Remove global REG_SETs

We currently maintain global REG_SET versions of fixed_reg_set
and regs_invalidated_by_call.  With bitmap_view, we can instead
operate directly on the underlying HARD_REG_SETs, avoiding the
need to keep the two pieces of data in sync.

I have a series of patches that removes the assumption that there's
a single global ABI for all functions in the translation unit,
which includes not relying on having a global regs_invalidated_by_call.
Removing the REG_SET equivalent is one step to doing that.

Note that the affected DF code is used for EH edges or dumping only,
so shouldn't be performance critical.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* regset.h (regs_invalidated_by_call_regset): Delete.
(fixed_reg_set_regset): Likewise.
* reginfo.c (regs_invalidated_by_call_regset): Likewise.
(fixed_reg_set_regset, persistent_obstack): Likewise.
(init_reg_sets_1, globalize_reg): Update accordingly.
* df.h (df_print_regset, df_print_word_regset): Take a const_bitmap
instead of a bitmap.
* df-core.c (df_print_regset, df_print_word_regset): Likewise.
* df-problems.c (df_rd_local_compute): Use regs_invalidated_by_call
instead of regs_invalidated_by_call_regset.
(df_lr_confluence_n, df_md_confluence_n): Likewise.
* df-scan.c (df_scan_start_dump): Likewise.
* dse.c (copy_fixed_regs): Likewise.
* config/sh/sh.c (sh_find_equiv_gbr_addr): Likewise.

From-SVN: r275537

Add "fast" conversions from arrays to bitmaps

This patch adds a bitmap_view<X> class that creates a read-only,
on-stack bitmap representation of an array-like object X.  The main
use case is to allow HARD_REG_SETs to be used in REG_SET (i.e. bitmap)
operations.

For now it only handles constant-sized arrays, but I've tried to
define the types in a way that could handle variable-sized arrays
in future (although less efficiently).  E.g. this might be useful
for combining bitmaps and sbitmaps.

For the read-only view to work as intended, I needed to make
bitmap_bit_p take a const_bitmap instead of a bitmap.  Logically
the bitmap really is read-only, but we update the "current" and
"indx" fields of the bitmap_head after doing a search.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* array-traits.h: New file.
* coretypes.h (array_traits, bitmap_view): New types.
* bitmap.h: Include "array-traits.h"
(bitmap_bit_p): Take a const_bitmap instead of a bitmap.
(base_bitmap_view, bitmap_view): New classes.
* bitmap.c (bitmap_bit_p): Take a const_bitmap instead of a bitmap.
* hard-reg-set.h: Include array-traits.h.
(array_traits<HARD_REG_SET>): New struct.
* regset.h (IOR_REG_SET_HRS): New macro.
* loop-iv.c (simplify_using_initial_values): Use IOR_REG_SET_HRS
rather than iterating over each hard register.
* sched-deps.c (sched_analyze_insn): Likewise.
* sel-sched-ir.c (setup_id_implicit_regs): Likewise.

From-SVN: r275536

Tweak interface to ira-build.c:ior_hard_reg_conflicts

This patch makes ior_hard_reg_conflicts take a const_hard_reg_set
rather than a pointer, so that it can be passed a temporary object
in later patches.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* ira-int.h (ior_hard_reg_conflicts): Take a const_hard_reg_set
instead of a HARD_REG_SET *.
* ira-build.c (ior_hard_reg_conflicts): Likewise.
(ira_build): Update call accordingly.
* ira-emit.c (add_range_and_copies_from_move_list): Likewise.

From-SVN: r275535

Remove hard_reg_set_equal_p

Use "x == y" instead of "hard_reg_set_equal_p (x, y)".

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (HARD_REG_SET::operator==): New function.
(HARD_REG_SET::operator!=): Likewise.
(hard_reg_set_equal_p): Delete.
* cfgcleanup.c (old_insns_match_p): Use == instead of
hard_reg_set_equal_p and != instead of !hard_reg_set_equal_p.
* ira-color.c (allocno_hard_regs_hasher::equal): Likewise.
(add_allocno_hard_regs_to_forest): Likewise.
(setup_allocno_available_regs_num): Likewise.
* ira.c (setup_pressure_classes): Likewise.
(setup_allocno_and_important_classes): Likewise.
(setup_reg_class_relations): Likewise.
* lra-lives.c (process_bb_lives): Likewise.
* reg-stack.c (change_stack, convert_regs_1): Likewise.

From-SVN: r275534

Remove IOR_COMPL_HARD_REG_SET

Use "x |= ~y" instead of "IOR_COMPL_HARD_REG_SET (x, y)", or just
"x | ~y" if the result is a temporary.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (IOR_COMPL_HARD_REG_SET): Delete.
* config/aarch64/cortex-a57-fma-steering.c (rename_single_chain):
Use "|~" instead of IOR_COMPL_HARD_REG_SET.
* config/aarch64/falkor-tag-collision-avoidance.c (init_unavailable):
Likewise.
* ira-build.c (ira_create_object, ira_set_allocno_class): Likewise.
* ira.c (setup_reg_renumber): Likewise.
* lra-assigns.c (find_hard_regno_for_1): Likewise.
* regrename.c (regrename_find_superclass): Likewise.
* reload1.c (find_reg): Likewise.

From-SVN: r275533

Remove AND_COMPL_HARD_REG_SET

Use "x &= ~y" instead of "AND_COMPL_HARD_REG_SET (x, y)", or just
"x & ~y" if the result is a temporary.  This means that we're splitting
it into two operations, but the compiler should be able to combine them
for reasonable values of FIRST_PSEUDO_REGISTER.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (AND_COMPL_HARD_REG_SET): Delete.
* caller-save.c (setup_save_areas): Use "&~" instead of
AND_COMPL_HARD_REG_SET.
(save_call_clobbered_regs): Likewise.
* config/epiphany/epiphany.c (epiphany_conditional_register_usage):
Likewise.
* config/frv/frv.c (frv_ifcvt_modify_tests): Likewise.
* config/gcn/gcn.c (gcn_md_reorg): Likewise.
* config/i386/i386.c (ix86_conditional_register_usage): Likewise.
* config/mips/mips.c (mips_class_max_nregs): Likewise.
(mips_conditional_register_usage): Likewise.
* config/sh/sh.c (output_stack_adjust): Likewise.
* ira-color.c (form_allocno_hard_regs_nodes_forest): Likewise.
(setup_profitable_hard_regs): Likewise.
(get_conflict_and_start_profitable_regs): Likewise.
* ira-conflicts.c (print_allocno_conflicts): Likewise.
(ira_build_conflicts): Likewise.
* ira-costs.c (restrict_cost_classes): Likewise.
(setup_regno_cost_classes_by_aclass): Likewise.
* ira-lives.c (process_bb_node_lives): Likewise.
* ira.c (setup_class_hard_regs, setup_reg_subclasses): Likewise.
(setup_class_subset_and_memory_move_costs, setup_pressure_classes)
(setup_allocno_and_important_classes, setup_class_translate_array)
(setup_reg_class_relations, setup_prohibited_class_mode_regs):
Likewise.
* lra-assigns.c (find_hard_regno_for_1): Likewise.
* lra-constraints.c (prohibited_class_reg_set_mode_p): Likewise.
(process_alt_operands, inherit_in_ebb): Likewise.
* lra-eliminations.c (update_reg_eliminate): Likewise.
* lra-lives.c (process_bb_lives): Likewise.
* reload1.c (update_eliminables_and_spill, reload_as_needed): Likewise.
* resource.c (find_dead_or_set_registers): Likewise.
(mark_target_live_regs): Likewise.
* sched-deps.c (get_implicit_reg_pending_clobbers): Likewise.
* sel-sched.c (mark_unavailable_hard_regs): Likewise.
(implicit_clobber_conflict_p): Likewise.
* shrink-wrap.c (requires_stack_frame_p): Likewise.
(try_shrink_wrapping): Likewise.

From-SVN: r275532

Remove IOR_HARD_REG_SET

Use "x |= y" instead of "IOR_HARD_REG_SET (x, y)" (or just "x | y"
if the result is a temporary).

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (HARD_REG_SET::operator|): New function.
(HARD_REG_SET::operator|=): Likewise.
(IOR_HARD_REG_SET): Delete.
* config/gcn/gcn.c (gcn_md_reorg): Use "|" instead of
IOR_HARD_REG_SET.
* config/m32c/m32c.c (m32c_register_move_cost): Likewise.
* config/s390/s390.c (s390_adjust_loop_scan_osc): Likewise.
* final.c (collect_fn_hard_reg_usage): Likewise.
* hw-doloop.c (scan_loop, optimize_loop): Likewise.
* ira-build.c (merge_hard_reg_conflicts): Likewise.
(ior_hard_reg_conflicts, create_cap_allocno, propagate_allocno_info)
(propagate_some_info_from_allocno): Likewise.
(copy_info_to_removed_store_destinations): Likewise.
* ira-color.c (add_allocno_hard_regs_to_forest, assign_hard_reg)
(allocno_reload_assign, ira_reassign_pseudos): Likewise.
(fast_allocation): Likewise.
* ira-conflicts.c (ira_build_conflicts): Likewise.
* ira-lives.c (make_object_dead, process_single_reg_class_operands)
(process_bb_node_lives): Likewise.
* ira.c (setup_pressure_classes, setup_reg_class_relations): Likewise.
* lra-assigns.c (find_hard_regno_for_1): Likewise.
(setup_live_pseudos_and_spill_after_risky_transforms): Likewise.
* lra-constraints.c (process_alt_operands, inherit_in_ebb): Likewise.
* lra-eliminations.c (spill_pseudos, update_reg_eliminate): Likewise.
* lra-lives.c (mark_pseudo_dead, check_pseudos_live_through_calls)
(process_bb_lives): Likewise.
* lra-spills.c (assign_spill_hard_regs): Likewise.
* postreload.c (reload_combine): Likewise.
* reginfo.c (init_reg_sets_1): Likewise.
* regrename.c (merge_overlapping_regs, find_rename_reg)
(merge_chains): Likewise.
* reload1.c (maybe_fix_stack_asms, order_regs_for_reload, find_reg)
(find_reload_regs, finish_spills, choose_reload_regs_init)
(emit_reload_insns): Likewise.
* reorg.c (redundant_insn): Likewise.
* resource.c (find_dead_or_set_registers, mark_set_resources)
(mark_target_live_regs): Likewise.
* rtlanal.c (find_all_hard_reg_sets): Likewise.
* sched-deps.c (sched_analyze_insn): Likewise.
* sel-sched.c (mark_unavailable_hard_regs): Likewise.
(find_best_reg_for_expr): Likewise.
* shrink-wrap.c (try_shrink_wrapping): Likewise.

From-SVN: r275531

Remove AND_HARD_REG_SET

Use "x &= y" instead of "AND_HARD_REG_SET (x, y)" (or just "x & y"
if the result is a temporary).

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (HARD_REG_SET::operator&): New function.
(HARD_REG_SET::operator&): Likewise.
(AND_HARD_REG_SET): Delete.
* caller-save.c (setup_save_areas): Use "&" instead of
AND_HARD_REG_SET.
(save_call_clobbered_regs): Likewise.
* config/gcn/gcn.c (gcn_md_reorg): Likewise.
* config/m32c/m32c.c (reduce_class): Likewise.
* config/rs6000/rs6000.c (rs6000_register_move_cost): Likewise.
* final.c (get_call_reg_set_usage): Likewise.
* ira-color.c (add_allocno_hard_regs_to_forest): Likewise.
(setup_left_conflict_sizes_p): Likewise.
* ira-conflicts.c (print_allocno_conflicts): Likewise.
(ira_build_conflicts): Likewise.
* ira-costs.c (restrict_cost_classes): Likewise.
* ira.c (setup_stack_reg_pressure_class, setup_class_translate_array)
(setup_reg_class_relations): Likewise.
* reginfo.c (init_reg_sets_1, record_subregs_of_mode): Likewise.
* reload1.c (maybe_fix_stack_asms, finish_spills): Likewise.
* resource.c (find_dead_or_set_registers): Likewise.
* sel-sched.c (mark_unavailable_hard_regs): Likewise.

From-SVN: r275530

Remove COMPL_HARD_REG_SET

"COMPL_HARD_REG_SET (x, y)" becomes "x = ~y".

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (HARD_REG_SET::operator~): New function.
(COMPL_HARD_REG_SET): Delete.
* config/c6x/c6x.c (c6x_call_saved_register_used): Use ~ instead
of COMPL_HARD_REG_SET.
(try_rename_operands): Likewise.
* config/sh/sh.c (push_regs): Likewise.
* lra-assigns.c (find_hard_regno_for_1): Likewise.
* lra-constraints.c (contains_reg_p): Likewise.
* reload1.c (finish_spills, choose_reload_regs_init): Likewise.

From-SVN: r275529

Remove COPY_HARD_REG_SET

This patch replaces "COPY_HARD_REG_SET (x, y)" with "x = y".

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (COPY_HARD_REG_SET): Delete.
* caller-save.c (save_call_clobbered_regs): Use assignment instead
of COPY_HARD_REG_SET.
* config/epiphany/epiphany.c (epiphany_compute_frame_size): Likewise.
(epiphany_conditional_register_usage): Likewise.
* config/frv/frv.c (frv_ifcvt_modify_tests): Likewise.
* config/gcn/gcn.c (gcn_md_reorg): Likewise.
* config/ia64/ia64.c (ia64_compute_frame_size): Likewise.
* config/m32c/m32c.c (m32c_register_move_cost): Likewise.
* config/m68k/m68k.c (m68k_conditional_register_usage): Likewise.
* config/mips/mips.c (mips_class_max_nregs): Likewise.
* config/pdp11/pdp11.c (pdp11_conditional_register_usage): Likewise.
* config/rs6000/rs6000.c (rs6000_register_move_cost): Likewise.
* config/sh/sh.c (output_stack_adjust): Likewise.
* final.c (collect_fn_hard_reg_usage): Likewise.
(get_call_reg_set_usage): Likewise.
* ira-build.c (ira_create_object, remove_low_level_allocnos)
(ira_flattening): Likewise.
* ira-color.c (add_allocno_hard_regs, add_allocno_hard_regs_to_forest)
(setup_left_conflict_sizes_p, setup_profitable_hard_regs)
(get_conflict_and_start_profitable_regs, allocno_reload_assign)
(ira_reassign_pseudos): Likewise.
* ira-conflicts.c (print_allocno_conflicts): Likewise.
(ira_build_conflicts): Likewise.
* ira-costs.c (restrict_cost_classes): Likewise.
(setup_regno_cost_classes_by_aclass): Likewise.
* ira.c (setup_class_hard_regs, setup_alloc_regs): Likewise.
(setup_reg_subclasses, setup_class_subset_and_memory_move_costs)
(setup_stack_reg_pressure_class, setup_pressure_classes)
(setup_allocno_and_important_classes, setup_class_translate_array)
(setup_reg_class_relations, setup_prohibited_class_mode_regs)
(ira_setup_eliminable_regset): Likewise.
* lra-assigns.c (find_hard_regno_for_1): Likewise.
(setup_live_pseudos_and_spill_after_risky_transforms): Likewise.
* lra-constraints.c (prohibited_class_reg_set_mode_p): Likewise.
(process_alt_operands, inherit_in_ebb): Likewise.
* lra-lives.c (process_bb_lives): Likewise.
* lra-spills.c (assign_spill_hard_regs): Likewise.
* lra.c (lra): Likewise.
* mode-switching.c (new_seginfo): Likewise.
* postreload.c (reload_combine): Likewise.
* reg-stack.c (straighten_stack): Likewise.
* reginfo.c (save_register_info, restore_register_info): Likewise.
(init_reg_sets_1, record_subregs_of_mode): Likewise
* regrename.c (create_new_chain, rename_chains): Likewise.
* reload1.c (order_regs_for_reload, find_reg): Likewise.
(find_reload_regs): Likewise.
* resource.c (find_dead_or_set_registers): Likewise.
(mark_target_live_regs): Likewise.
* sel-sched.c (mark_unavailable_hard_regs): Likewise.

From-SVN: r275528

Make note_stores take an rtx_insn

I have a series of patches that (as a side effect) makes all rtl
passes use the information collected by -fipa-ra.  This showed up a
latent bug in the liveness tracking in regrename.c, which doesn't take
CALL_INSN_FUNCTION_USAGE into account when processing clobbers.

This actually seems to be quite a common problem with passes that use
note_stores; only a handful remember to walk CALL_INSN_FUNCTION_USAGE
too.  I think it was just luck that I saw it with regrename first.

This patch tries to make things more robust by passing an insn rather
than a pattern to note_stores.  The old function is still available
as note_pattern_stores for the few places that need it.

When updating callers, I've erred on the side of using note_stores
rather than note_pattern_stores, because IMO note_stores should be
the default choice and we should only use note_pattern_stores if
there's a specific reason.  Specifically:

* For cselib.c, "body" may be a COND_EXEC_CODE instead of the main
  insn pattern.

* For ira.c, I wasn't sure whether extending no_equiv to
  CALL_INSN_FUNCTION_USAGE really made sense, since we don't do that
  for normal call-clobbered registers.  Same for mark_not_eliminable
  in reload1.c

* Some other places only have a pattern available, and since those
  places wouldn't benefit from walking CALL_INSN_FUNCTION_USAGE,
  it seemed better to alter the code as little as possible.

* In the config/ changes, quite a few callers have already weeded
  out CALL insns.  It still seemed better to use note_stores rather
  than prematurely optimise.  (note_stores should tail call to
  note_pattern_stores once it sees that the insn isn't a call.)

The patch also documents what SETs mean in CALL_INSN_FUNCTION_USAGE.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* rtl.h (CALL_INSN_FUNCTION_USAGE): Document what SETs mean.
(note_pattern_stores): Declare.
(note_stores): Take an rtx_insn *.
* rtlanal.c (set_of): Use note_pattern_stores instead of note_stores.
(find_all_hard_reg_sets): Pass the insn rather than its pattern to
note_stores.  Remove explicit handling of CALL_INSN_FUNCTION_USAGE.
(note_stores): Take an rtx_insn * as argument and process
CALL_INSN_FUNCTION_USAGE.  Rename old function to...
(note_pattern_stores): ...this.
(find_first_parameter_load): Pass the insn rather than
its pattern to note_stores.
* alias.c (memory_modified_in_insn_p, init_alias_analysis): Likewise.
* caller-save.c (setup_save_areas, save_call_clobbered_regs)
(insert_one_insn): Likewise.
* combine.c (combine_instructions): Likewise.
(likely_spilled_retval_p): Likewise.
(try_combine): Use note_pattern_stores instead of note_stores.
(record_dead_and_set_regs): Pass the insn rather than its pattern
to note_stores.
(reg_dead_at_p): Likewise.
* config/bfin/bfin.c (workaround_speculation): Likewise.
* config/c6x/c6x.c (maybe_clobber_cond): Likewise.  Take an rtx_insn *
rather than an rtx.
* config/frv/frv.c (frv_registers_update): Use note_pattern_stores
instead of note_stores.
(frv_optimize_membar_local): Pass the insn rather than its pattern
to note_stores.
* config/gcn/gcn.c (gcn_md_reorg): Likewise.
* config/i386/i386.c (ix86_avx_u128_mode_after): Likewise.
* config/mips/mips.c (vr4130_true_reg_dependence_p): Likewise.
(r10k_needs_protection_p, mips_sim_issue_insn): Likewise.
(mips_reorg_process_insns): Likewise.
* config/s390/s390.c (s390_regs_ever_clobbered): Likewise.
* config/sh/sh.c (flow_dependent_p): Likewise.  Take rtx_insn *s
rather than rtxes.
* cse.c (delete_trivially_dead_insns): Pass the insn rather than
its pattern to note_stores.
* cselib.c (cselib_record_sets): Use note_pattern_stores instead
of note_stores.
* dce.c (mark_nonreg_stores): Remove the "body" parameter and pass
the insn to note_stores.
(prescan_insns_for_dce): Update call accordingly.
* ddg.c (mem_write_insn_p): Pass the insn rather than its pattern
to note_stores.
* df-problems.c (can_move_insns_across): Likewise.
* dse.c (emit_inc_dec_insn_before, replace_read): Likewise.
* function.c (assign_parm_setup_reg): Likewise.
* gcse-common.c (record_last_mem_set_info_common): Likewise.
* gcse.c (load_killed_in_block_p, compute_hash_table_work): Likewise.
(single_set_gcse): Likewise.
* ira.c (validate_equiv_mem): Likewise.
(update_equiv_regs): Use note_pattern_stores rather than note_stores
for no_equiv.
* loop-doloop.c (doloop_optimize): Pass the insn rather than its
pattern to note_stores.
* loop-invariant.c (calculate_loop_reg_pressure): Likewise.
* loop-iv.c (simplify_using_initial_values): Likewise.
* mode-switching.c (optimize_mode_switching): Likewise.
* optabs.c (emit_libcall_block_1): Likewise.
(expand_atomic_compare_and_swap): Likewise.
* postreload-gcse.c (load_killed_in_block_p): Likewise.
(record_opr_changes): Likewise.  Remove explicit handling of
CALL_INSN_FUNCTION_USAGE.
* postreload.c (reload_combine, reload_cse_move2add): Likewise.
* regcprop.c (kill_clobbered_values): Likewise.
(copyprop_hardreg_forward_1): Pass the insn rather than its pattern
to note_stores.
* regrename.c (build_def_use): Likewise.
* reload1.c (reload):  Use note_pattern_stores instead of note_stores
for mark_not_eliminable.
(reload_as_needed): Pass the insn rather than its pattern
to note_stores.
(emit_output_reload_insns): Likewise.
* resource.c (mark_target_live_regs): Likewise.
* sched-deps.c (init_insn_reg_pressure_info): Likewise.
* sched-rgn.c (sets_likely_spilled): Use note_pattern_stores
instead of note_stores.
* shrink-wrap.c (try_shrink_wrapping): Pass the insn rather than
its pattern to note_stores.
* stack-ptr-mod.c (pass_stack_ptr_mod::execute): Likewise.
* var-tracking.c (adjust_insn, add_with_sets): Likewise.

From-SVN: r275527

Simplify the implementation of HARD_REG_SET

We have two styles of HARD_REG_SET: a single integer based on
HOST_WIDEST_FAST_INT (used when FIRST_PSEUDO_REGISTER is small enough)
or an array of integers.  One of the nice properties of this arrangement
is that:

  void foo (const HARD_REG_SET);

is passed by value as an integer when the set is small enough and
by reference otherwise.

(This is in constrast to "const HARD_REG_SET &", which would always
be passed by reference, and in contrast to passing a structure wrapper
like "struct s { T elts[1]; }" by value, where the structure might be
passed like a T or by reference, depending on the ABI.)

However, one of the disadvantages of using an array is that simple
assignment doesn't work.  We need to use COPY_HARD_REG_SET instead.

This patch uses a structure wrapper around the array, and preserves
the above "nice property" using a new const_hard_reg_set typedef.
The patch also removes the manual unrolling for small array sizes;
I think these days we can rely on the compiler to do that for us.

This meant fixing two port-specific quirks:

- epiphany passed NULL as a HARD_REG_SET whose value doesn't matter.
  The patch passes the NO_REGS set instead.

- ia64 reused TEST_HARD_REG_BIT and SET_HARD_REG_BIT for arrays that
  are bigger than HARD_REG_SET.  The patch just open-codes them.

The patch is probably being too conservative.  Very few places actually
take advantage of the "nice property" above, and we could have a
cleaner interface if we used a structure wrapper for all cases.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* hard-reg-set.h (HARD_REG_SET): Define using a typedef rather
than a #define.  Use a structure rather than an array as the
fallback definition.  Remove special cases for low array sizes.
(const_hard_reg_set): New typedef.
(hard_reg_set_subset_p): Use it instead of "const HARD_REG_SET".
(hard_reg_set_equal_p, hard_reg_set_intersect_p): Likewise.
(hard_reg_set_empty_p): Likewise.
(SET_HARD_REG_BIT): Use a function rather than a macro to
handle the case in which HARD_REG_SET is a structure.
(CLEAR_HARD_REG_BIT, TEST_HARD_REG_BIT, CLEAR_HARD_REG_SET)
(SET_HARD_REG_SET, COPY_HARD_REG_SET, COMPL_HARD_REG_SET)
(AND_HARD_REG_SET, AND_COMPL_HARD_REG_SET, IOR_HARD_REG_SET)
(IOR_COMPL_HARD_REG_SET): Likewise.
(hard_reg_set_iterator::pset): Constify the pointer target.
(hard_reg_set_iter_init): Take a const_hard_reg_set instead
of a "const HARD_REG_SET".  Update the handling of non-integer
HARD_REG_SETs.
* recog.h: Test HARD_CONST instead of CLEAR_HARD_REG_SET.
* reload.h: Likewise.
* rtl.h (choose_hard_reg_mode): Remove unnecessary line break.
* regs.h (in_hard_reg_set_p): Take a const_hard_reg_set instead
of a "const HARD_REG_SET".
(overlaps_hard_reg_set_p, range_overlaps_hard_reg_set_p): Likewise.
(range_in_hard_reg_set_p): Likewise.
* ira-costs.c (restrict_cost_classes): Likewise.
* shrink-wrap.c (move_insn_for_shrink_wrap): Likewise.
* config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes::execute):
Pass a NO_REGS HARD_REG_SET rather than NULL to emit_set_fp_mode.
* config/ia64/ia64.c (rws_insn): In the CHECKING_P version,
use unsigned HOST_WIDEST_FAST_INT rather than HARD_REG_ELT_TYPE.
(rws_insn_set, rws_insn_test): In the CHECKING_P version,
take an unsigned int and open-code the HARD_REG_SET operations.

From-SVN: r275526

rs6000: Update rlwinm-[012].c

gcc/testsuite/
* gcc.target/powerpc/rlwinm-0.c: Adjust expected instruction counts.
* gcc.target/powerpc/rlwinm-1.c: Ditto.
* gcc.target/powerpc/rlwinm-2.c: Ditto.

From-SVN: r275525

Forgot Changelog from latest commit.

From-SVN: r275524

Update docs for recent <span> and constexpr lib changes.

2019-09-09  Edward Smith-Rowland  <3dw4rd@verizon.net>

Update docs for recent <span> and constexpr lib changes.
* doc/xml/manual/status_cxx2020.xml: Update p0202r3, p0858r0, p0879r0,
p1023r0, p1085r2 status.
* doc/html/manual/status.html: Regenerate.

From-SVN: r275522

Remove bt-load.c

bt-load.c has AFAIK been dead code since the removal of the SH5 port
in 2016.  I have a patch series that would need to update the liveness
tracking in a nontrivial way, so it seemed better to remove the pass
rather than install an untested and probably bogus change.

2019-09-09  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* Makefile.in (OBJS): Remove bt-load.o.
* doc/invoke.texi (fbranch-target-load-optimize): Delete.
(fbranch-target-load-optimize2, fbtr-bb-exclusive): Likewise.
* common.opt (fbranch-target-load-optimize): Mark as Ignore and
document that the option no longer does anything.
(fbranch-target-load-optimize2, fbtr-bb-exclusive): Likewise.
* target.def (branch_target_register_class): Delete.
(branch_target_register_callee_saved): Likewise.
* doc/tm.texi.in (TARGET_BRANCH_TARGET_REGISTER_CLASS): Likewise.
(TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED): Likewise.
* doc/tm.texi: Regenerate.
* tree-pass.h (make_pass_branch_target_load_optimize1): Delete.
(make_pass_branch_target_load_optimize2): Likewise.
* passes.def (pass_branch_target_load_optimize1): Likewise.
(pass_branch_target_load_optimize2): Likewise.
* targhooks.h (default_branch_target_register_class): Likewise.
* targhooks.c (default_branch_target_register_class): Likewise.
* opt-suggestions.c (test_completion_valid_options): Remove
-fbtr-bb-exclusive from the list of test options.
* bt-load.c: Remove.

From-SVN: r275521

match.pd: Add flag_unsafe_math_optimizations check before deciding on the widest type in...

2019-09-09  Barnaby Wilks  <barnaby.wilks@arm.com>

* match.pd: Add flag_unsafe_math_optimizations check
before deciding on the widest type in a binary math operation.

* gcc.dg/fold-binary-math-casts.c: New test.

From-SVN: r275518

Update comment of removed options.

2019-09-09  Martin Liska  <mliska@suse.cz>

* config/i386/i386.opt: Update comment of removed
options that are preserved only for backward
compatibility.
2019-09-09  Martin Liska  <mliska@suse.cz>

* c.opt: Update comment of removed
options that are preserved only for backward
compatibility.

From-SVN: r275517

Minor std::to_chars optimisation for base 10

__to_chars_10_impl is quite fast. According to the IACA the main loop
takes only 6.0 cycles, the whole function with one iteration takes
10.0 cycles. Replacing the __first[pos] and __first[pos - 1] with
__first[0] and __first[1] drops the function time to 7.53 cycles.

2019-09-09  Antony Polukhin  <antoshkka@gmail.com>

* include/bits/charconv.h (__detail::__to_chars_10_impl): Replace
final offsets with constants.

From-SVN: r275514

Improve constraints for std::span constructors

This patch simplifies the constraints on the constructors from arrays by
removing the redundant checks that element_type and value_type are
convertible to element_type. The incorrect uses of __adl_data in those
constructors are removed as well (they should use std::data not
std::ranges::data, and the former doesn't use ADL).

The range/container constructors are now constrained to exclude all
specializations of std::span, not just the current instantiation. The
range constructor now also checks s subset of the contiguous_range
requirements.

All relevant constructor constraints now use the _Require helper in
order to short circuit and avoid unnecessary instantiations after the
first failed constraint.

A new constructor supports initialization from different specializations
of std::span<OtherType, OtherExtent>, as specified in the C++20 draft.

* include/bits/range_access.h (__adl_to_address): Remove.
* include/std/span (__is_base_derived_safe_convertible_v): Replace
with span::__is_compatible.
(__is_std_array_v): Replace with __is_std_array class template and
partial specializations.
(__is_std_array, __is_std_span): New class templates and partial
specializations.
(span::__is_compatible): New alias template for SFINAE constraints.
(span::span(element_type (&)[N])): Remove redundant constraints. Do
not use __adl_data to obtain a pointer.
(span::span(array<value_type, N>&)): Likewise.
(span::span(const array<value_type, N>&)): Likewise.
[_GLIBCXX_P1394] (span::iter_reference_t, span::iterator_t)
(span::iter_value_t, span::derived_from): New alias templates for
SFINAE constraints, until the equivalents are supported in <concepts>
and <iterator>.
[_GLIBCXX_P1394] (span::__is_compatible_iterator): New alias template
for SFINAE constraints.
[_GLIBCXX_P1394] (span::is_compatible_range): New class template for
SFINAE constraints.
[_GLIBCXX_P1394] (span::span(Range&&)): Improve constraints.
[_GLIBCXX_P1394] (span::span(ContiguousIterator, Sentinel)): Likewise.
Use std::to_address instead of __adl_to_address.
[_GLIBCXX_P1394] (span::span(ContiguousIterator, size_type)): Likewise.
[!_GLIBCXX_P1394] (span::__is_compatible_container): New alias
template for SFINAE constraints.
[!_GLIBCXX_P1394] (span::span(Container&))
(span::span(const Container&)): Improve constraints.
[!_GLIBCXX_P1394] (span::span(pointer, size_type))
(span::span(pointer, pointer)): Remove redundant cast of pointer.
(span(const span<OType, OExtent>&)): New constructor.

From-SVN: r275513

re PR target/87853 (_mm_cmpgt_epi8 broken with -funsigned-char)

PR target/87853
* config/i386/emmintrin.h (_mm_cmpeq_epi8): Use casts to __v16qi
instead of __v16qs.

* gcc.target/i386/pr87853.c: New test.

From-SVN: r275508

re PR target/91704 ([X86] Codegen for _mm256_cmpgt_epi8 is affected by -funsigned-char)

PR target/91704
* config/i386/avxintrin.h (__v32qs): New typedef.
* config/i386/avx2intrin.h (_mm256_cmpgt_epi8): Use casts to __v32qs
instead of __v32qi.

* gcc.target/i386/pr91704.c: New test.

From-SVN: r275507

GCC port for eBPF

This patch series introduces a port of GCC to eBPF, which is a virtual
machine that resides in the Linux kernel.  Initially intended for
user-level packet capture and filtering, eBPF is nowadays generalized
to serve as a general-purpose infrastructure also for non-networking
purposes.

The binutils support is already upstream.  See
https://sourceware.org/ml/binutils/2019-05/msg00306.html.

ChangeLog:

* MAINTAINERS: Add myself as the maintainer of the eBPF port.
Remove myself from Write After Approval section.
* configure.ac: Support for bpf-*-* targets.
* configure: Regenerate.

contrib/ChangeLog:

* config-list.mk (LIST): Disable go in bpf-*-* targets.

gcc/ChangeLog:

* doc/invoke.texi (Option Summary): Cover eBPF.
(eBPF Options): New section.
* doc/extend.texi (BPF Built-in Functions): Likewise.
(BPF Kernel Helpers): Likewise.
* config.gcc: Support for bpf-*-* targets.
* common/config/bpf/bpf-common.c: New file.
* config/bpf/t-bpf: Likewise.
* config/bpf/predicates.md: Likewise.
* config/bpf/constraints.md: Likewise.
* config/bpf/bpf.opt: Likewise.
* config/bpf/bpf.md: Likewise.
* config/bpf/bpf.h: Likewise.
* config/bpf/bpf.c: Likewise.
* config/bpf/bpf-protos.h: Likewise.
* config/bpf/bpf-opts.h: Likewise.
* config/bpf/bpf-helpers.h: Likewise.
* config/bpf/bpf-helpers.def: Likewise.

gcc/testsuite/ChangeLog:

* gcc.dg/builtins-config.h: eBPF doesn't support C99 standard
functions.
* gcc.c-torture/compile/20101217-1.c: Add a function prototype for
printf.
* gcc.c-torture/compile/20000211-1.c: Skip if target bpf-*-*.
* gcc.c-torture/compile/poor.c: Likewise.
* gcc.c-torture/compile/pr25311.c: Likewise.
* gcc.c-torture/compile/pr39928-1.c: Likewise.
* gcc.c-torture/compile/pr70061.c: Likewise.
* gcc.c-torture/compile/920501-7.c: Likewise.
* gcc.c-torture/compile/20000403-1.c: Likewise.
* gcc.c-torture/compile/20001226-1.c: Likewise.
* gcc.c-torture/compile/20030903-1.c: Likewise.
* gcc.c-torture/compile/20031125-1.c: Likewise.
* gcc.c-torture/compile/20040101-1.c: Likewise.
* gcc.c-torture/compile/20040317-2.c: Likewise.
* gcc.c-torture/compile/20040726-1.c: Likewise.
* gcc.c-torture/compile/20051216-1.c: Likewise.
* gcc.c-torture/compile/900313-1.c: Likewise.
* gcc.c-torture/compile/920625-1.c: Likewise.
* gcc.c-torture/compile/930421-1.c: Likewise.
* gcc.c-torture/compile/930623-1.c: Likewise.
* gcc.c-torture/compile/961004-1.c: Likewise.
* gcc.c-torture/compile/980504-1.c: Likewise.
* gcc.c-torture/compile/980816-1.c: Likewise.
* gcc.c-torture/compile/990625-1.c: Likewise.
* gcc.c-torture/compile/DFcmp.c: Likewise.
* gcc.c-torture/compile/HIcmp.c: Likewise.
* gcc.c-torture/compile/HIset.c: Likewise.
* gcc.c-torture/compile/QIcmp.c: Likewise.
* gcc.c-torture/compile/QIset.c: Likewise.
* gcc.c-torture/compile/SFset.c: Likewise.
* gcc.c-torture/compile/SIcmp.c: Likewise.
* gcc.c-torture/compile/SIset.c: Likewise.
* gcc.c-torture/compile/UHIcmp.c: Likewise.
* gcc.c-torture/compile/UQIcmp.c: Likewise.
* gcc.c-torture/compile/USIcmp.c: Likewise.
* gcc.c-torture/compile/consec.c: Likewise.
* gcc.c-torture/compile/limits-fndefn.c: Likewise.
* gcc.c-torture/compile/lll.c: Likewise.
* gcc.c-torture/compile/parms.c: Likewise.
* gcc.c-torture/compile/pass.c: Likewise.
* gcc.c-torture/compile/pp.c: Likewise.
* gcc.c-torture/compile/pr32399.c: Likewise.
* gcc.c-torture/compile/pr34091.c: Likewise.
* gcc.c-torture/compile/pr34688.c: Likewise.
* gcc.c-torture/compile/pr37258.c: Likewise.
* gcc.c-torture/compile/pr37327.c: Likewise.
* gcc.c-torture/compile/pr37381.c: Likewise.
* gcc.c-torture/compile/pr37669-2.c: Likewise.
* gcc.c-torture/compile/pr37669.c: Likewise.
* gcc.c-torture/compile/pr37742-3.c: Likewise.
* gcc.c-torture/compile/pr44063.c: Likewise.
* gcc.c-torture/compile/pr48596.c: Likewise.
* gcc.c-torture/compile/pr51856.c: Likewise.
* gcc.c-torture/compile/pr54428.c: Likewise.
* gcc.c-torture/compile/pr54713-1.c: Likewise.
* gcc.c-torture/compile/pr54713-2.c: Likewise.
* gcc.c-torture/compile/pr54713-3.c: Likewise.
* gcc.c-torture/compile/pr55921.c: Likewise.
* gcc.c-torture/compile/pr70240.c: Likewise.
* gcc.c-torture/compile/pr70355.c: Likewise.
* gcc.c-torture/compile/pr82052.c: Likewise.
* gcc.c-torture/compile/pr83487.c: Likewise.
* gcc.c-torture/compile/pr86122.c: Likewise.
* gcc.c-torture/compile/pret-arg.c: Likewise.
* gcc.c-torture/compile/regs-arg-size.c: Likewise.
* gcc.c-torture/compile/structret.c: Likewise.
* gcc.c-torture/compile/uuarg.c: Likewise.
* gcc.dg/20001009-1.c: Likewise.
* gcc.dg/20020418-1.c: Likewise.
* gcc.dg/20020426-2.c: Likewise.
* gcc.dg/20020430-1.c: Likewise.
* gcc.dg/20040306-1.c: Likewise.
* gcc.dg/20040622-2.c: Likewise.
* gcc.dg/20050603-2.c: Likewise.
* gcc.dg/20050629-1.c: Likewise.
* gcc.dg/20061026.c: Likewise.
* gcc.dg/Warray-bounds-3.c: Likewise.
* gcc.dg/Warray-bounds-30.c: Likewise.
* gcc.dg/Wframe-larger-than-2.c: Likewise.
* gcc.dg/Wframe-larger-than.c: Likewise.
* gcc.dg/Wrestrict-11.c: Likewise.
* gcc.c-torture/compile/20000804-1.c: Likewise.
* lib/target-supports.exp (check_effective_target_trampolines):
Adapt to eBPF.
(check_effective_target_indirect_jumps): Likewise.
(check_effective_target_nonlocal_goto): Likewise.
(check_effective_target_global_constructor): Likewise.
(check_effective_target_return_address): Likewise.
* gcc.target/bpf/bpf.exp: New file.
* gcc.target/bpf/builtin-load.c: Likewise.
* cc.target/bpf/constant-calls.c: Likewise.
* gcc.target/bpf/diag-funargs.c: Likewise.
* gcc.target/bpf/diag-funargs-2.c: Likewise.
* gcc.target/bpf/diag-funargs-3.c: Likewise.
* gcc.target/bpf/diag-indcalls.c: Likewise.
* gcc.target/bpf/helper-bind.c: Likewise.
* gcc.target/bpf/helper-bpf-redirect.c: Likewise.
* gcc.target/bpf/helper-clone-redirect.c: Likewise.
* gcc.target/bpf/helper-csum-diff.c: Likewise.
* gcc.target/bpf/helper-csum-update.c: Likewise.
* gcc.target/bpf/helper-current-task-under-cgroup.c: Likewise.
* gcc.target/bpf/helper-fib-lookup.c: Likewise.
* gcc.target/bpf/helper-get-cgroup-classid.c: Likewise.
* gcc.target/bpf/helper-get-current-cgroup-id.c: Likewise.
* gcc.target/bpf/helper-get-current-comm.c: Likewise.
* gcc.target/bpf/helper-get-current-pid-tgid.c: Likewise.
* gcc.target/bpf/helper-get-current-task.c: Likewise.
* gcc.target/bpf/helper-get-current-uid-gid.c: Likewise.
* gcc.target/bpf/helper-get-hash-recalc.c: Likewise.
* gcc.target/bpf/helper-get-listener-sock.c: Likewise.
* gcc.target/bpf/helper-get-local-storage.c: Likewise.
* gcc.target/bpf/helper-get-numa-node-id.c: Likewise.
* gcc.target/bpf/helper-get-prandom-u32.c: Likewise.
* gcc.target/bpf/helper-get-route-realm.c: Likewise.
* gcc.target/bpf/helper-get-smp-processor-id.c: Likewise.
* gcc.target/bpf/helper-get-socket-cookie.c: Likewise.
* gcc.target/bpf/helper-get-socket-uid.c: Likewise.
* gcc.target/bpf/helper-getsockopt.c: Likewise.
* gcc.target/bpf/helper-get-stack.c: Likewise.
* gcc.target/bpf/helper-get-stackid.c: Likewise.
* gcc.target/bpf/helper-ktime-get-ns.c: Likewise.
* gcc.target/bpf/helper-l3-csum-replace.c: Likewise.
* gcc.target/bpf/helper-l4-csum-replace.c: Likewise.
* gcc.target/bpf/helper-lwt-push-encap.c: Likewise.
* gcc.target/bpf/helper-lwt-seg6-action.c: Likewise.
* gcc.target/bpf/helper-lwt-seg6-adjust-srh.c: Likewise.
* gcc.target/bpf/helper-lwt-seg6-store-bytes.c: Likewise.
* gcc.target/bpf/helper-map-delete-elem.c: Likewise.
* gcc.target/bpf/helper-map-lookup-elem.c: Likewise.
* gcc.target/bpf/helper-map-peek-elem.c: Likewise.
* gcc.target/bpf/helper-map-pop-elem.c: Likewise.
* gcc.target/bpf/helper-map-push-elem.c: Likewise.
* gcc.target/bpf/helper-map-update-elem.c: Likewise.
* gcc.target/bpf/helper-msg-apply-bytes.c: Likewise.
* gcc.target/bpf/helper-msg-cork-bytes.c: Likewise.
* gcc.target/bpf/helper-msg-pop-data.c: Likewise.
* gcc.target/bpf/helper-msg-pull-data.c: Likewise.
* gcc.target/bpf/helper-msg-push-data.c: Likewise.
* gcc.target/bpf/helper-msg-redirect-hash.c: Likewise.
* gcc.target/bpf/helper-msg-redirect-map.c: Likewise.
* gcc.target/bpf/helper-override-return.c: Likewise.
* gcc.target/bpf/helper-perf-event-output.c: Likewise.
* gcc.target/bpf/helper-perf-event-read.c: Likewise.
* gcc.target/bpf/helper-perf-event-read-value.c: Likewise.
* gcc.target/bpf/helper-perf-prog-read-value.c: Likewise.
* gcc.target/bpf/helper-probe-read.c: Likewise.
* gcc.target/bpf/helper-probe-read-str.c: Likewise.
* gcc.target/bpf/helper-probe-write-user.c: Likewise.
* gcc.target/bpf/helper-rc-keydown.c: Likewise.
* gcc.target/bpf/helper-rc-pointer-rel.c: Likewise.
* gcc.target/bpf/helper-rc-repeat.c: Likewise.
* gcc.target/bpf/helper-redirect-map.c: Likewise.
* gcc.target/bpf/helper-set-hash.c: Likewise.
* gcc.target/bpf/helper-set-hash-invalid.c: Likewise.
* gcc.target/bpf/helper-setsockopt.c: Likewise.
* gcc.target/bpf/helper-skb-adjust-room.c: Likewise.
* gcc.target/bpf/helper-skb-cgroup-id.c: Likewise.
* gcc.target/bpf/helper-skb-change-head.c: Likewise.
* gcc.target/bpf/helper-skb-change-proto.c: Likewise.
* gcc.target/bpf/helper-skb-change-tail.c: Likewise.
* gcc.target/bpf/helper-skb-change-type.c: Likewise.
* gcc.target/bpf/helper-skb-ecn-set-ce.c: Likewise.
* gcc.target/bpf/helper-skb-get-tunnel-key.c: Likewise.
* gcc.target/bpf/helper-skb-get-tunnel-opt.c: Likewise.
* gcc.target/bpf/helper-skb-get-xfrm-state.c: Likewise.
* gcc.target/bpf/helper-skb-load-bytes.c: Likewise.
* gcc.target/bpf/helper-skb-load-bytes-relative.c: Likewise.
* gcc.target/bpf/helper-skb-pull-data.c: Likewise.
* gcc.target/bpf/helper-skb-set-tunnel-key.c: Likewise.
* gcc.target/bpf/helper-skb-set-tunnel-opt.c: Likewise.
* gcc.target/bpf/helper-skb-store-bytes.c: Likewise.
* gcc.target/bpf/helper-skb-under-cgroup.c: Likewise.
* gcc.target/bpf/helper-skb-vlan-pop.c: Likewise.
* gcc.target/bpf/helper-skb-vlan-push.c: Likewise.
* gcc.target/bpf/helper-skc-lookup-tcp.c: Likewise.
* gcc.target/bpf/helper-sk-fullsock.c: Likewise.
* gcc.target/bpf/helper-sk-lookup-tcp.c: Likewise.
* gcc.target/bpf/helper-sk-lookup-upd.c: Likewise.
* gcc.target/bpf/helper-sk-redirect-hash.c: Likewise.
* gcc.target/bpf/helper-sk-redirect-map.c: Likewise.
* gcc.target/bpf/helper-sk-release.c: Likewise.
* gcc.target/bpf/helper-sk-select-reuseport.c: Likewise.
* gcc.target/bpf/helper-sk-storage-delete.c: Likewise.
* gcc.target/bpf/helper-sk-storage-get.c: Likewise.
* gcc.target/bpf/helper-sock-hash-update.c: Likewise.
* gcc.target/bpf/helper-sock-map-update.c: Likewise.
* gcc.target/bpf/helper-sock-ops-cb-flags-set.c: Likewise.
* gcc.target/bpf/helper-spin-lock.c: Likewise.
* gcc.target/bpf/helper-spin-unlock.c: Likewise.
* gcc.target/bpf/helper-strtol.c: Likewise.
* gcc.target/bpf/helper-strtoul.c: Likewise.
* gcc.target/bpf/helper-sysctl-get-current-value.c: Likewise.
* gcc.target/bpf/helper-sysctl-get-name.c: Likewise.
* gcc.target/bpf/helper-sysctl-get-new-value.c: Likewise.
* gcc.target/bpf/helper-sysctl-set-new-value.c: Likewise.
* gcc.target/bpf/helper-tail-call.c: Likewise.
* gcc.target/bpf/helper-tcp-check-syncookie.c: Likewise.
* gcc.target/bpf/helper-tcp-sock.c: Likewise.
* gcc.target/bpf/helper-trace-printk.c: Likewise.
* gcc.target/bpf/helper-xdp-adjust-head.c: Likewise.
* gcc.target/bpf/helper-xdp-adjust-meta.c: Likewise.
* gcc.target/bpf/helper-xdp-adjust-tail.c: Likewise.
* gcc.target/bpf/skb-ancestor-cgroup-id.c: Likewise.
* gcc.target/bpf/sync-fetch-and-add.c: Likewise.

libgcc/ChangeLog:

* config.host: Set cpu_type for bpf-*-* targets.
* config/bpf/t-bpf: Likewise.
* config/bpf/crtn.S: Likewise.
* config/bpf/crti.S: New file.

From-SVN: r275506

testsuite: new require effective target indirect_calls

This patch adds a new dg_require_effective_target procedure to the
testsuite infrastructure: indirect_calls.  This new function tells
whether a target supports calls to non-constant call targets.

This patch also annotates the tests in the gcc.c-torture testuite that
require support for indirect calls.

    gcc/ChangeLog:

            * doc/sourcebuild.texi (Effective-Target Keywords): Document
            indirect_calls.

    gcc/testsuite/ChangeLog:

            * lib/target-supports.exp (check_effective_target_indirect_calls):
            New proc.
            * gcc.c-torture/compile/20010102-1.c: Annotate with
            dg-require-effective-target indirect_calls.
            * gcc.c-torture/compile/20010107-1.c: Likewise.
            * gcc.c-torture/compile/20011109-1.c: Likewise.
            * gcc.c-torture/compile/20011218-1.c: Likewise.
            * gcc.c-torture/compile/20011229-1.c: Likewise.
            * gcc.c-torture/compile/20020129-1.c: Likewise.
            * gcc.c-torture/compile/20020320-1.c: Likewise.
            * gcc.c-torture/compile/20020706-1.c: Likewise.
            * gcc.c-torture/compile/20020706-2.c: Likewise.
            * gcc.c-torture/compile/20021205-1.c: Likewise.
            * gcc.c-torture/compile/20030921-1.c: Likewise.
            * gcc.c-torture/compile/20031023-1.c: Likewise.
            * gcc.c-torture/compile/20031023-2.c: Likewise.
            * gcc.c-torture/compile/20031023-3.c: Likewise.
            * gcc.c-torture/compile/20031023-4.c: Likewise.
            * gcc.c-torture/compile/20040614-1.c: Likewise.
            * gcc.c-torture/compile/20040909-1.c: Likewise.
            * gcc.c-torture/compile/20050122-1.c: Likewise.
            * gcc.c-torture/compile/20050202-1.c: Likewise.
            * gcc.c-torture/compile/20060208-1.c: Likewise.
            * gcc.c-torture/compile/20081108-1.c: Likewise.
            * gcc.c-torture/compile/20150327.c: Likewise.
            * gcc.c-torture/compile/920428-2.c: Likewise.
            * gcc.c-torture/compile/920928-5.c: Likewise.
            * gcc.c-torture/compile/930117-1.c: Likewise.
            * gcc.c-torture/compile/930607-1.c: Likewise.
            * gcc.c-torture/compile/991213-2.c: Likewise.
            * gcc.c-torture/compile/callind.c: Likewise.
            * gcc.c-torture/compile/calls-void.c: Likewise.
            * gcc.c-torture/compile/calls.c: Likewise.
            * gcc.c-torture/compile/pr21840.c: Likewise.
            * gcc.c-torture/compile/pr32139.c: Likewise.
            * gcc.c-torture/compile/pr35607.c: Likewise.
            * gcc.c-torture/compile/pr37433-1.c: Likewise.
            * gcc.c-torture/compile/pr37433.c: Likewise.
            * gcc.c-torture/compile/pr39941.c: Likewise.
            * gcc.c-torture/compile/pr40080.c: Likewise.
            * gcc.c-torture/compile/pr43635.c: Likewise.
            * gcc.c-torture/compile/pr43791.c: Likewise.
            * gcc.c-torture/compile/pr43845.c: Likewise.
            * gcc.c-torture/compile/pr44043.c: Likewise.
            * gcc.c-torture/compile/pr51694.c: Likewise.
            * gcc.c-torture/compile/pr77754-2.c: Likewise.
            * gcc.c-torture/compile/pr77754-3.c: Likewise.
            * gcc.c-torture/compile/pr77754-4.c: Likewise.
            * gcc.c-torture/compile/pr89663-2.c: Likewise.
            * gcc.c-torture/compile/pta-1.c: Likewise.
            * gcc.c-torture/compile/stack-check-1.c: Likewise.
            * gcc.dg/Walloc-size-larger-than-18.c: Likewise.

From-SVN: r275505

testsuite: annotate c-torture/compile tests with dg-require-stack-size

This patch annotates tests that make use of a significant a mount of
stack space.  Embedded and other restricted targets may have problems
compiling and running these tests.  Note that the annotations are in
many cases not exact.

testsuite/ChangeLog:

        * gcc.c-torture/compile/20000609-1.c: Annotate with
        dg-require-stack-size.
        * gcc.c-torture/compile/20000804-1.c: Likewise.
        * gcc.c-torture/compile/20020304-1.c: Likewise.
        * gcc.c-torture/compile/20020604-1.c: Likewise.
        * gcc.c-torture/compile/20021015-1.c: Likewise.
        * gcc.c-torture/compile/20050303-1.c: Likewise.
        * gcc.c-torture/compile/20060421-1.c: Likewise.
        * gcc.c-torture/compile/20071207-1.c: Likewise.
        * gcc.c-torture/compile/20080903-1.c: Likewise.
        * gcc.c-torture/compile/20121027-1.c: Likewise.
        * gcc.c-torture/compile/20151204.c: Likewise.
        * gcc.c-torture/compile/920501-12.c: Likewise.
        * gcc.c-torture/compile/920501-4.c: Likewise.
        * gcc.c-torture/compile/920723-1.c: Likewise.
        * gcc.c-torture/compile/921202-1.c: Likewise.
        * gcc.c-torture/compile/931003-1.c: Likewise.
        * gcc.c-torture/compile/931004-1.c: Likewise.
        * gcc.c-torture/compile/950719-1.c: Likewise.
        * gcc.c-torture/compile/951222-1.c: Likewise.
        * gcc.c-torture/compile/990517-1.c: Likewise.
        * gcc.c-torture/compile/bcopy.c: Likewise.
        * gcc.c-torture/compile/pr23929.c: Likewise.
        * gcc.c-torture/compile/pr25310.c: Likewise.
        * gcc.c-torture/compile/pr34458.c: Likewise.
        * gcc.c-torture/compile/pr39937.c: Likewise.
        * gcc.c-torture/compile/pr41181.c: Likewise.
        * gcc.c-torture/compile/pr41634.c: Likewise.
        * gcc.c-torture/compile/pr43415.c: Likewise.
        * gcc.c-torture/compile/pr43417.c: Likewise.
        * gcc.c-torture/compile/pr44788.c: Likewise.
        * gcc.c-torture/compile/sound.c: Likewise.

From-SVN: r275504

opt-functions.awk: fix comparison of limit, begin and end

The function integer_range_info makes sure that, if provided, the
initial value fills in the especified range.  However, it is necessary
to convert the values to a numerical context before comparing, to make
sure awk is using arithmetical order and not lexicographical order.

gcc/ChangeLog:

      * opt-functions.awk (integer_range_info): Make sure values are in
      numeric context before operating with them.

From-SVN: r275503

Update config.sub and config.guess.

    * config.sub: Import upstream version 2019-06-30.
    * config.guess: Import upstream version 2019-07-24.

From-SVN: r275502

Daily bump.

From-SVN: r275501

genemit: Print file+line in the "Splitting with" message

It's tiresome to have to look in insn-emit.c to see where some split
came from, so let's print that info to the dump file as well.  But
don't print the full path, just the basename, for greater readability.

* genemit.c (gen_split): Print the filename and line number where the
splitter (or peephole2) was defined, to the dump file.

From-SVN: r275497

Daily bump.

From-SVN: r275493

pr91684.c: Use effective-target arm_prefer_ldrd_strd.

2019-09-07  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        * gcc.target/arm/pr91684.c: Use effective-target arm_prefer_ldrd_strd.

From-SVN: r275489

decomp2.C: Add TLS options.

        * g++.dg/cpp2a/decomp2.C: Add TLS options.
        * gcc.target/powerpc/pr88233.c: Limit to lp64.

From-SVN: r275488

re PR tree-optimization/91665 (ICE in build_vector_from_val, at tree.c:1904)

PR tree-optimization/91665
* tree-vect-loop.c (vectorizable_reduction): Punt if base has type
incompatible with the type of PHI result.

* gcc.dg/vect/pr91665.c: New test.

From-SVN: r275486

add PR target/91684 to ChangeLog

From-SVN: r275484

arm.c (arm_block_set_aligned_non_vect): Use gen_unaligned_storedi for 4-byte aligned addresses.

2019-09-07  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        * config/arm/arm.c (arm_block_set_aligned_non_vect): Use
        gen_unaligned_storedi for 4-byte aligned addresses.

testsuite:
2019-09-07  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        * gcc.target/arm/pr91684.c: New test.

From-SVN: r275483

Daily bump.

From-SVN: r275482

RISC-V: Re-enable -msave-restore for shared libraries.

This re-enables -msave-restore for shared libraries, and uses the
t-slibgcc-libgcc file to get the save-restore routines included directly
in shared libraries so that we don't need to indirect through the PLT
to reach them, which doesn't work.

gcc/
* config/riscv/riscv.c (riscv_option_override): Revert 2019-08-30
change.

libgcc/
* config.host (riscv*-*-linux*): Add t-slibgcc-libgcc to tmake_file.
(riscv*-*-freebsd*): Likewise.

From-SVN: r275478

rs6000: Delete UNSPEC_MV_CR_OV.

This isn't used since 2018.  (It's a remnant of paired single support).

* config/rs6000/rs6000.md (unspec): Delete UNSPEC_MV_CR_OV.

From-SVN: r275477

rs6000: Delete UNSPEC_FRSP

This isn't used since 2012.  (It's a remnant of RIOS support).

* config/rs6000/rs6000.c (rs6000_rtx_costs) <case UNSPEC>: Delete.
* config/rs6000/rs6000.md (unspec): Delete UNSPEC_FRSP.

From-SVN: r275476

re PR target/91654 (Regressions of SPEC2017 rate caused by r274994)

PR target/91654
* config/i386/x86-tune-costs.h (skylake_cost): Raise the
cost of SSE->integer and integer->SSE moves from 2 to 6.
(core_cost): Ditto.

From-SVN: r275475

libgo: update to Go 1.13beta1 release

    Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/193497

From-SVN: r275473

function.c (assign_parm_find_data_types): Use RECORD_OR_UNION_TYPE_P before testing TYPE_TRANSPARENT_AGGR.

* function.c (assign_parm_find_data_types): Use RECORD_OR_UNION_TYPE_P
before testing TYPE_TRANSPARENT_AGGR.
* calls.c (initialize_argument_information, load_register_parameters):
Likewise.

From-SVN: r275472

inf-compare-1.c: Skip AIX.

        * gcc.dg/torture/inf-compare-1.c: Skip AIX.
        * gcc.dg/torture/inf-compare-2.c: Skip AIX.
        * gcc.dg/torture/inf-compare-3.c: Skip AIX.
        * gcc.dg/torture/inf-compare-4.c: Skip AIX.
        * gcc.dg/torture/pr52451.c: Skip AIX.
        * gcc.dg/torture/pr68264.c: Skip AIX
        * gcc.dg/torture/pr91323.c: Skip AIX.

From-SVN: r275464

Warray-bounds-4.C: Skip AIX.

        * g++.dg/warn/Warray-bounds-4.C: Skip AIX.
        * g++.dg/warn/Warray-bounds-8.C: Skip AIX.
        * g++.dg/opt/flifetime-dse2.C: XFAIL AIX.
        * g++.dg/opt/flifetime-dse4.C: XFAIL AIX.

From-SVN: r275462

Fix testcase to not use LTO with -fvtable-verify.

2019-09-05  Caroline Tice  <cmtice@google.com>

        PR testsuite/91670
        * g++.dg/ubsan/pr59415.C: Disable LTO, since test uses
        -fvtable-verify, and the two options are no longer allowed
        together.

From-SVN: r275460

Define std::ssize for C++20 (P1227R2)

* include/bits/range_access.h (ssize): Define for C++20.
* testsuite/24_iterators/range_access_cpp20.cc: New test.
* doc/xml/manual/status_cxx2020.xml: Update P1227R2 status.
* doc/html/*: Regenerate.

From-SVN: r275458

[preprocessor] Popping "" file names

https://gcc.gnu.org/ml/gcc-patches/2019-08/msg02069.html
New # semantics for popping to "" name.
libcpp/
* directives.c (do_linemarker): Popping to "" name means get the
name from the include stack..

From-SVN: r275457

[PATCH] Deprecate -frepo option.

https://gcc.gnu.org/ml/gcc-patches/2019-09/msg00379.html
PR c++/91125
* cp-tree.h (IDENTIFIER_REPO_CHOSEN, DECL_REPO_AVAILABLE_P): Delete.
(struct lang_decl_base): Remove repo_available_p.
* decl.c (duplicate_decls): Don't copy DECL_REPO_AVAILABLE_P.

From-SVN: r275456

[C++ PATCH] Reserve a decl_lang bit

https://gcc.gnu.org/ml/gcc-patches/2019-09/msg00365.html
Reserve TREE_LANG_FLAG_3 for modules.
gcc/cp/
* cp-tree.h (DECL_CONSTRUCTION_VTABLE_P): Delete.
(DECL_NON_TRIVIALLY_INITIALIZED_P): Move to TREE_LANG_FLAG_6.
* class.c (build_ctor_vtbl_group): Don't set
DECL_CONSTRUCTION_VTABLE_P.
* decl2.c (determine_visibility_from_class): Don't check
DECL_CONSTRUCTION_VTABLE_P anymore.

From-SVN: r275455

Fix GCC_LINUX_FUTEX to work with C99 compilers

Without this change, libstdc++ is built without futex symbols if GCC
rejects implicit function declarations by default.

From-SVN: r275454

[arm] Add missing predicated-short-it variants to cmp_and and cmp_ior patterns

The cmp_and and cmp_ior patterns were missing a couple of short-it
variants for thumb2, where the comparisons are all using registers
some of which were HI_REGS.

* config/arm/arm.md (cmp_and): Add short-it variant for thumb2 with
high regs.
(cmp_ior): Likewise.

From-SVN: r275453

Improve documentation of for statement.

2019-09-06  Martin Liska  <mliska@suse.cz>

* doc/match-and-simplify.texi: Separate tuples with ;.

From-SVN: r275452

Remove support for repo files (PR c++/91125).

2019-09-06  Martin Liska  <mliska@suse.cz>

PR c++/91125
* Makefile.in: Remove tlink.o.
* collect2.c (do_link): New function isolated
from do_tlink.
(main): Use.
* collect2.h (do_tlink): Remove declaration of do_tlink.
* doc/extend.texi: Remove documentation of -frepo.
* doc/invoke.texi: Likewise.
* doc/sourcebuild.texi: Remove cleanup-repo-files.
* tlink.c: Remove.
2019-09-06  Martin Liska  <mliska@suse.cz>

PR c++/91125
* c-common.c: Remove definition of flag_use_repository.
* c-common.h: Likewise.
* c-opts.c (c_common_handle_option):
Do not handle OPT_frepo option.
* c.opt: Mark the option with Deprecated.
2019-09-06  Martin Liska  <mliska@suse.cz>

PR c++/91125
* Make-lang.in: Remove repo.o.
* config-lang.in: Likewise.
* cp-tree.h (init_repo): Remove declarations
of repo-related functions.
(repo_emit_p): Likewise.
(repo_export_class_p): Likewise.
(finish_repo): Likewise.
* decl2.c (import_export_class): Always
set -1 value/
(mark_needed): Remove -frepo from comment.
(import_export_decl): Similarly here.
(c_parse_final_cleanups): Remove call of finish_repo.
* lex.c (cxx_init): Remove call to init_repo.
* optimize.c (can_alias_cdtor): Remove dead condition.
* pt.c (push_template_decl_real): Update comment.
(instantiate_decl): Remove dead code used for -frepo.
* repo.c: Remove.
2019-09-06  Martin Liska  <mliska@suse.cz>

PR c++/91125
* g++.dg/parse/repo1.C: Remove.
* g++.dg/rtti/repo1.C: Remove.
* g++.dg/template/repo1.C: Remove.
* g++.dg/template/repo10.C: Remove.
* g++.dg/template/repo11.C: Remove.
* g++.dg/template/repo2.C: Remove.
* g++.dg/template/repo3.C: Remove.
* g++.dg/template/repo4.C: Remove.
* g++.dg/template/repo5.C: Remove.
* g++.dg/template/repo6.C: Remove.
* g++.dg/template/repo7.C: Remove.
* g++.dg/template/repo8.C: Remove.
* g++.dg/template/repo9.C: Remove.
* g++.old-deja/g++.pt/instantiate4.C: Remove.
* g++.old-deja/g++.pt/instantiate6.C: Remove.
* g++.old-deja/g++.pt/repo1.C: Remove.
* g++.old-deja/g++.pt/repo2.C: Remove.
* g++.old-deja/g++.pt/repo3.C: Remove.
* g++.old-deja/g++.pt/repo4.C: Remove.
* lib/g++.exp: Remove removal of repo files.
* lib/gcc-dg.exp: Likewise.
* lib/obj-c++.exp: Likewise.

From-SVN: r275450

Daily bump.

From-SVN: r275449

RISC-V: Fix bad insn splits with paradoxical subregs.

Shifting by more than the size of a SUBREG_REG doesn't work, so we either
need to disable splits if an input is paradoxical, or else we need to
generate a clean temporary for intermediate results.

Jakub wrote the first version of this patch, so gets primary credit for it.

gcc/
PR target/91635
* config/riscv/riscv.md (zero_extendsidi2, zero_extendhi<GPR:mode>2,
extend<SHORT:mode><SUPERQI:mode>2): Don't split if
paradoxical_subreg_p (operands[0]).
(*lshrsi3_zero_extend_3+1, *lshrsi3_zero_extend_3+2): Add clobber and
use as intermediate value.

gcc/testsuite/
PR target/91635
* gcc.c-torture/execute/pr91635.c: New test.
* gcc.target/riscv/shift-shift-4.c: New test.
* gcc.target/riscv/shift-shift-5.c: New test.

Co-Authored-By: Jim Wilson <jimw@sifive.com>
From-SVN: r275444

re PR fortran/91496 (!GCC$ directives error if mistyped or unknown)

2019-09-05  Harald Anlauf  <anlauf@gmx.de>

PR fortran/91496
* parse.c (parse_executable): Improve error messages for
improperly placed pragmas not preceeding a loop.

PR fortran/91496
* gfortran.dg/directive_unroll_5.f90: Adjust error message.

From-SVN: r275442

re PR fortran/91660 (Missing error on invalid type declaration)

2019-09-05  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/91660
* decl.c (gfc_match_decl_type_spec): Improve and restore error
message for malformed types-spec.

2019-09-05  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/91660
* gfortran.dg/pdt_4.f03: Fix invalid code.
* gfortran.dg/pr91660_1.f90: New test.
* gfortran.dg/pr91660_2.f90: Ditto.

From-SVN: r275426

PR c++/91644 - ICE with constinit in function template.

* decl.c (start_decl): Call retrofit_lang_decl for constinit variables.
* pt.c (tsubst_expr): Pass LOOKUP_CONSTINIT down to cp_finish_decl for
constinit variables.

* g++.dg/cpp2a/constinit13.C: New test.

From-SVN: r275421

Global GCN instructions need nops too.

2019-09-05  Andrew Stubbs  <ams@codesourcery.com>

gcc/
* config/gcn/gcn.md (*movti_insn): Set delayeduse for global_store.
(sync_compare_and_swap<mode>_insn): Likewise.

From-SVN: r275414

Improve precondition checks for std::span

* doc/xml/manual/status_cxx2020.xml: Update status for P0122R7 and
P1024R3. Remove entry for P0920R2.
* include/std/span  (__cpp_lib_span): Change value.
(__extent_storage, __extent_storage<dynamic_extent>): Remove default
constructor.
(span): Replace __extent_storage base class with data member.
(span::_S_subspan_extent): New function.
(span::empty()): Add nodiscard attribute.
(span::front, span::back, span::operator[]): Check preconditions.
(span::first, span::last, span::subspan): Add noexcept. Improve
precondition checks (LWG 3103).
(get): Remove redundant condition from static_assert.
(tuple_element<I, span<T, E>>): Fix static_assert message and simplify.
(as_writable_bytes): Add inline specifier.
* include/std/version (__cpp_lib_span): Change value.
* testsuite/23_containers/span/back_neg.cc: Remove stray semi-colon.
* testsuite/23_containers/span/front_neg.cc: Likewise.
* testsuite/23_containers/span/index_op_neg.cc: Likewise.
* testsuite/23_containers/span/last_neg.cc: Improve test.
* testsuite/23_containers/span/subspan_neg.cc: Likewise.
* testsuite/23_containers/span/1.cc: New test.
* testsuite/23_containers/span/2.cc: New test.
* testsuite/23_containers/span/back_assert_neg.cc: New test.
* testsuite/23_containers/span/first_2_assert_neg.cc: New test.
* testsuite/23_containers/span/first_assert_neg.cc: New test.
* testsuite/23_containers/span/first_neg.cc: New test.
* testsuite/23_containers/span/front_assert_neg.cc: New test.
* testsuite/23_containers/span/index_op_assert_neg.cc: New test.
* testsuite/23_containers/span/last_2_assert_neg.cc: New test.
* testsuite/23_containers/span/last_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_2_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_3_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_4_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_5_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_6_assert_neg.cc: New test.
* testsuite/23_containers/span/subspan_assert_neg.cc: New test.

From-SVN: r275411

re PR target/91615 ([armeb] ICEs since r274986)

2019-09-05  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR middle-end/91615
        * expr.c (expand_expr_real_1): Handle misaligned TARGET_MEM_REF
        without movmisalign optab.

From-SVN: r275409

re PR middle-end/91001 (internal compiler error: in extract_insn, at recog.c:2310)

PR middle-end/91001
PR middle-end/91105
PR middle-end/91106
* calls.c (load_register_parameters): For TYPE_TRANSPARENT_AGGR
types, use type of their first field instead of type of
args[i].tree_value.

* gcc.c-torture/compile/pr91001.c: New test.

From-SVN: r275408

re PR rtl-optimization/91656 (wrong code with -fgcse-after-reload)

2019-09-05  Richard Biener  <rguenther@suse.de>

PR rtl-optimization/91656
* postreload-gcse.c (record_last_mem_set_info): Revert addition
of early out.

* gcc.dg/torture/pr91656-1.c: New testcase.
* gcc.dg/torture/pr91656-2.c: Likewise.
* gcc.dg/torture/pr91656-3.c: Likewise.

From-SVN: r275406

[C++ PATCH] vtable decl marking

https://gcc.gnu.org/ml/gcc-patches/2019-08/msg02063.html
* cp-tree.h (DECL_VTABLE_OR_VTT_P): Forward to DECL_VIRTUAL_P.

From-SVN: r275404

Implement std::span for C++20

2019-09-05  JeanHeyd Meneide  <phdofthehouse@gmail.com>

* include/Makefile.am: Add <span> header.
* include/Makefile.in: Regenerate.
* include/bits/range_access.h (__adl_begin, __adl_end, __adl_cbegin)
(__adl_cend, __adl_rbegin, __adl_rend, __adl_crbegin, __adl_crend)
(__adl_data, __adl_cdata, __adl_size, __adl_empty, __adl_to_address):
New functions for performing argument-dependent lookup of range
customization points.
* include/bits/stl_iterator.h (__normal_iterator): Add
_GLIBCXX20_CONSTEXPR to all functions.
* include/std/span: New header.
* include/std/version (__cpp_lib_span): Define feature test macro.
* testsuite/23_containers/span/contiguous_range_neg.cc: New test.
* testsuite/23_containers/span/everything.cc: New test.
* testsuite/23_containers/span/get_neg.cc: New test.
* testsuite/23_containers/span/last_neg.cc: New test.
* testsuite/23_containers/span/subspan_neg.cc: New test.
* testsuite/23_containers/span/tuple_element_dynamic_neg.cc: New test.
* testsuite/23_containers/span/tuple_element_oob_neg.cc: New test.
* testsuite/23_containers/span/tuple_size_neg.cc: New test.

From-SVN: r275403

[preprocessor/91639] #includes at EOF

https://gcc.gnu.org/ml/gcc-patches/2019-09/msg00280.html
libcpp/
PR preprocessor/91639
* directives.c (do_include_common): Tell lexer we're a #include.
* files.c (_cpp_stack_file): Lexer will have always incremented.
* internal.h (struct cpp_context): Extend in_directive's
semantics.
* lex.c (_cpp_lex_direct): Increment line for final \n when lexing
for an ISO #include.
* line-map.c (linemap_line_start): Remember if we overflowed.

gcc/testsuite/
PR preprocessor/91639
* c-c++-common/cpp/pr91639.c: New.
* c-c++-common/cpp/pr91639-one.h: New.
* c-c++-common/cpp/pr91639-two.h: New.

From-SVN: r275402

re PR middle-end/90501 (ICE: address taken, but ADDRESSABLE bit not set)

2019-09-05  Richard Biener  <rguenther@suse.de>

PR middle-end/90501
* tree-inline.c (declare_return_variable): Mark the return
slot as addressable after building an address of it.

From-SVN: r275401

install.texi: Update and clarify requirements to build GNAT.

* doc/install.texi: Update and clarify requirements to build GNAT.

From-SVN: r275400

Force IFN_LOAD/STORE_LANES operands to be memory (PR91577)

This patch uses the workaround Richi suggested in the PR: make
discover_nonconstant_array_refs mark the source of an IFN_LOAD_LANES
call and the destination of an IFN_STORE_LANES call as addressable,
so that they don't end up being REG rtxes during expansion.

I had to move the discover_nonconstant_array_refs call outside the
currently_expanding_to_rtl block since otherwise mark_addressable
just queues the decision for later.

2019-09-05  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
PR middle-end/91577
* cfgexpand.c (discover_nonconstant_array_refs): Force the source
of an IFN_LOAD_LANES call and the destination of an IFN_STORE_LANES
call to be in memory.
(pass_expand::execute): Call discover_nonconstant_array_refs before
setting currently_expanding_to_rtl.

gcc/testsuite/
PR middle-end/91577
* gfortran.dg/pr91577.f90: New test, taken from temporary_1.f90.

From-SVN: r275399

Remove broken URL from libstdc++ manual

The URL for the "What Are Allocators Good For?" article has been a
recurring source of problems. It moved from the C/C++ Users Journal
website to the Dr Dobbs site after CUJ shut down, and the original
domain changed hands, leaving old links pointing to nefarious sites.

Now the URL to the copy on drdobbs.com no longer works either and I
can't find a (legal) copy of the article online. The simplest solution
is to remove the URL.

* doc/xml/manual/allocator.xml: Remove URL for bibliography entry.
* doc/html/*: Regenerate.

From-SVN: r275398

Improve PRNG jumping when using threads

Currently, when a new thread needs to use the RANDOM_NUMBER intrinsic,
the per-thread PRNG state is initialized by copying the master state
and then jumping forwards N*2**128 entries in the stream so that the
PRNG streams for different threads don't alias each other, where N is
the number of threads that have so far initialized the PRNG.

With this patch the master state itself is jumped forwards once each
time a new thread initializes the PRNG, thus obviating the need to
jump through all the N-1 previous streams. Effectively turning an O(N)
algorithm into an O(1) one.

Regtested on x86_64-pc-linux-gnu.

libgfortran/ChangeLog:

2019-09-05  Janne Blomqvist  <jb@gcc.gnu.org>

* intrinsics/random.c (master_init): Replace with
master_state.init.
(njumps): Remove variable.
(master_state): Make instance of struct prng_state.
(init_rand_state): When jumping, update the master_state once
instead of keeping track of how many jumps need to be done.
(SZU64): Modify to handle new master_state.
(SZ): Likewise.
(random_seed_i4): Likewise.
(random_seed_i8): Likewise.

From-SVN: r275397

re PR tree-optimization/91663 (split function can be re-inlined, leaving bad stack trace)

PR tree-optimization/91663
* go-lang.c (go_langhook_post_options): Clear
flag_partial_inlining.

From-SVN: r275396

Daily bump.

From-SVN: r275395

re PR fortran/91650 (ICE in gfc_conv_constant_to_tree, at fortran/trans-const.c:370)

2019-09-04  Steven G. Kargl  <kargl@gcvc.gnu.org>

PR fortran/91650
* io.c (match_io_element):  An output IO list item cannot be a BOZ.

2019-09-04  Steven G. Kargl  <kargl@gcvc.gnu.org>

PR fortran/91650
* gfortran.dg/pr91650_1.f90: New test.
* gfortran.dg/pr91650_2.f90: Ditto.

From-SVN: r275391

PR libstdc++/41861 Add full steady_clock support to condition_variable

The pthread_cond_clockwait function is available in glibc since the 2.30
release. If this function is available in the C library it can be used
to fix PR libstdc++/41861 by supporting std::chrono::steady_clock
properly with std::condition_variable.

This means that code using std::condition_variable::wait_for or
std::condition_variable::wait_until with std::chrono::steady_clock is no
longer subject to timing out early or potentially waiting for much
longer if the system clock is warped at an inopportune moment.

If pthread_cond_clockwait is available then std::chrono::steady_clock is
deemed to be the "best" clock available which means that it is used for
the relative wait_for calls and absolute wait_until calls using
user-defined clocks. Calls explicitly using std::chrono::system_clock
continue to use CLOCK_REALTIME via __gthread_cond_timedwait.

If pthread_cond_clockwait is not available then
std::chrono::system_clock is deemed to be the "best" clock available
which means that the previous suboptimal behaviour remains.

2019-09-04  Mike Crowe  <mac@mcrowe.com>

PR libstdc++/41861
* acinclude.m4 (GLIBCXX_CHECK_PTHREAD_COND_CLOCKWAIT): Check for new
pthread_cond_clockwait function.
* configure.ac: Use GLIBCXX_CHECK_PTHREAD_COND_CLOCKWAIT.
* configure: Regenerate.
* config.h.in: Regenerate.
* include/std/condition_variable: (condition_variable): Rename
__steady_clock_t typedef and add system_clock. Change __clock_t to be
a typedef for the preferred clock to convert arbitrary other clocks to.
[_GLIBCXX_USE_PTHREAD_COND_CLOCKWAIT] (wait_until): Add a steady_clock
overload.
(wait_until): Change __clock_t overload to use system_clock.
[_GLIBCXX_USE_PTHREAD_COND_CLOCKWAIT] (__wait_until_impl): Add
steady_clock overload that calls pthread_cond_clockwait.
(__wait_until_impl): Change __clock_t overload to use system_clock.
(condition_variable_any) [_GLIBCXX_USE_PTHREAD_COND_CLOCKWAIT]: Use
steady_clock for __clock_t if pthread_cond_clockwait is available.

From-SVN: r275390

Add user-defined clock to libstdc++ condition_variable tests

2019-09-04  Mike Crowe  <mac@mcrowe.com>

* testsuite/30_threads/condition_variable/members/2.cc (test01):
Parameterise so that test can be run against an arbitrary clock.
(main): Test using std::chrono::steady_clock and a user-defined
clock in addition to the previous std::chrono::system_clock.
* testsuite/30_threads/condition_variable_any/members/2.cc: Likewise.

From-SVN: r275389

Disallow -fvtable-verify and -flto to be used together.

Vtable verification currently does not work properly with
link time optimization.  Until this can be fixed users should
not be allowed to specify both options together.

2019-09-04  Caroline Tice  <cmtice@google.com>

        *  opts.c (finish_options):  Disallow -fvtable-verify and -flto to be
           specified together.

From-SVN: r275388

Remove -fdeduce-init-list.

From-SVN: r275387

re PR target/32413 (internal compiler error: in reload_cse_simplify_operands, at postreload.c:396)

PR target/32413
* config/i386/i386.c (inline_secondary_memory_needed): Return true
for QI and HImode moves between SSE and general registers.

From-SVN: r275377

Add warning Wenum-conversion for C and ObjC.

The patch enables warning with Wextra due to PR91593 and warnings with
allmodconfig kernel build. Once these issues are resolved, we could
consider promoting it to Wall.

2019-09-04  Prathamesh Kulkarni  <prathamesh.kulkarni@linaro.org>

PR c/78736
* doc/invoke.texi: Document -Wenum-conversion.

c-family
* c.opt (Wenum-conversion): New option.

c/
* c-typeck.c (convert_for_assignment): Handle Wenum-conversion.

testsuite/
* gcc.dg/Wenum-conversion.c: New test-case.

From-SVN: r275376

mklog: Do not print changed functions in testsuite

2019-09-04  Martin Liska  <mliska@suse.cz>

* mklog: Do not print changed functions for
testsuite files.

From-SVN: r275369

mklog: parse PR references from new test files

2019-09-04  Martin Liska  <mliska@suse.cz>

* mklog: Parse PR references from newly added
test files.

From-SVN: r275368

Use argparse.ArgumentParser for mklog.

2019-09-04  Martin Liska  <mliska@suse.cz>

* mklog: Use argparse instead of getopt.

From-SVN: r275367

re PR middle-end/36262 (Extreme memory usage of VRP compared to older versions)

2019-09-04  Richard Biener  <rguenther@suse.de>

PR rtl-optimization/36262
* postreload-gcse.c: Include intl.h and gcse.h.
(insert_expr_in_table): Insert at the head of cur_expr->avail_occr
to avoid linear list walk.
(record_last_mem_set_info): Gate off if not computing transparentness.
(get_bb_avail_insn): If transparentness isn't computed give up
early.
(gcse_after_reload_main): Skip compute_transp and extended PRE
if gcse_or_cprop_is_too_expensive says so.

From-SVN: r275365

gfortran.texi: Update documentation to catch up with BOZ changes.

2019-09-03  Steven G. Kargl  <kargl@gcc.gnu.org>

* gfortran.texi: Update documentation to catch up with BOZ changes.
* invoke.texi: Fix English from previous BOZ changes commit.

From-SVN: r275364

Daily bump.

From-SVN: r275362

compiler: only import variable into . if same package

    If we dot-import a package, we should only add an imported variable to
    the package bindings if the variable is in the package being imported.
    A test case for this is the 1.13 os package, in which ErrClosed and
    friends are defined both locally and in the imported internal/oserror package.

    Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/192718

From-SVN: r275358

MSP430: Use default_elf_select_section to determine sections for data

2019-09-03  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* config/msp430/msp430.c (msp430_init_sections): Remove handling of the
noinit section.
(msp430_select_section): Handle decls with the "noinit" attribute with
default_elf_select_section.
Handle SECCAT_RODATA_MERGE_* section types with
default_elf_select_section.
Add comments about handling of unsupported section types.
(msp430_section_type_flags): Remove handling of the noinit section.

From-SVN: r275357

MSP430: Setup exclusion tables for function and data attributes

gcc/ChangeLog:

2019-09-03  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* config/msp430/msp430.c (msp430_attr): Remove warnings about
conflicting msp430-specific attributes.
(msp430_section_attr): Likewise.
Add warnings about conflicts with generic "noinit" and "section"
attributes.
Fix grammar in -mlarge error message.
(msp430_data_attr): Rename to msp430_persist_attr.
Add warnings about conflicts with generic "noinit" and "section"
attributes.
Add warning for when variable is not initialized.
Chain conditionals which prevent the attribute being added.
(ATTR_EXCL): New helper.
(attr_reent_exclusions): New exclusion table.
(attr_naked_exclusions): Likewise.
(attr_crit_exclusions): Likewise.
(attr_lower_exclusions): Likewise.
(attr_upper_exclusions): Likewise.
(attr_either_exclusions): Likewise.
(attr_persist_exclusions): Likewise.
(msp430_attribute_table): Update with exclusion rules.
(msp430_output_aligned_decl_common): Don't output common symbol if decl
has a section.

gcc/testsuite/ChangeLog:

2019-09-03  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* gcc.target/msp430/data-attributes-2.c: New test.
* gcc.target/msp430/function-attributes-4.c: Update dg-warning
strings.
* gcc.target/msp430/region-attribute-misuse.c: Likewise.

From-SVN: r275356

Implement TARGET_HANDLE_GENERIC_ATTRIBUTE

gcc/ChangeLog:

2019-09-03  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* config/msp430/msp430.c (TARGET_HANDLE_GENERIC_ATTRIBUTE): Define.
(msp430_handle_generic_attribute): New function.
* doc/tm.texi: Regenerate.
* doc/tm.texi.in: Add TARGET_HANDLE_GENERIC_ATTRIBUTE.
* hooks.c (hook_tree_treeptr_tree_tree_int_boolptr_null): New.
* hooks.h (hook_tree_treeptr_tree_tree_int_boolptr_null): New.
* target.def: Define new hook TARGET_HANDLE_GENERIC_ATTRIBUTE.

gcc/c-family/ChangeLog:

2019-09-03  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* c-attribs.c (handle_section_attribute): Call the
handle_generic_attribute target hook after performing target
independent processing.
(handle_noinit_attribute): Likewise.

From-SVN: r275355

re PR tree-optimization/91504 (Inlining misses some logical operation folding)

PR tree-optimization/91504
* match.pd: Add ((~a & b) ^a) --> (a | b).

        PR tree-optimization/91504
gcc.dg/tree-ssa/pr91504.c: New test.

From-SVN: r275354

rust-demangle.c (unescape): Remove.

* rust-demangle.c (unescape): Remove.
(parse_lower_hex_nibble): New function.
(parse_legacy_escape): New function.
(is_prefixed_hash): Use parse_lower_hex_nibble.
(looks_like_rust): Use parse_legacy_escape.
(rust_demangle_sym): Use parse_legacy_escape.
* testsuite/rust-demangle-expected: Add 'llv$u6d$' test.

From-SVN: r275353

c-cppbuiltin.c (builtin_define_with_hex_fp_value): Always expand when using -fgo-dump-spec.

* c-cppbuiltin.c (builtin_define_with_hex_fp_value): Always expand
when using -fgo-dump-spec.

From-SVN: r275352

re PR target/91604 (ICE in extract_insn at recog.c:2310 since r272323)

PR target/91604
* config/i386/i386-expand.c (split_double_mode): If there is more than
one MEM operand and they are rtx_equal_p, reuse lo_half/hi_half from
already split matching MEM operand instead of calling adjust_address
again.

* gcc.target/i386/pr91604.c: New test.

From-SVN: r275344

Remove Cell Broadband Engine SPU targets

From-SVN: r275343

re PR middle-end/91603 (Unaligned access in expand_assignment)

2019-09-03  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR middle-end/91603
        PR middle-end/91612
        PR middle-end/91613
        * expr.c (expand_expr_real_1): Handle unaligned decl_rtl
        and SSA_NAME referring to CONSTANT_P correctly.

testsuite:
2019-09-03  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR middle-end/91603
        * testsuite/gcc.target/arm/pr91603.c: New test.

From-SVN: r275342

re PR other/79543 (Inappropriate "ld --version" checking)

2019-09-03  Chung-Lin Tang <cltang@codesourcery.com>

libatomic/
PR other/79543
* acinclude.m4 (LIBAT_CHECK_LINKER_FEATURES): Fix GNU ld --version
scanning to conform to the GNU Coding Standards.
* configure: Regenerate.

libffi/
PR other/79543
* acinclude.m4 (LIBAT_CHECK_LINKER_FEATURES): Fix GNU ld --version
scanning to conform to the GNU Coding Standards.
* configure: Regenerate.

libgomp/
PR other/79543
* acinclude.m4 (LIBGOMP_CHECK_LINKER_FEATURES): Fix GNU ld --version
scanning to conform to the GNU Coding Standards.
* configure: Regenerate.

libitm/
PR other/79543
* acinclude.m4 (LIBITM_CHECK_LINKER_FEATURES): Fix GNU ld --version
scanning to conform to the GNU Coding Standards.
* configure: Regenerate.

libstdc++-v3/
PR other/79543
* acinclude.m4 (GLIBCXX_CHECK_LINKER_FEATURES): Fix GNU ld --version
scanning to conform to the GNU Coding Standards.
* configure: Regenerate.

From-SVN: r275341

tree-ssa-sccvn.h (vn_nary_op_lookup): Remove.

2019-09-03  Richard Biener  <rguenther@suse.de>

* tree-ssa-sccvn.h (vn_nary_op_lookup): Remove.
(vn_nary_op_insert): Likewise.
* tree-ssa-sccvn.c (init_vn_nary_op_from_op): Remove.
(vn_nary_op_lookup): Likewise.
(vn_nary_op_insert): Likewise.

From-SVN: r275338

S/390: Commit forgotten test for r275336

gcc/testsuite/ChangeLog:

2019-09-03  Ilya Leoshkevich  <iii@linux.ibm.com>

* gcc.target/s390/sigfpe-eh.c: Forgotten test.

From-SVN: r275337

S/390: Fix failing RTL check in s390_canonicalize_comparison

The new sigfpe-eh.c fails with

    internal compiler error: RTL check: expected elt 0 type 'e' or 'u', have 'w' (rtx const_int)

This is most likely due to a typo: XEXP (*op1, 0) was used, when
XEXP (*op0, 1) was intended.  This did not cause any user-visible
problems, because reversed_comparison_code_parts ignores the
respective argument, and the release compiler is built without RTL
checks.

gcc/ChangeLog:

2019-09-03  Ilya Leoshkevich  <iii@linux.ibm.com>

* config/s390/s390.c (s390_canonicalize_comparison): Use XEXP
(*op0, 1) instead of XEXP (*op1, 0).

gcc/testsuite/ChangeLog:

2019-09-03  Ilya Leoshkevich  <iii@linux.ibm.com>

* gcc.target/s390/sigfpe-eh.c: New test.

From-SVN: r275336