Irreducible regions are not going to be optimized by ISL
so discard them early. Passes bootstrap and regtest.
gcc/ChangeLog:
2015-11-06 Aditya Kumar <aditya.k7@samsung.com>
* graphite-scop-detection.c (scop_detection::merge_sese): Entry and exit edges should not be a part of irreducible loop.
(scop_detection::can_represent_loop_1): Loops should not be irreducible.
(scop_detection::harmful_stmt_in_region): All the basic block should belong to reducible loops.
Co-Authored-By: Sebastian Pop <s.pop@samsung.com>
From-SVN: r229888
PR debug/66728
* dwarf2out.c (get_full_len): Return a value based upon the actual
precision needed for the value.
(add_const_value_attribute): Use a maximal wide-int for
CONST_WIDE_INTs, not VOIDmode.
(output_die): Don't ever output NULL with printf.
* rtl.h (get_precision of rtx_mode_t): Ensure we never process
BLKmode nor VOIDmode values.
From-SVN: r229885
gcc/ChangeLog:
* diagnostic-color.c (color_dict): Eliminate "caret"; add "range1"
and "range2".
(parse_gcc_colors): Update comment to describe default GCC_COLORS.
* diagnostic-core.h (warning_at_rich_loc): New declaration.
(error_at_rich_loc): New declaration.
(permerror_at_rich_loc): New declaration.
(inform_at_rich_loc): New declaration.
* diagnostic-show-locus.c (adjust_line): Delete.
(struct point_state): New struct.
(class colorizer): New class.
(class layout_point): New class.
(class layout_range): New class.
(struct line_bounds): New.
(class layout): New class.
(colorizer::colorizer): New ctor.
(colorizer::~colorizer): New dtor.
(layout::layout): New ctor.
(layout::print_source_line): New method.
(layout::print_annotation_line): New method.
(layout::get_state_at_point): New method.
(layout::get_x_bound_for_row): New method.
(diagnostic_show_locus): Reimplement in terms of class layout.
(diagnostic_print_caret_line): Delete.
* diagnostic.c (diagnostic_initialize): Replace
MAX_LOCATIONS_PER_MESSAGE with rich_location::MAX_RANGES.
(diagnostic_set_info_translated): Convert param from location_t
to rich_location *. Eliminate calls to set_location on the
message in favor of storing the rich_location ptr there.
(diagnostic_set_info): Convert param from location_t to
rich_location *.
(diagnostic_build_prefix): Break out array into...
(diagnostic_kind_color): New variable.
(diagnostic_get_color_for_kind): New function.
(diagnostic_report_diagnostic): Colorize the option_text
using the color for the severity.
(diagnostic_append_note): Update for change in signature of
diagnostic_set_info.
(diagnostic_append_note_at_rich_loc): New function.
(emit_diagnostic): Update for change in signature of
diagnostic_set_info.
(inform): Likewise.
(inform_at_rich_loc): New function.
(inform_n): Update for change in signature of diagnostic_set_info.
(warning): Likewise.
(warning_at): Likewise.
(warning_at_rich_loc): New function.
(warning_n): Update for change in signature of diagnostic_set_info.
(pedwarn): Likewise.
(permerror): Likewise.
(permerror_at_rich_loc): New function.
(error): Update for change in signature of diagnostic_set_info.
(error_n): Likewise.
(error_at): Likewise.
(error_at_rich_loc): New function.
(sorry): Update for change in signature of diagnostic_set_info.
(fatal_error): Likewise.
(internal_error): Likewise.
(internal_error_no_backtrace): Likewise.
(source_range::debug): New function.
* diagnostic.h (struct diagnostic_info): Eliminate field
"override_column". Add field "richloc".
(struct diagnostic_context): Add field "colorize_source_p".
(diagnostic_override_column): Delete.
(diagnostic_set_info): Convert param from location_t to
rich_location *.
(diagnostic_set_info_translated): Likewise.
(diagnostic_append_note_at_rich_loc): New function.
(diagnostic_num_locations): New function.
(diagnostic_expand_location): Get the location from the
rich_location.
(diagnostic_print_caret_line): Delete.
(diagnostic_get_color_for_kind): New declaration.
* genmatch.c (linemap_client_expand_location_to_spelling_point): New.
(error_cb): Update for change in signature of "error" callback.
(fatal_at): Likewise.
(warning_at): Likewise.
* input.c (linemap_client_expand_location_to_spelling_point): New.
* pretty-print.c (text_info::set_range): New method.
(text_info::get_location): New method.
* pretty-print.h (MAX_LOCATIONS_PER_MESSAGE): Eliminate this macro.
(struct text_info): Eliminate "locations" array in favor of
"m_richloc", a rich_location *.
(textinfo::set_location): Add a "caret_p" param, and reimplement
in terms of a call to set_range.
(textinfo::get_location): Eliminate inline implementation in favor of
an out-of-line reimplementation.
(textinfo::set_range): New method.
* rtl-error.c (diagnostic_for_asm): Update for change in signature
of diagnostic_set_info.
* tree-diagnostic.c (default_tree_printer): Update for new
"caret_p" param for textinfo::set_location.
* tree-pretty-print.c (percent_K_format): Likewise.
gcc/c-family/ChangeLog:
* c-common.c (c_cpp_error): Convert parameter from location_t to
rich_location *. Eliminate the "column_override" parameter and
the call to diagnostic_override_column.
Update the "done_lexing" clause to set range 0
on the rich_location, rather than overwriting a location_t.
* c-common.h (c_cpp_error): Convert parameter from location_t to
rich_location *. Eliminate the "column_override" parameter.
gcc/c/ChangeLog:
* c-decl.c (warn_defaults_to): Update for change in signature
of diagnostic_set_info.
* c-errors.c (pedwarn_c99): Likewise.
(pedwarn_c90): Likewise.
* c-objc-common.c (c_tree_printer): Update for new "caret_p" param
for textinfo::set_location.
gcc/cp/ChangeLog:
* error.c (cp_printer): Update for new "caret_p" param for
textinfo::set_location.
(pedwarn_cxx98): Update for change in signature of
diagnostic_set_info.
gcc/fortran/ChangeLog:
* cpp.c (cb_cpp_error): Convert parameter from location_t to
rich_location *. Eliminate the "column_override" parameter.
* error.c (gfc_warning): Update for change in signature of
diagnostic_set_info.
(gfc_format_decoder): Update handling of %C/%L for changes
to struct text_info.
(gfc_diagnostic_starter): Use richloc when determining whether to
print one locus or two. When handling a location that will
involve a call to diagnostic_show_locus, only attempt to print the
locus for the primary location, and don't call into
diagnostic_print_caret_line.
(gfc_warning_now_at): Update for change in signature of
diagnostic_set_info.
(gfc_warning_now): Likewise.
(gfc_error_now): Likewise.
(gfc_fatal_error): Likewise.
(gfc_error): Likewise.
(gfc_internal_error): Likewise.
gcc/testsuite/ChangeLog:
* gcc.dg/plugin/diagnostic-test-show-locus-bw.c: New file.
* gcc.dg/plugin/diagnostic-test-show-locus-color.c: New file.
* gcc.dg/plugin/diagnostic_plugin_test_show_locus.c: New file.
* gcc.dg/plugin/plugin.exp (plugin_test_list): Add the above.
* lib/gcc-dg.exp: Load multiline.exp.
libcpp/ChangeLog:
* errors.c (cpp_diagnostic): Update for change in signature
of "error" callback.
(cpp_diagnostic_with_line): Likewise, calling override_column
on the rich_location.
* include/cpplib.h (struct cpp_callbacks): Within "error"
callback, convert param from source_location to rich_location *,
and drop column_override param.
* include/line-map.h (struct source_range): New struct.
(struct location_range): New struct.
(class rich_location): New class.
(linemap_client_expand_location_to_spelling_point): New declaration.
* line-map.c (rich_location::rich_location): New ctors.
(rich_location::lazily_expand_location): New method.
(rich_location::override_column): New method.
(rich_location::add_range): New methods.
(rich_location::set_range): New method.
From-SVN: r229884
Only build net/hook_cloexec.go on GNU/Linux and FreeBSD, because those
are the only systems with accept4.
Add syscall/libcall_bsd.go to define sendfile for *BSD and Solaris.
Revert tcpsockopt_solaris.go back to the earlier version, so that it
works on Solaris 10.
Always pass the address of a Pid_t value to TIOCGPGRP and TIOCSPGRP.
Include <unistd.h> in runtime/go-varargs.c.
Reviewed-on: https://go-review.googlesource.com/16719
From-SVN: r229880
2015-11-06 Benedikt Huber <benedikt.huber@theobroma-systems.com>
Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
* config/aarch64/aarch64-builtins.c: Builtins for rsqrt and rsqrtf.
* config/aarch64/aarch64-protos.h: Declare.
* config/aarch64/aarch64-simd.md: Matching expressions for frsqrte and
frsqrts.
* config/aarch64/aarch64-tuning-flags.def: Added recip_sqrt.
* config/aarch64/aarch64.c: New functions. Emit rsqrt estimation code when
applicable.
* config/aarch64/aarch64.md: Added enum entries.
* config/aarch64/aarch64.opt: Added option -mlow-precision-recip-sqrt.
* testsuite/gcc.target/aarch64/rsqrt_asm_check_common.h: Common macros for
assembly checks.
* testsuite/gcc.target/aarch64/rsqrt_asm_check_negative_1.c: Make sure
frsqrts and frsqrte are not emitted.
* testsuite/gcc.target/aarch64/rsqrt_asm_check_1.c: Make sure frsqrts and
frsqrte are emitted.
* testsuite/gcc.target/aarch64/rsqrt_1.c: Functional tests for rsqrt.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Co-Authored-By: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
From-SVN: r229866
* parser.c (check_type_concept): Remove.
(cp_parser_maybe_constrained_type_specifier): Don't call it.
(synthesize_implicit_template_parm): Handle non-type and template
template parameters. Also compare extra args. Return the decl.
(cp_parser_template_argument): Handle constrained-type-specifiers for
non-type template parameters.
(finish_constrained_template_template_parm): Split out from
cp_parser_constrained_template_template_parm.
(cp_parser_nonclass_name): Move some logic into
cp_parser_maybe_concept_name.
(cp_parser_init_declarator): Fix error recovery.
(get_concept_from_constraint): Remove.
(cp_parser_simple_type_specifier): Adjust for
synthesize_implicit_template_parm returning the decl.
* constraint.cc (placeholder_extract_concept_and_args)
(equivalent_placeholder_constraints): Also handle TYPE_DECL
constrained parms.
From-SVN: r229860
PR ipa/68057
PR ipa/68220
* ipa-polymorphic-call.c
(ipa_polymorphic_call_context::restrict_to_inner_type): Fix ordering
issue when offset is out of range.
(contains_type_p): Fix out of range check, clear dynamic flag.
* g++.dg/lto/pr68057_0.C: New testcase.
* g++.dg/lto/pr68057_1.C: New testcase.
* g++.dg/torture/pr68220.C: New testcase.
From-SVN: r229859
2015-11-06 Tom de Vries <tom@codesourcery.com>
* tree-cfg.c (gimple_split_block_before_cond_jump): Split before
cond_jump, instead of split after last nondebug insn before cond_jump.
* tree-parloops.c (transform_to_exit_first_loop_alt): Verify ssa before
returning.
From-SVN: r229847
PR target/68088
* config/arm/aarch-common.c (aarch_accumulator_forwarding): Strip
subregs from accumulator and make sure it's a register.
* gcc.dg/pr68088_1.c: New test.
From-SVN: r229845
2015-11-06 Richard Biener <rguenther@suse.de>
* tree-vectorizer.h (struct _bb_vec_info): Add region_begin/end
members.
(vect_stmt_in_region_p): Declare.
* tree-vect-slp.c (new_bb_vec_info): Work on a region.
(destroy_bb_vec_info): Likewise.
(vect_bb_slp_scalar_cost): Use vect_stmt_in_region_p.
(vect_get_and_check_slp_defs): Likewise.
(vect_slp_analyze_bb_1): Refactor to make it work on sub-BBs.
(vect_slp_bb): Likewise.
* tree-vect-patterns.c (vect_same_loop_or_bb_p): Implement
in terms of vect_stmt_in_region_p.
(vect_pattern_recog): Iterate over the BB region.
* tree-vect-stmts.c (vect_is_simple_use): Use vect_stmt_in_region_p.
* tree-vectorizer.c (vect_stmt_in_region_p): New function.
(pass_slp_vectorize::execute): Initialize all stmt UIDs to -1.
* config/i386/i386.c: Include gimple-iterator.h.
* config/aarch64/aarch64.c: Likewise.
* gcc.dg/vect/bb-slp-38.c: New testcase.
From-SVN: r229842
In assign_parms_setup_block, the copy of args in PARALLELs from
entry_parm to stack_parm is deferred to the parm conversion insn seq,
but the copy from stack_parm to target_reg was inserted in the normal
copy seq, that is executed before the conversion insn seq. Oops.
We could do away with the need for an actual stack_parm in general,
which would have avoided the need for emitting the copy to target_reg
in the conversion seq, but at least on pa, due to the need for stack
to copy between SI and SF modes, it seems like using the reserved
stack slot is beneficial, so I put in logic to use a pre-reserved
stack slot when there is one, and emit the copy to target_reg in the
conversion seq if stack_parm was set up there.
for gcc/ChangeLog
PR rtl-optimization/67753
PR rtl-optimization/64164
* function.c (assign_parm_setup_block): Avoid allocating a
stack slot if we don't have an ABI-reserved one. Emit the
copy to target_reg in the conversion seq if the copy from
entry_parm is in it too. Don't use the conversion seq to copy
a PARALLEL to a REG or a CONCAT.
From-SVN: r229840
gcc/
* gimplify.c (gimplify_scan_omp_clauses): Add support for
OMP_CLAUSE_TILE. Update handling of OMP_CLAUSE_INDEPENDENT.
(gimplify_adjust_omp_clauses): Likewise.
* omp-low.c (scan_sharing_clauses): Add support for OMP_CLAUSE_TILE.
* tree-core.h (enum omp_clause_code): Add OMP_CLAUSE_TILE.
* tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE_TILE.
* tree.c (omp_clause_num_ops): Add an entry for OMP_CLAUSE_TILE.
(omp_clause_code_name): Likewise.
(walk_tree_1): Handle OMP_CLAUSE_TILE.
* tree.h (OMP_TILE_LIST): New macro.
gcc/c-family/
* c-omp.c (c_oacc_split_loop_clauses): Make TILE, GANG, WORKER, VECTOR,
AUTO, SEQ, INDEPENDENT and PRIVATE loop clauses. Associate REDUCTION
clauses with parallel and kernels and loops.
* c-pragma.h (enum pragma_omp_clause): Add entries for
PRAGMA_OACC_CLAUSE_{INDEPENDENT,TILE,DEFAULT}.
* pt.c (tsubst_omp_clauses): Add support for OMP_CLAUSE_{NUM_GANGS,
NUM_WORKERS,VECTOR_LENGTH,GANG,WORKER,VECTOR,ASYNC,WAIT,TILE,AUTO,
INDEPENDENT,SEQ}.
(tsubst_expr): Add support for OMP_CLAUSE_{KERNELS,PARALLEL,LOOP}.
gcc/c/
* c-parser.c (c_parser_omp_clause_name): Add support for
PRAGMA_OACC_CLAUSE_INDEPENDENT and PRAGMA_OACC_CLAUSE_TILE.
(c_parser_omp_clause_default): Add is_oacc argument. Handle
default(none) in OpenACC.
(c_parser_oacc_shape_clause): Allow pointer variables as gang static
arguments.
(c_parser_oacc_clause_tile): New function.
(c_parser_oacc_all_clauses): Add support for OMP_CLAUSE_DEFAULT,
OMP_CLAUSE_INDEPENDENT and OMP_CLAUSE_TILE.
(OACC_LOOP_CLAUSE_MASK): Add PRAGMA_OACC_CLAUSE_{PRIVATE,INDEPENDENT,
TILE}.
(OACC_KERNELS_MASK): Add PRAGMA_OACC_CLAUSE_DEFAULT.
(OACC_PARALLEL_MASK): Add PRAGMA_OACC_CLAUSE_{DEFAULT,PRIVATE,
FIRSTPRIVATE}.
(c_parser_omp_all_clauses): Update call to c_parser_omp_clause_default.
(c_parser_oacc_update): Update the error message for missing clauses.
* c-typeck.c (c_finish_omp_clauses): Add support for OMP_CLAUSE_TILE
and OMP_CLAUSE_INDEPENDENT.
gcc/cp/
* parser.c (cp_parser_omp_clause_name): Add support for
PRAGMA_OACC_CLAUSE_INDEPENDENT and PRAGMA_OACC_CLAUSE_TILE.
(cp_parser_oacc_shape_clause): Allow pointer variables as gang static
arguments.
(cp_parser_oacc_clause_tile): New function.
(cp_parser_omp_clause_default): Add is_oacc argument. Handle
default(none) in OpenACC.
(cp_parser_oacc_all_clauses): Add support for
(cp_parser_omp_all_clauses): Update call to
cp_parser_omp_clause_default.
PRAGMA_OACC_CLAUSE_{DEFAULT,INDEPENDENT,TILE,PRIVATE,FIRSTPRIVATE}.
(OACC_LOOP_CLAUSE_MASK): Add PRAGMA_OACC_CLAUSE_{PRIVATE,INDEPENDENT,
TILE}.
(OACC_KERNELS_MASK): Add PRAGMA_OACC_CLAUSE_DEFAULT.
(OACC_PARALLEL_MASK): Add PRAGMA_OACC_CLAUSE_{DEFAULT,PRIVATE,
FIRSTPRIVATE}.
(cp_parser_oacc_update): Update the error message for missing clauses.
* semantics.c (finish_omp_clauses): Add support for
OMP_CLAUSE_INDEPENDENT and OMP_CLAUSE_TILE.
gcc/fortran/
* openmp.c (gfc_match_omp_clauses): Update support for the tile
and default clauses in OpenACC.
(gfc_match_oacc_update): Error when data clauses are supplied.
(oacc_compatible_clauses): Delete.
(resolve_omp_clauses): Give special care for OpenACC reductions.
Also update error reporting for the tile clause.
(resolve_oacc_loop_blocks): Update error reporting for the tile clause.
* trans-openmp.c (gfc_trans_omp_clauses): Update OMP_CLAUSE_SEQ. Add
OMP_CLAUSE_{AUTO,TILE} and add support the the gang static argument.
(gfc_trans_oacc_combined_directive): Update the list of clauses which
are split to acc loops.
gcc/testsuite/
* c-c++-common/goacc/combined-directives.c: New test.
* c-c++-common/goacc/loop-clauses.c: New test.
* c-c++-common/goacc/tile.c: New test.
* c-c++-common/goacc/loop-shape.c: Add test for pointer variable
as gang static arguments.
* c-c++-common/goacc/update-1.c: Adjust expected error message.
* g++.dg/goacc/template.C: New test.
* gfortran.dg/goacc/combined-directives.f90: New test.
* gfortran.dg/goacc/default.f95: New test.
* gfortran.dg/goacc/default_none.f95: New test.
* gfortran.dg/goacc/firstprivate-1.f95: New test.
* gfortran.dg/goacc/gang-static.f95: New test.
* gfortran.dg/goacc/kernels-loop-inner.f95: New test.
* gfortran.dg/goacc/kernels-loops-adjacent.f95: New test.
* gfortran.dg/goacc/list.f95: Update test.
* gfortran.dg/goacc/loop-2.f95: Likewise.
* gfortran.dg/goacc/loop-4.f95: New test.
* gfortran.dg/goacc/loop-5.f95: New test.
* gfortran.dg/goacc/loop-6.f95: New test.
* gfortran.dg/goacc/loop-tree-1.f90: Update test.
* gfortran.dg/goacc/multi-clause.f90: New test.
* gfortran.dg/goacc/parallel-tree.f95: Update test.
* gfortran.dg/goacc/update.f95: New test.
Co-Authored-By: James Norris <jnorris@codesourcery.com>
Co-Authored-By: Nathan Sidwell <nathan@codesourcery.com>
Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
Co-Authored-By: Tom de Vries <tom@codesourcery.com>
From-SVN: r229832
gcc/cp/
* cp/init.c (warn_placement_new_too_small): Avoid assuming
the size of the first operand of placement new or its type
is known.
gcc/testsuite/
* g++.dg/warn/Wplacement-new-size.C: Exercise placement new
invocations where the size of the destination buffer object
or its type (or both) is unknown.
From-SVN: r229831
2015-11-05 Jakub Jelinek <jakub@redhat.com>
Ilya Verbin <ilya.verbin@intel.com>
* builtin-types.def
(BT_FN_VOID_INT_OMPFN_SIZE_PTR_PTR_PTR_UINT_PTR): Remove.
(BT_FN_VOID_INT_OMPFN_SIZE_PTR_PTR_PTR_UINT_PTR_INT_INT): New.
* cgraph.h (enum cgraph_simd_clone_arg_type): Add
SIMD_CLONE_ARG_TYPE_LINEAR_REF_VARIABLE_STEP,
SIMD_CLONE_ARG_TYPE_LINEAR_UVAL_VARIABLE_STEP and
SIMD_CLONE_ARG_TYPE_LINEAR_VAL_VARIABLE_STEP.
(struct cgraph_simd_clone_arg): Adjust comment.
* omp-builtins.def (BUILT_IN_GOMP_TARGET): Rename GOMP_target_41
to GOMP_target_ext. Add num_teams and thread_limit arguments.
(BUILT_IN_GOMP_TARGET_DATA): Rename GOMP_target_data_41
to GOMP_target_data_ext.
(BUILT_IN_GOMP_TARGET_UPDATE): Rename GOMP_target_update_41
to GOMP_target_update_ext.
(BUILT_IN_GOMP_LOOP_NONMONOTONIC_DYNAMIC_START,
BUILT_IN_GOMP_LOOP_NONMONOTONIC_GUIDED_START,
BUILT_IN_GOMP_LOOP_NONMONOTONIC_DYNAMIC_NEXT,
BUILT_IN_GOMP_LOOP_NONMONOTONIC_GUIDED_NEXT,
BUILT_IN_GOMP_LOOP_ULL_NONMONOTONIC_DYNAMIC_START,
BUILT_IN_GOMP_LOOP_ULL_NONMONOTONIC_GUIDED_START,
BUILT_IN_GOMP_LOOP_ULL_NONMONOTONIC_DYNAMIC_NEXT,
BUILT_IN_GOMP_LOOP_ULL_NONMONOTONIC_GUIDED_NEXT,
BUILT_IN_GOMP_PARALLEL_LOOP_NONMONOTONIC_DYNAMIC,
BUILT_IN_GOMP_PARALLEL_LOOP_NONMONOTONIC_GUIDED): New built-ins.
* tree-core.h (enum omp_clause_schedule_kind): Add
OMP_CLAUSE_SCHEDULE_MASK, OMP_CLAUSE_SCHEDULE_MONOTONIC,
OMP_CLAUSE_SCHEDULE_NONMONOTONIC and change
OMP_CLAUSE_SCHEDULE_LAST value.
* tree.def (OMP_SIMD, CILK_SIMD, CILK_FOR, OMP_DISTRIBUTE,
OMP_TASKLOOP, OACC_LOOP): Add OMP_FOR_ORIG_DECLS argument.
* tree.h (OMP_FOR_ORIG_DECLS): Use OMP_LOOP_CHECK instead of
OMP_FOR_CHECK. Remove comment.
* tree-pretty-print.c (dump_omp_clause): Handle
GOMP_MAP_FIRSTPRIVATE_REFERENCE and GOMP_MAP_ALWAYS_POINTER.
Simplify. Print schedule clause modifiers.
* tree-vect-stmts.c (vectorizable_simd_clone_call): Add
SIMD_CLONE_ARG_TYPE_LINEAR_{REF,VAL,UVAL}_VARIABLE_STEP
cases.
* gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP_ALWAYS_TO.
(omp_default_clause): Tweak for
private/firstprivate/is_device_ptr variables on target
construct and use_device_ptr on target data.
(omp_check_private): Likewise.
(omp_notice_variable): For references check whether what it refers
to has mappable type, rather than the reference itself.
(omp_is_private): Diagnose linear iteration variables on non-simd
constructs.
(omp_no_lastprivate): Return true only for Fortran.
(gimplify_scan_omp_clauses): Or in GOVD_MAP_ALWAYS_TO for
GOMP_MAP_ALWAYS_TO or GOMP_MAP_ALWAYS_TOFROM kinds.
Add support for GOMP_MAP_FIRSTPRIVATE_REFERENCE and
GOMP_MAP_ALWAYS_POINTER, remove old handling of structure element
based array sections. Use GOMP_MAP_ALWAYS_P. Fix up handling of
lastprivate and linear when combined with distribute. Gimplify
variable low-bound for array reduction. Look through
POINTER_PLUS_EXPR when looking for ADDR_EXPR for array section
reductions.
(gimplify_adjust_omp_clauses_1): For implicit references to
variables with reference type and when not ref to scalar or
ref to pointer, map what they refer to using tofrom and
use GOMP_MAP_FIRSTPRIVATE_REFERENCE for the reference.
(gimplify_adjust_omp_clauses): Remove GOMP_MAP_ALWAYS_POINTER
from target exit data. Handle GOMP_MAP_FIRSTPRIVATE_REFERENCE.
Drop OMP_CLAUSE_MAP_PRIVATE support. Use GOMP_MAP_ALWAYS_P.
Diagnose the same var on both firstprivate and lastprivate on
distribute construct.
(gimplify_omp_for): Fix up handling of predetermined
lastprivate or linear iter vars when combined with distribute.
(find_omp_teams, computable_teams_clause, optimize_target_teams): New
functions.
(gimplify_omp_workshare): Call optimize_target_teams.
* omp-low.c (struct omp_region): Add sched_modifiers field.
(struct omp_for_data): Likewise.
(omp_any_child_fn_dumped): New variable.
(extract_omp_for_data): Fill in sched_modifiers, and mask out
OMP_CLAUSE_SCHEDULE_KIND bits outside of OMP_CLAUSE_SCHEDULE_MASK
from sched_kind.
(determine_parallel_type): Use only OMP_CLAUSE_SCHEDULE_MASK
bits of OMP_CLAUSE_SCHED_KIND.
(scan_sharing_clauses): Handle GOMP_MAP_FIRSTPRIVATE_REFERENCE,
drop OMP_CLAUSE_MAP_PRIVATE support. Look through POINTER_PLUS_EXPR
for array section reductions.
(add_taskreg_looptemp_clauses): Add one extra _looptemp_ clause even
for distribute parallel for, if there are lastprivate clauses on the
for.
(lower_rec_input_clauses): Handle non-zero low-bound on array
section reductions.
(lower_reduction_clauses): Likewise.
(lower_send_clauses): Look through POINTER_PLUS_EXPR
for array section reductions.
(expand_parallel_call): Use nonmonotonic entrypoints for
nonmonotonic: dynamic/guided.
(expand_omp_taskreg): Call assign_assembler_name_if_neeeded on
child_fn if current_function_decl has assembler name set, but child_fn
does not. Dump the header and IL of the child function when not in SSA
form.
(expand_omp_target): Likewise. Pass num_teams and thread_limit
arguments to BUILT_IN_GOMP_TARGET.
(expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
Initialize the extra _looptemp_ clause to fd->loop.n2.
(expand_omp_for): Use nonmonotonic entrypoints for
nonmonotonic: dynamic/guided. Initialize region->sched_modifiers.
(expand_omp): Clear omp_any_child_fn_dumped. Dump function header
again if we have dumped any child functions.
(lower_omp_for_lastprivate): Determine the right count variable
for distribute simd, or distribute parallel for{, simd}.
(lower_omp_target): Handle GOMP_MAP_FIRSTPRIVATE_REFERENCE
and GOMP_MAP_ALWAYS_POINTER. Drop OMP_CLAUSE_MAP_PRIVATE
support.
(simd_clone_clauses_extract): Handle variable step
for references and arguments passed by reference.
(simd_clone_mangle): Mangle ref/uval/val variable steps.
(simd_clone_adjust_argument_types): Handle
SIMD_CLONE_ARG_TYPE_LINEAR_UVAL_VARIABLE_STEP like
SIMD_CLONE_ARG_TYPE_LINEAR_UVAL_CONSTANT_STEP and
SIMD_CLONE_ARG_TYPE_LINEAR_VAL_VARIABLE_STEP like
SIMD_CLONE_ARG_TYPE_LINEAR_VAL_CONSTANT_STEP.
(simd_clone_linear_addend): New function.
(simd_clone_adjust): Handle variable step like similarly
to constant step, use simd_clone_linear_addend to determine
the actual step at runtime.
gcc/c-family/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
* c-common.h (c_finish_omp_atomic): Add TEST argument.
(c_omp_check_loop_iv, c_omp_check_loop_iv_exprs): New prototypes.
* c-omp.c (c_finish_omp_atomic): Add TEST argument. Don't call
save_expr or create_tmp_var* if TEST is true.
(c_finish_omp_for): Store OMP_FOR_ORIG_DECLS always.
Don't call add_stmt here.
(struct c_omp_check_loop_iv_data): New type.
(c_omp_check_loop_iv_r, c_omp_check_loop_iv,
c_omp_check_loop_iv_exprs): New functions.
(c_omp_split_clauses): Adjust for lastprivate being allowed on
distribute.
(c_omp_declare_simd_clauses_to_numbers): Change
OMP_CLAUSE_LINEAR_VARIABLE_STRIDE OMP_CLAUSE_LINEAR_STEP into numbers.
(c_omp_declare_simd_clauses_to_decls): Similarly change those
from numbers to PARM_DECLs.
gcc/c/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
Ilya Verbin <ilya.verbin@intel.com>
* c-parser.c: Include context.h and gimple-expr.h.
(c_parser_omp_clause_schedule): Parse schedule modifiers, diagnose
monotonic together with nonmonotonic.
(c_parser_omp_for_loop): Call c_omp_check_loop_iv. Call add_stmt here.
(OMP_DISTRIBUTE_CLAUSE_MASK): Add lastprivate clause.
(c_parser_omp_target_data, c_parser_omp_target_enter_data,
c_parser_omp_target_exit_data): Allow GOMP_MAP_ALWAYS_POINTER.
(c_parser_omp_target): Likewise. Evaluate num_teams and thread_limit
expressions on combined target teams before the target.
(c_parser_omp_declare_target): If decl has "omp declare target" or
"omp declare target link" attribute, and cgraph or varpool node already
exists, then set corresponding flags. Call c_finish_omp_clauses
in the parenthesized extended-list syntax case.
* c-decl.c (c_decl_attributes): Don't diagnose block scope vars inside
declare target.
* c-typeck.c (handle_omp_array_sections_1): Allow non-zero low-bound
on OMP_CLAUSE_REDUCTION array sections.
(handle_omp_array_sections): Encode low-bound into the MEM_REF, either
into the constant offset, or for variable low-bound using
POINTER_PLUS_EXPR. For structure element based array sections use
GOMP_MAP_ALWAYS_POINTER instead of GOMP_MAP_FIRSTPRIVATE_POINTER.
(c_finish_omp_clauses): Drop generic_field_head, structure
elements are now always mapped even as array section bases,
diagnose same var in data sharing and mapping clauses. Diagnose if
linear step on declare simd is neither a constant nor a uniform
parameter. Look through POINTER_PLUS_EXPR for array section
reductions. Diagnose the same var or function appearing multiple
times on the same directive. Fix up wording for the to clause if t
is neither a FUNCTION_DECL nor a VAR_DECL. Diagnose nonmonotonic
modifier on kinds other than dynamic or guided or nonmonotonic
modifier together with ordered clause.
gcc/cp/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
Ilya Verbin <ilya.verbin@intel.com>
* cp-tree.h (finish_omp_for): Add ORIG_INITS argument.
(omp_privatize_field): Add SHARED argument.
* parser.c: Include context.h.
(cp_parser_omp_clause_schedule): Parse schedule
modifiers, diagnose monotonic together with nonmonotonic.
(cp_parser_omp_clause_linear): Add DECLARE_SIMD argument. Parse
parameter name as linear step as id-expression rather than expression.
(cp_parser_omp_all_clauses): Adjust caller.
(cp_parser_omp_for_loop_init): Add ORIG_INIT argument,
initialize it. Adjust omp_privatize_field caller.
(cp_parser_omp_for_loop): Compute orig_inits, pass it's address
to finish_omp_for.
(OMP_DISTRIBUTE_CLAUSE_MASK): Add lastprivate clause.
(cp_parser_omp_target_data,
cp_parser_omp_target_enter_data,
cp_parser_omp_target_exit_data): Allow GOMP_MAP_ALWAYS_POINTER
and GOMP_MAP_FIRSTPRIVATE_REFERENCE.
(cp_parser_omp_target): Likewise. Evaluate num_teams and
thread_limit expressions on combined target teams before the target.
(cp_parser_omp_declare_target): If decl has "omp declare target" or
"omp declare target link" attribute, and cgraph or varpool node already
exists, then set corresponding flags. Call finish_omp_clauses
in the parenthesized extended-list syntax case. Call
cp_parser_require_pragma_eol instead of cp_parser_skip_to_pragma_eol.
(cp_parser_omp_end_declare_target): Call cp_parser_require_pragma_eol
instead of cp_parser_skip_to_pragma_eol.
* decl2.c (cplus_decl_attributes): Don't diagnose block scope vars inside
declare target.
* pt.c (tsubst_omp_clauses): If OMP_CLAUSE_LINEAR_VARIABLE_STRIDE,
use tsubst_omp_clause_decl instead of tsubst_expr on
OMP_CLAUSE_LINEAR_STEP. Handle non-static data members in shared
clauses.
(tsubst_omp_for_iterator): Adjust omp_privatize_field caller.
(tsubst_find_omp_teams): New function.
(tsubst_expr): Evaluate num_teams and thread_limit expressions on
combined target teams before the target. Use OMP_FOR_ORIG_DECLS for
all OpenMP/OpenACC/Cilk+ looping constructs. Adjust finish_omp_for
caller.
* semantics.c (omp_privatize_field): Add SHARED argument, if true,
always create artificial var and never put it into the hash table
or vector.
(handle_omp_array_sections_1): Adjust omp_privatize_field caller.
Allow non-zero low-bound on OMP_CLAUSE_REDUCTION array sections.
(handle_omp_array_sections): For structure element
based array sections use GOMP_MAP_ALWAYS_POINTER instead of
GOMP_MAP_FIRSTPRIVATE_POINTER. Encode low-bound into the MEM_REF,
either into the constant offset, or for variable low-bound using
POINTER_PLUS_EXPR.
(finish_omp_clauses): Adjust omp_privatize_field caller. Drop
generic_field_head, structure elements are now always mapped even
as array section bases, diagnose same var in data sharing and
mapping clauses. For references map what they refer to using
GOMP_MAP_ALWAYS_POINTER for structure elements and
GOMP_MAP_FIRSTPRIVATE_REFERENCE otherwise. Diagnose if linear step
on declare simd is neither a constant nor a uniform parameter.
Allow non-static data members on shared clauses. Look through
POINTER_PLUS_EXPR for array section reductions. Diagnose nonmonotonic
modifier on kinds other than dynamic or guided or nonmonotonic
modifier together with ordered clause. Diagnose the same var or
function appearing multiple times on the same directive. Fix up
wording for the to clause if t is neither a FUNCTION_DECL nor a
VAR_DECL, use special wording for OVERLOADs and TEMPLATE_ID_EXPR.
(handle_omp_for_class_iterator): Add ORIG_DECLS argument. Call
c_omp_check_loop_iv_exprs on cond.
(finish_omp_for): Add ORIG_INITS argument. Call
c_omp_check_loop_iv_exprs on ORIG_INITS elements. Adjust
handle_omp_for_class_iterator caller. Call c_omp_check_loop_iv.
Call add_stmt.
(finish_omp_atomic): Adjust c_finish_omp_atomic caller.
gcc/fortran/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
* types.def (BT_FN_VOID_INT_OMPFN_SIZE_PTR_PTR_PTR_UINT_PTR): Remove.
(BT_FN_VOID_INT_OMPFN_SIZE_PTR_PTR_PTR_UINT_PTR_INT_INT): New.
gcc/testsuite/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
* c-c++-common/gomp/clauses-2.c (foo): Adjust for diagnostics
of variables in both data sharing and mapping clauses and for
structure element based array sections being mapped rather than
privatized.
* c-c++-common/gomp/declare-target-2.c: Add various new tests. Adjust
expected diagnostics wording in one case.
* c-c++-common/gomp/distribute-1.c: New test.
* c-c++-common/gomp/element-1.c: New test.
* c-c++-common/gomp/pr61486-2.c: Add #pragma omp declare target
and #pragma omp end declare target pair around the function.
Change s from a parameter to a file scope variable.
* c-c++-common/gomp/pr67521.c: Add dg-error directives.
* c-c++-common/gomp/reduction-1.c (foo): Don't expect diagnostics
on non-zero low-bound in reduction array sections. Add further
tests.
* c-c++-common/gomp/schedule-modifiers-1.c: New test.
* c-c++-common/gomp/target-teams-1.c: New test.
* gcc.dg/gomp/declare-simd-1.c: Add scan-assembler-times directives
for expected mangling on x86_64/i?86.
* gcc.dg/gomp/declare-simd-3.c: New test.
* gcc.dg/gomp/declare-simd-4.c: New test.
* gcc.dg/gomp/for-20.c: New test.
* gcc.dg/gomp/for-21.c: New test.
* gcc.dg/gomp/for-22.c: New test.
* gcc.dg/gomp/for-23.c: New test.
* gcc.dg/gomp/for-24.c: New test.
* gcc.dg/gomp/linear-1.c: New test.
* gcc.dg/gomp/loop-1.c: New test.
* g++.dg/gomp/atomic-17.C: New test.
* g++.dg/gomp/clause-1.C (T::test): Don't expect error on
non-static data member in shared clause. Add single construct.
* g++.dg/gomp/declare-simd-1.C: Add dg-options. Add
scan-assembler-times directives for expected mangling on x86_64/i?86.
* g++.dg/gomp/declare-simd-3.C: Likewise.
* g++.dg/gomp/declare-simd-4.C: New test.
* g++.dg/gomp/declare-simd-5.C: New test.
* g++.dg/gomp/declare-target-1.C: New test.
* g++.dg/gomp/linear-2.C: New test.
* g++.dg/gomp/loop-1.C: New test.
* g++.dg/gomp/loop-2.C: New test.
* g++.dg/gomp/loop-3.C: New test.
* g++.dg/gomp/member-2.C (B::m2, B::m4): Don't expect error on
non-static data member in shared clause.
* g++.dg/gomp/member-3.C: New test.
* g++.dg/gomp/member-4.C: New test.
* g++.dg/gomp/pr38639.C (foo): Adjust dg-error.
(bar): Remove dg-message.
* g++.dg/gomp/target-teams-1.C: New test.
include/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
Ilya Verbin <ilya.verbin@intel.com>
* gomp-constants.h (GOMP_MAP_FLAG_SPECIAL_2): Define.
(GOMP_MAP_FLAG_ALWAYS): Remove.
(enum gomp_map_kind): Use GOMP_MAP_FLAG_SPECIAL_2 instead of
GOMP_MAP_FLAG_ALWAYS for GOMP_MAP_ALWAYS_TO, GOMP_MAP_ALWAYS_FROM,
GOMP_MAP_ALWAYS_TOFROM, GOMP_MAP_STRUCT, GOMP_MAP_RELEASE.
Add GOMP_MAP_ALWAYS_POINTER and GOMP_MAP_FIRSTPRIVATE_REFERENCE.
(GOMP_MAP_ALWAYS_P): Define.
(GOMP_TARGET_FLAG_NOWAIT): Adjust comment.
libgomp/
2015-11-05 Jakub Jelinek <jakub@redhat.com>
Ilya Verbin <ilya.verbin@intel.com>
* libgomp_g.h (GOMP_loop_nonmonotonic_dynamic_next,
GOMP_loop_nonmonotonic_dynamic_start,
GOMP_loop_nonmonotonic_guided_next,
GOMP_loop_nonmonotonic_guided_start,
GOMP_loop_ull_nonmonotonic_dynamic_next,
GOMP_loop_ull_nonmonotonic_dynamic_start,
GOMP_loop_ull_nonmonotonic_guided_next,
GOMP_loop_ull_nonmonotonic_guided_start,
GOMP_parallel_loop_nonmonotonic_dynamic,
GOMP_parallel_loop_nonmonotonic_guided): New prototypes.
(GOMP_target_41): Renamed to ...
(GOMP_target_ext): ... this. Add num_teams and thread_limit
arguments.
(GOMP_target_data_41): Renamed to ...
(GOMP_target_data_ext): ... this.
(GOMP_target_update_41): Renamed to ...
(GOMP_target_update_ext): ... this.
* libgomp.map (GOMP_4.5): Export GOMP_target_ext,
GOMP_target_data_ext and GOMP_target_update_ext instead of
GOMP_target_41, GOMP_target_data_41 and GOMP_target_update_41.
Export GOMP_loop_nonmonotonic_dynamic_next,
GOMP_loop_nonmonotonic_dynamic_start,
GOMP_loop_nonmonotonic_guided_next,
GOMP_loop_nonmonotonic_guided_start,
GOMP_loop_ull_nonmonotonic_dynamic_next,
GOMP_loop_ull_nonmonotonic_dynamic_start,
GOMP_loop_ull_nonmonotonic_guided_next,
GOMP_loop_ull_nonmonotonic_guided_start,
GOMP_parallel_loop_nonmonotonic_dynamic and
GOMP_parallel_loop_nonmonotonic_guided.
* loop.c (GOMP_parallel_loop_nonmonotonic_dynamic,
GOMP_parallel_loop_nonmonotonic_guided,
GOMP_loop_nonmonotonic_dynamic_start,
GOMP_loop_nonmonotonic_guided_start,
GOMP_loop_nonmonotonic_dynamic_next,
GOMP_loop_nonmonotonic_guided_next): New aliases or functions.
* loop_ull.c (GOMP_loop_ull_nonmonotonic_dynamic_start,
GOMP_loop_ull_nonmonotonic_guided_start,
GOMP_loop_ull_nonmonotonic_dynamic_next,
GOMP_loop_ull_nonmonotonic_guided_next): Likewise.
* target.c (gomp_map_0len_lookup, gomp_map_val): New inline
functions.
(gomp_map_vars): Handle GOMP_MAP_ALWAYS_POINTER. For
GOMP_MAP_ZERO_LEN_ARRAY_SECTION use gomp_map_0len_lookup.
Use gomp_map_val function.
(gomp_target_fallback_firstprivate): New static function.
(GOMP_target_41): Renamed to ...
(GOMP_target_ext): ... this. Add num_teams and thread_limit
arguments. Move firstprivate fallback handling into a new
function.
(GOMP_target_data_41): Renamed to ...
(GOMP_target_data_ext): ... this.
(GOMP_target_update_41): Renamed to ...
(GOMP_target_update_ext): ... this.
(gomp_exit_data): For GOMP_MAP_*ZERO_LEN* use
gomp_map_0len_lookup instead of gomp_map_lookup.
(omp_target_is_present): Use gomp_map_0len_lookup instead of
gomp_map_lookup.
* testsuite/libgomp.c/target-28.c: Likewise.
* testsuite/libgomp.c/monotonic-1.c: New test.
* testsuite/libgomp.c/monotonic-2.c: New test.
* testsuite/libgomp.c/nonmonotonic-1.c: New test.
* testsuite/libgomp.c/nonmonotonic-2.c: New test.
* testsuite/libgomp.c/pr66199-5.c: New test.
* testsuite/libgomp.c/pr66199-6.c: New test.
* testsuite/libgomp.c/pr66199-7.c: New test.
* testsuite/libgomp.c/pr66199-8.c: New test.
* testsuite/libgomp.c/pr66199-9.c: New test.
* testsuite/libgomp.c/reduction-11.c: New test.
* testsuite/libgomp.c/reduction-12.c: New test.
* testsuite/libgomp.c/reduction-13.c: New test.
* testsuite/libgomp.c/reduction-14.c: New test.
* testsuite/libgomp.c/reduction-15.c: New test.
* testsuite/libgomp.c/target-12.c (main): Adjust for
omp_target_is_present change for one-past-last element.
* testsuite/libgomp.c/target-17.c (foo): Drop tests where
the same var is both mapped and privatized.
* testsuite/libgomp.c/target-19.c (foo): Adjust for different
handling of zero-length array sections.
* testsuite/libgomp.c/target-28.c: New test.
* testsuite/libgomp.c/target-29.c: New test.
* testsuite/libgomp.c/target-30.c: New test.
* testsuite/libgomp.c/target-teams-1.c: New test.
* testsuite/libgomp.c++/member-6.C: New test.
* testsuite/libgomp.c++/member-7.C: New test.
* testsuite/libgomp.c++/monotonic-1.C: New test.
* testsuite/libgomp.c++/monotonic-2.C: New test.
* testsuite/libgomp.c++/nonmonotonic-1.C: New test.
* testsuite/libgomp.c++/nonmonotonic-2.C: New test.
* testsuite/libgomp.c++/pr66199-3.C: New test.
* testsuite/libgomp.c++/pr66199-4.C: New test.
* testsuite/libgomp.c++/pr66199-5.C: New test.
* testsuite/libgomp.c++/pr66199-6.C: New test.
* testsuite/libgomp.c++/pr66199-7.C: New test.
* testsuite/libgomp.c++/pr66199-8.C: New test.
* testsuite/libgomp.c++/pr66199-9.C: New test.
* testsuite/libgomp.c++/reduction-11.C: New test.
* testsuite/libgomp.c++/reduction-12.C: New test.
* testsuite/libgomp.c++/target-13.C: New test.
* testsuite/libgomp.c++/target-14.C: New test.
* testsuite/libgomp.c++/target-15.C: New test.
* testsuite/libgomp.c++/target-16.C: New test.
* testsuite/libgomp.c++/target-17.C: New test.
* testsuite/libgomp.c++/target-18.C: New test.
* testsuite/libgomp.c++/target-19.C: New test.
Co-Authored-By: Ilya Verbin <ilya.verbin@intel.com>
From-SVN: r229814
Now that PR63304 is fixed and we have an option to address
any part of the memory using adrp / add or adrp / ldr instructions
it makes sense to switch out literal pools into their own
mergeable sections by default.
This would mean that by default we could now start getting
the benefits of constant sharing across the board, potentially
improving code size. The other advantage of doing so, for the
security conscious is that this prevents intermingling of literal
pools with code.
Wilco's kindly done some performance measurements and suggests that
there is not really a performance regression in doing this.
I've looked at the code size for SPEC2k6 today at -Ofast and
in general there is a good code size improvement as expected
by sharing said constants.
Tested on aarch64-none-elf with no regressions and bootstrapped
and regression tested in my tree for a number of days now.
2015-11-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
* config/aarch64/aarch64.c
(aarch64_can_use_per_function_literal_pools_p): New.
(aarch64_use_blocks_for_constant_p): Adjust declaration
and use aarch64_can_use_function_literal_pools_p.
(aarch64_select_rtx_section): Update.
From-SVN: r229795
