- Generated by
1.8.1.1
|
GCC Middle and Back End API Reference
|
Data Structures | |
| struct | omp_context |
| struct | omp_for_data_loop |
| struct | omp_for_data |
| struct | omp_taskcopy_context |
Typedefs | |
| typedef struct omp_context | omp_context |
Variables | |
| static splay_tree | all_contexts |
| static int | taskreg_nesting_level |
| struct omp_region * | root_omp_region |
| static bitmap | task_shared_vars |
| static unsigned int | tmp_ompfn_id_num |
| static splay_tree | critical_name_mutexes |
| static splay_tree | all_labels |
| typedef struct omp_context omp_context |
@verbatim Lowering pass for OpenMP directives. Converts OpenMP directives
into explicit calls to the runtime library (libgomp) and data marshalling to implement data sharing and copying clauses. Contributed by Diego Novillo dnovillo@redhat.com
Copyright (C) 2005-2013 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/.
Lowering of OpenMP parallel and workshare constructs proceeds in two phases. The first phase scans the function looking for OMP statements and then for variables that must be replaced to satisfy data sharing clauses. The second phase expands code for the constructs, as well as re-gimplifying things when variables have been replaced with complex expressions. Final code generation is done by pass_expand_omp. The flowgraph is scanned for parallel regions which are then moved to a new function, to be invoked by the thread library.
Context structure. Used to store information about each parallel directive in the code.
|
static |
Re-gimplification and code generation routines.
Build a call to GOMP_barrier.
References build_call_expr(), and builtin_decl_explicit().
Referenced by expand_omp_for_static_chunk(), expand_omp_for_static_nochunk(), expand_omp_single(), and lower_rec_input_clauses().
|
static |
Scan the CFG and build a tree of OMP regions. Return the root of the OMP region tree.
References build_omp_regions_1(), calculate_dominance_info(), and CDI_DOMINATORS.
Referenced by execute_expand_omp().
|
static |
Helper for build_omp_regions. Scan the dominator tree starting at block BB. PARENT is the region that contains BB. If SINGLE_TREE is true, the function ends once a single tree is built (otherwise, whole forest of OMP constructs may be built).
References CDI_DOMINATORS, omp_region::cont, omp_region::exit, first_dom_son(), gsi_end_p(), gsi_last_bb(), gsi_stmt(), is_gimple_omp(), new_omp_region(), next_dom_son(), omp_region::outer, and omp_region::type.
Referenced by build_omp_regions(), and build_omp_regions_root().
|
static |
Builds the tree of OMP regions rooted at ROOT, storing it to root_omp_region.
References build_omp_regions_1().
Referenced by omp_expand_local().
|
static |
Build tree nodes to access VAR in the scope outer to CTX. In the case of a parallel, this is a component reference; for workshare constructs this is some variable.
References build_receiver_ref(), GF_OMP_FOR_KIND_SIMD, gimple_omp_for_kind(), is_global_var(), is_reference(), is_taskreg_ctx(), is_variable_sized(), lookup_decl(), maybe_lookup_decl(), maybe_lookup_decl_in_outer_ctx(), omp_context::outer, omp_context::stmt, and use_pointer_for_field().
Referenced by lower_lastprivate_clauses(), lower_rec_input_clauses(), and lower_reduction_clauses().
|
static |
Build tree nodes to access the field for VAR on the receiver side.
References lookup_field(), maybe_lookup_field(), omp_build_component_ref(), and omp_context::receiver_decl.
Referenced by build_outer_var_ref(), lower_copyprivate_clauses(), and lower_rec_input_clauses().
|
static |
Build tree nodes to access the field for VAR on the sender side.
References lookup_sfield(), omp_build_component_ref(), and omp_context::sender_decl.
Referenced by lower_copyprivate_clauses(), lower_send_clauses(), and lower_send_shared_vars().
|
static |
Callback for walk_stmts. Check if the current statement only contains GIMPLE_OMP_FOR or GIMPLE_OMP_PARALLEL.
References gsi_stmt(), and walk_stmt_info::info.
Referenced by lower_omp_taskreg().
|
static |
Check OpenMP nesting restrictions.
References error_at(), find_omp_clause(), GF_OMP_FOR_KIND_SIMD, gimple_location(), gimple_omp_critical_name(), gimple_omp_for_clauses(), gimple_omp_for_kind(), is_gimple_call(), OMP_CLAUSE_ORDERED, omp_context::outer, and omp_context::stmt.
Referenced by scan_omp_1_stmt().
| tree copy_var_decl | ( | ) |
Create a new VAR_DECL and copy information from VAR to it.
References copy().
Referenced by omp_copy_decl_2(), and replace_by_duplicate_decl().
|
static |
Build a decl for the omp child function. It'll not contain a body yet, just the bare decl.
References build_function_type_list(), omp_context::cb, cfun, create_omp_child_function_name(), current_function_decl, function::decl, copy_body_data::dst_fn, function::function_end_locus, get_identifier(), gimple_location(), gimple_omp_task_set_copy_fn(), pop_cfun(), push_struct_function(), omp_context::receiver_decl, omp_context::stmt, and type().
Referenced by scan_omp_parallel(), and scan_omp_task().
|
static |
References clone_function_name(), and current_function_decl.
Referenced by create_omp_child_function().
|
static |
Create task copyfn.
References alloc_stmt_list(), append_to_statement_list(), build_fold_addr_expr_loc(), build_pointer_type(), build_simple_mem_ref_loc(), omp_context::cb, omp_taskcopy_context::cb, CB_CGE_MOVE, function::cfg, cgraph_get_node(), copy_body_data::copy_decl, omp_taskcopy_context::ctx, function::decl, copy_body_data::decl_map, lang_hooks::decls, df, copy_body_data::dst_fn, copy_body_data::dst_node, copy_body_data::eh_lp_nr, omp_context::field_map, gimple_location(), gimple_omp_task_clauses(), gimple_omp_task_copy_fn(), is_reference(), is_variable_sized(), memset(), omp_build_component_ref(), lang_hooks_for_decls::omp_clause_copy_ctor, OMP_CLAUSE_FIRSTPRIVATE, OMP_CLAUSE_PRIVATE, OMP_CLAUSE_SHARED, pointer_map_contains(), pointer_map_create(), pointer_map_destroy(), pop_cfun(), pop_gimplify_context(), push_cfun(), push_gimplify_context(), omp_context::record_type, record_type(), omp_context::sfield_map, copy_body_data::src_cfun, copy_body_data::src_fn, copy_body_data::src_node, omp_context::srecord_type, task_copyfn_copy_decl(), task_copyfn_remap_type(), copy_body_data::transform_call_graph_edges, use_pointer_for_field(), and variably_modified_type_p().
Referenced by lower_omp_taskreg().
| DEBUG_FUNCTION void debug_all_omp_regions | ( | void | ) |
References dump_omp_region().
| void debug_omp_region | ( | struct omp_region * | ) |
| DEBUG_FUNCTION void debug_omp_region | ( | ) |
References dump_omp_region().
|
static |
Destroy a omp_context data structures. Called through the splay tree value delete callback.
References omp_context::cb, copy_body_data::decl_map, omp_context::field_map, finalize_task_copyfn(), is_task_ctx(), pointer_map_destroy(), omp_context::record_type, omp_context::sfield_map, omp_context::srecord_type, and omp_context::stmt.
Referenced by execute_lower_omp().
|
static |
Discover whether REGION is a combined parallel+workshare region.
References omp_region::cont, omp_region::entry, omp_region::exit, find_omp_clause(), get_ws_args_for(), gimple_omp_for_clauses(), gimple_omp_parallel_combined_p(), omp_region::inner, omp_region::is_combined_parallel, last_and_only_stmt(), last_stmt(), OMP_CLAUSE_ORDERED, OMP_CLAUSE_SCHEDULE, OMP_CLAUSE_SCHEDULE_STATIC, single_succ(), omp_region::type, workshare_safe_to_combine_p(), and omp_region::ws_args.
Referenced by expand_omp().
|
static |
|
static |
Check for mismatched contexts and generate an error if needed. Return true if an error is detected.
References error(), gimple_build_nop(), and gsi_replace().
Referenced by diagnose_sb_2().
|
static |
Pass 1: Create a minimal tree of OpenMP structured blocks, and record where each label is found.
References gimple_label_label(), gimple_omp_body(), gimple_omp_for_pre_body(), gsi_stmt(), walk_stmt_info::info, and walk_gimple_seq().
Referenced by diagnose_omp_structured_block_errors().
|
static |
Pass 2: Check each branch and see if its context differs from that of the destination label's context.
References diagnose_sb_0(), gimple_cond_false_label(), gimple_cond_true_label(), gimple_goto_dest(), gimple_omp_body_ptr(), gimple_omp_for_pre_body_ptr(), gimple_switch_label(), gimple_switch_num_labels(), gsi_stmt(), walk_stmt_info::info, and walk_gimple_seq_mod().
Referenced by diagnose_omp_structured_block_errors().
| void dump_omp_region | ( | FILE * | , |
| struct omp_region * | , | ||
| int | |||
| ) |
Return the parallel region associated with STMT.
Debugging dumps for parallel regions.
Referenced by debug_all_omp_regions(), debug_omp_region(), dump_omp_region(), execute_expand_omp(), and omp_expand_local().
| void dump_omp_region | ( | ) |
Dump the parallel region tree rooted at REGION.
References omp_region::cont, dump_omp_region(), omp_region::entry, omp_region::exit, gimple_code_name, basic_block_def::index, omp_region::inner, omp_region::next, and omp_region::type.
|
static |
Main entry point for expanding OMP-GIMPLE into runtime calls.
References build_omp_regions(), cleanup_tree_cfg(), dump_file, dump_omp_region(), expand_omp(), free_omp_regions(), and remove_exit_barriers().
|
static |
Main entry point.
References all_contexts, current_function_decl, delete_omp_context(), gimple_body(), lower_omp(), pop_gimplify_context(), push_gimplify_context(), scan_omp(), and taskreg_nesting_level.
|
static |
Expand the parallel region tree rooted at REGION. Expansion proceeds in depth-first order. Innermost regions are expanded first. This way, parallel regions that require a new function to be created (e.g., GIMPLE_OMP_PARALLEL) can be expanded without having any internal dependencies in their body.
References determine_parallel_type(), omp_region::entry, expand_omp_atomic(), expand_omp_for(), expand_omp_sections(), expand_omp_single(), expand_omp_synch(), expand_omp_taskreg(), gimple_has_location(), gimple_location(), omp_region::inner, input_location, last_stmt(), omp_region::next, and omp_region::type.
Referenced by execute_expand_omp(), and omp_expand_local().
|
static |
Expand an GIMPLE_OMP_ATOMIC statement. We try to expand using expand_omp_atomic_fetch_op. If it failed, we try to call expand_omp_atomic_pipeline, and if it fails too, the ultimate fallback is wrapping the operation in a mutex (expand_omp_atomic_mutex). REGION is the atomic region built by build_omp_regions_1().
References omp_region::entry, exact_log2(), omp_region::exit, expand_omp_atomic_fetch_op(), expand_omp_atomic_load(), expand_omp_atomic_mutex(), expand_omp_atomic_pipeline(), expand_omp_atomic_store(), first_stmt(), gimple_omp_atomic_load_lhs(), gimple_omp_atomic_load_rhs(), gimple_omp_atomic_store_val(), HOST_WIDE_INT, last_stmt(), single_succ(), and tree_low_cst().
Referenced by expand_omp().
|
static |
A subroutine of expand_omp_atomic. Attempt to implement the atomic operation as a __atomic_fetch_op builtin. INDEX is log2 of the size of the data type, and thus usable to find the index of the builtin decl. Returns false if the expression is not of the proper form.
References build_call_expr_loc(), build_int_cst(), builtin_decl_explicit(), can_compare_and_swap_p(), cfun, commutative_tree_code(), fold_convert_loc(), force_gimple_operand_gsi(), gimple_assign_lhs(), gimple_assign_rhs1(), gimple_assign_rhs2(), gimple_assign_rhs_code(), gimple_in_ssa_p(), gimple_location(), gimple_omp_atomic_need_value_p(), gsi_after_labels(), gsi_last_bb(), gsi_next(), gsi_remove(), GSI_SAME_STMT, gsi_stmt(), is_gimple_assign(), last_stmt(), MEMMODEL_RELAXED, operand_equal_p(), single_succ(), and update_ssa().
Referenced by expand_omp_atomic().
|
static |
A subroutine of expand_omp_atomic. Attempt to implement the atomic operation as a normal volatile load.
References build_call_expr_loc(), build_int_cst(), builtin_decl_explicit(), cfun, force_gimple_operand_gsi(), gimple_in_ssa_p(), gimple_location(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_stmt(), MEMMODEL_RELAXED, single_succ(), type(), update_ssa(), and useless_type_conversion_p().
Referenced by expand_omp_atomic().
|
static |
A subroutine of expand_omp_atomic. Implement the atomic operation as:
GOMP_atomic_start ();
*addr = rhs;
GOMP_atomic_end ();
The result is not globally atomic, but works so long as all parallel
references are within #pragma omp atomic directives. According to
responses received from omp@openmp.org, appears to be within spec.
Which makes sense, since that's how several other compilers handle
this situation as well.
LOADED_VAL and ADDR are the operands of GIMPLE_OMP_ATOMIC_LOAD we're
expanding. STORED_VAL is the operand of the matching
GIMPLE_OMP_ATOMIC_STORE.
We replace
GIMPLE_OMP_ATOMIC_LOAD (loaded_val, addr) with
loaded_val = *addr;
and replace
GIMPLE_OMP_ATOMIC_STORE (stored_val) with
*addr = stored_val;
References build_call_expr(), builtin_decl_explicit(), cfun, force_gimple_operand_gsi(), gimple_in_ssa_p(), gsi_insert_before(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_stmt(), si, unshare_expr(), and update_ssa().
Referenced by expand_omp_atomic().
|
static |
A subroutine of expand_omp_atomic. Implement the atomic operation as:
oldval = *addr;
repeat:
newval = rhs; // with oldval replacing *addr in rhs
oldval = __sync_val_compare_and_swap (addr, oldval, newval);
if (oldval != newval)
goto repeat;
INDEX is log2 of the size of the data type, and thus usable to find the
index of the builtin decl.
References add_loop(), alloc_loop(), build_call_expr(), build_int_cst(), build_pointer_type_for_mode(), builtin_decl_explicit(), can_compare_and_swap_p(), cfun, create_phi_node(), create_tmp_reg(), create_tmp_var(), edge_def::flags, force_gimple_operand_gsi(), gimple_build_cond_empty(), gimple_in_ssa_p(), gimple_seq_empty_p(), gimple_seq_first_stmt(), gsi_insert_before(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_stmt(), loop::header, loop::latch, basic_block_def::loop_father, make_edge(), make_ssa_name(), phi_nodes(), ptr_mode, si, single_succ(), single_succ_edge(), type(), and update_ssa().
Referenced by expand_omp_atomic().
|
static |
A subroutine of expand_omp_atomic. Attempt to implement the atomic operation as a normal volatile store.
References build_call_expr_loc(), build_int_cst(), builtin_decl_explicit(), can_atomic_exchange_p(), cfun, force_gimple_operand_gsi(), gimple_in_ssa_p(), gimple_location(), gimple_omp_atomic_need_value_p(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_stmt(), MEMMODEL_RELAXED, single_succ(), type(), update_ssa(), and useless_type_conversion_p().
Referenced by expand_omp_atomic().
|
static |
Prepend TO = FROM assignment before *GSI_P.
References expand_omp_regimplify_p(), force_gimple_operand_gsi(), gimple_regimplify_operands(), walk_stmt_info::gsi, gsi_for_stmt(), gsi_insert_before(), and GSI_SAME_STMT.
Referenced by expand_omp_for_init_counts(), and expand_omp_simd().
|
static |
Expand the OpenMP loop defined by REGION.
References cfun, omp_for_data::chunk_size, omp_for_data::collapse, omp_region::cont, omp_region::entry, expand_omp_for_generic(), expand_omp_for_static_chunk(), expand_omp_for_static_nochunk(), expand_omp_simd(), extract_omp_for_data(), omp_for_data::for_stmt, GF_OMP_FOR_KIND_FOR, GF_OMP_FOR_KIND_SIMD, gimple_in_ssa_p(), gimple_omp_for_collapse(), gimple_omp_for_kind(), omp_for_data::have_ordered, omp_for_data::iter_type, last_stmt(), LOOPS_NEED_FIXUP, loops_state_set(), OMP_CLAUSE_SCHEDULE_AUTO, OMP_CLAUSE_SCHEDULE_RUNTIME, OMP_CLAUSE_SCHEDULE_STATIC, omp_for_data::sched_kind, omp_region::sched_kind, basic_block_def::succs, and update_ssa().
Referenced by expand_omp().
|
static |
A subroutine of expand_omp_for. Generate code for a parallel
loop with any schedule. Given parameters:
for (V = N1; V cond N2; V += STEP) BODY;
where COND is "<" or ">", we generate pseudocode
more = GOMP_loop_foo_start (N1, N2, STEP, CHUNK, &istart0, &iend0);
if (more) goto L0; else goto L3;
L0:
V = istart0;
iend = iend0;
L1:
BODY;
V += STEP;
if (V cond iend) goto L1; else goto L2;
L2:
if (GOMP_loop_foo_next (&istart0, &iend0)) goto L0; else goto L3;
L3:
If this is a combined omp parallel loop, instead of the call to
GOMP_loop_foo_start, we call GOMP_loop_foo_next.
For collapsed loops, given parameters:
collapse(3)
for (V1 = N11; V1 cond1 N12; V1 += STEP1)
for (V2 = N21; V2 cond2 N22; V2 += STEP2)
for (V3 = N31; V3 cond3 N32; V3 += STEP3)
BODY;
we generate pseudocode
if (__builtin_expect (N32 cond3 N31, 0)) goto Z0;
if (cond3 is <)
adj = STEP3 - 1;
else
adj = STEP3 + 1;
count3 = (adj + N32 - N31) / STEP3;
if (__builtin_expect (N22 cond2 N21, 0)) goto Z0;
if (cond2 is <)
adj = STEP2 - 1;
else
adj = STEP2 + 1;
count2 = (adj + N22 - N21) / STEP2;
if (__builtin_expect (N12 cond1 N11, 0)) goto Z0;
if (cond1 is <)
adj = STEP1 - 1;
else
adj = STEP1 + 1;
count1 = (adj + N12 - N11) / STEP1;
count = count1 * count2 * count3;
goto Z1;
Z0:
count = 0;
Z1:
more = GOMP_loop_foo_start (0, count, 1, CHUNK, &istart0, &iend0);
if (more) goto L0; else goto L3;
L0:
V = istart0;
T = V;
V3 = N31 + (T % count3) * STEP3;
T = T / count3;
V2 = N21 + (T % count2) * STEP2;
T = T / count2;
V1 = N11 + T * STEP1;
iend = iend0;
L1:
BODY;
V += 1;
if (V < iend) goto L10; else goto L2;
L10:
V3 += STEP3;
if (V3 cond3 N32) goto L1; else goto L11;
L11:
V3 = N31;
V2 += STEP2;
if (V2 cond2 N22) goto L1; else goto L12;
L12:
V2 = N21;
V1 += STEP1;
goto L1;
L2:
if (GOMP_loop_foo_next (&istart0, &iend0)) goto L0; else goto L3;
L3:
References add_bb_to_loop(), add_loop(), alloc_loop(), build_call_expr(), build_int_cst(), builtin_decl_explicit(), CDI_DOMINATORS, omp_for_data::chunk_size, omp_for_data::collapse, omp_for_data_loop::cond_code, omp_region::cont, create_empty_bb(), create_tmp_var(), edge_def::dest, omp_region::entry, omp_region::exit, expand_omp_for_init_counts(), expand_omp_for_init_vars(), extract_omp_for_update_vars(), find_edge(), edge_def::flags, force_gimple_operand_gsi(), get_immediate_dominator(), gimple_build_call(), gimple_build_cond_empty(), gimple_omp_continue_control_def(), gimple_omp_continue_control_use(), gimple_omp_return_nowait_p(), GSI_CONTINUE_LINKING, gsi_end_p(), gsi_insert_after(), gsi_insert_before(), gsi_last_bb(), gsi_next(), gsi_prev(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_stmt(), loop::header, is_combined_parallel(), omp_for_data::iter_type, loop::latch, omp_for_data::loop, basic_block_def::loop_father, make_edge(), omp_for_data_loop::n1, omp_for_data_loop::n2, phi_nodes(), phis, edge_def::probability, recompute_dominator(), remove_edge(), set_immediate_dominator(), signed_type_for(), single_succ(), split_block(), split_edge(), omp_for_data_loop::step, basic_block_def::succs, tree_int_cst_sgn(), type(), and omp_for_data_loop::v.
Referenced by expand_omp_for().
|
static |
Helper function for expand_omp_{for_*,simd}. If this is the outermost
of the combined collapse > 1 loop constructs, generate code like:
if (__builtin_expect (N32 cond3 N31, 0)) goto ZERO_ITER_BB;
if (cond3 is <)
adj = STEP3 - 1;
else
adj = STEP3 + 1;
count3 = (adj + N32 - N31) / STEP3;
if (__builtin_expect (N22 cond2 N21, 0)) goto ZERO_ITER_BB;
if (cond2 is <)
adj = STEP2 - 1;
else
adj = STEP2 + 1;
count2 = (adj + N22 - N21) / STEP2;
if (__builtin_expect (N12 cond1 N11, 0)) goto ZERO_ITER_BB;
if (cond1 is <)
adj = STEP1 - 1;
else
adj = STEP1 + 1;
count1 = (adj + N12 - N11) / STEP1;
count = count1 * count2 * count3;
Furthermore, if ZERO_ITER_BB is NULL, create a BB which does:
count = 0;
and set ZERO_ITER_BB to that bb. NOTE: It *could* be better to moosh all of the BBs together, creating one larger BB with all the computation and the unexpected jump at the end. I.e. bool zero3, zero2, zero1, zero; zero3 = N32 c3 N31; count3 = (N32 - N31) /[cl] STEP3; zero2 = N22 c2 N21; count2 = (N22 - N21) /[cl] STEP2; zero1 = N12 c1 N11; count1 = (N12 - N11) /[cl] STEP1; zero = zero3 || zero2 || zero1; count = count1 * count2 * count3; if (__builtin_expect(zero, false)) goto zero_iter_bb; After all, we expect the zero=false, and thus we expect to have to evaluate all of the comparison expressions, so short-circuiting oughtn't be a win. Since the condition isn't protecting a denominator, we're not concerned about divide-by-zero, so we can fully evaluate count even if a numerator turned out to be wrong. It seems like putting this all together would create much better scheduling opportunities, and less pressure on the chip's branch predictor.
References add_bb_to_loop(), build_int_cst(), build_zero_cst(), CDI_DOMINATORS, cfun, omp_for_data::collapse, omp_for_data_loop::cond_code, create_empty_bb(), create_tmp_reg(), edge_def::dest, expand_omp_build_assign(), expand_omp_regimplify_p(), edge_def::flags, force_gimple_operand_gsi(), gimple_build_cond(), gimple_cond_lhs_ptr(), gimple_cond_rhs_ptr(), gimple_in_ssa_p(), gimple_regimplify_operands(), gsi_after_labels(), gsi_for_stmt(), gsi_insert_before(), gsi_last_bb(), GSI_SAME_STMT, integer_onep(), omp_for_data::loop, basic_block_def::loop_father, omp_for_data::loops, make_edge(), omp_for_data_loop::n1, omp_for_data_loop::n2, edge_def::probability, set_immediate_dominator(), signed_type_for(), split_block(), omp_for_data_loop::step, unshare_expr(), and omp_for_data_loop::v.
Referenced by expand_omp_for_generic(), and expand_omp_simd().
|
static |
Helper function for expand_omp_{for_*,simd}. Generate code like:
T = V;
V3 = N31 + (T % count3) * STEP3;
T = T / count3;
V2 = N21 + (T % count2) * STEP2;
T = T / count2;
V1 = N11 + T * STEP1;
if this loop doesn't have an inner loop construct combined with it.
References omp_for_data::collapse, create_tmp_reg(), force_gimple_operand_gsi(), GSI_CONTINUE_LINKING, gsi_insert_after(), omp_for_data::loop, omp_for_data::loops, omp_for_data_loop::n1, signed_type_for(), omp_for_data_loop::step, and omp_for_data_loop::v.
Referenced by expand_omp_for_generic().
|
static |
A subroutine of expand_omp_for. Generate code for a parallel
loop with static schedule and a specified chunk size. Given
parameters:
for (V = N1; V cond N2; V += STEP) BODY;
where COND is "<" or ">", we generate pseudocode
if ((__typeof (V)) -1 > 0 && N2 cond N1) goto L2;
if (cond is <)
adj = STEP - 1;
else
adj = STEP + 1;
if ((__typeof (V)) -1 > 0 && cond is >)
n = -(adj + N2 - N1) / -STEP;
else
n = (adj + N2 - N1) / STEP;
trip = 0;
V = threadid * CHUNK * STEP + N1; -- this extra definition of V is
here so that V is defined
if the loop is not entered
L0:
s0 = (trip * nthreads + threadid) * CHUNK;
e0 = min(s0 + CHUNK, n);
if (s0 < n) goto L1; else goto L4;
L1:
V = s0 * STEP + N1;
e = e0 * STEP + N1;
L2:
BODY;
V += STEP;
if (V cond e) goto L2; else goto L3;
L3:
trip += 1;
goto L0;
L4:
References add_loop(), add_phi_arg(), alloc_loop(), build_call_expr(), build_int_cst(), build_omp_barrier(), builtin_decl_explicit(), CDI_DOMINATORS, cfun, omp_for_data::chunk_size, omp_for_data_loop::cond_code, omp_region::cont, create_phi_node(), create_tmp_reg(), edge_def::dest, edge_def::dest_idx, omp_region::entry, omp_region::exit, expand_omp_regimplify_p(), find_edge(), edge_def::flags, force_gimple_operand_gsi(), gimple_build_cond(), gimple_build_cond_empty(), gimple_cond_lhs_ptr(), gimple_cond_rhs_ptr(), gimple_in_ssa_p(), gimple_omp_continue_control_def(), gimple_omp_continue_control_use(), gimple_omp_return_nowait_p(), gimple_phi_arg_def(), gimple_phi_arg_location_from_edge(), gimple_phi_result(), gimple_regimplify_operands(), GSI_CONTINUE_LINKING, gsi_end_p(), gsi_for_stmt(), gsi_insert_after(), gsi_insert_before(), gsi_last_bb(), gsi_next(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_start_phis(), gsi_stmt(), loop::header, integer_onep(), last_stmt(), loop::latch, omp_for_data::loop, basic_block_def::loop_father, make_edge(), make_ssa_name(), omp_for_data_loop::n1, omp_for_data_loop::n2, edge_def::probability, recompute_dominator(), redirect_edge_and_branch(), redirect_edge_var_map_clear(), redirect_edge_var_map_def(), redirect_edge_var_map_location(), redirect_edge_var_map_result(), redirect_edge_var_map_vector(), remove_phi_node(), set_immediate_dominator(), si, signed_type_for(), single_pred_edge(), single_succ(), single_succ_edge(), split_block(), split_edge(), edge_def::src, omp_for_data_loop::step, basic_block_def::succs, type(), unshare_expr(), and omp_for_data_loop::v.
Referenced by expand_omp_for().
|
static |
A subroutine of expand_omp_for. Generate code for a parallel
loop with static schedule and no specified chunk size. Given
parameters:
for (V = N1; V cond N2; V += STEP) BODY;
where COND is "<" or ">", we generate pseudocode
if ((__typeof (V)) -1 > 0 && N2 cond N1) goto L2;
if (cond is <)
adj = STEP - 1;
else
adj = STEP + 1;
if ((__typeof (V)) -1 > 0 && cond is >)
n = -(adj + N2 - N1) / -STEP;
else
n = (adj + N2 - N1) / STEP;
q = n / nthreads;
tt = n % nthreads;
if (threadid < tt) goto L3; else goto L4;
L3:
tt = 0;
q = q + 1;
L4:
s0 = q * threadid + tt;
e0 = s0 + q;
V = s0 * STEP + N1;
if (s0 >= e0) goto L2; else goto L0;
L0:
e = e0 * STEP + N1;
L1:
BODY;
V += STEP;
if (V cond e) goto L1;
L2:
References add_loop(), add_phi_arg(), alloc_loop(), build_call_expr(), build_int_cst(), build_omp_barrier(), builtin_decl_explicit(), CDI_DOMINATORS, cfun, omp_for_data_loop::cond_code, omp_region::cont, create_tmp_reg(), edge_def::dest, edge_def::dest_idx, omp_region::entry, omp_region::exit, expand_omp_regimplify_p(), find_edge(), edge_def::flags, force_gimple_operand_gsi(), gimple_build_assign_with_ops(), gimple_build_cond(), gimple_build_cond_empty(), gimple_cond_lhs_ptr(), gimple_cond_rhs_ptr(), gimple_in_ssa_p(), gimple_omp_continue_control_def(), gimple_omp_continue_control_use(), gimple_omp_return_nowait_p(), gimple_phi_arg_def(), gimple_regimplify_operands(), GSI_CONTINUE_LINKING, gsi_end_p(), gsi_for_stmt(), gsi_insert_after(), gsi_insert_before(), gsi_last_bb(), gsi_next(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_start_phis(), gsi_stmt(), loop::header, integer_onep(), loop::latch, omp_for_data::loop, basic_block_def::loop_father, make_edge(), omp_for_data_loop::n1, omp_for_data_loop::n2, edge_def::probability, recompute_dominator(), set_immediate_dominator(), signed_type_for(), single_succ(), split_block(), split_edge(), omp_for_data_loop::step, basic_block_def::succs, type(), unshare_expr(), and omp_for_data_loop::v.
Referenced by expand_omp_for().
|
static |
Callback for expand_omp_build_assign. Return non-NULL if *tp needs to be regimplified.
References recompute_tree_invariant_for_addr_expr().
Referenced by expand_omp_build_assign(), expand_omp_for_init_counts(), expand_omp_for_static_chunk(), expand_omp_for_static_nochunk(), and expand_omp_simd().
|
static |
Expand code for an OpenMP sections directive. In pseudo code, we generate
v = GOMP_sections_start (n);
L0:
switch (v)
{
case 0:
goto L2;
case 1:
section 1;
goto L1;
case 2:
...
case n:
...
default:
abort ();
}
L1:
v = GOMP_sections_next ();
goto L0;
L2:
reduction;
If this is a combined parallel sections, replace the call to
GOMP_sections_start with call to GOMP_sections_next.
References add_bb_to_loop(), build_case_label(), build_int_cst(), builtin_decl_explicit(), CDI_DOMINATORS, omp_region::cont, create_empty_bb(), edge_def::dest, omp_region::entry, omp_region::exit, edge_def::flags, gimple_block_label(), gimple_build_call(), gimple_build_switch(), gimple_call_set_lhs(), gimple_omp_continue_control_def(), gimple_omp_continue_control_use(), gimple_omp_return_nowait_p(), gimple_omp_section_last_p(), gimple_omp_sections_control(), GSI_CONTINUE_LINKING, gsi_end_p(), gsi_insert_after(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_stmt(), omp_region::inner, is_combined_parallel(), last_stmt(), len, make_edge(), omp_region::next, basic_block_def::prev_bb, set_immediate_dominator(), si, single_pred(), single_pred_p(), single_succ(), single_succ_edge(), basic_block_def::succs, and omp_region::type.
Referenced by expand_omp().
|
static |
A subroutine of expand_omp_for. Generate code for a simd non-worksharing
loop. Given parameters:
for (V = N1; V cond N2; V += STEP) BODY;
where COND is "<" or ">", we generate pseudocode
V = N1;
goto L1;
L0:
BODY;
V += STEP;
L1:
if (V cond N2) goto L0; else goto L2;
L2:
For collapsed loops, given parameters:
collapse(3)
for (V1 = N11; V1 cond1 N12; V1 += STEP1)
for (V2 = N21; V2 cond2 N22; V2 += STEP2)
for (V3 = N31; V3 cond3 N32; V3 += STEP3)
BODY;
we generate pseudocode
if (cond3 is <)
adj = STEP3 - 1;
else
adj = STEP3 + 1;
count3 = (adj + N32 - N31) / STEP3;
if (cond2 is <)
adj = STEP2 - 1;
else
adj = STEP2 + 1;
count2 = (adj + N22 - N21) / STEP2;
if (cond1 is <)
adj = STEP1 - 1;
else
adj = STEP1 + 1;
count1 = (adj + N12 - N11) / STEP1;
count = count1 * count2 * count3;
V = 0;
V1 = N11;
V2 = N21;
V3 = N31;
goto L1;
L0:
BODY;
V += 1;
V3 += STEP3;
V2 += (V3 cond3 N32) ? 0 : STEP2;
V3 = (V3 cond3 N32) ? V3 : N31;
V1 += (V2 cond2 N22) ? 0 : STEP1;
V2 = (V2 cond2 N22) ? V2 : N21;
L1:
if (V < count) goto L0; else goto L2;
L2:
References add_loop(), alloc_loop(), build_int_cst(), CDI_DOMINATORS, cfun, omp_for_data::collapse, omp_for_data_loop::cond_code, omp_region::cont, edge_def::dest, omp_region::entry, omp_region::exit, expand_omp_build_assign(), expand_omp_for_init_counts(), expand_omp_regimplify_p(), find_omp_clause(), edge_def::flags, basic_block_def::flags, omp_for_data::for_stmt, force_gimple_operand_gsi(), loop::force_vect, gimple_build_cond_empty(), gimple_cond_lhs_ptr(), gimple_cond_rhs_ptr(), gimple_in_ssa_p(), gimple_omp_for_clauses(), gimple_regimplify_operands(), GSI_CONTINUE_LINKING, gsi_for_stmt(), gsi_insert_after(), gsi_last_bb(), gsi_remove(), gsi_start_bb(), gsi_stmt(), function::has_force_vect_loops, function::has_simduid_loops, loop::header, host_integerp(), HOST_WIDE_INT, last_stmt(), loop::latch, omp_for_data::loop, basic_block_def::loop_father, omp_for_data::loops, make_edge(), omp_for_data_loop::n1, omp_for_data_loop::n2, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_SAFELEN, edge_def::probability, remove_edge(), loop::safelen, set_immediate_dominator(), signed_type_for(), loop::simduid, single_succ(), single_succ_edge(), split_block(), split_edge(), omp_for_data_loop::step, basic_block_def::succs, tree_low_cst(), type(), and omp_for_data_loop::v.
Referenced by expand_omp_for().
|
static |
Expand code for an OpenMP single directive. We've already expanded much of the code, here we simply place the GOMP_barrier call.
References build_omp_barrier(), omp_region::entry, omp_region::exit, find_omp_clause(), edge_def::flags, force_gimple_operand_gsi(), gimple_omp_return_nowait_p(), gimple_omp_single_clauses(), gsi_last_bb(), gsi_remove(), GSI_SAME_STMT, gsi_stmt(), OMP_CLAUSE_COPYPRIVATE, si, and single_succ_edge().
Referenced by expand_omp().
|
static |
Generic expansion for OpenMP synchronization directives: master, ordered and critical. All we need to do here is remove the entry and exit markers for REGION.
References omp_region::entry, omp_region::exit, edge_def::flags, gsi_last_bb(), gsi_remove(), gsi_stmt(), si, and single_succ_edge().
Referenced by expand_omp().
|
static |
Expand the OpenMP parallel or task directive starting at REGION.
References function::cfg, cfun, cgraph_add_new_function(), changed, cleanup_tree_cfg(), function::curr_properties, function::decl, edge_def::dest, omp_region::entry, omp_region::exit, expand_parallel_call(), expand_task_call(), edge_def::flags, gimple_assign_lhs(), gimple_assign_rhs1(), gimple_assign_set_rhs1(), gimple_block(), gimple_build_nop(), gimple_build_return(), function::gimple_df, gimple_in_ssa_p(), gimple_num_ops(), gimple_omp_set_subcode(), gimple_omp_taskreg_child_fn(), gimple_omp_taskreg_data_arg(), gimple_purge_dead_eh_edges(), gimple_set_body(), gsi_end_p(), gsi_insert_after(), gsi_last_bb(), gsi_next(), gsi_remove(), GSI_SAME_STMT, gsi_start_bb(), gsi_stmt(), gimple_df::in_ssa_p, init_ssa_operands(), init_tree_ssa(), is_combined_parallel(), last_stmt(), function::local_decls, LOOPS_NEED_FIXUP, loops_state_satisfies_p(), make_edge(), make_ssa_name(), move_sese_region_to_fn(), optimize_omp_library_calls(), pop_cfun(), push_cfun(), rebuild_cgraph_edges(), remove_edge_and_dominated_blocks(), set_ssa_default_def(), single_succ(), single_succ_edge(), split_block(), ssa_default_def(), loops::state, update_ssa(), update_stmt(), varpool_finalize_decl(), vec2chain(), vec_safe_length(), vec_safe_truncate(), omp_region::ws_args, and function::x_current_loops.
Referenced by expand_omp().
|
static |
Build the function calls to GOMP_parallel_start etc to actually generate the parallel operation. REGION is the parallel region being expanded. BB is the block where to insert the code. WS_ARGS will be set if this is a call to a combined parallel+workshare construct, it contains the list of additional arguments needed by the workshare construct.
References add_bb_to_loop(), add_phi_arg(), build_call_expr_loc(), build_call_expr_loc_vec(), build_int_cst(), builtin_decl_explicit(), CDI_DOMINATORS, cfun, create_empty_bb(), create_phi_node(), create_tmp_var(), edge_def::dest, find_omp_clause(), fold_convert_loc(), force_gimple_operand_gsi(), gimple_boolify(), gimple_build_cond_empty(), gimple_in_ssa_p(), gimple_location(), gimple_omp_parallel_child_fn(), gimple_omp_parallel_clauses(), gimple_omp_parallel_data_arg(), walk_stmt_info::gsi, GSI_CONTINUE_LINKING, gsi_insert_after(), gsi_last_bb(), gsi_start_bb(), omp_region::inner, integer_zerop(), is_combined_parallel(), basic_block_def::loop_father, make_edge(), make_ssa_name(), OMP_CLAUSE_IF, OMP_CLAUSE_NUM_THREADS, OMP_CLAUSE_SCHEDULE_AUTO, OMP_CLAUSE_SCHEDULE_RUNTIME, remove_edge(), omp_region::sched_kind, set_immediate_dominator(), split_block(), edge_def::src, omp_region::type, vec_alloc(), and vec_safe_length().
Referenced by expand_omp_taskreg().
|
static |
Build the function call to GOMP_task to actually generate the task operation. BB is the block where to insert the code.
References build_call_expr(), build_fold_addr_expr_loc(), build_int_cst(), builtin_decl_explicit(), find_omp_clause(), force_gimple_operand_gsi(), gimple_boolify(), gimple_location(), gimple_omp_task_arg_align(), gimple_omp_task_arg_size(), gimple_omp_task_child_fn(), gimple_omp_task_clauses(), gimple_omp_task_copy_fn(), gimple_omp_task_data_arg(), walk_stmt_info::gsi, GSI_CONTINUE_LINKING, gsi_last_bb(), OMP_CLAUSE_FINAL, OMP_CLAUSE_IF, OMP_CLAUSE_MERGEABLE, and OMP_CLAUSE_UNTIED.
Referenced by expand_omp_taskreg().
|
static |
Extract the header elements of parallel loop FOR_STMT and store them into *FD.
References build_int_cst(), build_nonstandard_integer_type(), build_zero_cst(), omp_for_data::chunk_size, omp_for_data::collapse, omp_for_data_loop::cond_code, count, create_tmp_var(), fold_build_pointer_plus_hwi_loc(), fold_convert_loc(), omp_for_data::for_stmt, GF_OMP_FOR_KIND_SIMD, gimple_location(), gimple_omp_for_clauses(), gimple_omp_for_collapse(), gimple_omp_for_cond(), gimple_omp_for_final(), gimple_omp_for_incr(), gimple_omp_for_index(), gimple_omp_for_initial(), gimple_omp_for_kind(), omp_for_data::have_nowait, omp_for_data::have_ordered, integer_zerop(), omp_for_data::iter_type, omp_for_data::loop, omp_for_data::loops, omp_for_data_loop::n1, omp_for_data_loop::n2, OMP_CLAUSE_COLLAPSE, OMP_CLAUSE_NOWAIT, OMP_CLAUSE_ORDERED, OMP_CLAUSE_SCHEDULE, OMP_CLAUSE_SCHEDULE_AUTO, OMP_CLAUSE_SCHEDULE_RUNTIME, OMP_CLAUSE_SCHEDULE_STATIC, omp_for_data::pre, omp_for_data::sched_kind, signed_type_for(), omp_for_data_loop::step, tree_int_cst_lt(), and omp_for_data_loop::v.
Referenced by expand_omp_for(), get_ws_args_for(), lower_omp_for(), and workshare_safe_to_combine_p().
|
static |
Helper function for expand_omp_for_*. Generate code like:
L10:
V3 += STEP3;
if (V3 cond3 N32) goto BODY_BB; else goto L11;
L11:
V3 = N31;
V2 += STEP2;
if (V2 cond2 N22) goto BODY_BB; else goto L12;
L12:
V2 = N21;
V1 += STEP1;
goto BODY_BB;
References add_bb_to_loop(), CDI_DOMINATORS, omp_for_data::collapse, omp_for_data_loop::cond_code, create_empty_bb(), force_gimple_operand_gsi(), gimple_build_cond_empty(), GSI_CONTINUE_LINKING, gsi_insert_after(), gsi_start_bb(), last_bb, basic_block_def::loop_father, omp_for_data::loops, make_edge(), omp_for_data_loop::n1, omp_for_data_loop::n2, edge_def::probability, set_immediate_dominator(), omp_for_data_loop::step, and omp_for_data_loop::v.
Referenced by expand_omp_for_generic().
|
static |
Finalize task copyfn.
References cfun, cgraph_add_new_function(), function::curr_properties, gimple_build_bind(), gimple_omp_task_copy_fn(), gimple_seq_add_stmt(), gimple_set_body(), gimplify_body(), maybe_catch_exception(), pop_cfun(), and push_cfun().
Referenced by delete_omp_context().
| tree find_omp_clause | ( | ) |
Find an OpenMP clause of type KIND within CLAUSES.
Referenced by check_omp_nesting_restrictions(), determine_parallel_type(), expand_omp_simd(), expand_omp_single(), expand_parallel_call(), expand_task_call(), gimplify_omp_task(), lower_lastprivate_clauses(), lower_omp_sections(), lower_omp_single(), lower_rec_input_clauses(), lower_rec_simd_input_clauses(), optimize_omp_library_calls(), and scan_omp_parallel().
|
static |
Fix up RECEIVER_DECL with a type that has been remapped to the child context.
References build_pointer_type(), omp_context::cb, copy_node(), copy_tree_body_r(), omp_context::field_map, layout_type(), lang_hooks_for_types::make_type, nreverse(), omp_context::receiver_decl, omp_context::record_type, remap_type(), copy_body_data::src_fn, type(), lang_hooks::types, and variably_modified_type_p().
Referenced by scan_omp_parallel(), and scan_omp_task().
|
static |
Adjust the replacement for DECL in CTX for the new context. This means copying the DECL_VALUE_EXPR, and fixing up the type.
References omp_context::cb, copy_tree_body_r(), lookup_decl(), remap_decl(), and remap_type().
Referenced by scan_sharing_clauses().
|
static |
Release the memory associated with the region tree rooted at REGION.
References free(), omp_region::inner, and omp_region::next.
Referenced by free_omp_regions().
| void free_omp_regions | ( | void | ) |
Release the memory for the entire omp region tree.
References free_omp_region_1(), and omp_region::next.
Referenced by execute_expand_omp(), make_edges(), and omp_expand_local().
|
static |
|
static |
OMP expansion -- the default pass, run before creation of SSA form.
References seen_error().
Collect additional arguments needed to emit a combined parallel+workshare call. WS_STMT is the workshare directive being expanded.
References build_int_cst(), omp_for_data::chunk_size, extract_omp_for_data(), fold_convert_loc(), gimple_location(), omp_for_data::loop, omp_for_data_loop::n1, omp_for_data_loop::n2, single_succ(), omp_for_data_loop::step, basic_block_def::succs, and vec_alloc().
Referenced by determine_parallel_type().
|
static |
A convenience function to build an empty GIMPLE_COND with just the condition.
References gimple_build_cond(), and gimple_cond_get_ops_from_tree().
Referenced by expand_omp_atomic_pipeline(), expand_omp_for_generic(), expand_omp_for_static_chunk(), expand_omp_for_static_nochunk(), expand_omp_simd(), expand_parallel_call(), and extract_omp_for_update_vars().
|
static |
Add a new field for VAR inside the structure CTX->SENDER_DECL.
References build_pointer_type(), omp_context::field_map, insert_field_into_struct(), is_reference(), lang_hooks_for_types::make_type, omp_context::record_type, omp_context::sfield_map, omp_context::srecord_type, type(), and lang_hooks::types.
Referenced by scan_sharing_clauses().
|
static |
References omp_context::cb, insert_decl_map(), and omp_copy_decl_1().
Referenced by scan_sharing_clauses().
|
inlinestatic |
Return true if REGION is a combined parallel+workshare region.
References omp_region::is_combined_parallel.
Referenced by expand_omp_for_generic(), expand_omp_sections(), expand_omp_taskreg(), and expand_parallel_call().
|
inlinestatic |
Return true if CTX is for an omp parallel.
References omp_context::stmt.
Referenced by lower_lastprivate_clauses().
|
inlinestatic |
Return true if DECL is a reference type.
References lang_hooks::decls, and lang_hooks_for_decls::omp_privatize_by_reference.
Referenced by build_outer_var_ref(), create_task_copyfn(), install_var_field(), lower_copyprivate_clauses(), lower_lastprivate_clauses(), lower_rec_input_clauses(), lower_reduction_clauses(), lower_send_clauses(), and scan_sharing_clauses().
|
inlinestatic |
Return true if CTX is for an omp task.
References omp_context::stmt.
Referenced by delete_omp_context(), lower_rec_input_clauses(), lower_send_clauses(), scan_sharing_clauses(), and use_pointer_for_field().
|
inlinestatic |
Return true if CTX is for an omp parallel or omp task.
References omp_context::stmt.
Referenced by build_outer_var_ref(), omp_copy_decl(), scan_sharing_clauses(), and use_pointer_for_field().
|
inlinestatic |
Return true if EXPR is variable sized.
Referenced by build_outer_var_ref(), create_task_copyfn(), lower_rec_input_clauses(), lower_send_clauses(), and scan_sharing_clauses().
|
inlinestatic |
Lookup variables in the decl or field splay trees. The "maybe" form allows for the variable form to not have been entered, otherwise we assert that the variable must have been entered.
References omp_context::cb, copy_body_data::decl_map, and pointer_map_contains().
Referenced by build_outer_var_ref(), fixup_remapped_decl(), lower_lastprivate_clauses(), lower_rec_input_clauses(), and lower_reduction_clauses().
|
static |
Referenced by lower_copyprivate_clauses(), lower_send_clauses(), and lower_send_shared_vars().
|
static |
Find the mapping for DECL in CTX or the immediately enclosing
context that has a mapping for DECL.
If CTX is a nested parallel directive, we may have to use the decl
mappings created in CTX's parent context. Suppose that we have the
following parallel nesting (variable UIDs showed for clarity):
iD.1562 = 0;
#omp parallel shared(iD.1562) -> outer parallel
iD.1562 = iD.1562 + 1;
#omp parallel shared (iD.1562) -> inner parallel
iD.1562 = iD.1562 - 1;
Each parallel structure will create a distinct .omp_data_s structure
for copying iD.1562 in/out of the directive:
outer parallel .omp_data_s.1.i -> iD.1562
inner parallel .omp_data_s.2.i -> iD.1562
A shared variable mapping will produce a copy-out operation before
the parallel directive and a copy-in operation after it. So, in
this case we would have:
iD.1562 = 0;
.omp_data_o.1.i = iD.1562;
#omp parallel shared(iD.1562) -> outer parallel
.omp_data_i.1 = &.omp_data_o.1
.omp_data_i.1->i = .omp_data_i.1->i + 1;
.omp_data_o.2.i = iD.1562; -> **
#omp parallel shared(iD.1562) -> inner parallel
.omp_data_i.2 = &.omp_data_o.2
.omp_data_i.2->i = .omp_data_i.2->i - 1;
** This is a problem. The symbol iD.1562 cannot be referenced
inside the body of the outer parallel region. But since we are
emitting this copy operation while expanding the inner parallel
directive, we need to access the CTX structure of the outer
parallel directive to get the correct mapping:
.omp_data_o.2.i = .omp_data_i.1->i
Since there may be other workshare or parallel directives enclosing
the parallel directive, it may be necessary to walk up the context
parent chain. This is not a problem in general because nested
parallelism happens only rarely.
References is_global_var(), omp_context::is_nested, maybe_lookup_decl(), and omp_context::outer.
|
inlinestatic |
References omp_context::field_map.
Referenced by build_receiver_ref().
|
inlinestatic |
References omp_context::field_map, and omp_context::sfield_map.
Referenced by build_sender_ref().
|
static |
Generate code to implement the COPYPRIVATE clauses.
References build_fold_addr_expr_loc(), build_fold_indirect_ref_loc(), build_pointer_type(), build_receiver_ref(), build_sender_ref(), build_simple_mem_ref_loc(), lang_hooks::decls, fold_convert_loc(), gimplify_and_add(), gimplify_assign(), is_reference(), lookup_decl_in_outer_ctx(), lang_hooks_for_decls::omp_clause_assign_op, OMP_CLAUSE_COPYPRIVATE, and use_pointer_for_field().
Referenced by lower_omp_single_copy().
|
static |
Generate code to implement the LASTPRIVATE clauses. This is used for both parallel and workshare constructs. PREDICATE may be NULL if it's always true.
References build_outer_var_ref(), build_simple_mem_ref_loc(), create_artificial_label(), create_tmp_var(), lang_hooks::decls, fb_rvalue, find_omp_clause(), g, GF_OMP_FOR_KIND_SIMD, gimple_build_call_internal(), gimple_build_cond(), gimple_build_label(), gimple_call_set_lhs(), gimple_omp_for_kind(), gimple_omp_parallel_clauses(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimplify_and_add(), gimplify_expr(), is_gimple_val(), is_parallel_ctx(), is_reference(), lookup_attribute(), lookup_decl(), lower_omp(), OMP_CLAUSE__SIMDUID_, lang_hooks_for_decls::omp_clause_assign_op, OMP_CLAUSE_LASTPRIVATE, OMP_CLAUSE_LINEAR, omp_context::outer, and omp_context::stmt.
Referenced by lower_omp_for_lastprivate(), and lower_omp_sections().
|
static |
|
static |
References walk_stmt_info::gsi, gsi_end_p(), gsi_next(), input_location, and lower_omp_1().
|
static |
References gimple_bind_body_ptr(), gimple_build_nop(), gimple_catch_handler_ptr(), gimple_cond_lhs_ptr(), gimple_cond_rhs_ptr(), gimple_eh_filter_failure_ptr(), gimple_has_location(), gimple_location(), gimple_omp_atomic_load_rhs_ptr(), gimple_regimplify_operands(), gimple_transaction_body_ptr(), gimple_try_cleanup_ptr(), gimple_try_eval_ptr(), gsi_replace(), gsi_stmt(), input_location, is_gimple_omp(), lower_omp(), lower_omp_critical(), lower_omp_for(), lower_omp_master(), lower_omp_ordered(), lower_omp_regimplify_p(), lower_omp_sections(), lower_omp_single(), lower_omp_taskreg(), maybe_lookup_ctx(), memset(), seen_error(), and walk_gimple_op().
Referenced by lower_omp().
|
static |
References build_call_expr_loc(), build_fold_addr_expr_loc(), builtin_decl_explicit(), create_tmp_var_raw(), get_identifier(), gimple_bind_add_seq(), gimple_bind_add_stmt(), gimple_bind_append_vars(), gimple_bind_body(), gimple_bind_set_body(), gimple_bind_vars(), gimple_build_bind(), gimple_build_omp_return(), gimple_location(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_critical_name(), gimple_omp_set_body(), gimplify_and_add(), gsi_replace(), gsi_stmt(), lower_omp(), maybe_catch_exception(), pop_gimplify_context(), push_gimplify_context(), and varpool_finalize_decl().
Referenced by lower_omp_1().
|
static |
Lower code for an OpenMP loop directive.
References omp_context::block_vars, extract_omp_for_data(), get_formal_tmp_var(), gimple_bind_append_vars(), gimple_bind_set_body(), gimple_bind_vars(), gimple_build_bind(), gimple_build_omp_continue(), gimple_build_omp_return(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_for_clauses(), gimple_omp_for_collapse(), gimple_omp_for_final_ptr(), gimple_omp_for_incr(), gimple_omp_for_initial_ptr(), gimple_omp_for_pre_body(), gimple_omp_for_pre_body_ptr(), gimple_omp_for_set_pre_body(), gimple_omp_set_body(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimple_seq_empty_p(), gimple_seq_first_stmt(), gsi_replace(), gsi_stmt(), omp_for_data::have_nowait, is_gimple_min_invariant(), omp_for_data::loop, lower_omp(), lower_omp_for_lastprivate(), lower_rec_input_clauses(), lower_reduction_clauses(), maybe_catch_exception(), pop_gimplify_context(), push_gimplify_context(), and omp_for_data_loop::v.
Referenced by lower_omp_1().
|
static |
A subroutine of lower_omp_for. Generate code to emit the predicate for a lastprivate clause. Given a loop control predicate of (V cond N2), we gate the clause on (!(V cond N2)). The lowered form is appended to *DLIST, iterator initialization is appended to *BODY_P.
References build_int_cst(), omp_for_data_loop::cond_code, omp_for_data::for_stmt, gimple_omp_for_clauses(), gimple_seq_add_seq(), gimple_seq_empty_p(), gimplify_assign(), host_integerp(), HOST_WIDE_INT, integer_zerop(), omp_for_data::loop, lower_lastprivate_clauses(), omp_for_data_loop::n1, omp_for_data_loop::n2, omp_for_data_loop::step, unshare_expr(), and omp_for_data_loop::v.
Referenced by lower_omp_for().
|
static |
Expand code for an OpenMP master directive.
References omp_context::block_vars, build_and_jump(), build_call_expr_loc(), builtin_decl_explicit(), gimple_bind_add_seq(), gimple_bind_add_stmt(), gimple_bind_append_vars(), gimple_build_bind(), gimple_build_label(), gimple_build_omp_return(), gimple_location(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_set_body(), gimplify_and_add(), gsi_replace(), gsi_stmt(), lower_omp(), maybe_catch_exception(), pop_gimplify_context(), and push_gimplify_context().
Referenced by lower_omp_1().
|
static |
Expand code for an OpenMP ordered directive.
References omp_context::block_vars, builtin_decl_explicit(), gimple_bind_add_seq(), gimple_bind_add_stmt(), gimple_bind_append_vars(), gimple_bind_vars(), gimple_build_bind(), gimple_build_call(), gimple_build_omp_return(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_set_body(), gsi_replace(), gsi_stmt(), lower_omp(), maybe_catch_exception(), pop_gimplify_context(), and push_gimplify_context().
Referenced by lower_omp_1().
Callback for lower_omp_1. Return non-NULL if *tp needs to be regimplified. If DATA is non-NULL, lower_omp_1 is outside of OpenMP context, but with task_shared_vars set.
References bitmap_bit_p(), and recompute_tree_invariant_for_addr_expr().
Referenced by lower_omp_1().
|
static |
Routines to lower OpenMP directives into OMP-GIMPLE.
Lower the OpenMP sections directive in the current statement in GSI_P. CTX is the enclosing OMP context for the current statement.
References omp_context::block_vars, create_tmp_var(), find_omp_clause(), gimple_bind_append_vars(), gimple_bind_set_body(), gimple_bind_vars(), gimple_build_bind(), gimple_build_omp_continue(), gimple_build_omp_return(), gimple_build_omp_sections_switch(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_section_set_last(), gimple_omp_sections_clauses(), gimple_omp_sections_set_control(), gimple_omp_set_body(), gimple_seq_add_seq(), gimple_seq_add_stmt(), GSI_CONTINUE_LINKING, gsi_end_p(), gsi_insert_after(), gsi_insert_seq_after(), gsi_next(), gsi_one_before_end_p(), gsi_replace(), gsi_stmt(), lower_lastprivate_clauses(), lower_omp(), lower_rec_input_clauses(), lower_reduction_clauses(), maybe_catch_exception(), maybe_lookup_ctx(), OMP_CLAUSE_NOWAIT, pop_gimplify_context(), and push_gimplify_context().
Referenced by lower_omp_1().
|
static |
Expand code for an OpenMP single directive.
References omp_context::block_vars, find_omp_clause(), gimple_bind_append_vars(), gimple_bind_set_body(), gimple_build_bind(), gimple_build_omp_return(), gimple_omp_body_ptr(), gimple_omp_set_body(), gimple_omp_single_clauses(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gsi_replace(), gsi_stmt(), lower_omp(), lower_omp_single_copy(), lower_omp_single_simple(), lower_rec_input_clauses(), maybe_catch_exception(), OMP_CLAUSE_NOWAIT, pop_gimplify_context(), push_gimplify_context(), and omp_context::record_type.
Referenced by lower_omp_1().
|
static |
A subroutine of lower_omp_single. Expand the simple form of
a GIMPLE_OMP_SINGLE, with a copyprivate clause:
#pragma omp single copyprivate (a, b, c)
Create a new structure to hold copies of 'a', 'b' and 'c' and emit:
{
if ((copyout_p = GOMP_single_copy_start ()) == NULL)
{
BODY;
copyout.a = a;
copyout.b = b;
copyout.c = c;
GOMP_single_copy_end (©out);
}
else
{
a = copyout_p->a;
b = copyout_p->b;
c = copyout_p->c;
}
GOMP_barrier ();
}
FIXME. It may be better to delay expanding the logic of this until
pass_expand_omp. The expanded logic may make the job more difficult
to a synchronization analysis pass.
References build_and_jump(), build_call_expr_loc(), build_fold_addr_expr_loc(), build_int_cst(), build_pointer_type(), builtin_decl_explicit(), create_artificial_label(), create_tmp_var(), fold_convert_loc(), gimple_build_label(), gimple_location(), gimple_omp_body(), gimple_omp_single_clauses(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimplify_and_add(), gimplify_assign(), lower_copyprivate_clauses(), omp_context::receiver_decl, omp_context::record_type, and omp_context::sender_decl.
Referenced by lower_omp_single().
|
static |
A subroutine of lower_omp_single. Expand the simple form of
a GIMPLE_OMP_SINGLE, without a copyprivate clause:
if (GOMP_single_start ())
BODY;
[ GOMP_barrier (); ] -> unless 'nowait' is present.
FIXME. It may be better to delay expanding the logic of this until
pass_expand_omp. The expanded logic may make the job more difficult
to a synchronization analysis pass.
References builtin_decl_explicit(), create_artificial_label(), create_tmp_var(), fold_convert_loc(), gimple_build_call(), gimple_build_cond(), gimple_build_label(), gimple_call_set_lhs(), gimple_location(), gimple_omp_body(), gimple_seq_add_seq(), and gimple_seq_add_stmt().
Referenced by lower_omp_single().
|
static |
Lower the OpenMP parallel or task directive in the current statement in GSI_P. CTX holds context information for the directive.
References omp_context::block_vars, build_fold_addr_expr_loc(), omp_context::cb, check_combined_parallel(), create_task_copyfn(), create_tmp_var(), copy_body_data::dst_fn, fold_convert_loc(), gimple_bind_add_seq(), gimple_bind_add_stmt(), gimple_bind_block(), gimple_bind_body(), gimple_bind_vars(), gimple_build_bind(), gimple_build_omp_return(), gimple_location(), gimple_omp_body(), gimple_omp_parallel_combined_p(), gimple_omp_parallel_set_combined_p(), gimple_omp_set_body(), gimple_omp_taskreg_clauses(), gimple_omp_taskreg_set_data_arg(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimple_seq_first_stmt(), gsi_replace(), gsi_stmt(), walk_stmt_info::info, lower_omp(), lower_rec_input_clauses(), lower_reduction_clauses(), lower_send_clauses(), lower_send_shared_vars(), maybe_catch_exception(), memset(), pop_gimplify_context(), push_gimplify_context(), omp_context::receiver_decl, omp_context::record_type, record_vars_into(), omp_context::sender_decl, omp_context::srecord_type, walk_stmt_info::val_only, and walk_gimple_seq().
Referenced by lower_omp_1().
|
static |
Generate code to implement the input clauses, FIRSTPRIVATE and COPYIN, from the receiver (aka child) side and initializers for REFERENCE_TYPE private variables. Initialization statements go in ILIST, while calls to destructors go in DLIST.
References append_to_statement_list(), build_call_expr(), build_call_expr_loc(), build_fold_addr_expr_loc(), build_int_cst(), build_omp_barrier(), build_omp_clause(), build_outer_var_ref(), build_receiver_ref(), build_simple_mem_ref_loc(), builtin_decl_explicit(), c_kind, compare_tree_int(), create_artificial_label(), create_tmp_var(), create_tmp_var_raw(), lang_hooks::decls, find_omp_clause(), fold_convert_loc(), g, GF_OMP_FOR_KIND_FOR, GF_OMP_FOR_KIND_SIMD, gimple_add_tmp_var(), gimple_build_assign_with_ops(), gimple_build_call(), gimple_build_call_internal(), gimple_build_cond(), gimple_build_goto(), gimple_build_label(), gimple_call_set_lhs(), gimple_omp_body_ptr(), gimple_omp_for_clauses(), gimple_omp_for_kind(), gimple_omp_for_set_clauses(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimplify_and_add(), gimplify_assign(), gimplify_stmt(), walk_stmt_info::gsi, gsi_insert_before_without_update(), GSI_SAME_STMT, gsi_start_1(), is_global_var(), is_reference(), is_task_ctx(), is_variable_sized(), lookup_decl(), lower_omp(), lower_rec_simd_input_clauses(), maybe_lookup_decl(), maybe_lookup_decl_in_outer_ctx(), OMP_CLAUSE__SIMDUID_, lang_hooks_for_decls::omp_clause_assign_op, lang_hooks_for_decls::omp_clause_copy_ctor, OMP_CLAUSE_COPYIN, lang_hooks_for_decls::omp_clause_default_ctor, lang_hooks_for_decls::omp_clause_dtor, OMP_CLAUSE_FIRSTPRIVATE, OMP_CLAUSE_LASTPRIVATE, OMP_CLAUSE_LINEAR, OMP_CLAUSE_PRIVATE, OMP_CLAUSE_REDUCTION, OMP_CLAUSE_SAFELEN, OMP_CLAUSE_SHARED, omp_reduction_init(), omp_context::stmt, unshare_expr(), and use_pointer_for_field().
Referenced by lower_omp_for(), lower_omp_sections(), lower_omp_single(), and lower_omp_taskreg().
|
static |
Helper function of lower_rec_input_clauses, used for #pragma omp simd privatization.
References build_array_type_nelts(), compare_tree_int(), create_tmp_var(), create_tmp_var_raw(), find_omp_clause(), get_identifier(), gimple_add_tmp_var(), gimple_omp_for_clauses(), OMP_CLAUSE_SAFELEN, omp_max_vf(), omp_context::stmt, and tree_low_cst().
Referenced by lower_rec_input_clauses().
|
static |
Generate code to implement the REDUCTION clauses.
References build_fold_addr_expr_loc(), build_outer_var_ref(), build_simple_mem_ref_loc(), builtin_decl_explicit(), count, GF_OMP_FOR_KIND_SIMD, gimple_build_call(), gimple_omp_for_kind(), gimple_seq_add_seq(), gimple_seq_add_stmt(), gimplify_and_add(), gimplify_assign(), is_reference(), lookup_decl(), lower_omp(), OMP_CLAUSE_REDUCTION, save_expr(), and omp_context::stmt.
Referenced by lower_omp_for(), lower_omp_sections(), and lower_omp_taskreg().
|
static |
Generate code to implement the clauses, FIRSTPRIVATE, COPYIN, LASTPRIVATE, and REDUCTION from the sender (aka parent) side.
References build_fold_addr_expr_loc(), build_sender_ref(), lang_hooks::decls, gimplify_assign(), is_global_var(), is_reference(), is_task_ctx(), is_variable_sized(), lookup_decl_in_outer_ctx(), OMP_CLAUSE_COPYIN, OMP_CLAUSE_FIRSTPRIVATE, OMP_CLAUSE_LASTPRIVATE, OMP_CLAUSE_PRIVATE, OMP_CLAUSE_REDUCTION, lang_hooks_for_decls::omp_private_outer_ref, and use_pointer_for_field().
Referenced by lower_omp_taskreg().
|
static |
Generate code to implement SHARED from the sender (aka parent) side. This is trickier, since GIMPLE_OMP_PARALLEL_CLAUSES doesn't list things that got automatically shared.
References build_sender_ref(), gimplify_assign(), lookup_decl_in_outer_ctx(), maybe_lookup_decl(), omp_context::record_type, record_type(), omp_context::srecord_type, and use_pointer_for_field().
Referenced by lower_omp_taskreg().
| gimple_opt_pass* make_pass_diagnose_omp_blocks | ( | ) |
| gimple_opt_pass* make_pass_expand_omp | ( | ) |
| gimple_opt_pass* make_pass_lower_omp | ( | ) |
|
static |
|
static |
If exceptions are enabled, wrap the statements in BODY in a MUST_NOT_THROW catch handler and return it. This prevents programs from violating the structured block semantics with throws.
References builtin_decl_explicit(), lang_hooks::eh_protect_cleanup_actions, g, gimple_build_eh_must_not_throw(), gimple_build_try(), gimple_seq_alloc_with_stmt(), and GIMPLE_TRY_CATCH.
|
static |
If a context was created for STMT when it was scanned, return it.
References all_contexts.
Referenced by lower_omp_1(), and lower_omp_sections().
|
inlinestatic |
|
static |
Referenced by build_outer_var_ref(), lower_rec_input_clauses(), scan_sharing_clauses(), and use_pointer_for_field().
|
static |
Similar to lookup_decl_in_outer_ctx, but return DECL if not found in outer contexts.
References maybe_lookup_decl(), and omp_context::outer.
|
inlinestatic |
References omp_context::field_map.
Referenced by build_receiver_ref().
|
static |
Create a new context, with OUTER_CTX being the surrounding context.
References all_contexts, copy_body_data::block, omp_context::cb, CB_CGE_MOVE, cfun, cgraph_get_node(), current_function_decl, omp_context::depth, omp_copy_decl(), and pointer_map_create().
Referenced by scan_omp_1_stmt(), scan_omp_for(), scan_omp_parallel(), scan_omp_sections(), scan_omp_single(), and scan_omp_task().
|
read |
Create a new parallel region starting at STMT inside region PARENT.
References omp_region::entry, omp_region::inner, omp_region::next, omp_region::outer, root_omp_region, omp_region::type, and type().
Referenced by build_omp_regions_1(), and make_edges().
|
static |
Build COMPONENT_REF and set TREE_THIS_VOLATILE and TREE_READONLY on it as appropriate.
Referenced by build_receiver_ref(), build_sender_ref(), and create_task_copyfn().
|
static |
The callback for remap_decl. Search all containing contexts for a mapping of the variable; this avoids having to duplicate the splay tree ahead of time. We know a mapping doesn't already exist in the given context. Create new mappings to implement default semantics.
References omp_context::cb, create_artificial_label(), current_function_decl, decl_function_context(), insert_decl_map(), is_global_var(), is_taskreg_ctx(), maybe_lookup_decl(), omp_context::outer, and copy_body_data::src_fn.
Referenced by new_omp_context().
|
static |
References omp_copy_decl_2().
Referenced by install_var_local().
|
static |
Construct a new automatic decl similar to VAR.
References omp_context::block_vars, copy(), copy_var_decl(), and current_function_decl.
Referenced by omp_copy_decl_1().
| void omp_expand_local | ( | ) |
Expands omp construct (and its subconstructs) starting in HEAD.
References build_omp_regions_root(), dump_file, dump_flags, dump_omp_region(), expand_omp(), free_omp_regions(), and remove_exit_barriers().
Referenced by gen_parallel_loop().
|
static |
Return maximum possible vectorization factor for the target.
References floor_log2(), and targetm.
Referenced by lower_rec_simd_input_clauses().
| tree omp_reduction_init | ( | ) |
Construct the initialization value for reduction CLAUSE.
References build_real(), build_zero_cst(), fold_convert_loc(), real_arithmetic(), real_inf(), and real_maxval().
|
static |
Optimize omp_get_thread_num () and omp_get_num_threads ()
calls. These can't be declared as const functions, but
within one parallel body they are constant, so they can be
transformed there into __builtin_omp_get_{thread_num,num_threads} ()
which are declared const. Similarly for task body, except
that in untied task omp_get_thread_num () can change at any task
scheduling point.
References builtin_decl_explicit(), find_omp_clause(), gimple_call_fndecl(), gimple_call_num_args(), gimple_call_set_fndecl(), gimple_omp_task_clauses(), walk_stmt_info::gsi, gsi_end_p(), gsi_next(), gsi_start_bb(), gsi_stmt(), is_gimple_call(), OMP_CLAUSE_UNTIED, and types_compatible_p().
Referenced by expand_omp_taskreg().
|
static |
Remove barriers in REGION->EXIT's block. Note that this is only valid for GIMPLE_OMP_PARALLEL regions. Since the end of a parallel region is an implicit barrier, any workshare inside the GIMPLE_OMP_PARALLEL that left a barrier at the end of the GIMPLE_OMP_PARALLEL region can now be removed.
References omp_region::entry, omp_region::exit, gimple_block(), gimple_omp_parallel_child_fn(), gimple_omp_return_nowait_p(), gimple_omp_return_set_nowait(), walk_stmt_info::gsi, gsi_end_p(), gsi_last_bb(), gsi_prev(), gsi_stmt(), last_stmt(), basic_block_def::preds, and edge_def::src.
Referenced by remove_exit_barriers().
|
static |
References omp_region::inner, omp_region::next, remove_exit_barrier(), and omp_region::type.
Referenced by execute_expand_omp(), and omp_expand_local().
|
static |
|
static |
Scan all the statements starting at the current statement. CTX contains context information about the OpenMP directives and clauses found during the scan.
References walk_stmt_info::info, input_location, memset(), scan_omp_1_op(), scan_omp_1_stmt(), walk_gimple_seq_mod(), and walk_stmt_info::want_locations.
Referenced by scan_omp(), and scan_omp_op().
|
static |
Helper function scan_omp. Callback for walk_tree or operators in walk_gimple_stmt used to scan for OpenMP directives in TP.
References build_int_cst_wide(), omp_context::cb, walk_stmt_info::info, remap_decl(), and remap_type().
|
static |
Helper function for scan_omp. Callback for walk_gimple_stmt used to scan for OpenMP directives in the current statement in GSI.
References BUILT_IN_NORMAL, omp_context::cb, check_omp_nesting_restrictions(), gimple_bind_vars(), gimple_build_nop(), gimple_call_fndecl(), gimple_has_location(), gimple_location(), gimple_omp_body_ptr(), gsi_replace(), gsi_stmt(), walk_stmt_info::info, input_location, insert_decl_map(), is_gimple_call(), is_gimple_omp(), new_omp_context(), scan_omp(), scan_omp_for(), scan_omp_parallel(), scan_omp_sections(), scan_omp_single(), scan_omp_task(), and taskreg_nesting_level.
Referenced by scan_omp().
|
static |
Scan an OpenMP loop directive.
References gimple_omp_body_ptr(), gimple_omp_for_clauses(), gimple_omp_for_collapse(), gimple_omp_for_final_ptr(), gimple_omp_for_incr_ptr(), gimple_omp_for_index_ptr(), gimple_omp_for_initial_ptr(), gimple_omp_for_pre_body_ptr(), new_omp_context(), scan_omp(), scan_omp_op(), and scan_sharing_clauses().
Referenced by scan_omp_1_stmt().
|
inlinestatic |
Convenience function for calling scan_omp_1_op on tree operands.
References walk_stmt_info::info, memset(), scan_omp_1_op(), and walk_stmt_info::want_locations.
Referenced by scan_omp_for(), and scan_sharing_clauses().
|
static |
Scan an OpenMP parallel directive.
References omp_context::cb, create_omp_child_function(), create_tmp_var_name(), omp_context::default_kind, copy_body_data::dst_fn, empty_body_p(), omp_context::field_map, find_omp_clause(), fixup_child_record_type(), gimple_build_nop(), gimple_location(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_parallel_clauses(), gimple_omp_parallel_set_child_fn(), gsi_replace(), gsi_stmt(), omp_context::is_nested, layout_type(), lang_hooks_for_types::make_type, new_omp_context(), OMP_CLAUSE_COPYIN, OMP_CLAUSE_DEFAULT_SHARED, omp_context::receiver_decl, omp_context::record_type, scan_omp(), scan_sharing_clauses(), taskreg_nesting_level, and lang_hooks::types.
Referenced by scan_omp_1_stmt().
|
static |
Scan an OpenMP sections directive.
References gimple_omp_body_ptr(), gimple_omp_sections_clauses(), new_omp_context(), scan_omp(), and scan_sharing_clauses().
Referenced by scan_omp_1_stmt().
|
static |
Scan an OpenMP single directive.
References create_tmp_var_name(), omp_context::field_map, gimple_location(), gimple_omp_body_ptr(), gimple_omp_single_clauses(), layout_type(), lang_hooks_for_types::make_type, new_omp_context(), omp_context::record_type, scan_omp(), scan_sharing_clauses(), and lang_hooks::types.
Referenced by scan_omp_1_stmt().
|
static |
Scan an OpenMP task directive.
References build_int_cst(), omp_context::cb, create_omp_child_function(), create_tmp_var_name(), omp_context::default_kind, copy_body_data::dst_fn, empty_body_p(), omp_context::field_map, fixup_child_record_type(), fold_convert_loc(), gimple_build_nop(), gimple_location(), gimple_omp_body(), gimple_omp_body_ptr(), gimple_omp_task_clauses(), gimple_omp_task_set_arg_align(), gimple_omp_task_set_arg_size(), gimple_omp_task_set_child_fn(), gsi_replace(), gsi_stmt(), omp_context::is_nested, layout_type(), lang_hooks_for_types::make_type, new_omp_context(), OMP_CLAUSE_DEFAULT_SHARED, omp_context::receiver_decl, omp_context::record_type, scan_omp(), scan_sharing_clauses(), omp_context::srecord_type, taskreg_nesting_level, and lang_hooks::types.
Referenced by scan_omp_1_stmt().
|
static |
Instantiate decls as necessary in CTX to satisfy the data sharing specified by CLAUSES.
References function::decl, omp_context::default_kind, fixup_remapped_decl(), install_var_field(), install_var_local(), is_global_var(), is_reference(), is_task_ctx(), is_taskreg_ctx(), is_variable_sized(), maybe_lookup_decl_in_outer_ctx(), OMP_CLAUSE_COLLAPSE, OMP_CLAUSE_COPYIN, OMP_CLAUSE_COPYPRIVATE, OMP_CLAUSE_DEFAULT, OMP_CLAUSE_FINAL, OMP_CLAUSE_FIRSTPRIVATE, OMP_CLAUSE_IF, OMP_CLAUSE_LASTPRIVATE, OMP_CLAUSE_LINEAR, OMP_CLAUSE_MERGEABLE, OMP_CLAUSE_NOWAIT, OMP_CLAUSE_NUM_THREADS, OMP_CLAUSE_ORDERED, OMP_CLAUSE_PRIVATE, OMP_CLAUSE_REDUCTION, OMP_CLAUSE_SAFELEN, OMP_CLAUSE_SCHEDULE, OMP_CLAUSE_SHARED, OMP_CLAUSE_UNTIED, omp_context::outer, scan_omp(), scan_omp_op(), and use_pointer_for_field().
Referenced by scan_omp_for(), scan_omp_parallel(), scan_omp_sections(), scan_omp_single(), and scan_omp_task().
|
static |
References create_tmp_var(), omp_taskcopy_context::ctx, and omp_context::sfield_map.
Referenced by create_task_copyfn().
|
static |
References omp_taskcopy_context::cb, copy_node(), copy_tree_body_r(), omp_taskcopy_context::ctx, copy_body_data::decl_map, gimple_location(), layout_type(), lang_hooks_for_types::make_type, nreverse(), pointer_map_insert(), remap_type(), omp_context::stmt, type(), and lang_hooks::types.
Referenced by create_task_copyfn().
|
static |
Return true if DECL should be copied by pointer. SHARED_CTX is the parallel context if DECL is to be shared.
References bitmap_set_bit(), gimple_omp_taskreg_clauses(), is_gimple_reg(), omp_context::is_nested, is_task_ctx(), is_taskreg_ctx(), maybe_lookup_decl(), maybe_lookup_decl_in_outer_ctx(), OMP_CLAUSE_SHARED, omp_context::outer, and omp_context::stmt.
Referenced by build_outer_var_ref(), create_task_copyfn(), lower_copyprivate_clauses(), lower_rec_input_clauses(), lower_send_clauses(), lower_send_shared_vars(), and scan_sharing_clauses().
|
static |
Chain all the DECLs in LIST by their TREE_CHAIN fields.
Referenced by expand_omp_taskreg().
|
static |
Given two blocks PAR_ENTRY_BB and WS_ENTRY_BB such that WS_ENTRY_BB
is the immediate dominator of PAR_ENTRY_BB, return true if there
are no data dependencies that would prevent expanding the parallel
directive at PAR_ENTRY_BB as a combined parallel+workshare region.
When expanding a combined parallel+workshare region, the call to
the child function may need additional arguments in the case of
GIMPLE_OMP_FOR regions. In some cases, these arguments are
computed out of variables passed in from the parent to the child
via 'struct .omp_data_s'. For instance:
#pragma omp parallel for schedule (guided, i * 4)
for (j ...)
Is lowered into:
# BLOCK 2 (PAR_ENTRY_BB)
.omp_data_o.i = i;
#pragma omp parallel [child fn: bar.omp_fn.0 ( ..., D.1598)
# BLOCK 3 (WS_ENTRY_BB)
.omp_data_i = &.omp_data_o;
D.1667 = .omp_data_i->i;
D.1598 = D.1667 * 4;
#pragma omp for schedule (guided, D.1598)
When we outline the parallel region, the call to the child function
'bar.omp_fn.0' will need the value D.1598 in its argument list, but
that value is computed *after* the call site. So, in principle we
cannot do the transformation.
To see whether the code in WS_ENTRY_BB blocks the combined
parallel+workshare call, we collect all the variables used in the
GIMPLE_OMP_FOR header check whether they appear on the LHS of any
statement in WS_ENTRY_BB. If so, then we cannot emit the combined
call.
FIXME. If we had the SSA form built at this point, we could merely
hoist the code in block 3 into block 2 and be done with it. But at
this point we don't have dataflow information and though we could
hack something up here, it is really not worth the aggravation.
References omp_for_data::chunk_size, omp_for_data::collapse, extract_omp_for_data(), is_gimple_min_invariant(), omp_for_data::iter_type, last_stmt(), omp_for_data::loop, omp_for_data_loop::n1, omp_for_data_loop::n2, and omp_for_data_loop::step.
Referenced by determine_parallel_type().
|
static |
Referenced by execute_lower_omp(), maybe_lookup_ctx(), and new_omp_context().
|
static |
The following is a utility to diagnose OpenMP structured block violations. It is not part of the "omplower" pass, as that's invoked too late. It should be invoked by the respective front ends after gimplification.
|
static |
Gimplify a GIMPLE_OMP_CRITICAL statement. This is a relatively simple substitution of a couple of function calls. But in the NAMED case, requires that languages coordinate a symbol name. It is therefore best put here in common code.
| struct omp_region* root_omp_region |
Referenced by make_edges(), and new_omp_region().
|
static |
|
static |
Referenced by execute_lower_omp(), scan_omp_1_stmt(), scan_omp_parallel(), and scan_omp_task().
|
static |
Create a new name for omp child function. Returns an identifier.
1.8.1.1