone_summary () { if (prev_clone) edge_clone_summaries->get (prev_clone)->next_clone = next_clone; if (next_clone) edge_clone_summaries->get (next_clone)->prev_clone = prev_clone; } cgraph_edge *prev_clone; cgraph_edge *next_clone; }; class edge_clone_summary_t: public call_summary { public: edge_clone_summary_t (symbol_table *symtab): call_summary (symtab) { m_initialize_when_cloning = true; } void duplicate (cgraph_edge *src_edge, cgraph_edge *dst_edge, edge_clone_summary *src_data, edge_clone_summary *dst_data) final override; }; /* Edge duplication hook. */ void edge_clone_summary_t::duplicate (cgraph_edge *src_edge, cgraph_edge *dst_edge, edge_clone_summary *src_data, edge_clone_summary *dst_data) { if (src_data->next_clone) edge_clone_summaries->get (src_data->next_clone)->prev_clone = dst_edge; dst_data->prev_clone = src_edge; dst_data->next_clone = src_data->next_clone; src_data->next_clone = dst_edge; } /* Return true is CS calls DEST or its clone for all contexts. When ALLOW_RECURSION_TO_CLONE is false, also return false for self-recursive edges from/to an all-context clone. */ static bool calls_same_node_or_its_all_contexts_clone_p (cgraph_edge *cs, cgraph_node *dest, bool allow_recursion_to_clone) { enum availability availability; cgraph_node *callee = cs->callee->function_symbol (&availability); if (availability <= AVAIL_INTERPOSABLE) return false; if (callee == dest) return true; if (!allow_recursion_to_clone && cs->caller == callee) return false; ipa_node_params *info = ipa_node_params_sum->get (callee); return info->is_all_contexts_clone && info->ipcp_orig_node == dest; } /* Return true if edge CS does bring about the value described by SRC to DEST_VAL of node DEST or its clone for all contexts. */ static bool cgraph_edge_brings_value_p (cgraph_edge *cs, ipcp_value_source *src, cgraph_node *dest, ipcp_value *dest_val) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); if (!calls_same_node_or_its_all_contexts_clone_p (cs, dest, !src->val) || caller_info->node_dead) return false; if (!src->val) return true; if (caller_info->ipcp_orig_node) { tree t = NULL_TREE; if (src->offset == -1) t = caller_info->known_csts[src->index]; else if (ipcp_transformation *ts = ipcp_get_transformation_summary (cs->caller)) { ipa_argagg_value_list avl (ts); t = avl.get_value (src->index, src->offset / BITS_PER_UNIT); } return (t != NULL_TREE && values_equal_for_ipcp_p (src->val->value, t)); } else { if (src->val == dest_val) return true; struct ipcp_agg_lattice *aglat; class ipcp_param_lattices *plats = ipa_get_parm_lattices (caller_info, src->index); if (src->offset == -1) return (plats->itself.is_single_const () && values_equal_for_ipcp_p (src->val->value, plats->itself.values->value)); else { if (plats->aggs_bottom || plats->aggs_contain_variable) return false; for (aglat = plats->aggs; aglat; aglat = aglat->next) if (aglat->offset == src->offset) return (aglat->is_single_const () && values_equal_for_ipcp_p (src->val->value, aglat->values->value)); } return false; } } /* Return true if edge CS does bring about the value described by SRC to DST_VAL of node DEST or its clone for all contexts. */ static bool cgraph_edge_brings_value_p (cgraph_edge *cs, ipcp_value_source *src, cgraph_node *dest, ipcp_value *) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); if (!calls_same_node_or_its_all_contexts_clone_p (cs, dest, true) || caller_info->node_dead) return false; if (!src->val) return true; if (caller_info->ipcp_orig_node) return (caller_info->known_contexts.length () > (unsigned) src->index) && values_equal_for_ipcp_p (src->val->value, caller_info->known_contexts[src->index]); class ipcp_param_lattices *plats = ipa_get_parm_lattices (caller_info, src->index); return plats->ctxlat.is_single_const () && values_equal_for_ipcp_p (src->val->value, plats->ctxlat.values->value); } /* Get the next clone in the linked list of clones of an edge. */ static inline struct cgraph_edge * get_next_cgraph_edge_clone (struct cgraph_edge *cs) { edge_clone_summary *s = edge_clone_summaries->get (cs); return s != NULL ? s->next_clone : NULL; } /* Given VAL that is intended for DEST, iterate over all its sources and if any of them is viable and hot, return true. In that case, for those that still hold, add their edge frequency and their number and cumulative profile counts of self-ecursive and other edges into *FREQUENCY, *CALLER_COUNT, REC_COUNT_SUM and NONREC_COUNT_SUM respectively. */ template static bool get_info_about_necessary_edges (ipcp_value *val, cgraph_node *dest, sreal *freq_sum, int *caller_count, profile_count *rec_count_sum, profile_count *nonrec_count_sum) { ipcp_value_source *src; sreal freq = 0; int count = 0; profile_count rec_cnt = profile_count::zero (); profile_count nonrec_cnt = profile_count::zero (); bool hot = false; bool non_self_recursive = false; for (src = val->sources; src; src = src->next) { struct cgraph_edge *cs = src->cs; while (cs) { if (cgraph_edge_brings_value_p (cs, src, dest, val)) { count++; freq += cs->sreal_frequency (); hot |= cs->maybe_hot_p (); if (cs->caller != dest) { non_self_recursive = true; if (cs->count.ipa ().initialized_p ()) rec_cnt += cs->count.ipa (); } else if (cs->count.ipa ().initialized_p ()) nonrec_cnt += cs->count.ipa (); } cs = get_next_cgraph_edge_clone (cs); } } /* If the only edges bringing a value are self-recursive ones, do not bother evaluating it. */ if (!non_self_recursive) return false; *freq_sum = freq; *caller_count = count; *rec_count_sum = rec_cnt; *nonrec_count_sum = nonrec_cnt; if (!hot && ipa_node_params_sum->get (dest)->node_within_scc) { struct cgraph_edge *cs; /* Cold non-SCC source edge could trigger hot recursive execution of function. Consider the case as hot and rely on following cost model computation to further select right one. */ for (cs = dest->callers; cs; cs = cs->next_caller) if (cs->caller == dest && cs->maybe_hot_p ()) return true; } return hot; } /* Given a NODE, and a set of its CALLERS, try to adjust order of the callers to let a non-self-recursive caller be the first element. Thus, we can simplify intersecting operations on values that arrive from all of these callers, especially when there exists self-recursive call. Return true if this kind of adjustment is possible. */ static bool adjust_callers_for_value_intersection (vec &callers, cgraph_node *node) { for (unsigned i = 0; i < callers.length (); i++) { cgraph_edge *cs = callers[i]; if (cs->caller != node) { if (i > 0) { callers[i] = callers[0]; callers[0] = cs; } return true; } } return false; } /* Return a vector of incoming edges that do bring value VAL to node DEST. It is assumed their number is known and equal to CALLER_COUNT. */ template static vec gather_edges_for_value (ipcp_value *val, cgraph_node *dest, int caller_count) { ipcp_value_source *src; vec ret; ret.create (caller_count); for (src = val->sources; src; src = src->next) { struct cgraph_edge *cs = src->cs; while (cs) { if (cgraph_edge_brings_value_p (cs, src, dest, val)) ret.quick_push (cs); cs = get_next_cgraph_edge_clone (cs); } } if (caller_count > 1) adjust_callers_for_value_intersection (ret, dest); return ret; } /* Construct a replacement map for a know VALUE for a formal parameter PARAM. Return it or NULL if for some reason it cannot be created. FORCE_LOAD_REF should be set to true when the reference created for the constant should be a load one and not an address one because the corresponding parameter p is only used as *p. */ static struct ipa_replace_map * get_replacement_map (class ipa_node_params *info, tree value, int parm_num, bool force_load_ref) { struct ipa_replace_map *replace_map; replace_map = ggc_alloc (); if (dump_file) { fprintf (dump_file, " replacing "); ipa_dump_param (dump_file, info, parm_num); fprintf (dump_file, " with const "); print_generic_expr (dump_file, value); if (force_load_ref) fprintf (dump_file, " - forcing load reference\n"); else fprintf (dump_file, "\n"); } replace_map->parm_num = parm_num; replace_map->new_tree = value; replace_map->force_load_ref = force_load_ref; return replace_map; } /* Dump new profiling counts of NODE. SPEC is true when NODE is a specialzied one, otherwise it will be referred to as the original node. */ static void dump_profile_updates (cgraph_node *node, bool spec) { if (spec) fprintf (dump_file, " setting count of the specialized node %s to ", node->dump_name ()); else fprintf (dump_file, " setting count of the original node %s to ", node->dump_name ()); node->count.dump (dump_file); fprintf (dump_file, "\n"); for (cgraph_edge *cs = node->callees; cs; cs = cs->next_callee) { fprintf (dump_file, " edge to %s has count ", cs->callee->dump_name ()); cs->count.dump (dump_file); fprintf (dump_file, "\n"); } } /* With partial train run we do not want to assume that original's count is zero whenever we redurect all executed edges to clone. Simply drop profile to local one in this case. In eany case, return the new value. ORIG_NODE is the original node and its count has not been updaed yet. */ profile_count lenient_count_portion_handling (profile_count remainder, cgraph_node *orig_node) { if (remainder.ipa_p () && !remainder.ipa ().nonzero_p () && orig_node->count.ipa_p () && orig_node->count.ipa ().nonzero_p () && opt_for_fn (orig_node->decl, flag_profile_partial_training)) remainder = remainder.guessed_local (); return remainder; } /* Structure to sum counts coming from nodes other than the original node and its clones. */ struct gather_other_count_struct { cgraph_node *orig; profile_count other_count; }; /* Worker callback of call_for_symbol_thunks_and_aliases summing the number of counts that come from non-self-recursive calls.. */ static bool gather_count_of_non_rec_edges (cgraph_node *node, void *data) { gather_other_count_struct *desc = (gather_other_count_struct *) data; for (cgraph_edge *cs = node->callers; cs; cs = cs->next_caller) if (cs->caller != desc->orig && cs->caller->clone_of != desc->orig) desc->other_count += cs->count.ipa (); return false; } /* Structure to help analyze if we need to boost counts of some clones of some non-recursive edges to match the new callee count. */ struct desc_incoming_count_struct { cgraph_node *orig; hash_set *processed_edges; profile_count count; unsigned unproc_orig_rec_edges; }; /* Go over edges calling NODE and its thunks and gather information about incoming counts so that we know if we need to make any adjustments. */ static void analyze_clone_icoming_counts (cgraph_node *node, desc_incoming_count_struct *desc) { for (cgraph_edge *cs = node->callers; cs; cs = cs->next_caller) if (cs->caller->thunk) { analyze_clone_icoming_counts (cs->caller, desc); continue; } else { if (cs->count.initialized_p ()) desc->count += cs->count.ipa (); if (!desc->processed_edges->contains (cs) && cs->caller->clone_of == desc->orig) desc->unproc_orig_rec_edges++; } } /* If caller edge counts of a clone created for a self-recursive arithmetic jump function must be adjusted because it is coming from a the "seed" clone for the first value and so has been excessively scaled back as if it was not a recursive call, adjust it so that the incoming counts of NODE match its count. NODE is the node or its thunk. */ static void adjust_clone_incoming_counts (cgraph_node *node, desc_incoming_count_struct *desc) { for (cgraph_edge *cs = node->callers; cs; cs = cs->next_caller) if (cs->caller->thunk) { adjust_clone_incoming_counts (cs->caller, desc); profile_count sum = profile_count::zero (); for (cgraph_edge *e = cs->caller->callers; e; e = e->next_caller) if (e->count.initialized_p ()) sum += e->count.ipa (); cs->count = cs->count.combine_with_ipa_count (sum); } else if (!desc->processed_edges->contains (cs) && cs->caller->clone_of == desc->orig) { cs->count += desc->count; if (dump_file) { fprintf (dump_file, " Adjusted count of an incoming edge of " "a clone %s -> %s to ", cs->caller->dump_name (), cs->callee->dump_name ()); cs->count.dump (dump_file); fprintf (dump_file, "\n"); } } } /* When ORIG_NODE has been cloned for values which have been generated fora self-recursive call as a result of an arithmetic pass-through jump-functions, adjust its count together with counts of all such clones in SELF_GEN_CLONES which also at this point contains ORIG_NODE itself. The function sums the counts of the original node and all its clones that cannot be attributed to a specific clone because it comes from a non-recursive edge. This sum is then evenly divided between the clones and on top of that each one gets all the counts which can be attributed directly to it. */ static void update_counts_for_self_gen_clones (cgraph_node *orig_node, const vec &self_gen_clones) { profile_count redist_sum = orig_node->count.ipa (); if (!(redist_sum > profile_count::zero ())) return; if (dump_file) fprintf (dump_file, " Updating profile of self recursive clone " "series\n"); gather_other_count_struct gocs; gocs.orig = orig_node; gocs.other_count = profile_count::zero (); auto_vec other_edges_count; for (cgraph_node *n : self_gen_clones) { gocs.other_count = profile_count::zero (); n->call_for_symbol_thunks_and_aliases (gather_count_of_non_rec_edges, &gocs, false); other_edges_count.safe_push (gocs.other_count); redist_sum -= gocs.other_count; } hash_set processed_edges; unsigned i = 0; for (cgraph_node *n : self_gen_clones) { profile_count orig_count = n->count; profile_count new_count = (redist_sum / self_gen_clones.length () + other_edges_count[i]); new_count = lenient_count_portion_handling (new_count, orig_node); n->count = new_count; profile_count::adjust_for_ipa_scaling (&new_count, &orig_count); for (cgraph_edge *cs = n->callees; cs; cs = cs->next_callee) { cs->count = cs->count.apply_scale (new_count, orig_count); processed_edges.add (cs); } for (cgraph_edge *cs = n->indirect_calls; cs; cs = cs->next_callee) cs->count = cs->count.apply_scale (new_count, orig_count); i++; } /* There are still going to be edges to ORIG_NODE that have one or more clones coming from another node clone in SELF_GEN_CLONES and which we scaled by the same amount, which means that the total incoming sum of counts to ORIG_NODE will be too high, scale such edges back. */ for (cgraph_edge *cs = orig_node->callees; cs; cs = cs->next_callee) { if (cs->callee->ultimate_alias_target () == orig_node) { unsigned den = 0; for (cgraph_edge *e = cs; e; e = get_next_cgraph_edge_clone (e)) if (e->callee->ultimate_alias_target () == orig_node && processed_edges.contains (e)) den++; if (den > 0) for (cgraph_edge *e = cs; e; e = get_next_cgraph_edge_clone (e)) if (e->callee->ultimate_alias_target () == orig_node && processed_edges.contains (e)) e->count /= den; } } /* Edges from the seeds of the valus generated for arithmetic jump-functions along self-recursive edges are likely to have fairly low count and so edges from them to nodes in the self_gen_clones do not correspond to the artificially distributed count of the nodes, the total sum of incoming edges to some clones might be too low. Detect this situation and correct it. */ for (cgraph_node *n : self_gen_clones) { if (!(n->count.ipa () > profile_count::zero ())) continue; desc_incoming_count_struct desc; desc.orig = orig_node; desc.processed_edges = &processed_edges; desc.count = profile_count::zero (); desc.unproc_orig_rec_edges = 0; analyze_clone_icoming_counts (n, &desc); if (n->count.differs_from_p (desc.count)) { if (n->count > desc.count && desc.unproc_orig_rec_edges > 0) { desc.count = n->count - desc.count; desc.count = desc.count /= desc.unproc_orig_rec_edges; adjust_clone_incoming_counts (n, &desc); } else if (dump_file) fprintf (dump_file, " Unable to fix up incoming counts for %s.\n", n->dump_name ()); } } if (dump_file) for (cgraph_node *n : self_gen_clones) dump_profile_updates (n, n != orig_node); return; } /* After a specialized NEW_NODE version of ORIG_NODE has been created, update their profile information to reflect this. This function should not be used for clones generated for arithmetic pass-through jump functions on a self-recursive call graph edge, that situation is handled by update_counts_for_self_gen_clones. */ static void update_profiling_info (struct cgraph_node *orig_node, struct cgraph_node *new_node) { struct caller_statistics stats; profile_count new_sum; profile_count remainder, orig_node_count = orig_node->count.ipa (); if (!(orig_node_count > profile_count::zero ())) return; if (dump_file) { fprintf (dump_file, " Updating profile from original count: "); orig_node_count.dump (dump_file); fprintf (dump_file, "\n"); } init_caller_stats (&stats, new_node); new_node->call_for_symbol_thunks_and_aliases (gather_caller_stats, &stats, false); new_sum = stats.count_sum; bool orig_edges_processed = false; if (new_sum > orig_node_count) { /* TODO: Profile has alreay gone astray, keep what we have but lower it to global0 category. */ remainder = orig_node->count.global0 (); for (cgraph_edge *cs = orig_node->callees; cs; cs = cs->next_callee) cs->count = cs->count.global0 (); for (cgraph_edge *cs = orig_node->indirect_calls; cs; cs = cs->next_callee) cs->count = cs->count.global0 (); orig_edges_processed = true; } else if (stats.rec_count_sum.nonzero_p ()) { int new_nonrec_calls = stats.n_nonrec_calls; /* There are self-recursive edges which are likely to bring in the majority of calls but which we must divide in between the original and new node. */ init_caller_stats (&stats, orig_node); orig_node->call_for_symbol_thunks_and_aliases (gather_caller_stats, &stats, false); int orig_nonrec_calls = stats.n_nonrec_calls; profile_count orig_nonrec_call_count = stats.count_sum; if (orig_node->local) { if (!orig_nonrec_call_count.nonzero_p ()) { if (dump_file) fprintf (dump_file, " The original is local and the only " "incoming edges from non-dead callers with nonzero " "counts are self-recursive, assuming it is cold.\n"); /* The NEW_NODE count and counts of all its outgoing edges are still unmodified copies of ORIG_NODE's. Just clear the latter and bail out. */ profile_count zero; if (opt_for_fn (orig_node->decl, flag_profile_partial_training)) zero = profile_count::zero ().guessed_local (); else zero = profile_count::adjusted_zero (); orig_node->count = zero; for (cgraph_edge *cs = orig_node->callees; cs; cs = cs->next_callee) cs->count = zero; for (cgraph_edge *cs = orig_node->indirect_calls; cs; cs = cs->next_callee) cs->count = zero; return; } } else { /* Let's behave as if there was another caller that accounts for all the calls that were either indirect or from other compilation units. */ orig_nonrec_calls++; profile_count pretend_caller_count = (orig_node_count - new_sum - orig_nonrec_call_count - stats.rec_count_sum); orig_nonrec_call_count += pretend_caller_count; } /* Divide all "unexplained" counts roughly proportionally to sums of counts of non-recursive calls. We put rather arbitrary limits on how many counts we claim because the number of non-self-recursive incoming count is only a rough guideline and there are cases (such as mcf) where using it blindly just takes too many. And if lattices are considered in the opposite order we could also take too few. */ profile_count unexp = orig_node_count - new_sum - orig_nonrec_call_count; int limit_den = 2 * (orig_nonrec_calls + new_nonrec_calls); profile_count new_part = MAX(MIN (unexp.apply_scale (new_sum, new_sum + orig_nonrec_call_count), unexp.apply_scale (limit_den - 1, limit_den)), unexp.apply_scale (new_nonrec_calls, limit_den)); if (dump_file) { fprintf (dump_file, " Claiming "); new_part.dump (dump_file); fprintf (dump_file, " of unexplained "); unexp.dump (dump_file); fprintf (dump_file, " counts because of self-recursive " "calls\n"); } new_sum += new_part; remainder = lenient_count_portion_handling (orig_node_count - new_sum, orig_node); } else remainder = lenient_count_portion_handling (orig_node_count - new_sum, orig_node); new_sum = orig_node_count.combine_with_ipa_count (new_sum); new_node->count = new_sum; orig_node->count = remainder; profile_count orig_new_node_count = orig_node_count; profile_count::adjust_for_ipa_scaling (&new_sum, &orig_new_node_count); for (cgraph_edge *cs = new_node->callees; cs; cs = cs->next_callee) cs->count = cs->count.apply_scale (new_sum, orig_new_node_count); for (cgraph_edge *cs = new_node->indirect_calls; cs; cs = cs->next_callee) cs->count = cs->count.apply_scale (new_sum, orig_new_node_count); if (!orig_edges_processed) { profile_count::adjust_for_ipa_scaling (&remainder, &orig_node_count); for (cgraph_edge *cs = orig_node->callees; cs; cs = cs->next_callee) cs->count = cs->count.apply_scale (remainder, orig_node_count); for (cgraph_edge *cs = orig_node->indirect_calls; cs; cs = cs->next_callee) cs->count = cs->count.apply_scale (remainder, orig_node_count); } if (dump_file) { dump_profile_updates (new_node, true); dump_profile_updates (orig_node, false); } } /* Update the respective profile of specialized NEW_NODE and the original ORIG_NODE after additional edges with cumulative count sum REDIRECTED_SUM have been redirected to the specialized version. */ static void update_specialized_profile (struct cgraph_node *new_node, struct cgraph_node *orig_node, profile_count redirected_sum) { struct cgraph_edge *cs; profile_count new_node_count, orig_node_count = orig_node->count.ipa (); if (dump_file) { fprintf (dump_file, " the sum of counts of redirected edges is "); redirected_sum.dump (dump_file); fprintf (dump_file, "\n old ipa count of the original node is "); orig_node_count.dump (dump_file); fprintf (dump_file, "\n"); } if (!(orig_node_count > profile_count::zero ())) return; new_node_count = new_node->count; new_node->count += redirected_sum; orig_node->count = lenient_count_portion_handling (orig_node->count - redirected_sum, orig_node); for (cs = new_node->callees; cs; cs = cs->next_callee) cs->count += cs->count.apply_scale (redirected_sum, new_node_count); for (cs = orig_node->callees; cs; cs = cs->next_callee) { profile_count dec = cs->count.apply_scale (redirected_sum, orig_node_count); cs->count -= dec; } if (dump_file) { dump_profile_updates (new_node, true); dump_profile_updates (orig_node, false); } } static void adjust_references_in_caller (cgraph_edge *cs, symtab_node *symbol, int index); /* Simple structure to pass a symbol and index (with same meaning as parameters of adjust_references_in_caller) through a void* parameter of a call_for_symbol_thunks_and_aliases callback. */ struct symbol_and_index_together { symtab_node *symbol; int index; }; /* Worker callback of call_for_symbol_thunks_and_aliases to recursively call adjust_references_in_caller on edges up in the call-graph, if necessary. */ static bool adjust_refs_in_act_callers (struct cgraph_node *node, void *data) { symbol_and_index_together *pack = (symbol_and_index_together *) data; for (cgraph_edge *cs = node->callers; cs; cs = cs->next_caller) if (!cs->caller->thunk) adjust_references_in_caller (cs, pack->symbol, pack->index); return false; } /* At INDEX of a function being called by CS there is an ADDR_EXPR of a variable which is only dereferenced and which is represented by SYMBOL. See if we can remove ADDR reference in callers assosiated witht the call. */ static void adjust_references_in_caller (cgraph_edge *cs, symtab_node *symbol, int index) { ipa_edge_args *args = ipa_edge_args_sum->get (cs); ipa_jump_func *jfunc = ipa_get_ith_jump_func (args, index); if (jfunc->type == IPA_JF_CONST) { ipa_ref *to_del = cs->caller->find_reference (symbol, cs->call_stmt, cs->lto_stmt_uid, IPA_REF_ADDR); if (!to_del) return; to_del->remove_reference (); ipa_zap_jf_refdesc (jfunc); if (dump_file) fprintf (dump_file, " Removed a reference from %s to %s.\n", cs->caller->dump_name (), symbol->dump_name ()); return; } if (jfunc->type != IPA_JF_PASS_THROUGH || ipa_get_jf_pass_through_operation (jfunc) != NOP_EXPR || ipa_get_jf_pass_through_refdesc_decremented (jfunc)) return; int fidx = ipa_get_jf_pass_through_formal_id (jfunc); cgraph_node *caller = cs->caller; ipa_node_params *caller_info = ipa_node_params_sum->get (caller); /* TODO: This consistency check may be too big and not really that useful. Consider removing it. */ tree cst; if (caller_info->ipcp_orig_node) cst = caller_info->known_csts[fidx]; else { ipcp_lattice *lat = ipa_get_scalar_lat (caller_info, fidx); gcc_assert (lat->is_single_const ()); cst = lat->values->value; } gcc_assert (TREE_CODE (cst) == ADDR_EXPR && (symtab_node::get (get_base_address (TREE_OPERAND (cst, 0))) == symbol)); int cuses = ipa_get_controlled_uses (caller_info, fidx); if (cuses == IPA_UNDESCRIBED_USE) return; gcc_assert (cuses > 0); cuses--; ipa_set_controlled_uses (caller_info, fidx, cuses); ipa_set_jf_pass_through_refdesc_decremented (jfunc, true); if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Controlled uses of parameter %i of %s dropped " "to %i.\n", fidx, caller->dump_name (), cuses); if (cuses) return; if (caller_info->ipcp_orig_node) { /* Cloning machinery has created a reference here, we need to either remove it or change it to a read one. */ ipa_ref *to_del = caller->find_reference (symbol, NULL, 0, IPA_REF_ADDR); if (to_del) { to_del->remove_reference (); if (dump_file) fprintf (dump_file, " Removed a reference from %s to %s.\n", cs->caller->dump_name (), symbol->dump_name ()); if (ipa_get_param_load_dereferenced (caller_info, fidx)) { caller->create_reference (symbol, IPA_REF_LOAD, NULL); if (dump_file) fprintf (dump_file, " ...and replaced it with LOAD one.\n"); } } } symbol_and_index_together pack; pack.symbol = symbol; pack.index = fidx; if (caller->can_change_signature) caller->call_for_symbol_thunks_and_aliases (adjust_refs_in_act_callers, &pack, true); } /* Return true if we would like to remove a parameter from NODE when cloning it with KNOWN_CSTS scalar constants. */ static bool want_remove_some_param_p (cgraph_node *node, vec known_csts) { auto_vec surviving; bool filled_vec = false; ipa_node_params *info = ipa_node_params_sum->get (node); int i, count = ipa_get_param_count (info); for (i = 0; i < count; i++) { if (!known_csts[i] && ipa_is_param_used (info, i)) continue; if (!filled_vec) { clone_info *info = clone_info::get (node); if (!info || !info->param_adjustments) return true; info->param_adjustments->get_surviving_params (&surviving); filled_vec = true; } if (surviving.length() < (unsigned) i && surviving[i]) return true; } return false; } /* Create a specialized version of NODE with known constants in KNOWN_CSTS, known contexts in KNOWN_CONTEXTS and known aggregate values in AGGVALS and redirect all edges in CALLERS to it. */ static struct cgraph_node * create_specialized_node (struct cgraph_node *node, vec known_csts, vec known_contexts, vec *aggvals, vec &callers) { ipa_node_params *new_info, *info = ipa_node_params_sum->get (node); vec *replace_trees = NULL; vec *new_params = NULL; struct cgraph_node *new_node; int i, count = ipa_get_param_count (info); clone_info *cinfo = clone_info::get (node); ipa_param_adjustments *old_adjustments = cinfo ? cinfo->param_adjustments : NULL; ipa_param_adjustments *new_adjustments; gcc_assert (!info->ipcp_orig_node); gcc_assert (node->can_change_signature || !old_adjustments); if (old_adjustments) { /* At the moment all IPA optimizations should use the number of parameters of the prevailing decl as the m_always_copy_start. Handling any other value would complicate the code below, so for the time bing let's only assert it is so. */ gcc_assert (old_adjustments->m_always_copy_start == count || old_adjustments->m_always_copy_start < 0); int old_adj_count = vec_safe_length (old_adjustments->m_adj_params); for (i = 0; i < old_adj_count; i++) { ipa_adjusted_param *old_adj = &(*old_adjustments->m_adj_params)[i]; if (!node->can_change_signature || old_adj->op != IPA_PARAM_OP_COPY || (!known_csts[old_adj->base_index] && ipa_is_param_used (info, old_adj->base_index))) { ipa_adjusted_param new_adj = *old_adj; new_adj.prev_clone_adjustment = true; new_adj.prev_clone_index = i; vec_safe_push (new_params, new_adj); } } bool skip_return = old_adjustments->m_skip_return; new_adjustments = (new (ggc_alloc ()) ipa_param_adjustments (new_params, count, skip_return)); } else if (node->can_change_signature && want_remove_some_param_p (node, known_csts)) { ipa_adjusted_param adj; memset (&adj, 0, sizeof (adj)); adj.op = IPA_PARAM_OP_COPY; for (i = 0; i < count; i++) if (!known_csts[i] && ipa_is_param_used (info, i)) { adj.base_index = i; adj.prev_clone_index = i; vec_safe_push (new_params, adj); } new_adjustments = (new (ggc_alloc ()) ipa_param_adjustments (new_params, count, false)); } else new_adjustments = NULL; auto_vec self_recursive_calls; for (i = callers.length () - 1; i >= 0; i--) { cgraph_edge *cs = callers[i]; if (cs->caller == node) { self_recursive_calls.safe_push (cs); callers.unordered_remove (i); } } replace_trees = cinfo ? vec_safe_copy (cinfo->tree_map) : NULL; for (i = 0; i < count; i++) { tree t = known_csts[i]; if (!t) continue; gcc_checking_assert (TREE_CODE (t) != TREE_BINFO); bool load_ref = false; symtab_node *ref_symbol; if (TREE_CODE (t) == ADDR_EXPR) { tree base = get_base_address (TREE_OPERAND (t, 0)); if (TREE_CODE (base) == VAR_DECL && ipa_get_controlled_uses (info, i) == 0 && ipa_get_param_load_dereferenced (info, i) && (ref_symbol = symtab_node::get (base))) { load_ref = true; if (node->can_change_signature) for (cgraph_edge *caller : callers) adjust_references_in_caller (caller, ref_symbol, i); } } ipa_replace_map *replace_map = get_replacement_map (info, t, i, load_ref); if (replace_map) vec_safe_push (replace_trees, replace_map); } unsigned &suffix_counter = clone_num_suffixes->get_or_insert ( IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME ( node->decl))); new_node = node->create_virtual_clone (callers, replace_trees, new_adjustments, "constprop", suffix_counter); suffix_counter++; bool have_self_recursive_calls = !self_recursive_calls.is_empty (); for (unsigned j = 0; j < self_recursive_calls.length (); j++) { cgraph_edge *cs = get_next_cgraph_edge_clone (self_recursive_calls[j]); /* Cloned edges can disappear during cloning as speculation can be resolved, check that we have one and that it comes from the last cloning. */ if (cs && cs->caller == new_node) cs->redirect_callee_duplicating_thunks (new_node); /* Any future code that would make more than one clone of an outgoing edge would confuse this mechanism, so let's check that does not happen. */ gcc_checking_assert (!cs || !get_next_cgraph_edge_clone (cs) || get_next_cgraph_edge_clone (cs)->caller != new_node); } if (have_self_recursive_calls) new_node->expand_all_artificial_thunks (); ipa_set_node_agg_value_chain (new_node, aggvals); for (const ipa_argagg_value &av : aggvals) new_node->maybe_create_reference (av.value, NULL); if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " the new node is %s.\n", new_node->dump_name ()); if (known_contexts.exists ()) { for (i = 0; i < count; i++) if (!known_contexts[i].useless_p ()) { fprintf (dump_file, " known ctx %i is ", i); known_contexts[i].dump (dump_file); } } if (aggvals) { fprintf (dump_file, " Aggregate replacements:"); ipa_argagg_value_list avs (aggvals); avs.dump (dump_file); } } new_info = ipa_node_params_sum->get (new_node); new_info->ipcp_orig_node = node; new_node->ipcp_clone = true; new_info->known_csts = known_csts; new_info->known_contexts = known_contexts; ipcp_discover_new_direct_edges (new_node, known_csts, known_contexts, aggvals); return new_node; } /* Return true if JFUNC, which describes a i-th parameter of call CS, is a pass-through function to itself when the cgraph_node involved is not an IPA-CP clone. When SIMPLE is true, further check if JFUNC is a simple no-operation pass-through. */ static bool self_recursive_pass_through_p (cgraph_edge *cs, ipa_jump_func *jfunc, int i, bool simple = true) { enum availability availability; if (cs->caller == cs->callee->function_symbol (&availability) && availability > AVAIL_INTERPOSABLE && jfunc->type == IPA_JF_PASS_THROUGH && (!simple || ipa_get_jf_pass_through_operation (jfunc) == NOP_EXPR) && ipa_get_jf_pass_through_formal_id (jfunc) == i && ipa_node_params_sum->get (cs->caller) && !ipa_node_params_sum->get (cs->caller)->ipcp_orig_node) return true; return false; } /* Return true if JFUNC, which describes a part of an aggregate represented or pointed to by the i-th parameter of call CS, is a pass-through function to itself when the cgraph_node involved is not an IPA-CP clone.. When SIMPLE is true, further check if JFUNC is a simple no-operation pass-through. */ static bool self_recursive_agg_pass_through_p (const cgraph_edge *cs, const ipa_agg_jf_item *jfunc, int i, bool simple = true) { enum availability availability; if (cs->caller == cs->callee->function_symbol (&availability) && availability > AVAIL_INTERPOSABLE && jfunc->jftype == IPA_JF_LOAD_AGG && jfunc->offset == jfunc->value.load_agg.offset && (!simple || jfunc->value.pass_through.operation == NOP_EXPR) && jfunc->value.pass_through.formal_id == i && useless_type_conversion_p (jfunc->value.load_agg.type, jfunc->type) && ipa_node_params_sum->get (cs->caller) && !ipa_node_params_sum->get (cs->caller)->ipcp_orig_node) return true; return false; } /* Given a NODE, and a subset of its CALLERS, try to populate blanks slots in KNOWN_CSTS with constants that are also known for all of the CALLERS. */ static void find_more_scalar_values_for_callers_subset (struct cgraph_node *node, vec &known_csts, const vec &callers) { ipa_node_params *info = ipa_node_params_sum->get (node); int i, count = ipa_get_param_count (info); for (i = 0; i < count; i++) { struct cgraph_edge *cs; tree newval = NULL_TREE; int j; bool first = true; tree type = ipa_get_type (info, i); if (ipa_get_scalar_lat (info, i)->bottom || known_csts[i]) continue; FOR_EACH_VEC_ELT (callers, j, cs) { struct ipa_jump_func *jump_func; tree t; ipa_edge_args *args = ipa_edge_args_sum->get (cs); if (!args || i >= ipa_get_cs_argument_count (args) || (i == 0 && call_passes_through_thunk (cs))) { newval = NULL_TREE; break; } jump_func = ipa_get_ith_jump_func (args, i); /* Besides simple pass-through jump function, arithmetic jump function could also introduce argument-direct-pass-through for self-feeding recursive call. For example, fn (int i) { fn (i & 1); } Given that i is 0, recursive propagation via (i & 1) also gets 0. */ if (self_recursive_pass_through_p (cs, jump_func, i, false)) { gcc_assert (newval); t = ipa_get_jf_arith_result ( ipa_get_jf_pass_through_operation (jump_func), newval, ipa_get_jf_pass_through_operand (jump_func), type); } else t = ipa_value_from_jfunc (ipa_node_params_sum->get (cs->caller), jump_func, type); if (!t || (newval && !values_equal_for_ipcp_p (t, newval)) || (!first && !newval)) { newval = NULL_TREE; break; } else newval = t; first = false; } if (newval) { if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " adding an extra known scalar value "); print_ipcp_constant_value (dump_file, newval); fprintf (dump_file, " for "); ipa_dump_param (dump_file, info, i); fprintf (dump_file, "\n"); } known_csts[i] = newval; } } } /* Given a NODE and a subset of its CALLERS, try to populate plank slots in KNOWN_CONTEXTS with polymorphic contexts that are also known for all of the CALLERS. */ static void find_more_contexts_for_caller_subset (cgraph_node *node, vec *known_contexts, const vec &callers) { ipa_node_params *info = ipa_node_params_sum->get (node); int i, count = ipa_get_param_count (info); for (i = 0; i < count; i++) { cgraph_edge *cs; if (ipa_get_poly_ctx_lat (info, i)->bottom || (known_contexts->exists () && !(*known_contexts)[i].useless_p ())) continue; ipa_polymorphic_call_context newval; bool first = true; int j; FOR_EACH_VEC_ELT (callers, j, cs) { ipa_edge_args *args = ipa_edge_args_sum->get (cs); if (!args || i >= ipa_get_cs_argument_count (args)) return; ipa_jump_func *jfunc = ipa_get_ith_jump_func (args, i); ipa_polymorphic_call_context ctx; ctx = ipa_context_from_jfunc (ipa_node_params_sum->get (cs->caller), cs, i, jfunc); if (first) { newval = ctx; first = false; } else newval.meet_with (ctx); if (newval.useless_p ()) break; } if (!newval.useless_p ()) { if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " adding an extra known polymorphic " "context "); print_ipcp_constant_value (dump_file, newval); fprintf (dump_file, " for "); ipa_dump_param (dump_file, info, i); fprintf (dump_file, "\n"); } if (!known_contexts->exists ()) known_contexts->safe_grow_cleared (ipa_get_param_count (info), true); (*known_contexts)[i] = newval; } } } /* Push all aggregate values coming along edge CS for parameter number INDEX to RES. If INTERIM is non-NULL, it contains the current interim state of collected aggregate values which can be used to compute values passed over self-recursive edges. This basically one iteration of push_agg_values_from_edge over one parameter, which allows for simpler early returns. */ static void push_agg_values_for_index_from_edge (struct cgraph_edge *cs, int index, vec *res, const ipa_argagg_value_list *interim) { bool agg_values_from_caller = false; bool agg_jf_preserved = false; unsigned unit_delta = UINT_MAX; int src_idx = -1; ipa_jump_func *jfunc = ipa_get_ith_jump_func (ipa_edge_args_sum->get (cs), index); if (jfunc->type == IPA_JF_PASS_THROUGH && ipa_get_jf_pass_through_operation (jfunc) == NOP_EXPR) { agg_values_from_caller = true; agg_jf_preserved = ipa_get_jf_pass_through_agg_preserved (jfunc); src_idx = ipa_get_jf_pass_through_formal_id (jfunc); unit_delta = 0; } else if (jfunc->type == IPA_JF_ANCESTOR && ipa_get_jf_ancestor_agg_preserved (jfunc)) { agg_values_from_caller = true; agg_jf_preserved = true; src_idx = ipa_get_jf_ancestor_formal_id (jfunc); unit_delta = ipa_get_jf_ancestor_offset (jfunc) / BITS_PER_UNIT; } ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); if (agg_values_from_caller) { if (caller_info->ipcp_orig_node) { struct cgraph_node *orig_node = caller_info->ipcp_orig_node; ipcp_transformation *ts = ipcp_get_transformation_summary (cs->caller); ipa_node_params *orig_info = ipa_node_params_sum->get (orig_node); ipcp_param_lattices *orig_plats = ipa_get_parm_lattices (orig_info, src_idx); if (ts && orig_plats->aggs && (agg_jf_preserved || !orig_plats->aggs_by_ref)) { ipa_argagg_value_list src (ts); src.push_adjusted_values (src_idx, index, unit_delta, res); return; } } else { ipcp_param_lattices *src_plats = ipa_get_parm_lattices (caller_info, src_idx); if (src_plats->aggs && !src_plats->aggs_bottom && (agg_jf_preserved || !src_plats->aggs_by_ref)) { if (interim && self_recursive_pass_through_p (cs, jfunc, index)) { interim->push_adjusted_values (src_idx, index, unit_delta, res); return; } if (!src_plats->aggs_contain_variable) { push_agg_values_from_plats (src_plats, index, unit_delta, res); return; } } } } if (!jfunc->agg.items) return; bool first = true; unsigned prev_unit_offset = 0; for (const ipa_agg_jf_item &agg_jf : *jfunc->agg.items) { tree value, srcvalue; /* Besides simple pass-through aggregate jump function, arithmetic aggregate jump function could also bring same aggregate value as parameter passed-in for self-feeding recursive call. For example, fn (int *i) { int j = *i & 1; fn (&j); } Given that *i is 0, recursive propagation via (*i & 1) also gets 0. */ if (interim && self_recursive_agg_pass_through_p (cs, &agg_jf, index, false) && (srcvalue = interim->get_value(index, agg_jf.offset / BITS_PER_UNIT))) value = ipa_get_jf_arith_result (agg_jf.value.pass_through.operation, srcvalue, agg_jf.value.pass_through.operand, agg_jf.type); else value = ipa_agg_value_from_jfunc (caller_info, cs->caller, &agg_jf); if (value) { struct ipa_argagg_value iav; iav.value = value; iav.unit_offset = agg_jf.offset / BITS_PER_UNIT; iav.index = index; iav.by_ref = jfunc->agg.by_ref; gcc_assert (first || iav.unit_offset > prev_unit_offset); prev_unit_offset = iav.unit_offset; first = false; res->safe_push (iav); } } return; } /* Push all aggregate values coming along edge CS to RES. DEST_INFO is the description of ultimate callee of CS or the one it was cloned from (the summary where lattices are). If INTERIM is non-NULL, it contains the current interim state of collected aggregate values which can be used to compute values passed over self-recursive edges (if OPTIMIZE_SELF_RECURSION is true) and to skip values which clearly will not be part of intersection with INTERIM. */ static void push_agg_values_from_edge (struct cgraph_edge *cs, ipa_node_params *dest_info, vec *res, const ipa_argagg_value_list *interim, bool optimize_self_recursion) { ipa_edge_args *args = ipa_edge_args_sum->get (cs); if (!args) return; int count = MIN (ipa_get_param_count (dest_info), ipa_get_cs_argument_count (args)); unsigned interim_index = 0; for (int index = 0; index < count; index++) { if (interim) { while (interim_index < interim->m_elts.size () && interim->m_elts[interim_index].value && interim->m_elts[interim_index].index < index) interim_index++; if (interim_index >= interim->m_elts.size () || interim->m_elts[interim_index].index > index) continue; } ipcp_param_lattices *plats = ipa_get_parm_lattices (dest_info, index); if (!ipa_is_param_used (dest_info, index) || plats->aggs_bottom) continue; push_agg_values_for_index_from_edge (cs, index, res, optimize_self_recursion ? interim : NULL); } } /* Look at edges in CALLERS and collect all known aggregate values that arrive from all of them. Return nullptr if there are none. */ static struct vec * find_aggregate_values_for_callers_subset (struct cgraph_node *node, const vec &callers) { ipa_node_params *dest_info = ipa_node_params_sum->get (node); if (dest_info->ipcp_orig_node) dest_info = ipa_node_params_sum->get (dest_info->ipcp_orig_node); /* gather_edges_for_value puts a non-recursive call into the first element of callers if it can. */ auto_vec interim; push_agg_values_from_edge (callers[0], dest_info, &interim, NULL, true); unsigned valid_entries = interim.length (); if (!valid_entries) return nullptr; unsigned caller_count = callers.length(); for (unsigned i = 1; i < caller_count; i++) { auto_vec last; ipa_argagg_value_list avs (&interim); push_agg_values_from_edge (callers[i], dest_info, &last, &avs, true); valid_entries = intersect_argaggs_with (interim, last); if (!valid_entries) return nullptr; } vec *res = NULL; vec_safe_reserve_exact (res, valid_entries); for (const ipa_argagg_value &av : interim) if (av.value) res->quick_push(av); gcc_checking_assert (res->length () == valid_entries); return res; } /* Determine whether CS also brings all scalar values that the NODE is specialized for. */ static bool cgraph_edge_brings_all_scalars_for_node (struct cgraph_edge *cs, struct cgraph_node *node) { ipa_node_params *dest_info = ipa_node_params_sum->get (node); int count = ipa_get_param_count (dest_info); class ipa_node_params *caller_info; class ipa_edge_args *args; int i; caller_info = ipa_node_params_sum->get (cs->caller); args = ipa_edge_args_sum->get (cs); for (i = 0; i < count; i++) { struct ipa_jump_func *jump_func; tree val, t; val = dest_info->known_csts[i]; if (!val) continue; if (i >= ipa_get_cs_argument_count (args)) return false; jump_func = ipa_get_ith_jump_func (args, i); t = ipa_value_from_jfunc (caller_info, jump_func, ipa_get_type (dest_info, i)); if (!t || !values_equal_for_ipcp_p (val, t)) return false; } return true; } /* Determine whether CS also brings all aggregate values that NODE is specialized for. */ static bool cgraph_edge_brings_all_agg_vals_for_node (struct cgraph_edge *cs, struct cgraph_node *node) { ipcp_transformation *ts = ipcp_get_transformation_summary (node); if (!ts || vec_safe_is_empty (ts->m_agg_values)) return true; const ipa_argagg_value_list existing (ts->m_agg_values); auto_vec edge_values; ipa_node_params *dest_info = ipa_node_params_sum->get (node); gcc_checking_assert (dest_info->ipcp_orig_node); dest_info = ipa_node_params_sum->get (dest_info->ipcp_orig_node); push_agg_values_from_edge (cs, dest_info, &edge_values, &existing, false); const ipa_argagg_value_list avl (&edge_values); return avl.superset_of_p (existing); } /* Given an original NODE and a VAL for which we have already created a specialized clone, look whether there are incoming edges that still lead into the old node but now also bring the requested value and also conform to all other criteria such that they can be redirected the special node. This function can therefore redirect the final edge in a SCC. */ template static void perhaps_add_new_callers (cgraph_node *node, ipcp_value *val) { ipcp_value_source *src; profile_count redirected_sum = profile_count::zero (); for (src = val->sources; src; src = src->next) { struct cgraph_edge *cs = src->cs; while (cs) { if (cgraph_edge_brings_value_p (cs, src, node, val) && cgraph_edge_brings_all_scalars_for_node (cs, val->spec_node) && cgraph_edge_brings_all_agg_vals_for_node (cs, val->spec_node)) { if (dump_file) fprintf (dump_file, " - adding an extra caller %s of %s\n", cs->caller->dump_name (), val->spec_node->dump_name ()); cs->redirect_callee_duplicating_thunks (val->spec_node); val->spec_node->expand_all_artificial_thunks (); if (cs->count.ipa ().initialized_p ()) redirected_sum = redirected_sum + cs->count.ipa (); } cs = get_next_cgraph_edge_clone (cs); } } if (redirected_sum.nonzero_p ()) update_specialized_profile (val->spec_node, node, redirected_sum); } /* Return true if KNOWN_CONTEXTS contain at least one useful context. */ static bool known_contexts_useful_p (vec known_contexts) { ipa_polymorphic_call_context *ctx; int i; FOR_EACH_VEC_ELT (known_contexts, i, ctx) if (!ctx->useless_p ()) return true; return false; } /* Return a copy of KNOWN_CSTS if it is not empty, otherwise return vNULL. */ static vec copy_useful_known_contexts (const vec &known_contexts) { if (known_contexts_useful_p (known_contexts)) return known_contexts.copy (); else return vNULL; } /* Copy known scalar values from AVALS into KNOWN_CSTS and modify the copy according to VAL and INDEX. If non-empty, replace KNOWN_CONTEXTS with its copy too. */ static void copy_known_vectors_add_val (ipa_auto_call_arg_values *avals, vec *known_csts, vec *known_contexts, ipcp_value *val, int index) { *known_csts = avals->m_known_vals.copy (); *known_contexts = copy_useful_known_contexts (avals->m_known_contexts); (*known_csts)[index] = val->value; } /* Copy known scalar values from AVALS into KNOWN_CSTS. Similarly, copy contexts to KNOWN_CONTEXTS and modify the copy according to VAL and INDEX. */ static void copy_known_vectors_add_val (ipa_auto_call_arg_values *avals, vec *known_csts, vec *known_contexts, ipcp_value *val, int index) { *known_csts = avals->m_known_vals.copy (); *known_contexts = avals->m_known_contexts.copy (); (*known_contexts)[index] = val->value; } /* Return true if OFFSET indicates this was not an aggregate value or there is a replacement equivalent to VALUE, INDEX and OFFSET among those in the AGGVALS list. */ DEBUG_FUNCTION bool ipcp_val_agg_replacement_ok_p (vec *aggvals, int index, HOST_WIDE_INT offset, tree value) { if (offset == -1) return true; const ipa_argagg_value_list avl (aggvals); tree v = avl.get_value (index, offset / BITS_PER_UNIT); return v && values_equal_for_ipcp_p (v, value); } /* Return true if offset is minus one because source of a polymorphic context cannot be an aggregate value. */ DEBUG_FUNCTION bool ipcp_val_agg_replacement_ok_p (vec *, int , HOST_WIDE_INT offset, ipa_polymorphic_call_context) { return offset == -1; } /* Decide whether to create a special version of NODE for value VAL of parameter at the given INDEX. If OFFSET is -1, the value is for the parameter itself, otherwise it is stored at the given OFFSET of the parameter. AVALS describes the other already known values. SELF_GEN_CLONES is a vector which contains clones created for self-recursive calls with an arithmetic pass-through jump function. */ template static bool decide_about_value (struct cgraph_node *node, int index, HOST_WIDE_INT offset, ipcp_value *val, ipa_auto_call_arg_values *avals, vec *self_gen_clones) { int caller_count; sreal freq_sum; profile_count count_sum, rec_count_sum; vec callers; if (val->spec_node) { perhaps_add_new_callers (node, val); return false; } else if (val->local_size_cost + overall_size > get_max_overall_size (node)) { if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " Ignoring candidate value because " "maximum unit size would be reached with %li.\n", val->local_size_cost + overall_size); return false; } else if (!get_info_about_necessary_edges (val, node, &freq_sum, &caller_count, &rec_count_sum, &count_sum)) return false; if (!dbg_cnt (ipa_cp_values)) return false; if (val->self_recursion_generated_p ()) { /* The edge counts in this case might not have been adjusted yet. Nevertleless, even if they were it would be only a guesswork which we can do now. The recursive part of the counts can be derived from the count of the original node anyway. */ if (node->count.ipa ().nonzero_p ()) { unsigned dem = self_gen_clones->length () + 1; rec_count_sum = node->count.ipa () / dem; } else rec_count_sum = profile_count::zero (); } /* get_info_about_necessary_edges only sums up ipa counts. */ count_sum += rec_count_sum; if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " - considering value "); print_ipcp_constant_value (dump_file, val->value); fprintf (dump_file, " for "); ipa_dump_param (dump_file, ipa_node_params_sum->get (node), index); if (offset != -1) fprintf (dump_file, ", offset: " HOST_WIDE_INT_PRINT_DEC, offset); fprintf (dump_file, " (caller_count: %i)\n", caller_count); } if (!good_cloning_opportunity_p (node, val->local_time_benefit, freq_sum, count_sum, val->local_size_cost) && !good_cloning_opportunity_p (node, val->prop_time_benefit, freq_sum, count_sum, val->prop_size_cost)) return false; if (dump_file) fprintf (dump_file, " Creating a specialized node of %s.\n", node->dump_name ()); vec known_csts; vec known_contexts; callers = gather_edges_for_value (val, node, caller_count); if (offset == -1) copy_known_vectors_add_val (avals, &known_csts, &known_contexts, val, index); else { known_csts = avals->m_known_vals.copy (); known_contexts = copy_useful_known_contexts (avals->m_known_contexts); } find_more_scalar_values_for_callers_subset (node, known_csts, callers); find_more_contexts_for_caller_subset (node, &known_contexts, callers); vec *aggvals = find_aggregate_values_for_callers_subset (node, callers); gcc_checking_assert (ipcp_val_agg_replacement_ok_p (aggvals, index, offset, val->value)); val->spec_node = create_specialized_node (node, known_csts, known_contexts, aggvals, callers); if (val->self_recursion_generated_p ()) self_gen_clones->safe_push (val->spec_node); else update_profiling_info (node, val->spec_node); callers.release (); overall_size += val->local_size_cost; if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, " overall size reached %li\n", overall_size); /* TODO: If for some lattice there is only one other known value left, make a special node for it too. */ return true; } /* Like irange::contains_p(), but convert VAL to the range of R if necessary. */ static inline bool ipa_range_contains_p (const vrange &r, tree val) { if (r.undefined_p ()) return false; tree type = r.type (); if (!wi::fits_to_tree_p (wi::to_wide (val), type)) return false; val = fold_convert (type, val); return r.contains_p (val); } /* Decide whether and what specialized clones of NODE should be created. */ static bool decide_whether_version_node (struct cgraph_node *node) { ipa_node_params *info = ipa_node_params_sum->get (node); int i, count = ipa_get_param_count (info); bool ret = false; if (count == 0) return false; if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, "\nEvaluating opportunities for %s.\n", node->dump_name ()); auto_vec self_gen_clones; ipa_auto_call_arg_values avals; gather_context_independent_values (info, &avals, false, NULL); for (i = 0; i < count;i++) { if (!ipa_is_param_used (info, i)) continue; class ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); ipcp_lattice *lat = &plats->itself; ipcp_lattice *ctxlat = &plats->ctxlat; if (!lat->bottom && !avals.m_known_vals[i]) { ipcp_value *val; for (val = lat->values; val; val = val->next) { /* If some values generated for self-recursive calls with arithmetic jump functions fall outside of the known range for the parameter, we can skip them. */ if (TREE_CODE (val->value) == INTEGER_CST && !plats->m_value_range.bottom_p () && !ipa_range_contains_p (plats->m_value_range.m_vr, val->value)) { /* This can happen also if a constant present in the source code falls outside of the range of parameter's type, so we cannot assert. */ if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, " - skipping%s value ", val->self_recursion_generated_p () ? " self_recursion_generated" : ""); print_ipcp_constant_value (dump_file, val->value); fprintf (dump_file, " because it is outside known " "value range.\n"); } continue; } ret |= decide_about_value (node, i, -1, val, &avals, &self_gen_clones); } } if (!plats->aggs_bottom) { struct ipcp_agg_lattice *aglat; ipcp_value *val; for (aglat = plats->aggs; aglat; aglat = aglat->next) if (!aglat->bottom && aglat->values /* If the following is false, the one value has been considered for cloning for all contexts. */ && (plats->aggs_contain_variable || !aglat->is_single_const ())) for (val = aglat->values; val; val = val->next) ret |= decide_about_value (node, i, aglat->offset, val, &avals, &self_gen_clones); } if (!ctxlat->bottom && avals.m_known_contexts[i].useless_p ()) { ipcp_value *val; for (val = ctxlat->values; val; val = val->next) ret |= decide_about_value (node, i, -1, val, &avals, &self_gen_clones); } } if (!self_gen_clones.is_empty ()) { self_gen_clones.safe_push (node); update_counts_for_self_gen_clones (node, self_gen_clones); } if (info->do_clone_for_all_contexts) { if (!dbg_cnt (ipa_cp_values)) { info->do_clone_for_all_contexts = false; return ret; } struct cgraph_node *clone; auto_vec callers = node->collect_callers (); for (int i = callers.length () - 1; i >= 0; i--) { cgraph_edge *cs = callers[i]; ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); if (caller_info && caller_info->node_dead) callers.unordered_remove (i); } if (!adjust_callers_for_value_intersection (callers, node)) { /* If node is not called by anyone, or all its caller edges are self-recursive, the node is not really in use, no need to do cloning. */ info->do_clone_for_all_contexts = false; return ret; } if (dump_file) fprintf (dump_file, " - Creating a specialized node of %s " "for all known contexts.\n", node->dump_name ()); vec known_csts = avals.m_known_vals.copy (); vec known_contexts = copy_useful_known_contexts (avals.m_known_contexts); find_more_scalar_values_for_callers_subset (node, known_csts, callers); find_more_contexts_for_caller_subset (node, &known_contexts, callers); vec *aggvals = find_aggregate_values_for_callers_subset (node, callers); if (!known_contexts_useful_p (known_contexts)) { known_contexts.release (); known_contexts = vNULL; } clone = create_specialized_node (node, known_csts, known_contexts, aggvals, callers); info->do_clone_for_all_contexts = false; ipa_node_params_sum->get (clone)->is_all_contexts_clone = true; ret = true; } return ret; } /* Transitively mark all callees of NODE within the same SCC as not dead. */ static void spread_undeadness (struct cgraph_node *node) { struct cgraph_edge *cs; for (cs = node->callees; cs; cs = cs->next_callee) if (ipa_edge_within_scc (cs)) { struct cgraph_node *callee; class ipa_node_params *info; callee = cs->callee->function_symbol (NULL); info = ipa_node_params_sum->get (callee); if (info && info->node_dead) { info->node_dead = 0; spread_undeadness (callee); } } } /* Return true if NODE has a caller from outside of its SCC that is not dead. Worker callback for cgraph_for_node_and_aliases. */ static bool has_undead_caller_from_outside_scc_p (struct cgraph_node *node, void *data ATTRIBUTE_UNUSED) { struct cgraph_edge *cs; for (cs = node->callers; cs; cs = cs->next_caller) if (cs->caller->thunk && cs->caller->call_for_symbol_thunks_and_aliases (has_undead_caller_from_outside_scc_p, NULL, true)) return true; else if (!ipa_edge_within_scc (cs)) { ipa_node_params *caller_info = ipa_node_params_sum->get (cs->caller); if (!caller_info /* Unoptimized caller are like dead ones. */ || !caller_info->node_dead) return true; } return false; } /* Identify nodes within the same SCC as NODE which are no longer needed because of new clones and will be removed as unreachable. */ static void identify_dead_nodes (struct cgraph_node *node) { struct cgraph_node *v; for (v = node; v; v = ((struct ipa_dfs_info *) v->aux)->next_cycle) if (v->local) { ipa_node_params *info = ipa_node_params_sum->get (v); if (info && !v->call_for_symbol_thunks_and_aliases (has_undead_caller_from_outside_scc_p, NULL, true)) info->node_dead = 1; } for (v = node; v; v = ((struct ipa_dfs_info *) v->aux)->next_cycle) { ipa_node_params *info = ipa_node_params_sum->get (v); if (info && !info->node_dead) spread_undeadness (v); } if (dump_file && (dump_flags & TDF_DETAILS)) { for (v = node; v; v = ((struct ipa_dfs_info *) v->aux)->next_cycle) if (ipa_node_params_sum->get (v) && ipa_node_params_sum->get (v)->node_dead) fprintf (dump_file, " Marking node as dead: %s.\n", v->dump_name ()); } } /* The decision stage. Iterate over the topological order of call graph nodes TOPO and make specialized clones if deemed beneficial. */ static void ipcp_decision_stage (class ipa_topo_info *topo) { int i; if (dump_file) fprintf (dump_file, "\nIPA decision stage:\n\n"); for (i = topo->nnodes - 1; i >= 0; i--) { struct cgraph_node *node = topo->order[i]; bool change = false, iterate = true; while (iterate) { struct cgraph_node *v; iterate = false; for (v = node; v; v = ((struct ipa_dfs_info *) v->aux)->next_cycle) if (v->has_gimple_body_p () && ipcp_versionable_function_p (v)) iterate |= decide_whether_version_node (v); change |= iterate; } if (change) identify_dead_nodes (node); } } /* Look up all VR and bits information that we have discovered and copy it over to the transformation summary. */ static void ipcp_store_vr_results (void) { cgraph_node *node; FOR_EACH_FUNCTION_WITH_GIMPLE_BODY (node) { ipa_node_params *info = ipa_node_params_sum->get (node); bool dumped_sth = false; bool found_useful_result = false; bool do_vr = true; bool do_bits = true; if (!info || !opt_for_fn (node->decl, flag_ipa_vrp)) { if (dump_file) fprintf (dump_file, "Not considering %s for VR discovery " "and propagate; -fipa-ipa-vrp: disabled.\n", node->dump_name ()); do_vr = false; } if (!info || !opt_for_fn (node->decl, flag_ipa_bit_cp)) { if (dump_file) fprintf (dump_file, "Not considering %s for ipa bitwise " "propagation ; -fipa-bit-cp: disabled.\n", node->dump_name ()); do_bits = false; } if (!do_bits && !do_vr) continue; if (info->ipcp_orig_node) info = ipa_node_params_sum->get (info->ipcp_orig_node); if (!info->lattices) /* Newly expanded artificial thunks do not have lattices. */ continue; unsigned count = ipa_get_param_count (info); for (unsigned i = 0; i < count; i++) { ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); if (do_vr && !plats->m_value_range.bottom_p () && !plats->m_value_range.top_p ()) { found_useful_result = true; break; } if (do_bits && plats->bits_lattice.constant_p ()) { found_useful_result = true; break; } } if (!found_useful_result) continue; ipcp_transformation_initialize (); ipcp_transformation *ts = ipcp_transformation_sum->get_create (node); vec_safe_reserve_exact (ts->m_vr, count); for (unsigned i = 0; i < count; i++) { ipcp_param_lattices *plats = ipa_get_parm_lattices (info, i); ipcp_bits_lattice *bits = NULL; if (do_bits && plats->bits_lattice.constant_p () && dbg_cnt (ipa_cp_bits)) bits = &plats->bits_lattice; if (do_vr && !plats->m_value_range.bottom_p () && !plats->m_value_range.top_p () && dbg_cnt (ipa_cp_vr)) { if (bits) { Value_Range tmp = plats->m_value_range.m_vr; tree type = ipa_get_type (info, i); irange &r = as_a (tmp); irange_bitmask bm (wide_int::from (bits->get_value (), TYPE_PRECISION (type), TYPE_SIGN (type)), wide_int::from (bits->get_mask (), TYPE_PRECISION (type), TYPE_SIGN (type))); r.update_bitmask (bm); ipa_vr vr (tmp); ts->m_vr->quick_push (vr); } else { ipa_vr vr (plats->m_value_range.m_vr); ts->m_vr->quick_push (vr); } } else if (bits) { tree type = ipa_get_type (info, i); Value_Range tmp; tmp.set_varying (type); irange &r = as_a (tmp); irange_bitmask bm (wide_int::from (bits->get_value (), TYPE_PRECISION (type), TYPE_SIGN (type)), wide_int::from (bits->get_mask (), TYPE_PRECISION (type), TYPE_SIGN (type))); r.update_bitmask (bm); ipa_vr vr (tmp); ts->m_vr->quick_push (vr); } else { ipa_vr vr; ts->m_vr->quick_push (vr); } if (!dump_file || !bits) continue; if (!dumped_sth) { fprintf (dump_file, "Propagated bits info for function %s:\n", node->dump_name ()); dumped_sth = true; } fprintf (dump_file, " param %i: value = ", i); print_hex (bits->get_value (), dump_file); fprintf (dump_file, ", mask = "); print_hex (bits->get_mask (), dump_file); fprintf (dump_file, "\n"); } } } /* The IPCP driver. */ static unsigned int ipcp_driver (void) { class ipa_topo_info topo; if (edge_clone_summaries == NULL) edge_clone_summaries = new edge_clone_summary_t (symtab); ipa_check_create_node_params (); ipa_check_create_edge_args (); clone_num_suffixes = new hash_map; if (dump_file) { fprintf (dump_file, "\nIPA structures before propagation:\n"); if (dump_flags & TDF_DETAILS) ipa_print_all_params (dump_file); ipa_print_all_jump_functions (dump_file); } /* Topological sort. */ build_toporder_info (&topo); /* Do the interprocedural propagation. */ ipcp_propagate_stage (&topo); /* Decide what constant propagation and cloning should be performed. */ ipcp_decision_stage (&topo); /* Store results of value range and bits propagation. */ ipcp_store_vr_results (); /* Free all IPCP structures. */ delete clone_num_suffixes; free_toporder_info (&topo); delete edge_clone_summaries; edge_clone_summaries = NULL; ipa_free_all_structures_after_ipa_cp (); if (dump_file) fprintf (dump_file, "\nIPA constant propagation end\n"); return 0; } /* Initialization and computation of IPCP data structures. This is the initial intraprocedural analysis of functions, which gathers information to be propagated later on. */ static void ipcp_generate_summary (void) { struct cgraph_node *node; if (dump_file) fprintf (dump_file, "\nIPA constant propagation start:\n"); ipa_register_cgraph_hooks (); FOR_EACH_FUNCTION_WITH_GIMPLE_BODY (node) ipa_analyze_node (node); } namespace { const pass_data pass_data_ipa_cp = { IPA_PASS, /* type */ "cp", /* name */ OPTGROUP_NONE, /* optinfo_flags */ TV_IPA_CONSTANT_PROP, /* tv_id */ 0, /* properties_required */ 0, /* properties_provided */ 0, /* properties_destroyed */ 0, /* todo_flags_start */ ( TODO_dump_symtab | TODO_remove_functions ), /* todo_flags_finish */ }; class pass_ipa_cp : public ipa_opt_pass_d { public: pass_ipa_cp (gcc::context *ctxt) : ipa_opt_pass_d (pass_data_ipa_cp, ctxt, ipcp_generate_summary, /* generate_summary */ NULL, /* write_summary */ NULL, /* read_summary */ ipcp_write_transformation_summaries, /* write_optimization_summary */ ipcp_read_transformation_summaries, /* read_optimization_summary */ NULL, /* stmt_fixup */ 0, /* function_transform_todo_flags_start */ ipcp_transform_function, /* function_transform */ NULL) /* variable_transform */ {} /* opt_pass methods: */ bool gate (function *) final override { /* FIXME: We should remove the optimize check after we ensure we never run IPA passes when not optimizing. */ return (flag_ipa_cp && optimize) || in_lto_p; } unsigned int execute (function *) final override { return ipcp_driver (); } }; // class pass_ipa_cp } // anon namespace ipa_opt_pass_d * make_pass_ipa_cp (gcc::context *ctxt) { return new pass_ipa_cp (ctxt); } /* Reset all state within ipa-cp.cc so that we can rerun the compiler within the same process. For use by toplev::finalize. */ void ipa_cp_cc_finalize (void) { base_count = profile_count::uninitialized (); overall_size = 0; orig_overall_size = 0; ipcp_free_transformation_sum (); } /* Given PARAM which must be a parameter of function FNDECL described by THIS, return its index in the DECL_ARGUMENTS chain, using a pre-computed DECL_UID-sorted vector if available (which is pre-computed only if there are many parameters). Can return -1 if param is static chain not represented among DECL_ARGUMENTS. */ int ipcp_transformation::get_param_index (const_tree fndecl, const_tree param) const { gcc_assert (TREE_CODE (param) == PARM_DECL); if (m_uid_to_idx) { unsigned puid = DECL_UID (param); const ipa_uid_to_idx_map_elt *res = std::lower_bound (m_uid_to_idx->begin(), m_uid_to_idx->end (), puid, [] (const ipa_uid_to_idx_map_elt &elt, unsigned uid) { return elt.uid < uid; }); if (res == m_uid_to_idx->end () || res->uid != puid) { gcc_assert (DECL_STATIC_CHAIN (fndecl)); return -1; } return res->index; } unsigned index = 0; for (tree p = DECL_ARGUMENTS (fndecl); p; p = DECL_CHAIN (p), index++) if (p == param) return (int) index; gcc_assert (DECL_STATIC_CHAIN (fndecl)); return -1; } /* Helper function to qsort a vector of ipa_uid_to_idx_map_elt elements according to the uid. */ static int compare_uids (const void *a, const void *b) { const ipa_uid_to_idx_map_elt *e1 = (const ipa_uid_to_idx_map_elt *) a; const ipa_uid_to_idx_map_elt *e2 = (const ipa_uid_to_idx_map_elt *) b; if (e1->uid < e2->uid) return -1; if (e1->uid > e2->uid) return 1; gcc_unreachable (); } /* Assuming THIS describes FNDECL and it has sufficiently many parameters to justify the overhead, create a DECL_UID-sorted vector to speed up mapping from parameters to their indices in DECL_ARGUMENTS chain. */ void ipcp_transformation::maybe_create_parm_idx_map (tree fndecl) { int c = count_formal_params (fndecl); if (c < 32) return; m_uid_to_idx = NULL; vec_safe_reserve (m_uid_to_idx, c, true); unsigned index = 0; for (tree p = DECL_ARGUMENTS (fndecl); p; p = DECL_CHAIN (p), index++) { ipa_uid_to_idx_map_elt elt; elt.uid = DECL_UID (p); elt.index = index; m_uid_to_idx->quick_push (elt); } m_uid_to_idx->qsort (compare_uids); }