URL
https://opencores.org/ocsvn/openrisc_me/openrisc_me/trunk
Subversion Repositories openrisc_me
[/] [openrisc/] [trunk/] [gnu-src/] [gcc-4.5.1/] [gcc/] [tree-ssa-copy.c] - Rev 310
Go to most recent revision | Compare with Previous | Blame | View Log
/* Copy propagation and SSA_NAME replacement support routines. Copyright (C) 2004, 2005, 2006, 2007, 2008, 2010 Free Software Foundation, Inc. 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 <http://www.gnu.org/licenses/>. */ #include "config.h" #include "system.h" #include "coretypes.h" #include "tm.h" #include "tree.h" #include "flags.h" #include "rtl.h" #include "tm_p.h" #include "ggc.h" #include "basic-block.h" #include "output.h" #include "expr.h" #include "function.h" #include "diagnostic.h" #include "timevar.h" #include "tree-dump.h" #include "tree-flow.h" #include "tree-pass.h" #include "tree-ssa-propagate.h" #include "langhooks.h" #include "cfgloop.h" /* This file implements the copy propagation pass and provides a handful of interfaces for performing const/copy propagation and simple expression replacement which keep variable annotations up-to-date. We require that for any copy operation where the RHS and LHS have a non-null memory tag the memory tag be the same. It is OK for one or both of the memory tags to be NULL. We also require tracking if a variable is dereferenced in a load or store operation. We enforce these requirements by having all copy propagation and replacements of one SSA_NAME with a different SSA_NAME to use the APIs defined in this file. */ /* Return true if we may propagate ORIG into DEST, false otherwise. */ bool may_propagate_copy (tree dest, tree orig) { tree type_d = TREE_TYPE (dest); tree type_o = TREE_TYPE (orig); /* If ORIG flows in from an abnormal edge, it cannot be propagated. */ if (TREE_CODE (orig) == SSA_NAME && SSA_NAME_OCCURS_IN_ABNORMAL_PHI (orig)) return false; /* If DEST is an SSA_NAME that flows from an abnormal edge, then it cannot be replaced. */ if (TREE_CODE (dest) == SSA_NAME && SSA_NAME_OCCURS_IN_ABNORMAL_PHI (dest)) return false; /* Do not copy between types for which we *do* need a conversion. */ if (!useless_type_conversion_p (type_d, type_o)) return false; /* Propagating virtual operands is always ok. */ if (TREE_CODE (dest) == SSA_NAME && !is_gimple_reg (dest)) { /* But only between virtual operands. */ gcc_assert (TREE_CODE (orig) == SSA_NAME && !is_gimple_reg (orig)); return true; } /* Anything else is OK. */ return true; } /* Like may_propagate_copy, but use as the destination expression the principal expression (typically, the RHS) contained in statement DEST. This is more efficient when working with the gimple tuples representation. */ bool may_propagate_copy_into_stmt (gimple dest, tree orig) { tree type_d; tree type_o; /* If the statement is a switch or a single-rhs assignment, then the expression to be replaced by the propagation may be an SSA_NAME. Fortunately, there is an explicit tree for the expression, so we delegate to may_propagate_copy. */ if (gimple_assign_single_p (dest)) return may_propagate_copy (gimple_assign_rhs1 (dest), orig); else if (gimple_code (dest) == GIMPLE_SWITCH) return may_propagate_copy (gimple_switch_index (dest), orig); /* In other cases, the expression is not materialized, so there is no destination to pass to may_propagate_copy. On the other hand, the expression cannot be an SSA_NAME, so the analysis is much simpler. */ if (TREE_CODE (orig) == SSA_NAME && SSA_NAME_OCCURS_IN_ABNORMAL_PHI (orig)) return false; if (is_gimple_assign (dest)) type_d = TREE_TYPE (gimple_assign_lhs (dest)); else if (gimple_code (dest) == GIMPLE_COND) type_d = boolean_type_node; else if (is_gimple_call (dest) && gimple_call_lhs (dest) != NULL_TREE) type_d = TREE_TYPE (gimple_call_lhs (dest)); else gcc_unreachable (); type_o = TREE_TYPE (orig); if (!useless_type_conversion_p (type_d, type_o)) return false; return true; } /* Similarly, but we know that we're propagating into an ASM_EXPR. */ bool may_propagate_copy_into_asm (tree dest) { /* Hard register operands of asms are special. Do not bypass. */ return !(TREE_CODE (dest) == SSA_NAME && TREE_CODE (SSA_NAME_VAR (dest)) == VAR_DECL && DECL_HARD_REGISTER (SSA_NAME_VAR (dest))); } /* Common code for propagate_value and replace_exp. Replace use operand OP_P with VAL. FOR_PROPAGATION indicates if the replacement is done to propagate a value or not. */ static void replace_exp_1 (use_operand_p op_p, tree val, bool for_propagation ATTRIBUTE_UNUSED) { #if defined ENABLE_CHECKING tree op = USE_FROM_PTR (op_p); gcc_assert (!(for_propagation && TREE_CODE (op) == SSA_NAME && TREE_CODE (val) == SSA_NAME && !may_propagate_copy (op, val))); #endif if (TREE_CODE (val) == SSA_NAME) SET_USE (op_p, val); else SET_USE (op_p, unsave_expr_now (val)); } /* Propagate the value VAL (assumed to be a constant or another SSA_NAME) into the operand pointed to by OP_P. Use this version for const/copy propagation as it will perform additional checks to ensure validity of the const/copy propagation. */ void propagate_value (use_operand_p op_p, tree val) { replace_exp_1 (op_p, val, true); } /* Replace *OP_P with value VAL (assumed to be a constant or another SSA_NAME). Use this version when not const/copy propagating values. For example, PRE uses this version when building expressions as they would appear in specific blocks taking into account actions of PHI nodes. */ void replace_exp (use_operand_p op_p, tree val) { replace_exp_1 (op_p, val, false); } /* Propagate the value VAL (assumed to be a constant or another SSA_NAME) into the tree pointed to by OP_P. Use this version for const/copy propagation when SSA operands are not available. It will perform the additional checks to ensure validity of the const/copy propagation, but will not update any operand information. Be sure to mark the stmt as modified. */ void propagate_tree_value (tree *op_p, tree val) { #if defined ENABLE_CHECKING gcc_assert (!(TREE_CODE (val) == SSA_NAME && *op_p && TREE_CODE (*op_p) == SSA_NAME && !may_propagate_copy (*op_p, val))); #endif if (TREE_CODE (val) == SSA_NAME) *op_p = val; else *op_p = unsave_expr_now (val); } /* Like propagate_tree_value, but use as the operand to replace the principal expression (typically, the RHS) contained in the statement referenced by iterator GSI. Note that it is not always possible to update the statement in-place, so a new statement may be created to replace the original. */ void propagate_tree_value_into_stmt (gimple_stmt_iterator *gsi, tree val) { gimple stmt = gsi_stmt (*gsi); if (is_gimple_assign (stmt)) { tree expr = NULL_TREE; if (gimple_assign_single_p (stmt)) expr = gimple_assign_rhs1 (stmt); propagate_tree_value (&expr, val); gimple_assign_set_rhs_from_tree (gsi, expr); stmt = gsi_stmt (*gsi); } else if (gimple_code (stmt) == GIMPLE_COND) { tree lhs = NULL_TREE; tree rhs = fold_convert (TREE_TYPE (val), integer_zero_node); propagate_tree_value (&lhs, val); gimple_cond_set_code (stmt, NE_EXPR); gimple_cond_set_lhs (stmt, lhs); gimple_cond_set_rhs (stmt, rhs); } else if (is_gimple_call (stmt) && gimple_call_lhs (stmt) != NULL_TREE) { gimple new_stmt; tree expr = NULL_TREE; propagate_tree_value (&expr, val); new_stmt = gimple_build_assign (gimple_call_lhs (stmt), expr); move_ssa_defining_stmt_for_defs (new_stmt, stmt); gsi_replace (gsi, new_stmt, false); } else if (gimple_code (stmt) == GIMPLE_SWITCH) propagate_tree_value (gimple_switch_index_ptr (stmt), val); else gcc_unreachable (); } /*--------------------------------------------------------------------------- Copy propagation ---------------------------------------------------------------------------*/ /* During propagation, we keep chains of variables that are copies of one another. If variable X_i is a copy of X_j and X_j is a copy of X_k, COPY_OF will contain: COPY_OF[i].VALUE = X_j COPY_OF[j].VALUE = X_k COPY_OF[k].VALUE = X_k After propagation, the copy-of value for each variable X_i is converted into the final value by walking the copy-of chains and updating COPY_OF[i].VALUE to be the last element of the chain. */ static prop_value_t *copy_of; /* Used in set_copy_of_val to determine if the last link of a copy-of chain has changed. */ static tree *cached_last_copy_of; /* Return true if this statement may generate a useful copy. */ static bool stmt_may_generate_copy (gimple stmt) { if (gimple_code (stmt) == GIMPLE_PHI) return !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (gimple_phi_result (stmt)); if (gimple_code (stmt) != GIMPLE_ASSIGN) return false; /* If the statement has volatile operands, it won't generate a useful copy. */ if (gimple_has_volatile_ops (stmt)) return false; /* Statements with loads and/or stores will never generate a useful copy. */ if (gimple_vuse (stmt)) return false; /* Otherwise, the only statements that generate useful copies are assignments whose RHS is just an SSA name that doesn't flow through abnormal edges. */ return (gimple_assign_rhs_code (stmt) == SSA_NAME && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (gimple_assign_rhs1 (stmt))); } /* Return the copy-of value for VAR. */ static inline prop_value_t * get_copy_of_val (tree var) { prop_value_t *val = ©_of[SSA_NAME_VERSION (var)]; if (val->value == NULL_TREE && !stmt_may_generate_copy (SSA_NAME_DEF_STMT (var))) { /* If the variable will never generate a useful copy relation, make it its own copy. */ val->value = var; } return val; } /* Return last link in the copy-of chain for VAR. */ static tree get_last_copy_of (tree var) { tree last; int i; /* Traverse COPY_OF starting at VAR until we get to the last link in the chain. Since it is possible to have cycles in PHI nodes, the copy-of chain may also contain cycles. To avoid infinite loops and to avoid traversing lengthy copy-of chains, we artificially limit the maximum number of chains we are willing to traverse. The value 5 was taken from a compiler and runtime library bootstrap and a mixture of C and C++ code from various sources. More than 82% of all copy-of chains were shorter than 5 links. */ #define LIMIT 5 last = var; for (i = 0; i < LIMIT; i++) { tree copy = copy_of[SSA_NAME_VERSION (last)].value; if (copy == NULL_TREE || copy == last) break; last = copy; } /* If we have reached the limit, then we are either in a copy-of cycle or the copy-of chain is too long. In this case, just return VAR so that it is not considered a copy of anything. */ return (i < LIMIT ? last : var); } /* Set FIRST to be the first variable in the copy-of chain for DEST. If DEST's copy-of value or its copy-of chain has changed, return true. MEM_REF is the memory reference where FIRST is stored. This is used when DEST is a non-register and we are copy propagating loads and stores. */ static inline bool set_copy_of_val (tree dest, tree first) { unsigned int dest_ver = SSA_NAME_VERSION (dest); tree old_first, old_last, new_last; /* Set FIRST to be the first link in COPY_OF[DEST]. If that changed, return true. */ old_first = copy_of[dest_ver].value; copy_of[dest_ver].value = first; if (old_first != first) return true; /* If FIRST and OLD_FIRST are the same, we need to check whether the copy-of chain starting at FIRST ends in a different variable. If the copy-of chain starting at FIRST ends up in a different variable than the last cached value we had for DEST, then return true because DEST is now a copy of a different variable. This test is necessary because even though the first link in the copy-of chain may not have changed, if any of the variables in the copy-of chain changed its final value, DEST will now be the copy of a different variable, so we have to do another round of propagation for everything that depends on DEST. */ old_last = cached_last_copy_of[dest_ver]; new_last = get_last_copy_of (dest); cached_last_copy_of[dest_ver] = new_last; return (old_last != new_last); } /* Dump the copy-of value for variable VAR to FILE. */ static void dump_copy_of (FILE *file, tree var) { tree val; sbitmap visited; print_generic_expr (file, var, dump_flags); if (TREE_CODE (var) != SSA_NAME) return; visited = sbitmap_alloc (num_ssa_names); sbitmap_zero (visited); SET_BIT (visited, SSA_NAME_VERSION (var)); fprintf (file, " copy-of chain: "); val = var; print_generic_expr (file, val, 0); fprintf (file, " "); while (copy_of[SSA_NAME_VERSION (val)].value) { fprintf (file, "-> "); val = copy_of[SSA_NAME_VERSION (val)].value; print_generic_expr (file, val, 0); fprintf (file, " "); if (TEST_BIT (visited, SSA_NAME_VERSION (val))) break; SET_BIT (visited, SSA_NAME_VERSION (val)); } val = get_copy_of_val (var)->value; if (val == NULL_TREE) fprintf (file, "[UNDEFINED]"); else if (val != var) fprintf (file, "[COPY]"); else fprintf (file, "[NOT A COPY]"); sbitmap_free (visited); } /* Evaluate the RHS of STMT. If it produces a valid copy, set the LHS value and store the LHS into *RESULT_P. If STMT generates more than one name (i.e., STMT is an aliased store), it is enough to store the first name in the VDEF list into *RESULT_P. After all, the names generated will be VUSEd in the same statements. */ static enum ssa_prop_result copy_prop_visit_assignment (gimple stmt, tree *result_p) { tree lhs, rhs; prop_value_t *rhs_val; lhs = gimple_assign_lhs (stmt); rhs = gimple_assign_rhs1 (stmt); gcc_assert (gimple_assign_rhs_code (stmt) == SSA_NAME); rhs_val = get_copy_of_val (rhs); if (TREE_CODE (lhs) == SSA_NAME) { /* Straight copy between two SSA names. First, make sure that we can propagate the RHS into uses of LHS. */ if (!may_propagate_copy (lhs, rhs)) return SSA_PROP_VARYING; /* Notice that in the case of assignments, we make the LHS be a copy of RHS's value, not of RHS itself. This avoids keeping unnecessary copy-of chains (assignments cannot be in a cycle like PHI nodes), speeding up the propagation process. This is different from what we do in copy_prop_visit_phi_node. In those cases, we are interested in the copy-of chains. */ *result_p = lhs; if (set_copy_of_val (*result_p, rhs_val->value)) return SSA_PROP_INTERESTING; else return SSA_PROP_NOT_INTERESTING; } return SSA_PROP_VARYING; } /* Visit the GIMPLE_COND STMT. Return SSA_PROP_INTERESTING if it can determine which edge will be taken. Otherwise, return SSA_PROP_VARYING. */ static enum ssa_prop_result copy_prop_visit_cond_stmt (gimple stmt, edge *taken_edge_p) { enum ssa_prop_result retval = SSA_PROP_VARYING; location_t loc = gimple_location (stmt); tree op0 = gimple_cond_lhs (stmt); tree op1 = gimple_cond_rhs (stmt); /* The only conditionals that we may be able to compute statically are predicates involving two SSA_NAMEs. */ if (TREE_CODE (op0) == SSA_NAME && TREE_CODE (op1) == SSA_NAME) { op0 = get_last_copy_of (op0); op1 = get_last_copy_of (op1); /* See if we can determine the predicate's value. */ if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, "Trying to determine truth value of "); fprintf (dump_file, "predicate "); print_gimple_stmt (dump_file, stmt, 0, 0); } /* We can fold COND and get a useful result only when we have the same SSA_NAME on both sides of a comparison operator. */ if (op0 == op1) { tree folded_cond = fold_binary_loc (loc, gimple_cond_code (stmt), boolean_type_node, op0, op1); if (folded_cond) { basic_block bb = gimple_bb (stmt); *taken_edge_p = find_taken_edge (bb, folded_cond); if (*taken_edge_p) retval = SSA_PROP_INTERESTING; } } } if (dump_file && (dump_flags & TDF_DETAILS) && *taken_edge_p) fprintf (dump_file, "\nConditional will always take edge %d->%d\n", (*taken_edge_p)->src->index, (*taken_edge_p)->dest->index); return retval; } /* Evaluate statement STMT. If the statement produces a new output value, return SSA_PROP_INTERESTING and store the SSA_NAME holding the new value in *RESULT_P. If STMT is a conditional branch and we can determine its truth value, set *TAKEN_EDGE_P accordingly. If the new value produced by STMT is varying, return SSA_PROP_VARYING. */ static enum ssa_prop_result copy_prop_visit_stmt (gimple stmt, edge *taken_edge_p, tree *result_p) { enum ssa_prop_result retval; if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, "\nVisiting statement:\n"); print_gimple_stmt (dump_file, stmt, 0, dump_flags); fprintf (dump_file, "\n"); } if (gimple_assign_single_p (stmt) && TREE_CODE (gimple_assign_lhs (stmt)) == SSA_NAME && TREE_CODE (gimple_assign_rhs1 (stmt)) == SSA_NAME) { /* If the statement is a copy assignment, evaluate its RHS to see if the lattice value of its output has changed. */ retval = copy_prop_visit_assignment (stmt, result_p); } else if (gimple_code (stmt) == GIMPLE_COND) { /* See if we can determine which edge goes out of a conditional jump. */ retval = copy_prop_visit_cond_stmt (stmt, taken_edge_p); } else retval = SSA_PROP_VARYING; if (retval == SSA_PROP_VARYING) { tree def; ssa_op_iter i; /* Any other kind of statement is not interesting for constant propagation and, therefore, not worth simulating. */ if (dump_file && (dump_flags & TDF_DETAILS)) fprintf (dump_file, "No interesting values produced.\n"); /* The assignment is not a copy operation. Don't visit this statement again and mark all the definitions in the statement to be copies of nothing. */ FOR_EACH_SSA_TREE_OPERAND (def, stmt, i, SSA_OP_ALL_DEFS) set_copy_of_val (def, def); } return retval; } /* Visit PHI node PHI. If all the arguments produce the same value, set it to be the value of the LHS of PHI. */ static enum ssa_prop_result copy_prop_visit_phi_node (gimple phi) { enum ssa_prop_result retval; unsigned i; prop_value_t phi_val = { 0, NULL_TREE }; tree lhs = gimple_phi_result (phi); if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, "\nVisiting PHI node: "); print_gimple_stmt (dump_file, phi, 0, dump_flags); fprintf (dump_file, "\n\n"); } for (i = 0; i < gimple_phi_num_args (phi); i++) { prop_value_t *arg_val; tree arg = gimple_phi_arg_def (phi, i); edge e = gimple_phi_arg_edge (phi, i); /* We don't care about values flowing through non-executable edges. */ if (!(e->flags & EDGE_EXECUTABLE)) continue; /* Constants in the argument list never generate a useful copy. Similarly, names that flow through abnormal edges cannot be used to derive copies. */ if (TREE_CODE (arg) != SSA_NAME || SSA_NAME_OCCURS_IN_ABNORMAL_PHI (arg)) { phi_val.value = lhs; break; } /* Avoid copy propagation from an inner into an outer loop. Otherwise, this may move loop variant variables outside of their loops and prevent coalescing opportunities. If the value was loop invariant, it will be hoisted by LICM and exposed for copy propagation. Not a problem for virtual operands though. */ if (is_gimple_reg (lhs) && loop_depth_of_name (arg) > loop_depth_of_name (lhs)) { phi_val.value = lhs; break; } /* If the LHS appears in the argument list, ignore it. It is irrelevant as a copy. */ if (arg == lhs || get_last_copy_of (arg) == lhs) continue; if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, "\tArgument #%d: ", i); dump_copy_of (dump_file, arg); fprintf (dump_file, "\n"); } arg_val = get_copy_of_val (arg); /* If the LHS didn't have a value yet, make it a copy of the first argument we find. Notice that while we make the LHS be a copy of the argument itself, we take the memory reference from the argument's value so that we can compare it to the memory reference of all the other arguments. */ if (phi_val.value == NULL_TREE) { phi_val.value = arg_val->value ? arg_val->value : arg; continue; } /* If PHI_VAL and ARG don't have a common copy-of chain, then this PHI node cannot be a copy operation. Also, if we are copy propagating stores and these two arguments came from different memory references, they cannot be considered copies. */ if (get_last_copy_of (phi_val.value) != get_last_copy_of (arg)) { phi_val.value = lhs; break; } } if (phi_val.value && may_propagate_copy (lhs, phi_val.value) && set_copy_of_val (lhs, phi_val.value)) retval = (phi_val.value != lhs) ? SSA_PROP_INTERESTING : SSA_PROP_VARYING; else retval = SSA_PROP_NOT_INTERESTING; if (dump_file && (dump_flags & TDF_DETAILS)) { fprintf (dump_file, "\nPHI node "); dump_copy_of (dump_file, lhs); fprintf (dump_file, "\nTelling the propagator to "); if (retval == SSA_PROP_INTERESTING) fprintf (dump_file, "add SSA edges out of this PHI and continue."); else if (retval == SSA_PROP_VARYING) fprintf (dump_file, "add SSA edges out of this PHI and never visit again."); else fprintf (dump_file, "do nothing with SSA edges and keep iterating."); fprintf (dump_file, "\n\n"); } return retval; } /* Initialize structures used for copy propagation. PHIS_ONLY is true if we should only consider PHI nodes as generating copy propagation opportunities. */ static void init_copy_prop (void) { basic_block bb; copy_of = XCNEWVEC (prop_value_t, num_ssa_names); cached_last_copy_of = XCNEWVEC (tree, num_ssa_names); FOR_EACH_BB (bb) { gimple_stmt_iterator si; int depth = bb->loop_depth; bool loop_exit_p = false; for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si)) { gimple stmt = gsi_stmt (si); ssa_op_iter iter; tree def; /* The only statements that we care about are those that may generate useful copies. We also need to mark conditional jumps so that their outgoing edges are added to the work lists of the propagator. Avoid copy propagation from an inner into an outer loop. Otherwise, this may move loop variant variables outside of their loops and prevent coalescing opportunities. If the value was loop invariant, it will be hoisted by LICM and exposed for copy propagation. */ if (stmt_ends_bb_p (stmt)) prop_set_simulate_again (stmt, true); else if (stmt_may_generate_copy (stmt) /* Since we are iterating over the statements in BB, not the phi nodes, STMT will always be an assignment. */ && loop_depth_of_name (gimple_assign_rhs1 (stmt)) <= depth) prop_set_simulate_again (stmt, true); else prop_set_simulate_again (stmt, false); /* Mark all the outputs of this statement as not being the copy of anything. */ FOR_EACH_SSA_TREE_OPERAND (def, stmt, iter, SSA_OP_ALL_DEFS) if (!prop_simulate_again_p (stmt)) set_copy_of_val (def, def); else cached_last_copy_of[SSA_NAME_VERSION (def)] = def; } /* In loop-closed SSA form do not copy-propagate through PHI nodes in blocks with a loop exit edge predecessor. */ if (current_loops && loops_state_satisfies_p (LOOP_CLOSED_SSA)) { edge_iterator ei; edge e; FOR_EACH_EDGE (e, ei, bb->preds) if (loop_exit_edge_p (e->src->loop_father, e)) loop_exit_p = true; } for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si)) { gimple phi = gsi_stmt (si); tree def; def = gimple_phi_result (phi); if (!is_gimple_reg (def) || loop_exit_p) prop_set_simulate_again (phi, false); else prop_set_simulate_again (phi, true); if (!prop_simulate_again_p (phi)) set_copy_of_val (def, def); else cached_last_copy_of[SSA_NAME_VERSION (def)] = def; } } } /* Deallocate memory used in copy propagation and do final substitution. */ static void fini_copy_prop (void) { size_t i; prop_value_t *tmp; /* Set the final copy-of value for each variable by traversing the copy-of chains. */ tmp = XCNEWVEC (prop_value_t, num_ssa_names); for (i = 1; i < num_ssa_names; i++) { tree var = ssa_name (i); if (!var || !copy_of[i].value || copy_of[i].value == var) continue; tmp[i].value = get_last_copy_of (var); /* In theory the points-to solution of all members of the copy chain is their intersection. For now we do not bother to compute this but only make sure we do not lose points-to information completely by setting the points-to solution of the representative to the first solution we find if it doesn't have one already. */ if (tmp[i].value != var && POINTER_TYPE_P (TREE_TYPE (var)) && SSA_NAME_PTR_INFO (var) && !SSA_NAME_PTR_INFO (tmp[i].value)) duplicate_ssa_name_ptr_info (tmp[i].value, SSA_NAME_PTR_INFO (var)); } substitute_and_fold (tmp, NULL, true); free (cached_last_copy_of); free (copy_of); free (tmp); } /* Main entry point to the copy propagator. PHIS_ONLY is true if we should only consider PHI nodes as generating copy propagation opportunities. The algorithm propagates the value COPY-OF using ssa_propagate. For every variable X_i, COPY-OF(X_i) indicates which variable is X_i created from. The following example shows how the algorithm proceeds at a high level: 1 a_24 = x_1 2 a_2 = PHI <a_24, x_1> 3 a_5 = PHI <a_2> 4 x_1 = PHI <x_298, a_5, a_2> The end result should be that a_2, a_5, a_24 and x_1 are a copy of x_298. Propagation proceeds as follows. Visit #1: a_24 is copy-of x_1. Value changed. Visit #2: a_2 is copy-of x_1. Value changed. Visit #3: a_5 is copy-of x_1. Value changed. Visit #4: x_1 is copy-of x_298. Value changed. Visit #1: a_24 is copy-of x_298. Value changed. Visit #2: a_2 is copy-of x_298. Value changed. Visit #3: a_5 is copy-of x_298. Value changed. Visit #4: x_1 is copy-of x_298. Stable state reached. When visiting PHI nodes, we only consider arguments that flow through edges marked executable by the propagation engine. So, when visiting statement #2 for the first time, we will only look at the first argument (a_24) and optimistically assume that its value is the copy of a_24 (x_1). The problem with this approach is that it may fail to discover copy relations in PHI cycles. Instead of propagating copy-of values, we actually propagate copy-of chains. For instance: A_3 = B_1; C_9 = A_3; D_4 = C_9; X_i = D_4; In this code fragment, COPY-OF (X_i) = { D_4, C_9, A_3, B_1 }. Obviously, we are only really interested in the last value of the chain, however the propagator needs to access the copy-of chain when visiting PHI nodes. To represent the copy-of chain, we use the array COPY_CHAINS, which holds the first link in the copy-of chain for every variable. If variable X_i is a copy of X_j, which in turn is a copy of X_k, the array will contain: COPY_CHAINS[i] = X_j COPY_CHAINS[j] = X_k COPY_CHAINS[k] = X_k Keeping copy-of chains instead of copy-of values directly becomes important when visiting PHI nodes. Suppose that we had the following PHI cycle, such that x_52 is already considered a copy of x_53: 1 x_54 = PHI <x_53, x_52> 2 x_53 = PHI <x_898, x_54> Visit #1: x_54 is copy-of x_53 (because x_52 is copy-of x_53) Visit #2: x_53 is copy-of x_898 (because x_54 is a copy of x_53, so it is considered irrelevant as a copy). Visit #1: x_54 is copy-of nothing (x_53 is a copy-of x_898 and x_52 is a copy of x_53, so they don't match) Visit #2: x_53 is copy-of nothing This problem is avoided by keeping a chain of copies, instead of the final copy-of value. Propagation will now only keep the first element of a variable's copy-of chain. When visiting PHI nodes, arguments are considered equal if their copy-of chains end in the same variable. So, as long as their copy-of chains overlap, we know that they will be a copy of the same variable, regardless of which variable that may be). Propagation would then proceed as follows (the notation a -> b means that a is a copy-of b): Visit #1: x_54 = PHI <x_53, x_52> x_53 -> x_53 x_52 -> x_53 Result: x_54 -> x_53. Value changed. Add SSA edges. Visit #1: x_53 = PHI <x_898, x_54> x_898 -> x_898 x_54 -> x_53 Result: x_53 -> x_898. Value changed. Add SSA edges. Visit #2: x_54 = PHI <x_53, x_52> x_53 -> x_898 x_52 -> x_53 -> x_898 Result: x_54 -> x_898. Value changed. Add SSA edges. Visit #2: x_53 = PHI <x_898, x_54> x_898 -> x_898 x_54 -> x_898 Result: x_53 -> x_898. Value didn't change. Stable state Once the propagator stabilizes, we end up with the desired result x_53 and x_54 are both copies of x_898. */ static unsigned int execute_copy_prop (void) { init_copy_prop (); ssa_propagate (copy_prop_visit_stmt, copy_prop_visit_phi_node); fini_copy_prop (); return 0; } static bool gate_copy_prop (void) { return flag_tree_copy_prop != 0; } struct gimple_opt_pass pass_copy_prop = { { GIMPLE_PASS, "copyprop", /* name */ gate_copy_prop, /* gate */ execute_copy_prop, /* execute */ NULL, /* sub */ NULL, /* next */ 0, /* static_pass_number */ TV_TREE_COPY_PROP, /* tv_id */ PROP_ssa | PROP_cfg, /* properties_required */ 0, /* properties_provided */ 0, /* properties_destroyed */ 0, /* todo_flags_start */ TODO_cleanup_cfg | TODO_dump_func | TODO_ggc_collect | TODO_verify_ssa | TODO_update_ssa /* todo_flags_finish */ } };
Go to most recent revision | Compare with Previous | Blame | View Log