* *.c: Check the value of STACK_GROWS_DOWNWARD rather than if it
is defined.
* config/**/*.h: Define STACK_GROWS_DOWNWARD to an integer.
* defaults.h: Provide default for STACK_GROWS_DOWNWARD.
* doc/tm.texi.in: Update references to STACK_GROWS_DOWNWARD.
* doc/tm.texi: Regenerate.
H.J. Lu [Thu, 21 May 2015 21:58:57 +0000 (21:58 +0000)]
Allow indirect branch via GOT slot for x32
X32 doesn't support indirect branch via 32-bit memory slot since
indirect branch will load 64-bit address from 64-bit memory slot.
Since x32 GOT slot is 64-bit, we should allow indirect branch via GOT
slot for x32.
gcc/
PR target/66232
* config/i386/constraints.md (Bg): New constraint for GOT memory
operand.
* config/i386/i386.md (*call_got_x32): New pattern.
(*call_value_got_x32): Likewise.
* config/i386/predicates.md (GOT_memory_operand): New predicate.
David Edelsohn [Thu, 21 May 2015 17:25:00 +0000 (17:25 +0000)]
inclhack.def (aix_externc): New fix.
* inclhack.def (aix_externc): New fix.
(aix_externcpp[12]): New fix.
* fixincl.x: Regenerate.
* test/base/ctype.h [AIX_EXTERNC_CHECK]: New test.
* test/base/sys/socket.h [AIX_EXTERNCPP[12]_CHECK]: New test.
* test/base/fcntl.h: New file.
David Edelsohn [Thu, 21 May 2015 17:18:25 +0000 (13:18 -0400)]
re PR target/66224 (PowerPC _GLIBCXX_READ_MEM_BARRIER too weak)
PR target/66224
* config/cpu/powerpc/atomic_word.h (_GLIBCXX_READ_MEM_BARRIER):
Don't use isync. Use lwsync if available.
* configure.host (atomic_word_dir) [aix[56789]*]: Delete to use
powerpc cpu definition.
* einfo.adb (Contract): This attribute now applies to constants.
(Set_Contract): This attribute now applies to constants.
(Write_Field34_Name): Add output for constants.
* einfo.ads Attribute Contract now applies to constants.
* sem_ch3.adb (Analyze_Object_Contract): Constants now have
their Part_Of indicator verified.
* sem_prag.adb (Analyze_Constituent): A constant is now a valid
constituent.
(Analyze_Global_Item): A constant cannot act as an output.
(Analyze_Initialization_Item): Constants are now a valid
initialization item.
(Analyze_Initializes_In_Decl_Part): Rename
global variable States_And_Vars to States_And_Objs and update
all its occurrences.
(Analyze_Input_Item): Constants are now a
valid initialization item. Remove SPARM RM references from error
messages.
(Analyze_Pragma): Indicator Part_Of can now apply to a constant.
(Collect_Body_States): Collect both source constants
and variables.
(Collect_States_And_Objects): Collect both source constants and
variables.
(Collect_States_And_Variables): Rename
to Collect_States_And_Objects and update all its occurrences.
(Collect_Visible_States): Do not collect constants and variables
used to map generic formals to actuals.
(Find_Role): The role of a constant is that of an input. Separate the
role of a variable from that of a constant.
(Report_Unused_Constituents): Add specialized wording for constants.
(Report_Unused_States): Add specialized wording for constants.
* sem_util.adb (Add_Contract_Item): Add processing for constants.
* sem_util.ads (Add_Contract_Item): Update the comment on usage.
(Find_Placement_In_State_Space): Update the comment on usage.
Jeff Law [Thu, 21 May 2015 12:44:35 +0000 (06:44 -0600)]
pa.c (pa_print_operand): New 'o' output modifier.
2015-05-20 Jeff Law <law@redhat.com>
* config/pa/pa.c (pa_print_operand): New 'o' output modifier.
(pa_mem_shadd_constant_p): Renamed from pa_shadd_constant_p.
(pa_shadd_constant_p): Allow constants for shadd insns rather
than valid scaling constants for memory addresses.
* config/pa/pa-protos.h (pa_mem_shadd_constant_p): Add prototype.
* config/pa/predicates.md (mem_shadd_operand): New predicate.
* config/pa/pa.md (shift-add insns using MULT): Use mem_shadd_operand.
(shift-add insns using ASHIFT): New patterns.
* gcc.target/hppa/hppa.exp: New target test driver.
* gcc.target/hppa/shadd-1.c: New test.
Arnaud Charlet [Thu, 21 May 2015 10:51:22 +0000 (12:51 +0200)]
[multiple changes]
2015-05-21 Robert Dewar <dewar@adacore.com>
* freeze.adb: Minor reformatting.
* cstand.adb (Print_Standard): Fix bad printing of Duration
low bound.
* a-reatim.adb (Time_Of): Complete rewrite to properly detect
out of range args.
2015-05-21 Ed Schonberg <schonberg@adacore.com>
* sem_ch5.adb: add (useless) initial value.
* sem_ch3.adb (Replace_Anonymous_Access_To_Protected_Subprogram):
Check whether the procedure has parameters before processing
formals in ASIS mode.
Ed Schonberg [Thu, 21 May 2015 10:47:34 +0000 (10:47 +0000)]
sem_ch13.adb (Check_Iterator_Functions): Emit error on Iterator aspect as well when indexing function is illegal.
2015-05-21 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Check_Iterator_Functions): Emit error on Iterator
aspect as well when indexing function is illegal.
(Valid_Default_Iterator): Handle properly somme illegal cases
to prevent compilation abandoned messages.
(Check_Primitive_Function): Verify that type and indexing function
are in the same scope.
* freeze.adb (Freeze_Record): Extend patch on the presence of
indexing aspects to aspect Default_Iterator.
PR c/52952
* c-format.c (location_column_from_byte_offset): New.
(location_from_offset): New.
(struct format_wanted_type): Add offset_loc field.
(check_format_info): Move handling of location for extra arguments
closer to the point of warning.
(check_format_info_main): Pass the result of location_from_offset
to warning_at.
(format_type_warning): Pass the result of location_from_offset
to warning_at.
Thomas Schwinge [Thu, 21 May 2015 06:46:55 +0000 (08:46 +0200)]
genrecog: Address -Wsign-compare diagnostics.
g++-4.6 [...] [...]/gcc/genrecog.c
[...]/gcc/genrecog.c: In function 'state_size find_subroutines(routine_type, state*, vec<state*>&)':
[...]/gcc/genrecog.c:3338:35: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
[...]/gcc/genrecog.c:3347:37: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
[...]/gcc/genrecog.c:3359:29: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
[...]/gcc/genrecog.c:3365:32: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
François Dumont [Wed, 20 May 2015 19:44:25 +0000 (19:44 +0000)]
formatter.h (_GLIBCXX_TYPEID): New macro to simplify usage of typeid.
2015-05-20 François Dumont <fdumont@gcc.gnu.org>
* include/debug/formatter.h (_GLIBCXX_TYPEID): New macro to simplify
usage of typeid.
(_Error_formatter::_M_print_type): New.
* src/c++11/debug.cc
(_Error_formatter::_Parameter::_M_print_field): Use latter.
(_Error_formatter::_M_print_type): Implement latter using
__cxaabiv1::__cxa_demangle to print demangled type name.
Mikhail Maltsev [Wed, 20 May 2015 19:39:42 +0000 (19:39 +0000)]
Promote types of RTL expressions to more derived ones.
* bb-reorder.c (set_edge_can_fallthru_flag): Use rtx_jump_insn where
feasible.
(fix_up_fall_thru_edges): Likewise.
(fix_crossing_conditional_branches): Likewise. Promote jump targets
from to rtx_insn to rtx_code_label where feasible.
* bt-load.c (move_btr_def): Remove as-a cast of the value returned by
gen_move_insn (returned type changed to rtx_insn).
* builtins.c (expand_errno_check): Fix arguments of
do_compare_rtx_and_jump (now expects rtx_code_label).
(expand_builtin_acc_on_device): Likewise.
* cfgcleanup.c (try_simplify_condjump): Add cast when calling
invert_jump (now exprects rtx_jump_insn).
* cfgexpand.c (label_rtx_for_bb): Promote return type to rtx_code_label.
(construct_init_block): Use rtx_code_label.
* cfgrtl.c (block_label): Promote return type to rtx_code_label.
(try_redirect_by_replacing_jump): Use cast to rtx_jump_insn when
calling redirect_jump.
(patch_jump_insn): Likewise.
(redirect_branch_edge): Likewise.
(force_nonfallthru_and_redirect): Likewise.
(fixup_reorder_chain): Explicitly use rtx_jump_insn instead of rtx_insn
when suitable.
(rtl_lv_add_condition_to_bb): Update call of do_compare_rtx_and_jump.
* cfgrtl.h: Promote return type of block_label to rtx_code_label.
* config/bfin/bfin.c (hwloop_optimize): Fix call of emit_label_before.
* config/i386/i386.c (ix86_emit_cmove): Explicitly use rtx_code_label
to store the value retured by gen_label_rtx.
* config/mips/mips.c (mips16_split_long_branches): Promote rtx_insn to
rtx_jump_insn.
* config/sh/sh.c (gen_far_branch): Likewise. Fix call of invert_jump.
(split_branches): Fix calls of redirect_jump.
* dojump.c (jumpifnot): Promote argument type from rtx to rtx_code_label.
(jumpifnot_1): Likewise.
(jumpif): Likewise.
(jumpif_1): Likewise.
(do_jump_1): Likewise.
(do_jump): Likewise. Use rtx_code_label when feasible.
(do_jump_by_parts_greater_rtx): Likewise.
(do_jump_by_parts_zero_rtx): Likewise.
(do_jump_by_parts_equality_rtx): Likewise.
(do_compare_rtx_and_jump): Likewise.
* dojump.h: Update function prototypes.
* dse.c (emit_inc_dec_insn_before): Remove case (gen_move_insn now
returns rtx_insn).
* emit-rtl.c (emit_jump_insn_before_noloc): Promote return type to
rtx_jump_insn.
(emit_label_before): Likewise.
(emit_jump_insn_after_noloc): Likewise.
(emit_jump_insn_after_setloc): Likewise.
(emit_jump_insn_after): Likewise
(emit_jump_insn_before_setloc): Likewise.
(emit_jump_insn_before): Likewise.
(emit_label_before): Promote return type to rtx_code_label.
(emit_label): Likewise.
* except.c (sjlj_emit_dispatch_table): Use jump_target_rtx.
* explow.c (emit_stack_save): Use gen_move_insn_uncast instead of
gen_move_insn.
(emit_stack_restore): Likewise.
* expmed.c (emit_store_flag_force): Fix calls of do_compare_rtx_and_jump.
(do_cmp_and_jump): Likewise.
* expr.c (expand_expr_real_2): Likewise. Promote some local variables
from rtx to rtx_code_label.
(gen_move_insn_uncast): New function.
* expr.h: Update return type of gen_move_insn (promote to rtx_insn).
* function.c (convert_jumps_to_returns): Fix call of redirect_jump.
* gcse.c (pre_insert_copy_insn): Use rtx_insn instead of rtx.
* ifcvt.c (dead_or_predicable): Use rtx_jump_insn when calling
invert_jump_1 and redirect_jump_1.
* internal-fn.c (expand_arith_overflow_result_store): Fix call of
do_compare_rtx_and_jump.
(expand_addsub_overflow): Likewise.
(expand_neg_overflow): Likewise.
(expand_mul_overflow): Likewise.
* ira.c (split_live_ranges_for_shrink_wrap): Use rtx_insn for
return value of gen_move_insn.
* jump.c (redirect_jump): Promote argument from rtx to rtx_jump_insn.
* loop-doloop.c (add_test): Use rtx_code_label.
(doloop_modify): Likewise.
(doloop_optimize): Likewise.
* loop-unroll.c (compare_and_jump_seq): Promote rtx to rtx_code_label.
* lra-constraints.c (emit_spill_move): Remove cast of value returned
by gen_move_insn.
(inherit_reload_reg): Add cast when calling dump_insn_slim.
(split_reg): Likewise.
* modulo-sched.c (schedule_reg_moves): Remove cast of value returned by
gen_move_insn.
* optabs.c (expand_binop_directly): Remove casts of values returned by
maybe_gen_insn.
(expand_unop_direct): Likewise.
(expand_abs): Likewise.
(maybe_emit_unop_insn): Likewise.
(maybe_gen_insn): Promote return type to rtx_insn.
* optabs.h: Update prototype of maybe_gen_insn.
* postreload-gcse.c (eliminate_partially_redundant_load): Remove
redundant cast.
* recog.c (struct peep2_insn_data): Promote type of insn field to
rtx_insn.
(peep2_reinit_state): Use NULL instead of NULL_RTX.
(peep2_attempt): Remove casts of insn in peep2_insn_data.
(peep2_fill_buffer): Promote argument from rtx to rtx_insn
* recog.h (struct insn_gen_fn): Promote return types of function
pointers and operator ().from rtx to rtx_insn.
* reorg.c (fill_simple_delay_slots): Promote rtx_insn to rtx_jump_insn.
(fill_eager_delay_slots): Likewise.
(relax_delay_slots): Likewise.
(make_return_insns): Likewise.
(dbr_schedule): Likewise.
(optimize_skips): Likewise.
(reorg_redirect_jump): Likewise.
(fill_slots_from_thread): Likewise.
* reorg.h: Update prototypes.
* resource.c (find_dead_or_set_registers): Use dyn_cast to
rtx_jump_insn instead of check. Use it's jump_target method.
* rtl.h (rtx_jump_insn::jump_label): Define new method.
(rtx_jump_insn::jump_target): Define new method.
(rtx_jump_insn::set_jump_target): Define new method.
* rtlanal.c (tablejump_p): Promote type of one local variable.
* sched-deps.c (sched_analyze_2): Promote rtx to rtx_insn_list.
(sched_analyze_insn): Likewise.
* sched-vis.c (print_insn_with_notes): Promote rtx to rtx_insn.
(print_insn): Likewise.
* stmt.c (label_rtx): Promote return type to rtx_insn.
(force_label_rtx): Likewise.
(jump_target_rtx): Define new function.
(expand_label): Use it, get rid of one cast.
(expand_naked_return): Promote rtx to rtx_code_label.
(do_jump_if_equal): Fix do_compare_rtx_and_jump call.
(expand_case): Use rtx_code_label instread of rtx where feasible.
(expand_sjlj_dispatch_table): Likewise.
(emit_case_nodes): Likewise.
* stmt.h: Declare jump_target_rtx. Update prototypes. Fix comments.
* store-motion.c (insert_store): Make use of new return type of
gen_move_insn and remove a cast.
(replace_store_insn): Likewise.
Max Filippov [Wed, 20 May 2015 18:56:14 +0000 (18:56 +0000)]
re PR target/65730 (xtensa: ICE in libstdc++-v3/include/bits/atomic_base.h: In function ‘bool std::atomic_flag_test_and_set_explicit(std::__atomic_flag_base*, std::memory_order)’)
Fix PR target/65730
2015-05-20 Max Filippov <jcmvbkbc@gmail.com>
gcc/
* config/xtensa/xtensa.c (init_alignment_context): Replace MULT
by BITS_PER_UNIT with ASHIFT by exact_log2 (BITS_PER_UNIT).
PR fortran/65548
* trans-stmt.c (gfc_trans_allocate): Always retrieve the
descriptor or a reference to a source= expression for
arrays and non-arrays, respectively. Use a temporary
symbol and gfc_trans_assignment for all source=
assignments to allocated objects besides for class and
derived types.
David Malcolm [Wed, 20 May 2015 09:23:36 +0000 (09:23 +0000)]
Introduce TV_INITIALIZE_RTL
gcc/ChangeLog:
* timevar.def (TV_INITIALIZE_RTL): New.
* toplev.c (initialize_rtl): Use an auto_timevar to account this
function's time to TV_INITIALIZE_RTL.
David Malcolm [Wed, 20 May 2015 09:13:19 +0000 (09:13 +0000)]
libcpp: Eliminate most of the non-const/reference-returning inline fns
gcc/java/ChangeLog:
* jcf-parse.c (set_source_filename): Replace write through
ORDINARY_MAP_FILE_NAME with direct access to "to_file".
libcpp/ChangeLog:
* include/line-map.h (MAP_START_LOCATION): Eliminate the non-const
variant, and tweak comment for the const variant.
(ORDINARY_MAP_STARTING_LINE_NUMBER): Drop the non-const variant.
(ORDINARY_MAP_INCLUDER_FILE_INDEX): Likewise.
(ORDINARY_MAP_IN_SYSTEM_HEADER_P): Likewise.
(SET_ORDINARY_MAP_NUMBER_OF_COLUMN_BITS): Delete.
(ORDINARY_MAP_FILE_NAME): Drop the non-const variant.
(MACRO_MAP_MACRO): Likewise.
(MACRO_MAP_NUM_MACRO_TOKENS): Likewise.
(MACRO_MAP_LOCATIONS): Likewise.
(MACRO_MAP_EXPANSION_POINT_LOCATION): Likewise.
* line-map.c (linemap_add): Replace writes through macros with
direct field accesses.
(linemap_enter_macro): Likewise.
(linemap_line_start): Likewise.
Bin Cheng [Wed, 20 May 2015 05:15:56 +0000 (05:15 +0000)]
re PR tree-optimization/65447 (AArch64: iv-opt causes bad addressing)
PR tree-optimization/65447
* tree-ssa-loop-ivopts.c (struct iv_use): New fields.
(dump_use, dump_uses): Support to dump sub use.
(record_use): New parameters to support sub use. Remove call to
dump_use.
(record_sub_use, record_group_use): New functions.
(compute_max_addr_offset, split_all_small_groups): New functions.
(group_address_uses, rewrite_use_address): New functions.
(strip_offset): New declaration.
(find_interesting_uses_address): Call record_group_use.
(add_candidate): New assertion.
(infinite_cost_p): Move definition forward.
(add_costs): Check INFTY cost and return immediately.
(get_computation_cost_at): Clear setup cost and dependent bitmap
for sub uses.
(determine_use_iv_cost_address): Compute cost for sub uses.
(rewrite_use_address_1): Rename from old rewrite_use_address.
(free_loop_data): Free sub uses.
(tree_ssa_iv_optimize_loop): Call group_address_uses.
gcc/testsuite
PR tree-optimization/65447
* gcc.dg/tree-ssa/pr65447.c: New test.
Jeff Law [Wed, 20 May 2015 02:21:29 +0000 (20:21 -0600)]
tree-ssa-threadupdate.c (thread_single_edge): Use delete_jump_thread instead of open-coded version.
* tree-ssa-threadupdate.c (thread_single_edge): Use delete_jump_thread
instead of open-coded version. Also delete the jump thread created
within this function.
* config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Return
stack adjusting insn. Formatting.
(rs6000_emit_prologue): Track stack adjusting insn, and use of
r12. If possible, emit first -fsplit-stack arg pointer insn
before stack adjust. Don't use r12 to save cr if split-stack.
Alan Modra [Wed, 20 May 2015 01:17:45 +0000 (10:47 +0930)]
rs6000.c (struct rs6000_stack): Correct comments.
* config/rs6000/rs6000.c (struct rs6000_stack): Correct comments.
(rs6000_stack_info): Don't zero offsets when not saving registers.
(debug_stack_info): Adjust to omit printing unused offsets,
as before.
(direct_return): Test vrsave_size rather than vrsave_mask.
(rs6000_emit_prologue): Likewise. Remove redundant altivec tests.
(rs6000_emit_epilogue): Likewise.
Alan Modra [Wed, 20 May 2015 01:15:59 +0000 (10:45 +0930)]
rs6000.c (rs6000_stack_info): Don't zero offsets when not saving registers.
* config/rs6000/rs6000.c (rs6000_stack_info): Don't zero offsets
when not saving registers.
(debug_stack_info): Adjust to omit printing unused offsets,
as before.
(rs6000_emit_epilogue): Adjust use_backchain_to_restore_sp
expression.
Jonathan Wakely [Tue, 19 May 2015 22:24:50 +0000 (23:24 +0100)]
stl_list.h (_M_resize_pos(size_type&)): Declare.
* include/bits/stl_list.h (_M_resize_pos(size_type&)): Declare.
(operator==(const list&, const list&)): If size() is O(1) compare
sizes before comparing each element.
* include/bits/list.tcc (list::_M_resize_pos(size_type&)): Define.
(list::resize): Use _M_resize_pos.
Jan Hubicka [Tue, 19 May 2015 21:02:06 +0000 (23:02 +0200)]
ipa-devirt.c (type_in_anonymous_namespace_p): Return true or implicit declarations.
* ipa-devirt.c (type_in_anonymous_namespace_p): Return true
or implicit declarations.
(odr_type_p): Check that TYPE_NAME is TYPE_DECL before looking
into it.
(get_odr_type): Check type has linkage before adding bases.
(register_odr_type): Check that type has linkage before adding it.
(type_known_to_have_no_deriavations_p): Rename to ..
(type_known_to_have_no_derivations_p): This one.
* ipa-utils.h (type_known_to_have_no_deriavations_p): Rename to ..
(type_known_to_have_no_derivations_p): This one.
* ipa-polymorphic-call.c
(ipa_polymorphic_call_context::restrict_to_inner_type): Check that
type has linkage.
Andreas Krebbel [Tue, 19 May 2015 17:41:21 +0000 (17:41 +0000)]
S/390 Invalid vector binary ops
This is a first try to implement at least some of the requirements
regarding the vector bool type documented for IBM XLC.
With this patch error messages will be issued for invalid uses of
vector bool types in binary operators.
vector bool types are being marked opaque in order to prevent the
front-end from complaining about "vector bool long" vs "vector bool
long long" combinations on 64 bit. The opaque flag basically
suppresses any type checking. However, we still want vector bool to be
accepted only in contexts specified in the documentation (to be
published soon). Implementing the invalid binary op hook does this
for binary operators at least. But this is far from being complete :(
gcc/
* config/s390/s390.c (s390_vector_bool_type_p): New function.
(s390_invalid_binary_op): New function.
(TARGET_INVALID_BINARY_OP): Define macro.
David Sherwood [Tue, 19 May 2015 17:37:45 +0000 (17:37 +0000)]
loop-invariant.c (create_new_invariant): Don't calculate address cost if mode is not a scalar integer.
2015-05-19 David Sherwood <david.sherwood@arm.com>
* loop-invariant.c (create_new_invariant): Don't calculate address cost
if mode is not a scalar integer.
(get_inv_cost): Increase computational cost for unused invariants.
Andreas Krebbel [Tue, 19 May 2015 17:35:08 +0000 (17:35 +0000)]
S/390 Add zvector testcases.
gcc/testsuite/
* gcc.target/s390/zvector/vec-dbl-math-compile-1.c: New test.
* gcc.target/s390/zvector/vec-genbytemask-1.c: New test.
* gcc.target/s390/zvector/vec-genmask-1.c: New test.
* gcc.target/s390/zvector/vec-lcbb-1.c: New test.
* gcc.target/s390/zvector/vec-overloading-1.c: New test.
* gcc.target/s390/zvector/vec-overloading-2.c: New test.
* gcc.target/s390/zvector/vec-overloading-3.c: New test.
* gcc.target/s390/zvector/vec-overloading-4.c: New test.
* gcc.target/s390/zvector/vec-test-mask-1.c: New test.
* gcc.target/s390/zvector/vec-elem-1.c: New test.
Andreas Krebbel [Tue, 19 May 2015 17:32:07 +0000 (17:32 +0000)]
S/390 zvector builtin support.
With this patch GCC implements an Altivec style set of builtins to
make use of vector instructions in C/C++ code. This is provided for
compatibility with the IBM XL compiler.
gcc/
* config.gcc: Add vecintrin.h to extra_headers. Add s390-c.o to
c_target_objs and cxx_target_objs. Add t-s390 to tmake_file.
* config/s390/s390-builtin-types.def: New file.
* config/s390/s390-builtins.def: New file.
* config/s390/s390-builtins.h: New file.
* config/s390/s390-c.c: New file.
* config/s390/s390-modes.def: Add modes CCVEQANY, CCVH,
CCVHANY, CCVHU, CCVHUANY, CCVFHANY, CCVFHEANY.
* config/s390/s390-protos.h (s390_expand_vec_compare_cc)
(s390_cpu_cpp_builtins, s390_register_target_pragmas): Add
prototypes.
* config/s390/s390.c (s390-builtins.h, s390-builtins.def):
Include.
(flags_builtin, flags_overloaded_builtin_var, s390_builtin_types)
(s390_builtin_fn_types, s390_builtin_decls, code_for_builtin): New
variable definitions.
(s390_const_operand_ok): New function.
(s390_expand_builtin): Rewrite.
(s390_init_builtins): New function.
(s390_handle_vectorbool_attribute): New function.
(s390_attribute_table): Add s390_vector_bool attribute.
(s390_match_ccmode_set): Handle new cc modes CCVH, CCVHU.
(s390_branch_condition_mask): Generate masks for new modes.
(s390_expand_vec_compare_cc): New function.
(s390_mangle_type): Add mangling for vector bool types.
(enum s390_builtin): Remove.
(s390_atomic_assign_expand_fenv): Rename constants for sfpc and
efpc builtins.
* config/s390/s390.h (TARGET_CPU_CPP_BUILTINS): Call
s390_cpu_cpp_builtins.
(REGISTER_TARGET_PRAGMAS): New macro.
* config/s390/s390.md: Define more UNSPEC_VEC_* constants.
(insn_cmp mode attribute): Add new CC modes.
(s390_sfpc, s390_efpc): Rename patterns to sfpc and efpc.
(lcbb): New pattern definition.
* config/s390/s390intrin.h: Include vecintrin.h.
* config/s390/t-s390: New file.
* config/s390/vecintrin.h: New file.
* config/s390/vector.md: Include vx-builtins.md.
* config/s390/vx-builtins.md: New file.S/390 zvector builtin support.
Andreas Krebbel [Tue, 19 May 2015 17:30:25 +0000 (17:30 +0000)]
S/390 Add vector scalar instruction support.
With this patch GCC makes use of the vector instruction which are
available in single element mode. By using these instructions scalar
double operations can use 32 registers.
gcc/
* config/s390/s390-modes.def: Add new modes CCVEQ, CCVFH, and
CCVFHE.
* config/s390/s390.c (s390_match_ccmode_set): Handle new modes.
(s390_select_ccmode): Likewise.
(s390_canonicalize_comparison): Swap operands if necessary.
(s390_expand_vec_compare_scalar): Expand DFmode compare using
single element vector instructions.
(s390_emit_compare): Call s390_expand_vec_compare_scalar.
(s390_branch_condition_mask): Generate CC masks for the new modes.
* config/s390/s390.md (v0, vf, vd): New mode attributes.
(VFCMP, asm_fcmp, insn_cmp): New mode iterator and attributes.
(*vec_cmp<insn_cmp>df_cconly, *fixuns_truncdfdi2_z13)
(*fix_trunc<BFP:mode><GPR:mode>2_bfp, *floatunsdidf2_z13)
(*floatuns<GPR:mode><FP:mode>2, *extendsfdf2_z13)
(*extend<DSF:mode><BFP:mode>2): New insn definition.
(fix_trunc<BFP:mode><GPR:mode>2_bfp, loatuns<GPR:mode><FP:mode>2)
(extend<DSF:mode><BFP:mode>2): Turn into expander.
(floatdi<mode>2, truncdfsf2, add<mode>3, sub<mode>3, mul<mode>3)
(div<mode>3, *neg<mode>2, *abs<mode>2, *negabs<mode>2)
(sqrt<mode>2): Add vector instruction.
gcc/testsuite/
* gcc.target/s390/vector/vec-scalar-cmp-1.c: New test.
Andreas Krebbel [Tue, 19 May 2015 17:28:42 +0000 (17:28 +0000)]
S/390: Vector base support - testcases
gcc/testsuite/
* gcc.target/s390/s390.exp
(check_effective_target_vector): New check.
* gcc.target/s390/vector/vec-abi-1.c: New test.
* gcc.target/s390/vector/vec-abi-2.c: New test.
* gcc.target/s390/vector/vec-abi-3.c: New test.
* gcc.target/s390/vector/vec-abi-4.c: New test.
* gcc.target/s390/vector/vec-abi-align-1.c: New test.
* gcc.target/s390/vector/vec-abi-single-1.c: New test.
* gcc.target/s390/vector/vec-abi-single-2.c: New test.
* gcc.target/s390/vector/vec-abi-struct-1.c: New test.
* gcc.target/s390/vector/vec-abi-vararg-1.c: New test.
* gcc.target/s390/vector/vec-abi-vararg-2.c: New test.
* gcc.target/s390/vector/vec-clobber-1.c: New test.
* gcc.target/s390/vector/vec-cmp-1.c: New test.
* gcc.target/s390/vector/vec-cmp-2.c: New test.
* gcc.target/s390/vector/vec-dbl-math-compile-1.c: New test.
* gcc.target/s390/vector/vec-genbytemask-1.c: New test.
* gcc.target/s390/vector/vec-genbytemask-2.c: New test.
* gcc.target/s390/vector/vec-genmask-1.c: New test.
* gcc.target/s390/vector/vec-genmask-2.c: New test.
* gcc.target/s390/vector/vec-init-1.c: New test.
* gcc.target/s390/vector/vec-int-math-compile-1.c: New test.
* gcc.target/s390/vector/vec-shift-1.c: New test.
* gcc.target/s390/vector/vec-sub-1.c: New test.
Jakub Jelinek [Tue, 19 May 2015 16:16:15 +0000 (18:16 +0200)]
re PR middle-end/66199 (lastprivate/linear clause issues on combined constructs)
PR middle-end/66199
* tree.h (OMP_TEAMS_COMBINED): Define.
* gimplify.c (enum gimplify_omp_var_data): Add
GOVD_LINEAR_LASTPRIVATE_NO_OUTER.
(enum omp_region_type): Add ORT_COMBINED_TEAMS.
(omp_notice_variable): Accept both ORT_TEAMS
and ORT_COMBINED_TEAMS. Don't recurse if
GOVD_LINEAR_LASTPRIVATE_NO_OUTER is set and either
GOVD_LINEAR is set, or GOVD_LASTPRIVATE without
GOVD_FIRSTPRIVATE.
(omp_no_lastprivate): New function.
(gimplify_scan_omp_clauses): For OMP_CLAUSE_LASTPRIVATE
and OMP_CLAUSE_LINEAR, if omp_no_lastprivate, don't
notice_outer and set appropriate bits, otherwise make
sure default(none) combined constructs won't complain.
(gimplify_adjust_omp_clauses): Remove OMP_CLAUSE_LINEAR
outer special casing, for OMP_CLAUSE_LASTPRIVATE if
omp_no_lastprivate either remove the clause or turn it
into OMP_CLAUSE_PRIVATE.
(gimplify_omp_for): Fix up handling of implicit
lastprivate or linear iterators.
(gimplify_omp_workshare): For OMP_TEAMS_COMBINED use
ORT_COMBINED_TEAMS.
* omp-low.c (lower_omp_for_lastprivate): For combined
for simd use fd.loop.n2 from the for rather than simd.
gcc/c/
* c-parser.c (c_parser_omp_for_loop): Don't add
OMP_CLAUSE_SHARED to OMP_PARALLEL_CLAUSES when moving
OMP_CLAUSE_LASTPRIVATE clause to OMP_FOR_CLAUSES.
(c_parser_omp_teams): Set OMP_TEAMS_COMBINED for combined
constructs.
gcc/cp/
* parser.c (cp_parser_omp_for_loop): Don't add
OMP_CLAUSE_SHARED to OMP_PARALLEL_CLAUSES when moving
OMP_CLAUSE_LASTPRIVATE clause to OMP_FOR_CLAUSES.
(cp_parser_omp_teams): Set OMP_TEAMS_COMBINED for combined
constructs.
gcc/fortran/
* trans-openmp.c (gfc_trans_omp_teams): Set OMP_TEAMS_COMBINED for
combined constructs.
(gfc_trans_omp_target): Make sure BIND_EXPR has non-NULL
BIND_EXPR_BLOCK.
libgomp/
* testsuite/libgomp.c/pr66199-1.c: New test.
* testsuite/libgomp.c/pr66199-2.c: New test.
* testsuite/libgomp.c++/pr66199-1.C: New test.
* testsuite/libgomp.c++/pr66199-2.C: New test.
* testsuite/libgomp.fortran/pr66199-1.f90: New test.
* testsuite/libgomp.fortran/pr66199-2.f90: New test.
projects / thirdparty / gcc.git / log
