jb [Sun, 16 Nov 2014 01:56:54 +0000 (01:56 +0000)]
PR 60324 VLA related fixes to random number generator.
2014-11-16 Janne Blomqvist <jb@gcc.gnu.org>
PR libfortran/60324
* intrinsics/random.c (kiss_size): Rename to KISS_SIZE, make it a
macro instead of a variable.
(random_seed_i4): Make seed correct size, remove assert, KISS_SIZE
related changes.
(random_seed_i8): KISS_SIZE related changes.
ebotcazou [Sat, 15 Nov 2014 12:34:20 +0000 (12:34 +0000)]
* tree-cfg.c (replace_loop_annotate_in_block): New function extracted
from...
(replace_loop_annotate): ...here. Call it on the header and on the
latch block, if any. Restore proper behavior of final cleanup.
pinskia [Fri, 14 Nov 2014 21:21:25 +0000 (21:21 +0000)]
[AARCH64] Add scheduler for ThunderX
This adds the schedule model for ThunderX. There are a few TODOs in that
not all of the SIMD is model currently. Also the idea of a simple
shift/extend is not modeled and all cases where there is a shift/extend
is considered as non simple and take up two cycles rather than correct
value of one cycle. Also the 32bit divide and the 64bit divide
have different cycle counts but there is no way to model that currently.
Also multiply high takes one cycle more than the normal multiply but
there is no way to model that currently either.
Build and tested for aarch64-elf with no regressions.
ChangeLog:
* config/aarch64/aarch64-cores.def (thunderx): Change the scheduler
over to thunderx.
* config/aarch64/aarch64.md: Include thunderx.md.
(generic_sched): Set to no for thunderx.
* config/aarch64/thunderx.md: New file.
meissner [Fri, 14 Nov 2014 20:45:21 +0000 (20:45 +0000)]
[gcc]
2014-11-14 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/predicates.md (easy_fp_constant): Delete redunant
tests for 0.0.
* config/rs6000/vector.md (VEC_R): Move secondary reload support
insns to rs6000.md from vector.md.
(reload_<VEC_R:mode>_<P:mptrsize>_store): Likewise.
(reload_<VEC_R:mode>_<P:mptrsize>_load): Likewise.
(vec_reload_and_plus_<mptrsize>): Likewise.
* config/rs6000/rs6000.md (Fa): New mode attribute to give
constraint for the Altivec registers for a type.
(RELOAD): New mode iterator for all of the types that have
secondary reload address support to load up a base register.
(extendsfdf2_fpr): Use correct constraint.
(copysign<mode>3_fcpsgn): For SFmode, use correct xscpsgndp
instruction.
(floatsi<mode>2_lfiwax): Add support for -mupper-regs-{sf,df}.
Generate the non-VSX instruction if all registers were FPRs. Do
not use the patterns in vsx.md for scalar operations.
(floatsi<mode>2_lfiwax_mem): Likewise.
(floatunssi<mode>2_lfiwzx): Likewise.
(floatunssi<mode>2_lfiwzx_mem): Likewise.
(fix_trunc<mode>di2_fctidz): Likewise.
(fixuns_trunc<mode>di2_fctiduz): Likewise.
(fctiwz_<mode>): Likewise.
(fctiwuz_<mode>): Likewise.
(friz): Likewise.
(floatdidf2_fpr): Likewise.
(floatdidf2_mem): Likewise.
(floatunsdidf2): Likewise.
(floatunsdidf2_fcfidu): Likewise.
(floatunsdidf2_mem): Likewise.
(floatdisf2_fcfids): Likewise.
(floatdisf2_mem): Likewise.
(floatdisf2_internal1): Add explicit test for not FCFIDS to make
it more obvious that the code is for pre-ISA 2.06 machines.
(floatdisf2_internal2): Likewise.
(floatunsdisf2_fcfidus): Add support for -mupper-regs-{sf,df}.
Generate the non-VSX instruction if all registers were FPRs. Do
not use the patterns in vsx.md for scalar operations.
(floatunsdisf2_mem): Likewise.
(reload_<RELOAD:mode>_<P:mptrsize>_store): Move the reload
handlers here from vector.md, and expand the types we generate
reload handlers for.
(reload_<RELOAD:mode>_<P:mptrsize>_load): Likewise.
(vec_reload_and_plus_<mptrsize>): Likewise.
* config/rs6000/vsx.md (vsx_float<VSi><mode>2): Only provide the
vector forms of the instructions. Move VSX scalar forms to
rs6000.md, and add support for -mupper-regs-sf.
(vsx_floatuns<VSi><mode>2): Likewise.
(vsx_fix_trunc<mode><VSi>2): Likewise.
(vsx_fixuns_trunc<mode><VSi>2): Likewise.
(vsx_float_fix_<mode>2): Delete DF version, rename to
vsx_float_fix_v2df2.
(vsx_float_fix_v2df2): Likewise.
[gcc/testsuite]
2014-11-14 Michael Meissner <meissner@linux.vnet.ibm.com>
jamborm [Fri, 14 Nov 2014 20:07:39 +0000 (20:07 +0000)]
2014-11-14 Martin Jambor <mjambor@suse.cz>
* ipa-prop.h (jump_func_type): Removed value IPA_JF_KNOWN_TYPE.
(ipa_pass_through_data): Removed field type_preserved.
(ipa_ancestor_jf_data): removed fields type and type_preserved.
(ipa_jump_func): Removed field known_type.
(ipa_get_jf_known_type_offset): Removed.
(ipa_get_jf_known_type_base_type): Likewise.
(ipa_get_jf_known_type_component_type): Likewise.
(ipa_get_jf_ancestor_type): Likewise.
* ipa-cp.c (print_ipcp_constant_value): Removed BINFO handling.
(ipa_get_jf_pass_through_result): Likewise.
(ipa_get_jf_ancestor_result): Always build ptr_node_type accesses.
(values_equal_for_ipcp_p): Removed BINFO handling.
(ipa_get_indirect_edge_target_1): Updated comment.
* ipa-prop.c (ipa_print_node_jump_functions_for_edge): Removed handling
of IPA_JF_KNOWN_TYPE jump functions. Do not print removed fields.
(ipa_set_jf_known_type): Removed.
(ipa_set_jf_simple_pass_through): Do not set removed fields. Update
all callers.
(ipa_set_jf_arith_pass_through): Likewise.
(ipa_set_ancestor_jf): Likewise.
(ipa_binfo_from_known_type_jfunc): Removed.
(prop_type_change_info): Removed fields known_current_type and
multiple_types_encountered.
(extr_type_from_vtbl_ptr_store): Removed.
(check_stmt_for_type_change): Do not attempt to identify changed type.
(detect_type_change_from_memory_writes): Do not set the removed fields,
always set jfunc to unknown.
(compute_complex_assign_jump_func): Do not detect dynamic type change.
(compute_complex_ancestor_jump_func): Likewise.
(compute_known_type_jump_func): Removed.
(ipa_compute_jump_functions_for_edge): Do not detect dynamic type
change. Do not comute known type jump functions.
(combine_known_type_and_ancestor_jfs): Removed.
(update_jump_functions_after_inlining): Removed handling of
IPA_JF_KNOWN_TYPE jump functions. Do not set removed fields.
(ipa_write_jump_function): Do not stream removed fields or known type
jump functions.
(ipa_read_jump_function): Likewise.
vmakarov [Fri, 14 Nov 2014 20:06:11 +0000 (20:06 +0000)]
2014-11-14 Vladimir Makarov <vmakarov@redhat.com>
* lra-int.h (lra_create_live_ranges): Add parameter.
* lra-lives.c (temp_bitmap): Move higher.
(initiate_live_solver): Move temp_bitmap initialization into
lra_live_ranges_init.
(finish_live_solver): Move temp_bitmap clearing into
live_ranges_finish.
(process_bb_lives): Add parameter. Use it to control live info
update and dead insn elimination. Pass it to mark_regno_live and
mark_regno_dead.
(lra_create_live_ranges): Add parameter. Pass it to
process_bb_lives.
(lra_live_ranges_init, lra_live_ranges_finish): See changes in
initiate_live_solver and finish_live_solver.
* lra-remat.c (do_remat): Process insn non-operand hard regs too.
Use temp_bitmap to update avail_cands.
* lra.c (lra): Pass new parameter to lra_create_live_ranges. Move
check with lra_need_for_spill_p after live range pass. Switch on
rematerialization pass.
jamborm [Fri, 14 Nov 2014 20:03:19 +0000 (20:03 +0000)]
2014-11-14 Martin Jambor <mjambor@suse.cz>
* ipa-prop.h (ipa_get_jf_pass_through_type_preserved): use
agg_preserved flag instead.
(ipa_get_jf_ancestor_type_preserved): Likewise.
(ipa_node_params): Rename known_vals to known_csts, update all users.
New field known_contexts.
(ipa_get_indirect_edge_target): Update prototype.
(ipcp_poly_ctx_values_pool): Declare.
(ipa_context_from_jfunc): Likewise.
* ipa-inline.h (estimate_ipcp_clone_size_and_time): Updated prototype.
* cgraph.h (ipa_polymorphic_call_context): New method equal_to. New
parameter newline of method dump.
* ipa-cp.c (ctxlat): New field.
(ipcp_values_pool): Renamed to ipcp_cst_values_pool, updated all users.
(ipcp_poly_ctx_values_pool):New variable.
(ipa_get_poly_ctx_lat): New function.
(print_ipcp_constant_value): New overloaded function for contexts.
(print_all_lattices): Also print contexts.
(ipa_topo_info): New field contexts;
(set_all_contains_variable): Also set the flag in the context lattice.
(initialize_node_lattices): Likewise for flag bottom.
(ipa_get_jf_ancestor_result): Removed BINFO handling.
(ipa_value_from_jfunc): Likewise.
(ipa_context_from_jfunc): New function.
(values_equal_for_ipcp_p): New overloaded function for contexts.
(allocate_and_init_ipcp_value): Construct the value.
(allocate_and_init_ipcp_value): New overloaded function for contexts.
(propagate_scalar_accross_jump_function): Removed handling of
KNOWN_TYPE jump functions.
(propagate_context_accross_jump_function): New function.
(propagate_constants_accross_call): Also propagate contexts.
(ipa_get_indirect_edge_target_1): Work on contexts rather than BINFOs.
(ipa_get_indirect_edge_target): Likewise.
(devirtualization_time_bonus): Likewise.
(gather_context_independent_values): Create and populate known_contexts
vector rather than known_binfos.
(perform_estimation_of_a_value): Work on contexts rather than BINFOs.
(estimate_local_effects): Likewise.
(add_all_node_vals_to_toposort): Also add contexts to teir topological
sort.
(ipcp_propagate_stage): Also propagate effects of contexts.
(ipcp_discover_new_direct_edges): Receive and pass known_contexts to
ipa_get_indirect_edge_target_1.
(cgraph_edge_brings_value_p): New overloaded function for contexts.
(create_specialized_node): Work on contexts rather than BINFOs.
(find_more_contexts_for_caller_subset): New function.
(known_contexts_useful_p): New function.
(copy_useful_known_contexts): Likewise.
(modify_known_vectors_with_val): Likewise.
(ipcp_val_in_agg_replacements_p): Renamed to
ipcp_val_agg_replacement_ok_p, return true for all offset indicating
non-aggregate.
(ipcp_val_agg_replacement_ok_p): New overloaded function for contexts.
(decide_about_value): Work on contexts rather than BINFOs.
(decide_whether_version_node): Likewise.
(ipcp_driver): Initialize the new alloc pool.
* ipa-prop.c (ipa_print_node_jump_functions_for_edge): Prettify
printing of edge contexts.
(ipa_set_ancestor_jf): Replace assert with conditional setting of
type_preserved to false.
(update_jump_functions_after_inlining): Use access function instead of
reading agg_preserved directly. Store combined context in the ancestor
case.
(try_make_edge_direct_virtual_call): Work on contexts rather than
BINFOs.
(update_indirect_edges_after_inlining): Get context from
ipa_context_from_jfunc.
(ipa_free_node_params_substructures): Free also known_contexts.
(ipa_free_all_structures_after_ipa_cp): Free the new alloc pool.
(ipa_free_all_structures_after_iinln): Likewise.
* ipa-inline-analysis.c (evaluate_properties_for_edge): Work on
contexts rather than BINFOs.
(estimate_edge_devirt_benefit): Likewise.
(estimate_edge_size_and_time): Likewise.
(estimate_calls_size_and_time): Likewise.
(estimate_node_size_and_time): Likewise.
(estimate_ipcp_clone_size_and_time): Likewise.
(do_estimate_edge_time): Likewise.
(do_estimate_edge_size): Likewise.
(do_estimate_edge_hints): Likewise.
* ipa-polymorphic-call.c (ipa_polymorphic_call_context::dump): New
parameter newline, ouput newline only when it is set.
(ipa_polymorphic_call_context::equal_to): New method.
testsuite/
* g++.dg/ipa/devirt-11.C: Dont't run ipa-cp, remove times constraint
from the dump scan.
* g++.dg/ipa/devirt-21.C: Xfail.
* g++.dg/ipa/devirt-24.C: Likewise.
* g++.dg/ipa/devirt-10.C: Removed times constraint from the dump scan.
* g++.dg/ipa/devirt-41.C: Updated the dump scan.
* g++.dg/ipa/devirt-44.C: Likewise.
* g++.dg/ipa/devirt-43.C: Xfail.
jamborm [Fri, 14 Nov 2014 19:58:18 +0000 (19:58 +0000)]
2014-11-14 Martin Jambor <mjambor@suse.cz>
* ipa-cp.c (ipcp_value_source): Converted to a template class. All
users converted to the same specialization as the using class/function
or specialization on tree.
(ipcp_value): Likewise.
(ipcp_lattice): Likewise.
(ipcp_agg_lattice): Now derived from tree specialization of
ipcp_lattice.
(values_topo): Moved to new class value_topo_info.
(ipa_lat_is_single_const): Turned into ipcp_lattice::is_single_const.
Updated all callers.
(print_lattice): Turned into ipcp_lattice::print. Updated all
callers.
(value_topo_info): New class template.
(ipa_topo_info): New field constants. New constructor.
(build_toporder_info): Do not clear stack_top, only checkign assert
it.
(set_lattice_to_bottom): Turned into ipcp_lattice::set_to_bottom.
Updated all callers.
(set_lattice_contains_variable): Turned into
ipcp_lattice::set_contains_variable. Updated all callers.
(add_value_source): Turned into ipcp_value::add_source. Updated all
callers.
(allocate_and_init_ipcp_value): New function.
(add_value_to_lattice): Turned into ipcp_lattice::add_value. Last
parameter got default a value. Updated all callers.
(add_scalar_value_to_lattice): Removed, users converted to using
ipcp_lattice::add_value with default value of the last parameter.
(add_val_to_toposort): Turned to value_topo_info::add_val. Updated
all callers.
(propagate_effects): Made method of value_topo_info.
(cgraph_edge_brings_value_p): Now a template function.
(get_info_about_necessary_edges): Likewise.
(gather_edges_for_value): Likewise.
(perhaps_add_new_callers): Likewise.
(decide_about_value): Likewise.
* ipa-prop.h (ipcp_lattice): Remove fowrward declaration.
ian [Fri, 14 Nov 2014 18:02:50 +0000 (18:02 +0000)]
runtime: set library name based on compiler name
Different compilers may have different release cadences or ABI
incompatibilities, so it does not make sense to use the same library
name for runtime libraries intended for different compilers.
This CL causes a libgo built by llgo to receive the name libgo-llgo.
Likewise, libgobegin is named libgobegin-llgo.
jakub [Fri, 14 Nov 2014 17:46:50 +0000 (17:46 +0000)]
* ipa-pure-const.c (struct funct_state_d): Add can_free field.
(varying_state): Add true for can_free.
(check_call): For builtin or internal !nonfreeing_call_p set
local->can_free.
(check_stmt): For asm volatile and asm with "memory" set
local->can_free.
(analyze_function): Clear local->can_free initially, continue
calling check_stmt until all flags are computed, dump can_free
flag.
(pure_const_write_summary): Write can_free flag.
(pure_const_read_summary): Read it back.
(propagate_pure_const): Propagate also can_free flag, set
w->nonfreeing_fn if it is false after propagation.
* cgraph.h (cgraph_node): Add nonfreeing_fn member.
* gimple.c: Include ipa-ref.h, lto-streamer.h and cgraph.h.
(nonfreeing_call_p): Return cgraph nonfreeing_fn flag if set.
Also return true for IFN_ABNORMAL_DISPATCHER.
* cgraph.c (cgraph_node::dump): Dump nonfreeing_fn flag.
* lto-cgraph.c (lto_output_node): Write nonfreeing_fn flag.
(input_overwrite_node): Read it back.
jakub [Fri, 14 Nov 2014 17:19:55 +0000 (17:19 +0000)]
* libgomp.c/examples-4/e.54.2.c (main): Use N / 8 instead
of 32 as block_size.
* libgomp.fortran/examples-4/e.54.2.f90 (e_54_1): Use n / 8
instead of 32 as block_size.
rguenth [Fri, 14 Nov 2014 13:32:56 +0000 (13:32 +0000)]
2014-11-14 Richard Biener <rguenther@suse.de>
* gimple-fold.h (gimple_fold_stmt_to_constant_1): Add 2nd
valueization hook defaulted to no_follow_ssa_edges.
* gimple-fold.c (gimple_fold_stmt_to_constant_1): Pass
2nd valueization hook to gimple_simplify.
* tree-ssa-ccp.c (valueize_op_1): New function to be
used for gimple_simplify called via gimple_fold_stmt_to_constant_1.
(ccp_fold): Adjust.
* tree-vrp.c (vrp_valueize_1): New function to be
used for gimple_simplify called via gimple_fold_stmt_to_constant_1.
(vrp_visit_assignment_or_call): Adjust.
mpolacek [Fri, 14 Nov 2014 11:57:05 +0000 (11:57 +0000)]
* fold-const.c (fold_negate_expr): Don't fold INTEGER_CST if
that overflows when SANITIZE_SI_OVERFLOW is on. Guard -(-A)
folding with TYPE_OVERFLOW_SANITIZED.
* c-c++-common/ubsan/overflow-negate-3.c: New test.
rguenth [Fri, 14 Nov 2014 08:37:41 +0000 (08:37 +0000)]
2014-11-14 Richard Biener <rguenther@suse.de>
* genmatch.c (add_operator): Allow CONSTRUCTOR.
(dt_node::gen_kids): Handle CONSTRUCTOR not as GENERIC.
(parser::parse_op): Allow to iterate over predicates.
timshen [Fri, 14 Nov 2014 04:23:06 +0000 (04:23 +0000)]
2014-11-13 Daniel Kruegler <daniel.kruegler@gmail.com>
* include/bits/regex.h: Support embedded zeros in sub_match
comparisons (DR 2217)
* testsuite/28_regex/sub_match/embedded_zeros_cmp.cc:
Add test cases for embedded zeros
jamborm [Fri, 14 Nov 2014 00:22:28 +0000 (00:22 +0000)]
2014-11-14 Martin Jambor <mjambor@suse.cz>
* cgraph.h (clear_outer_type): Make public. Fix comment.
* ipa-devirt.c (possible_polymorphic_call_targets): Use
clear_outer_type when resetting the context.
alalaw01 [Thu, 13 Nov 2014 16:27:37 +0000 (16:27 +0000)]
[Vectorizer] Use a VEC_PERM_EXPR instead of VEC_RSHIFT_EXPR; expand appropriate VEC_PERM_EXPRs using vec_shr_optab
* optabs.c (can_vec_perm_p): Update comment, does not consider vec_shr.
(shift_amt_for_vec_perm_mask): New.
(expand_vec_perm_1): Use vec_shr_optab if second vector is const0_rtx
and mask appropriate.
* tree-vect-loop.c (calc_vec_perm_mask_for_shift): New.
(have_whole_vector_shift): New.
(vect_model_reduction_cost): Call have_whole_vector_shift instead of
looking for vec_shr_optab.
(vect_create_epilog_for_reduction): Likewise; also rename local variable
have_whole_vector_shift to reduce_with_shift; output VEC_PERM_EXPRs
instead of VEC_RSHIFT_EXPRs.
kyukhin [Thu, 13 Nov 2014 14:05:49 +0000 (14:05 +0000)]
[PATCH 3/4] OpenMP 4.0 offloading to Intel MIC: libgomp plugin.
liboffloadmic/
* Makefile.in: Regenerate.
* configure: Regenerate.
* configure.ac: Add subdirectory 'plugin'.
* plugin/Makefile.am: New file.
* plugin/Makefile.in: New file, generated by automake.
* plugin/aclocal.m4: New file, generated by aclocal.
* plugin/configure: New file, generated by autoconf.
* plugin/configure.ac: New file.
* plugin/libgomp-plugin-intelmic.cpp: New file.
* plugin/offload_target_main.cpp: New file.
kyukhin [Thu, 13 Nov 2014 14:03:17 +0000 (14:03 +0000)]
[PATCH 2/4] OpenMP 4.0 offloading to Intel MIC: liboffloadmic.
* Makefile.def: Add liboffloadmic to target_modules. Make
liboffloadmic depend on libgomp's configure, libstdc++ and libgcc.
* Makefile.in: Regenerate.
* configure: Regenerate.
* configure.ac: Add liboffloadmic to target binaries.
Restrict liboffloadmic for POSIX and i*86, and x86_64 architectures.
Add liboffloadmic to noconfig list when C++ is not supported.
config/
* target-posix: New file.
libcilkrts/
* configure.tgt: Use config/target-posix.
rguenth [Thu, 13 Nov 2014 13:58:59 +0000 (13:58 +0000)]
2014-11-13 Richard Biener <rguenther@suse.de>
* match.pd: Add tcc_comparison, inverted_tcc_comparison
and inverted_tcc_comparison_with_nans operator lists.
Use tcc_comparison in the truth_valued_p predicate definition.
Restrict logical_inverted_value with bit_xor to integral types.
Build a boolean true for simplifying x |^ !x because of
vector types. Implement patterns from forward_propagate_comparison
* tree-ssa-forwprop.c (forward_propagate_comparison): Remove.
(get_prop_dest_stmt): Likewise.
(pass_forwprop::execute): Do not call it.
* fold-const.c (fold_unary_loc): Remove the pattern here.
kyukhin [Thu, 13 Nov 2014 13:54:07 +0000 (13:54 +0000)]
[PATCH 6/7] OpenMP 4.0 offloading infrastructure: option handling.
gcc/
* common.opt (foffload, foffload-abi): New options.
* config/i386/i386.c (ix86_offload_options): New static function.
(TARGET_OFFLOAD_OPTIONS): Define.
* coretypes.h (enum offload_abi): New enum.
* doc/tm.texi: Regenerate.
* doc/tm.texi.in (TARGET_OFFLOAD_OPTIONS): Document.
* gcc.c (offload_targets): New static variable.
(handle_foffload_option): New static function.
(driver_handle_option): Handle OPT_foffload_.
(driver::maybe_putenv_OFFLOAD_TARGETS): Set OFFLOAD_TARGET_NAMES
according to offload_targets.
* hooks.c (hook_charptr_void_null): New hook.
* hooks.h (hook_charptr_void_null): Declare.
* lto-opts.c: Include lto-section-names.h.
(lto_write_options): Append options from target offload_options hook and
store them to offload_lto section. Do not store target-specific,
driver and diagnostic options in offload_lto section.
* lto-wrapper.c (merge_and_complain): Handle OPT_foffload_ and
OPT_foffload_abi_.
(append_compiler_options, append_linker_options)
(append_offload_options): New static functions.
(compile_offload_image): Add new arguments with options.
Call append_compiler_options and append_offload_options.
(compile_images_for_offload_targets): Add new arguments with options.
(find_and_merge_options): New static function.
(run_gcc): Outline options handling into the new functions:
find_and_merge_options, append_compiler_options, append_linker_options.
* opts.c (common_handle_option): Don't handle OPT_foffload_.
Forbid OPT_foffload_abi_ for non-offload compiler.
* target.def (offload_options): New target hook.
libgomp/
* libgomp.map (GOMP_4.0.1): New symbol version.
Add GOMP_offload_register.
* libgomp_target.h: New file.
* splay-tree.h: New file.
* target.c: Include config.h, libgomp_target.h, dlfcn.h, splay-tree.h.
(gomp_target_init): New forward declaration.
(gomp_is_initialized): New static variable.
(splay_tree_node, splay_tree, splay_tree_key): New typedefs.
(struct target_mem_desc, struct splay_tree_key_s, offload_image_descr):
New structures.
(offload_images, num_offload_images, devices, num_devices): New static
variables.
(splay_compare): New static function.
(struct gomp_device_descr): New structure.
(gomp_get_num_devices): Call gomp_target_init.
(resolve_device, gomp_map_vars_existing, gomp_map_vars, gomp_unmap_tgt)
(gomp_unmap_vars, gomp_update, gomp_init_device): New static functions.
(GOMP_offload_register): New function.
(GOMP_target): Arrange for host callback to be performed in a separate
initial thread and contention group, inheriting ICVs from
gomp_global_icv etc. Call gomp_map_vars and gomp_unmap_vars.
Add device initialization and lookup for target function in splay tree.
(GOMP_target_data): Add device initialization and call gomp_map_vars.
(GOMP_target_end_data): Call gomp_unmap_vars.
(GOMP_target_update): Add device initialization and call gomp_update.
(gomp_load_plugin_for_device, gomp_register_images_for_device)
(gomp_target_init): New static functions.
gcc/
* gcc.c (spec_host_machine, accel_dir_suffix): New variables.
(process_command): Tweak path construction for the possibility
of being configured as an offload compiler.
(driver::maybe_putenv_OFFLOAD_TARGETS): New function.
(driver::main): Call maybe_putenv_OFFLOAD_TARGETS.
(driver::set_up_specs): Tweak path construction for the possibility of
being configured as an offload compiler.
* lto-wrapper.c (OFFLOAD_TARGET_NAMES_ENV): Define.
(offload_names, offloadbegin, offloadend): New static variables.
(free_array_of_ptrs, parse_env_var, access_check, compile_offload_image)
(compile_images_for_offload_targets, copy_file, find_offloadbeginend):
New static functions.
(run_gcc): Determine whether offload sections are present. If so, run
compile_images_for_offload_targets and return the names of new generated
objects to linker. If there are offload sections, but no LTO sections,
then return the copies of input objects without link-time recompilation.
lto-plugin/
* lto-plugin.c (OFFLOAD_SECTION, OFFLOAD_SECTION_LEN): Define.
(struct plugin_objfile): Add new field "offload".
(process_offload_section): New static function.
(claim_file_handler): Claim file if it contains offload sections.
gcc/
* cgraph.c: Include context.h.
(cgraph_node::create): Set node->offloadable and g->have_offload if
decl have "omp declare target" attribute.
* cgraph.h (symtab_node): Add need_lto_streaming and offloadable flags.
* cgraphunit.c: Include lto-section-names.h.
(ipa_passes): Call ipa_write_summaries if there is something to write to
OFFLOAD_SECTION_NAME_PREFIX sections.
(symbol_table::compile): Set flag_generate_lto if there is something to
offload.
Replace flag_lto with flag_generate_lto before lto_streamer_hooks_init.
* context.c (gcc::context::context): Initialize have_offload with false.
* context.h (class context): Add have_offload flag.
* ipa-inline-analysis.c (inline_generate_summary): Do not exit under
flag_generate_lto.
(inline_free_summary): Always remove hooks.
* lto-cgraph.c (referenced_from_other_partition_p): Ignore references
from non-offloadable nodes while streaming a node into offload section.
(reachable_from_other_partition_p): Likewise.
(select_what_to_stream): New function.
(compute_ltrans_boundary): Do not call
lto_set_symtab_encoder_in_partition if the node should not be streamed.
* lto-section-names.h (OFFLOAD_SECTION_NAME_PREFIX): Define.
(section_name_prefix): Declare.
* lto-streamer.c (section_name_prefix): New variable.
(lto_get_section_name): Use section_name_prefix instead of
LTO_SECTION_NAME_PREFIX.
* lto-streamer.h (select_what_to_stream): Declare.
* omp-low.c: Include context.h.
(is_targetreg_ctx): New function.
(scan_sharing_clauses): Use offloadable flag, instead of an attribute.
(create_omp_child_function, check_omp_nesting_restrictions): Use new
is_targetreg_ctx function. Replace usage of "omp declare target"
attribute with a cgraph_node flag offloadable.
(expand_omp_target): Set mark_force_output for offloadable functions.
(lower_omp_critical): Set offloadable flag for omp critical symbol.
* passes.c (ipa_write_summaries): New argument offload_lto_mode. Call
select_what_to_stream. Do not call lto_set_symtab_encoder_in_partition
if the node should not be streamed out.
* tree-pass.h (ipa_write_summaries): New bool argument.
* varpool.c: Include context.h.
(varpool_node::get_create): Set node->offloadable and g->have_offload if
decl have "omp declare target" attribute.
gcc/lto/
* lto-object.c (lto_obj_add_section): Use section_name_prefix instead of
LTO_SECTION_NAME_PREFIX.
* lto-partition.c (lto_promote_cross_file_statics): Call
select_what_to_stream.
* lto.c (lto_section_with_id): Use section_name_prefix instead of
LTO_SECTION_NAME_PREFIX.
(read_cgraph_and_symbols): Read OFFLOAD_SECTION_NAME_PREFIX sections, if
being built as an offload compiler.
kyukhin [Thu, 13 Nov 2014 13:28:56 +0000 (13:28 +0000)]
[PATCH 1/7] OpenMP 4.0 offloading infrastructure: configure and make
* configure: Regenerate.
* configure.ac (--enable-as-accelerator-for)
(--enable-offload-targets): New configure options.
gcc/
* Makefile.in (real_target_noncanonical, accel_dir_suffix)
(enable_as_accelerator): New variables substituted by configure.
(libsubdir, libexecsubdir, unlibsubdir): Tweak for the possibility of
being configured as an offload compiler.
(DRIVER_DEFINES): Pass new defines DEFAULT_REAL_TARGET_MACHINE and
ACCEL_DIR_SUFFIX.
(install-cpp, install-common, install_driver, install-gcc-ar): Do not
install for the offload compiler.
* config.in: Regenerate.
* configure: Regenerate.
* configure.ac (real_target_noncanonical, accel_dir_suffix)
(enable_as_accelerator): Compute new variables.
(ACCEL_COMPILER): Define if the compiler is built as the accel compiler.
(OFFLOAD_TARGETS): List of target names suitable for offloading.
(ENABLE_OFFLOADING): Define if list of offload targets is not empty.
gcc/cp/
* Make-lang.in (c++.install-common): Do not install for the offload
compiler.
gcc/doc/
* install.texi (Options specification): Document
--enable-as-accelerator-for and --enable-offload-targets.
gcc/fortran/
* Make-lang.in (fortran.install-common): Do not install for the offload
compiler.
libgcc/
* Makefile.in (crtoffloadbegin$(objext)): New rule.
(crtoffloadend$(objext)): Likewise.
* configure: Regenerate.
* configure.ac (accel_dir_suffix): Compute new variable.
(extra_parts): Add crtoffloadbegin.o and crtoffloadend.o
if enable_offload_targets is not empty.
* offloadstuff.c: New file.
libgomp/
* config.h.in: Regenerate.
* configure: Regenerate.
* configure.ac: Check for libdl, required for plugin support.
(PLUGIN_SUPPORT): Define if plugins are supported.
(enable_offload_targets): Support Intel MIC targets.
(OFFLOAD_TARGETS): List of target names suitable for offloading.
lto-plugin/
* Makefile.am (libexecsubdir): Tweak for the possibility of being
configured for offload compiler.
(accel_dir_suffix, real_target_noncanonical): New variables substituted
by configure.
* Makefile.in: Regenerate.
* configure: Regenerate.
* configure.ac (accel_dir_suffix, real_target_noncanonical): Compute new
variables.
jb [Thu, 13 Nov 2014 12:05:01 +0000 (12:05 +0000)]
PR 60324 Unbounded stack allocations in libgfortran.
2014-11-13 Janne Blomqvist <jb@gcc.gnu.org>
PR libfortran/60324
* configure: Regenerated.
* configure.ac (AM_CFLAGS): Add Werror=vla.
* libgfortran.h (gfc_alloca): Remove macro.
(fc_strdup_notrim): New prototype.
* intrinsics/access.c (access_func): Use fc_strdup rather than
stack allocation.
* intrinsics/chdir.c (chdir_i4_sub): Likewise.
(chdir_i8_sub): Likewise.
* intrinsics/chmod.c (chmod_internal): New function, move logic
here.
(chmod_func): Call chmod_internal.
* intrinsics/env.c (getenv): Use fc_strdup rather than stack
allocation.
(get_environment_variable_i4): Likewise.
* intrinsics/execute_command_line.c (execute_command_line):
Likewise.
* intrinsics/hostnm.c (hostnm_0): New function, use static buffer
rather than VLA.
(hostnm_i4_sub): Call hostnm_0.
(hostnm_i8_sub): Likewise.
(hostnm): Likewise.
* intrinsics/link.c (link_internal): New function, use fc_strdup
rather than stack allocation.
(link_i4_sub): Call link_internal.
(link_i8_sub): Likewise.
(link_i4): Likewise.
(link_i8): Likewise.
* intrinsics/perror.c (perror_sub): Use fc_strdup rather than
stack allocation.
* intrinsics/random.c (random_seed_i4): Use static buffer rather
than VLA, use _Static_assert to make sure it's big enough.
* intrinsics/rename.c (rename_internal): New function, use
fc_strdup rather than stack allocation.
(rename_i4_sub): Call rename_internal.
(rename_i8_sub): Likewise.
(rename_i4): Likewise.
(rename_i8): Likewise.
* intrinsics/stat.c (stat_i4_sub_0): Use fc_strdup rather than
stack allocation.
(stat_i8_sub_0): Likewise.
* intrinsics/symlink.c (symlnk_internal): New function, use
fc_strdup rather than stack allocation.
(symlnk_i4_sub): Call symlnk_internal.
(symlnk_i8_sub): Likewise.
(symlnk_i4): Likewise.
(symlnk_i8): Likewise.
* intrinsics/system.c (system_sub): Use fc_strdup rather than
stack allocation.
* intrinsics/unlink.c (unlink_i4_sub): Likewise.
* io/file_pos.c (READ_CHUNK): Make it a macro rather than variable.
* io/list_read.c (nml_get_obj_data): Use fixed stack buffer, fall
back to xmalloc/free for large sizes.
* io/read.c (read_f): Likewise.
* io/transfer.c (MAX_READ): Make it a macro rather than variable.
(WRITE_CHUNK): Likewise.
* io/write_float.def (write_float): Use fixed stack buffer, fall
back to xmalloc/free for large sizes.
* runtime/string.c (fc_strdup_notrim): New function.
vries [Thu, 13 Nov 2014 10:51:58 +0000 (10:51 +0000)]
Run pass_expand_omp_ssa after pass_paralellize_loops
2014-11-13 Tom de Vries <tom@codesourcery.com>
* omp-low.c (pass_data_expand_omp): Set properties_provided to
PROP_gimple_eomp.
(pass_expand_omp::gate): Remove function. Move gate expression to ...
(pass_expand_omp::execute): ... here, as new variable gate. Add early
exit if gate is false.
(pass_data pass_data_expand_omp_ssa): New pass_data.
(class pass_expand_omp_ssa): New pass.
(make_pass_expand_omp_ssa): New function.
* passes.def (pass_parallelize_loops): Use PUSH_INSERT_PASSES_WITHIN
instead of NEXT_PASS.
(pass_expand_omp_ssa): Add after pass_parallelize_loops.
* tree-parloops.c (gen_parallel_loop): Remove call to omp_expand_local.
(pass_parallelize_loops::execute): Don't do cleanups TODO_cleanup_cfg
and TODO_rebuild_alias yet. Add TODO_update_ssa. Set
cfun->omp_expand_needed.
* tree-pass.h: Add define PROP_gimple_eomp.
(make_pass_expand_omp_ssa): Declare.
projects / thirdparty / gcc.git / log
