poly_int: MEM_OFFSET and MEM_SIZE

This patch changes the MEM_OFFSET and MEM_SIZE memory attributes
from HOST_WIDE_INT to poly_int64.  Most of it is mechanical,
but there is one nonbovious change in widen_memory_access.
Previously the main while loop broke with:

      /* Similarly for the decl.  */
      else if (DECL_P (attrs.expr)
               && DECL_SIZE_UNIT (attrs.expr)
               && TREE_CODE (DECL_SIZE_UNIT (attrs.expr)) == INTEGER_CST
               && compare_tree_int (DECL_SIZE_UNIT (attrs.expr), size) >= 0
               && (! attrs.offset_known_p || attrs.offset >= 0))
        break;

but it seemed wrong to optimistically assume the best case
when the offset isn't known (and thus might be negative).
As it happens, the "! attrs.offset_known_p" condition was
always false, because we'd already nullified attrs.expr in
that case:

  /* If we don't know what offset we were at within the expression, then
     we can't know if we've overstepped the bounds.  */
  if (! attrs.offset_known_p)
    attrs.expr = NULL_TREE;

The patch therefore drops "! attrs.offset_known_p ||" when
converting the offset check to the may/must interface.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* rtl.h (mem_attrs): Add a default constructor.  Change size and
offset from HOST_WIDE_INT to poly_int64.
* emit-rtl.h (set_mem_offset, set_mem_size, adjust_address_1)
(adjust_automodify_address_1, set_mem_attributes_minus_bitpos)
(widen_memory_access): Take the sizes and offsets as poly_int64s
rather than HOST_WIDE_INTs.
* alias.c (ao_ref_from_mem): Handle the new form of MEM_OFFSET.
(offset_overlap_p): Take poly_int64s rather than HOST_WIDE_INTs
and ints.
(adjust_offset_for_component_ref): Change the offset from a
HOST_WIDE_INT to a poly_int64.
(nonoverlapping_memrefs_p): Track polynomial offsets and sizes.
* cfgcleanup.c (merge_memattrs): Update after mem_attrs changes.
* dce.c (find_call_stack_args): Likewise.
* dse.c (record_store): Likewise.
* dwarf2out.c (tls_mem_loc_descriptor, dw_sra_loc_expr): Likewise.
* print-rtl.c (rtx_writer::print_rtx): Likewise.
* read-rtl-function.c (test_loading_mem): Likewise.
* rtlanal.c (may_trap_p_1): Likewise.
* simplify-rtx.c (delegitimize_mem_from_attrs): Likewise.
* var-tracking.c (int_mem_offset, track_expr_p): Likewise.
* emit-rtl.c (mem_attrs_eq_p, get_mem_align_offset): Likewise.
(mem_attrs::mem_attrs): New function.
(set_mem_attributes_minus_bitpos): Change bitpos from a
HOST_WIDE_INT to poly_int64.
(set_mem_alias_set, set_mem_addr_space, set_mem_align, set_mem_expr)
(clear_mem_offset, clear_mem_size, change_address)
(get_spill_slot_decl, set_mem_attrs_for_spill): Directly
initialize mem_attrs.
(set_mem_offset, set_mem_size, adjust_address_1)
(adjust_automodify_address_1, offset_address, widen_memory_access):
Likewise.  Take poly_int64s rather than HOST_WIDE_INT.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255875

poly_int: rtx_addr_can_trap_p_1

This patch changes the offset and size arguments of
rtx_addr_can_trap_p_1 from HOST_WIDE_INT to poly_int64.  It also
uses a size of -1 rather than 0 to represent an unknown size and
BLKmode rather than VOIDmode to represent an unknown mode.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* rtlanal.c (rtx_addr_can_trap_p_1): Take the offset and size
as poly_int64s rather than HOST_WIDE_INTs.  Use a size of -1
rather than 0 to represent an unknown size.  Assert that the size
is known when the mode isn't BLKmode.
(may_trap_p_1): Use -1 for unknown sizes.
(rtx_addr_can_trap_p): Likewise.  Pass BLKmode rather than VOIDmode.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255874

poly_int: dse.c

This patch makes RTL DSE use poly_int for offsets and sizes.
The local phase can optimise them normally but the global phase
treats them as wild accesses.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* dse.c (store_info): Change offset and width from HOST_WIDE_INT
to poly_int64.  Update commentary for positions_needed.large.
(read_info_type): Change offset and width from HOST_WIDE_INT
to poly_int64.
(set_usage_bits): Likewise.
(canon_address): Return the offset as a poly_int64 rather than
a HOST_WIDE_INT.  Use strip_offset_and_add.
(set_all_positions_unneeded, any_positions_needed_p): Use
positions_needed.large to track stores with non-constant widths.
(all_positions_needed_p): Likewise.  Take the offset and width
as poly_int64s rather than ints.  Assert that rhs is nonnull.
(record_store): Cope with non-constant offsets and widths.
Nullify the rhs of an earlier store if we can't tell which bytes
of it are needed.
(find_shift_sequence): Take the access_size and shift as poly_int64s
rather than ints.
(get_stored_val): Take the read_offset and read_width as poly_int64s
rather than HOST_WIDE_INTs.
(check_mem_read_rtx, scan_stores, scan_reads, dse_step5): Handle
non-constant offsets and widths.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255873

poly_int: ao_ref and vn_reference_op_t

This patch changes the offset, size and max_size fields
of ao_ref from HOST_WIDE_INT to poly_int64 and propagates
the change through the code that references it.  This includes
changing the off field of vn_reference_op_struct in the same way.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* inchash.h (inchash::hash::add_poly_int): New function.
* tree-ssa-alias.h (ao_ref::offset, ao_ref::size, ao_ref::max_size):
Use poly_int64 rather than HOST_WIDE_INT.
(ao_ref::max_size_known_p): New function.
* tree-ssa-sccvn.h (vn_reference_op_struct::off): Use poly_int64_pod
rather than HOST_WIDE_INT.
* tree-ssa-alias.c (ao_ref_base): Apply get_ref_base_and_extent
to temporaries until its interface is adjusted to match.
(ao_ref_init_from_ptr_and_size): Handle polynomial offsets and sizes.
(aliasing_component_refs_p, decl_refs_may_alias_p)
(indirect_ref_may_alias_decl_p, indirect_refs_may_alias_p): Take
the offsets and max_sizes as poly_int64s instead of HOST_WIDE_INTs.
(refs_may_alias_p_1, stmt_kills_ref_p): Adjust for changes to
ao_ref fields.
* alias.c (ao_ref_from_mem): Likewise.
* tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Likewise.
* tree-ssa-dse.c (valid_ao_ref_for_dse, normalize_ref)
(clear_bytes_written_by, setup_live_bytes_from_ref, compute_trims)
(maybe_trim_complex_store, maybe_trim_constructor_store)
(live_bytes_read, dse_classify_store): Likewise.
* tree-ssa-sccvn.c (vn_reference_compute_hash, vn_reference_eq):
(copy_reference_ops_from_ref, ao_ref_init_from_vn_reference)
(fully_constant_vn_reference_p, valueize_refs_1): Likewise.
(vn_reference_lookup_3): Likewise.
* tree-ssa-uninit.c (warn_uninitialized_vars): Likewise.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255872

poly_int: indirect_refs_may_alias_p

This patch makes indirect_refs_may_alias_p use ranges_may_overlap_p
rather than ranges_overlap_p.  Unlike the former, the latter can handle
negative offsets, so the fix for PR44852 should no longer be necessary.
It can also handle offset_int, so avoids unchecked truncations to
HOST_WIDE_INT.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* tree-ssa-alias.c (indirect_ref_may_alias_decl_p)
(indirect_refs_may_alias_p): Use ranges_may_overlap_p
instead of ranges_overlap_p.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255871

poly_int: same_addr_size_stores_p

This patch makes tree-ssa-alias.c:same_addr_size_stores_p handle
poly_int sizes and offsets.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* tree-ssa-alias.c (same_addr_size_stores_p): Take the offsets and
sizes as poly_int64s rather than HOST_WIDE_INTs.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255870

poly_int: fold_ctor_reference

This patch changes the offset and size arguments to
fold_ctor_reference from unsigned HOST_WIDE_INT to poly_uint64.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* gimple-fold.h (fold_ctor_reference): Take the offset and size
as poly_uint64 rather than unsigned HOST_WIDE_INT.
* gimple-fold.c (fold_ctor_reference): Likewise.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255869

poly_int: DWARF locations

This patch adds support for DWARF location expressions
that involve polynomial offsets.  It adds a target hook that
says how the runtime invariants used in the offsets should be
represented in DWARF.  SVE vectors have to be a multiple of
128 bits in size, so the GCC port uses the number of 128-bit
blocks minus one as the runtime invariant.  However, in DWARF,
the vector length is exposed via a pseudo "VG" register that
holds the number of 64-bit elements in a vector.  Thus:

  indeterminate 1 == (VG / 2) - 1

The hook needs to be general enough to express this.
Note that in most cases the division and subtraction fold
away into surrounding expressions.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* target.def (dwarf_poly_indeterminate_value): New hook.
* targhooks.h (default_dwarf_poly_indeterminate_value): Declare.
* targhooks.c (default_dwarf_poly_indeterminate_value): New function.
* doc/tm.texi.in (TARGET_DWARF_POLY_INDETERMINATE_VALUE): Document.
* doc/tm.texi: Regenerate.
* dwarf2out.h (build_cfa_loc, build_cfa_aligned_loc): Take the
offset as a poly_int64.
* dwarf2out.c (new_reg_loc_descr): Move later in file.  Take the
offset as a poly_int64.
(loc_descr_plus_const, loc_list_plus_const, build_cfa_aligned_loc):
Take the offset as a poly_int64.
(build_cfa_loc): Likewise.  Use loc_descr_plus_const.
(frame_pointer_fb_offset): Change to a poly_int64.
(int_loc_descriptor): Take the offset as a poly_int64.  Use
targetm.dwarf_poly_indeterminate_value for polynomial offsets.
(based_loc_descr): Take the offset as a poly_int64.
Use strip_offset_and_add to handle (plus X (const)).
Use new_reg_loc_descr instead of an open-coded version of the
previous implementation.
(mem_loc_descriptor): Handle CONST_POLY_INT.
(compute_frame_pointer_to_fb_displacement): Take the offset as a
poly_int64.  Use strip_offset_and_add to handle (plus X (const)).

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255868

poly_int: REG_OFFSET

This patch changes the type of the reg_attrs offset field
from HOST_WIDE_INT to poly_int64 and updates uses accordingly.
This includes changing reg_attr_hasher::hash to use inchash.
(Doing this has no effect on code generation since the only
use of the hasher is to avoid creating duplicate objects.)

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
            Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* rtl.h (reg_attrs::offset): Change from HOST_WIDE_INT to poly_int64.
(gen_rtx_REG_offset): Take the offset as a poly_int64.
* inchash.h (inchash::hash::add_poly_hwi): New function.
* gengtype.c (main): Register poly_int64.
* emit-rtl.c (reg_attr_hasher::hash): Use inchash.  Treat the
offset as a poly_int.
(reg_attr_hasher::equal): Use must_eq to compare offsets.
(get_reg_attrs, update_reg_offset, gen_rtx_REG_offset): Take the
offset as a poly_int64.
(set_reg_attrs_from_value): Treat the offset as a poly_int64.
* print-rtl.c (print_poly_int): New function.
(rtx_writer::print_rtx_operand_code_r): Treat REG_OFFSET as
a poly_int.
* var-tracking.c (track_offset_p, get_tracked_reg_offset): New
functions.
(var_reg_set, var_reg_delete_and_set, var_reg_delete): Use them.
(same_variable_part_p, track_loc_p): Take the offset as a poly_int64.
(vt_get_decl_and_offset): Return the offset as a poly_int64.
Enforce track_offset_p for parts of a PARALLEL.
(vt_add_function_parameter): Use const_offset for the final
offset to track.  Use get_tracked_reg_offset for the parts
of a PARALLEL.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255867

poly_int: TRULY_NOOP_TRUNCATION

This patch makes TRULY_NOOP_TRUNCATION take the mode sizes as
poly_uint64s instead of unsigned ints.  The function bodies
don't need to change.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
            Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* target.def (truly_noop_truncation): Take poly_uint64s instead of
unsigned ints.  Change default to hook_bool_puint64_puint64_true.
* doc/tm.texi: Regenerate.
* hooks.h (hook_bool_uint_uint_true): Delete.
(hook_bool_puint64_puint64_true): Declare.
* hooks.c (hook_bool_uint_uint_true): Delete.
(hook_bool_puint64_puint64_true): New function.
* config/mips/mips.c (mips_truly_noop_truncation): Take poly_uint64s
instead of unsigned ints.
* config/spu/spu.c (spu_truly_noop_truncation): Likewise.
* config/tilegx/tilegx.c (tilegx_truly_noop_truncation): Likewise.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255866

poly_int: create_integer_operand

This patch generalises create_integer_operand so that it accepts
poly_int64s rather than HOST_WIDE_INTs.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* optabs.h (expand_operand): Add an int_value field.
(create_expand_operand): Add an int_value parameter and use it
to initialize the new expand_operand field.
(create_integer_operand): Replace with a declaration of a function
that accepts poly_int64s.  Move the implementation to...
* optabs.c (create_integer_operand): ...here.
(maybe_legitimize_operand): For EXPAND_INTEGER, check whether
the mode preserves the value of int_value, instead of calling
const_int_operand on the rtx.  Use gen_int_mode to generate
the new rtx.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255865

poly_int: dump routines

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* dumpfile.h (dump_dec): Declare.
* dumpfile.c (dump_dec): New function.
* pretty-print.h (pp_wide_integer): Turn into a function and
declare a poly_int version.
* pretty-print.c (pp_wide_integer): New function for poly_ints.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255864

poly_int: tree constants

This patch adds a tree representation for poly_ints.  Unlike the
rtx version, the coefficients are INTEGER_CSTs rather than plain
integers, so that we can easily access them as poly_widest_ints
and poly_offset_ints.

The patch also adjusts some places that previously
relied on "constant" meaning "INTEGER_CST".  It also makes
sure that the TYPE_SIZE agrees with the TYPE_SIZE_UNIT for
vector booleans, given the existing:

/* Several boolean vector elements may fit in a single unit.  */
if (VECTOR_BOOLEAN_TYPE_P (type)
    && type->type_common.mode != BLKmode)
  TYPE_SIZE_UNIT (type)
    = size_int (GET_MODE_SIZE (type->type_common.mode));
else
  TYPE_SIZE_UNIT (type) = int_const_binop (MULT_EXPR,
   TYPE_SIZE_UNIT (innertype),
   size_int (nunits));

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* doc/generic.texi (POLY_INT_CST): Document.
* tree.def (POLY_INT_CST): New tree code.
* treestruct.def (TS_POLY_INT_CST): New tree layout.
* tree-core.h (tree_poly_int_cst): New struct.
(tree_node): Add a poly_int_cst field.
* tree.h (POLY_INT_CST_P, POLY_INT_CST_COEFF): New macros.
(wide_int_to_tree, force_fit_type): Take a poly_wide_int_ref
instead of a wide_int_ref.
(build_int_cst, build_int_cst_type): Take a poly_int64 instead
of a HOST_WIDE_INT.
(build_int_cstu, build_array_type_nelts): Take a poly_uint64
instead of an unsigned HOST_WIDE_INT.
(build_poly_int_cst, tree_fits_poly_int64_p, tree_fits_poly_uint64_p)
(ptrdiff_tree_p): Declare.
(tree_to_poly_int64, tree_to_poly_uint64): Likewise.  Provide
extern inline implementations if the target doesn't use POLY_INT_CST.
(poly_int_tree_p): New function.
(wi::unextended_tree): New class.
(wi::int_traits <unextended_tree>): New override.
(wi::extended_tree): Add a default constructor.
(wi::extended_tree::get_tree): New function.
(wi::widest_extended_tree, wi::offset_extended_tree): New typedefs.
(wi::tree_to_widest_ref, wi::tree_to_offset_ref): Use them.
(wi::tree_to_poly_widest_ref, wi::tree_to_poly_offset_ref)
(wi::tree_to_poly_wide_ref): New typedefs.
(wi::ints_for): Provide overloads for extended_tree and
unextended_tree.
(poly_int_cst_value, wi::to_poly_widest, wi::to_poly_offset)
(wi::to_wide): New functions.
(wi::fits_to_boolean_p, wi::fits_to_tree_p): Handle poly_ints.
* tree.c (poly_int_cst_hasher): New struct.
(poly_int_cst_hash_table): New variable.
(tree_node_structure_for_code, tree_code_size, simple_cst_equal)
(valid_constant_size_p, add_expr, drop_tree_overflow): Handle
POLY_INT_CST.
(initialize_tree_contains_struct): Handle TS_POLY_INT_CST.
(init_ttree): Initialize poly_int_cst_hash_table.
(build_int_cst, build_int_cst_type, build_invariant_address): Take
a poly_int64 instead of a HOST_WIDE_INT.
(build_int_cstu, build_array_type_nelts): Take a poly_uint64
instead of an unsigned HOST_WIDE_INT.
(wide_int_to_tree): Rename to...
(wide_int_to_tree_1): ...this.
(build_new_poly_int_cst, build_poly_int_cst): New functions.
(force_fit_type): Take a poly_wide_int_ref instead of a wide_int_ref.
(wide_int_to_tree): New function that takes a poly_wide_int_ref.
(ptrdiff_tree_p, tree_to_poly_int64, tree_to_poly_uint64)
(tree_fits_poly_int64_p, tree_fits_poly_uint64_p): New functions.
* lto-streamer-out.c (DFS::DFS_write_tree_body, hash_tree): Handle
TS_POLY_INT_CST.
* tree-streamer-in.c (lto_input_ts_poly_tree_pointers): Likewise.
(streamer_read_tree_body): Likewise.
* tree-streamer-out.c (write_ts_poly_tree_pointers): Likewise.
(streamer_write_tree_body): Likewise.
* tree-streamer.c (streamer_check_handled_ts_structures): Likewise.
* asan.c (asan_protect_global): Require the size to be an INTEGER_CST.
* cfgexpand.c (expand_debug_expr): Handle POLY_INT_CST.
* expr.c (expand_expr_real_1, const_vector_from_tree): Likewise.
* gimple-expr.h (is_gimple_constant): Likewise.
* gimplify.c (maybe_with_size_expr): Likewise.
* print-tree.c (print_node): Likewise.
* tree-data-ref.c (data_ref_compare_tree): Likewise.
* tree-pretty-print.c (dump_generic_node): Likewise.
* tree-ssa-address.c (addr_for_mem_ref): Likewise.
* tree-vect-data-refs.c (dr_group_sort_cmp): Likewise.
* tree-vrp.c (compare_values_warnv): Likewise.
* tree-ssa-loop-ivopts.c (determine_base_object, constant_multiple_of)
(get_loop_invariant_expr, add_candidate_1, get_computation_aff_1)
(force_expr_to_var_cost): Likewise.
* tree-ssa-loop.c (for_each_index): Likewise.
* fold-const.h (build_invariant_address, size_int_kind): Take a
poly_int64 instead of a HOST_WIDE_INT.
* fold-const.c (fold_negate_expr_1, const_binop, const_unop)
(fold_convert_const, multiple_of_p, fold_negate_const): Handle
POLY_INT_CST.
(size_binop_loc): Likewise.  Allow int_const_binop_1 to fail.
(int_const_binop_2): New function, split out from...
(int_const_binop_1): ...here.  Handle POLY_INT_CST.
(size_int_kind): Take a poly_int64 instead of a HOST_WIDE_INT.
* expmed.c (make_tree): Handle CONST_POLY_INT_P.
* gimple-ssa-strength-reduction.c (slsr_process_add)
(slsr_process_mul): Check for INTEGER_CSTs before using them
as candidates.
* stor-layout.c (bits_from_bytes): New function.
(bit_from_pos): Use it.
(layout_type): Likewise.  For vectors, multiply the TYPE_SIZE_UNIT
by BITS_PER_UNIT to get the TYPE_SIZE.
* tree-cfg.c (verify_expr, verify_types_in_gimple_reference): Allow
MEM_REF and TARGET_MEM_REF offsets to be a POLY_INT_CST.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255863

poly_int: rtx constants

This patch adds an rtl representation of poly_int values.
There were three possible ways of doing this:

(1) Add a new rtl code for the poly_ints themselves and store the
    coefficients as trailing wide_ints.  This would give constants like:

      (const_poly_int [c0 c1 ... cn])

    The runtime value would be:

      c0 + c1 * x1 + ... + cn * xn

(2) Like (1), but use rtxes for the coefficients.  This would give
    constants like:

      (const_poly_int [(const_int c0)
                       (const_int c1)
                       ...
                       (const_int cn)])

    although the coefficients could be const_wide_ints instead
    of const_ints where appropriate.

(3) Add a new rtl code for the polynomial indeterminates,
    then use them in const wrappers.  A constant like c0 + c1 * x1
    would then look like:

      (const:M (plus:M (mult:M (const_param:M x1)
                               (const_int c1))
                       (const_int c0)))

There didn't seem to be that much to choose between them.  The main
advantage of (1) is that it's a more efficient representation and
that we can refer to the cofficients directly as wide_int_storage.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* doc/rtl.texi (const_poly_int): Document.  Also document the
rtl sharing behavior.
* gengenrtl.c (excluded_rtx): Return true for CONST_POLY_INT.
* rtl.h (const_poly_int_def): New struct.
(rtx_def::u): Add a cpi field.
(CASE_CONST_UNIQUE, CASE_CONST_ANY): Add CONST_POLY_INT.
(CONST_POLY_INT_P, CONST_POLY_INT_COEFFS): New macros.
(wi::rtx_to_poly_wide_ref): New typedef
(const_poly_int_value, wi::to_poly_wide, rtx_to_poly_int64)
(poly_int_rtx_p): New functions.
(trunc_int_for_mode): Declare a poly_int64 version.
(plus_constant): Take a poly_int64 instead of a HOST_WIDE_INT.
(immed_wide_int_const): Take a poly_wide_int_ref rather than
a wide_int_ref.
(strip_offset): Declare.
(strip_offset_and_add): New function.
* rtl.def (CONST_POLY_INT): New rtx code.
* rtl.c (rtx_size): Handle CONST_POLY_INT.
(shared_const_p): Use poly_int_rtx_p.
* emit-rtl.h (gen_int_mode): Take a poly_int64 instead of a
HOST_WIDE_INT.
(gen_int_shift_amount): Likewise.
* emit-rtl.c (const_poly_int_hasher): New class.
(const_poly_int_htab): New variable.
(init_emit_once): Initialize it when NUM_POLY_INT_COEFFS > 1.
(const_poly_int_hasher::hash): New function.
(const_poly_int_hasher::equal): Likewise.
(gen_int_mode): Take a poly_int64 instead of a HOST_WIDE_INT.
(immed_wide_int_const): Rename to...
(immed_wide_int_const_1): ...this and make static.
(immed_wide_int_const): New function, taking a poly_wide_int_ref
instead of a wide_int_ref.
(gen_int_shift_amount): Take a poly_int64 instead of a HOST_WIDE_INT.
(gen_lowpart_common): Handle CONST_POLY_INT.
* cse.c (hash_rtx_cb, equiv_constant): Likewise.
* cselib.c (cselib_hash_rtx): Likewise.
* dwarf2out.c (const_ok_for_output_1): Likewise.
* expr.c (convert_modes): Likewise.
* print-rtl.c (rtx_writer::print_rtx, print_value): Likewise.
* rtlhash.c (add_rtx): Likewise.
* explow.c (trunc_int_for_mode): Add a poly_int64 version.
(plus_constant): Take a poly_int64 instead of a HOST_WIDE_INT.
Handle existing CONST_POLY_INT rtxes.
* expmed.h (expand_shift): Take a poly_int64 instead of a
HOST_WIDE_INT.
* expmed.c (expand_shift): Likewise.
* rtlanal.c (strip_offset): New function.
(commutative_operand_precedence): Give CONST_POLY_INT the same
precedence as CONST_DOUBLE and put CONST_WIDE_INT between that
and CONST_INT.
* rtl-tests.c (const_poly_int_tests): New struct.
(rtl_tests_c_tests): Use it.
* simplify-rtx.c (simplify_const_unary_operation): Handle
CONST_POLY_INT.
(simplify_const_binary_operation): Likewise.
(simplify_binary_operation_1): Fold additions of symbolic constants
and CONST_POLY_INTs.
(simplify_subreg): Handle extensions and truncations of
CONST_POLY_INTs.
(simplify_const_poly_int_tests): New struct.
(simplify_rtx_c_tests): Use it.
* wide-int.h (storage_ref): Add default constructor.
(wide_int_ref_storage): Likewise.
(trailing_wide_ints): Use GTY((user)).
(trailing_wide_ints::operator[]): Add a const version.
(trailing_wide_ints::get_precision): New function.
(trailing_wide_ints::extra_size): Likewise.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255862

Add a gen_int_shift_amount helper function

This patch adds a helper routine that constructs rtxes
for constant shift amounts, given the mode of the value
being shifted.  As well as helping with the SVE patches, this
is one step towards allowing CONST_INTs to have a real mode.

One long-standing problem has been to decide what the mode
of a shift count should be for arbitrary rtxes (as opposed to those
directly tied to a target pattern).  Realistic choices would be
the mode of the shifted elements, word_mode, QImode, a 64-bit mode,
or the same mode as the shift optabs (in which case what should the
mode be when the target doesn't have a pattern?)

For now the patch picks a 64-bit mode, but with a ??? comment.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* emit-rtl.h (gen_int_shift_amount): Declare.
* emit-rtl.c (gen_int_shift_amount): New function.
* asan.c (asan_emit_stack_protection): Use gen_int_shift_amount
instead of GEN_INT.
* calls.c (shift_return_value): Likewise.
* cse.c (fold_rtx): Likewise.
* dse.c (find_shift_sequence): Likewise.
* expmed.c (init_expmed_one_mode, store_bit_field_1, expand_shift_1)
(expand_shift, expand_smod_pow2): Likewise.
* lower-subreg.c (shift_cost): Likewise.
* optabs.c (expand_superword_shift, expand_doubleword_mult)
(expand_unop, expand_binop, shift_amt_for_vec_perm_mask)
(expand_vec_perm_var): Likewise.
* simplify-rtx.c (simplify_unary_operation_1): Likewise.
(simplify_binary_operation_1): Likewise.
* combine.c (try_combine, find_split_point, force_int_to_mode)
(simplify_shift_const_1, simplify_shift_const): Likewise.
(change_zero_ext): Likewise.  Use simplify_gen_binary.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255861

Fix multiple_p for two non-poly_ints

Fix a stupid inversion.  This function is very rarely used and was
mostly to help split patches up, which is why it didn't get picked
up during initial testing.

2017-12-20  Richard Sandiford  <richard.sandiford@linaro.org>

gcc/
* poly-int.h (multiple_p): Fix handling of two non-poly_ints.

gcc/testsuite/
* gcc.dg/plugin/poly-int-tests.h (test_nonpoly_multiple_p): New
function.
(test_nonpoly_type): Call it.

From-SVN: r255860

[arm][doc] Document accepted -march=armv8.3-a extension options

I noticed that we helpfully list the extensions that are accepted
by the -march options on arm but we were missing the information
for 'armv8.3-a'.

This patchlet corrects that.
Built the documentation and it looked ok.

* doc/invoke.texi (ARM Options): Document accepted extension options
for -march=armv8.3-a.

From-SVN: r255859

[arm] PR target/83105: Minor change of default CPU for arm-linux-gnueabi

When GCC for ARM/linux is configured with --with-float=hard, or
--with-float=softfp the compiler will now die when trying to build the
support libraries because the baseline architecture is too old to
support VFP (older versions of GCC just emitted the VFP instructions
anyway, even though they wouldn't run on that version of the
architecture; but we're now more prickly about it).

This patch fixed the problem by raising the default architecture
(actually the default CPU) to ARMv5te (ARM10e) when we need to generate
HW floating-point code.

PR target/83105
* config.gcc (arm*-*-linux*): When configured with --with-float=hard
or --with-float=softfp, set the default CPU to arm10e.

From-SVN: r255858

[aarch64][libstdc++] Use __ARM_BIG_ENDIAN instead of __AARCH64EB__ in opt_random.h

As has been spotted at https://gcc.gnu.org/ml/gcc-patches/2017-12/msg01289.html
we check the __AARCH64EB__ macro for aarch64 big-endian
detection in config/cpu/aarch64/opt/ext/opt_random.h.
That works just fine with GCC but the standardised ACLE[1] macro
for that purpose is __ARM_BIG_ENDIAN so there is a possibility
that non-GCC compilers that include this header are not aware
of this predefine.

So this patch changes the use of __AARCH64EB__ to
the more portable __ARM_BIG_ENDIAN.

Tested on aarch64-none-elf and aarch64_be-none-elf.

Preapproved by Jeff at https://gcc.gnu.org/ml/gcc-patches/2017-12/msg01326.html

* config/cpu/aarch64/opt/ext/opt_random.h (__VEXT): Check
__ARM_BIG_ENDIAN instead of __AARCH64EB__.

From-SVN: r255857

constraints.md (J, K, L): Use IN_RANGE macro.

* config/visium/constraints.md (J, K, L): Use IN_RANGE macro.
* config/visium/predicates.md (const_shift_operand): Likewise.
* config/visium/visium.c (visium_legitimize_address): Fix oversight.
(visium_legitimize_reload_address): Likewise.

From-SVN: r255856

2017-12-20  Paolo Carlini  <paolo.carlini@oracle.com>

* Committing ChangeLog entry.

From-SVN: r255855

trans.c (Loop_Statement_to_gnu): Use IN_RANGE macro.

* gcc-interface/trans.c (Loop_Statement_to_gnu): Use IN_RANGE macro.
* gcc-interface/misc.c (gnat_get_array_descr_info): Likewise.
(default_pass_by_ref): Likewise.
* gcc-interface/decl.c (gnat_to_gnu_entity): Likewise.

From-SVN: r255854

[arm] PR target/82975: Guard against reg_renumber being NULL in arm.h

Commit missing hunk to arm.h TEST_REGNO comment.

PR target/82975
* config/arm/arm.h (TEST_REGNO): Adjust comment as expected in
r255830.

From-SVN: r255853

re PR c++/83490 (ICE in find_call_stack_args, at dce.c:392)

PR c++/83490
* calls.c (compute_argument_addresses): Ignore TYPE_EMPTY_P arguments.

* g++.dg/abi/empty29.C: New test.

From-SVN: r255852

Add two test-cases for (PR middle-end/82404).

2017-12-20  Martin Liska  <mliska@suse.cz>

PR middle-end/82404
* g++.dg/pr82404.C: New test.
* gcc.dg/pr82404.c: New test.

From-SVN: r255851

Enable VPCLMULQDQ support

gcc/
* common/config/i386/i386-common.c (OPTION_MASK_ISA_VPCLMULQDQ_SET,
OPTION_MASK_ISA_VPCLMULQDQ_UNSET): New.
(ix86_handle_option): Handle -mvpclmulqdq, move cx6 to flags2.
* config.gcc: Include vpclmulqdqintrin.h.
* config/i386/cpuid.h: Handle bit_VPCLMULQDQ.
* config/i386/driver-i386.c (host_detect_local_cpu): Handle -mvpclmulqdq.
* config/i386/i386-builtin.def (__builtin_ia32_vpclmulqdq_v2di,
__builtin_ia32_vpclmulqdq_v4di, __builtin_ia32_vpclmulqdq_v8di): New.
* config/i386/i386-c.c (__VPCLMULQDQ__): New.
* config/i386/i386.c (isa2_opts): Add -mcx16.
(isa_opts): Add -mpclmulqdq, remove -mcx16.
(ix86_option_override_internal): Move mcx16 to flags2.
(ix86_valid_target_attribute_inner_p): Add vpclmulqdq.
(ix86_expand_builtin): Handle OPTION_MASK_ISA_VPCLMULQDQ.
* config/i386/i386.h (TARGET_VPCLMULQDQ, TARGET_VPCLMULQDQ_P): New.
* config/i386/i386.opt: Add mvpclmulqdq, move mcx16 to flags2.
* config/i386/immintrin.h: Include vpclmulqdqintrin.h.
* config/i386/sse.md (vpclmulqdq_<mode>): New pattern.
* config/i386/vpclmulqdqintrin.h (_mm512_clmulepi64_epi128,
_mm_clmulepi64_epi128, _mm256_clmulepi64_epi128): New intrinsics.
* doc/invoke.texi: Add -mvpclmulqdq.

gcc/testsuite/
* gcc.target/i386/avx-1.c: Handle new intrinsics.
* gcc.target/i386/sse-13.c: Ditto.
* gcc.target/i386/sse-23.c: Ditto.
* gcc.target/i386/avx512-check.h: Handle bit_VPCLMULQDQ.
* gcc.target/i386/avx512f-vpclmulqdq-2.c: New test.
* gcc.target/i386/avx512vl-vpclmulqdq-2.c: Ditto.
* gcc.target/i386/vpclmulqdq.c: Ditto.
* gcc.target/i386/i386.exp (check_effective_target_vpclmulqdq): New.

From-SVN: r255850

Don't call targetm.calls.static_chain in non-static function

2017-12-20  Tom de Vries  <tom@codesourcery.com>

PR middle-end/83423
* config/i386/i386.c (ix86_static_chain): Move DECL_STATIC_CHAIN test ...
* calls.c (rtx_for_static_chain): ... here.  New function.
* calls.h (rtx_for_static_chain): Declare.
* builtins.c (expand_builtin_setjmp_receiver): Use rtx_for_static_chain
instead of targetm.calls.static_chain.
* df-scan.c (df_get_entry_block_def_set): Same.

From-SVN: r255849

Daily bump.

From-SVN: r255848

re PR c++/82593 (Internal compiler error: in process_init_constructor_array, at cp/typeck2.c:1294)

/cp
2017-12-19  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/82593
* decl.c (check_array_designated_initializer): Not static.
* cp-tree.h (check_array_designated_initializer): Declare.
* typeck2.c (process_init_constructor_array): Call the latter.
* parser.c (cp_parser_initializer_list): Check the return value
of require_potential_rvalue_constant_expression.

/testsuite
2017-12-19  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/82593
* g++.dg/cpp0x/desig2.C: New.
* g++.dg/cpp0x/desig3.C: Likewise.
* g++.dg/cpp0x/desig4.C: Likewise.

From-SVN: r255845

PR c++/83394 - always_inline vs. noinline no longer diagnosed

PR c++/83394 - always_inline vs. noinline no longer diagnosed
PR c++/83322 - ICE: tree check: expected class ‘type’, have ‘exceptional’

gcc/cp/ChangeLog:

PR c++/83394
PR c++/83322
* decl2.c (cplus_decl_attributes): Look up member functions
in the scope of their class.

gcc/testsuite/ChangeLog:

PR c++/83394
* g++.dg/Wattributes-3.C: New test.
* g++.dg/Wattributes-4.C: New test.
* g++.dg/Wattributes-5.C: New test.

From-SVN: r255844

re PR target/82975 (ICE in baseness at rtlanal.c:6220)

PR target/82975
* gcc.dg/pr82975.c: Only add -mtune=cortex-a57 on arm*/aarch64*
targets.

From-SVN: r255843

Fix sharing in translate_isl_ast_node_for

2017-12-19  Tom de Vries  <tom@codesourcery.com>

PR tree-optimization/83493
* graphite-isl-ast-to-gimple.c (translate_isl_ast_node_for): Unshare ub
and lb.

From-SVN: r255842

Fix last ChangeLog entry date.

From-SVN: r255841

gimple-ssa-sprintf.c (format_directive): Use inform_n instead of inform with hardcoded english plural handling.

* gimple-ssa-sprintf.c (format_directive): Use inform_n instead of
inform with hardcoded english plural handling.

From-SVN: r255840

re PR c++/82231 (ICE when deducing non-type template parameter value whose type depends on a non-type `auto` template parameter from function arguments)

PR c++/82231
* g++.dg/cpp1z/nontype-auto14.C: New test.

From-SVN: r255839

re PR middle-end/83477 (Wrong code w/ -O1)

PR tree-optimization/83477
* tree-ssa-threadedge.c (record_temporary_equivalences_from_phis): For
a non-virtual PHI, always push a new range.

PR tree-optimization/83477
* gcc.c-torture/execute/pr83477.c: New test.

From-SVN: r255837

PR middle-end/77608 - missing protection on trivially detectable runtime buffer overflow

gcc/ChangeLog:

PR middle-end/77608
* builtins.c (compute_objsize): Handle non-constant offsets.

gcc/testsuite/ChangeLog:

PR middle-end/77608
* gcc.dg/Wstringop-overflow.c: New test.
* gcc/testsuite/c-c++-common/Warray-bounds-3.c: Adjust.

From-SVN: r255836

re PR tree-optimization/83444 (missing strlen optimization on a member array of a local struct)

PR tree-optimization/83444
* tree-ssa-strlen.c (strlen_check_and_optimize_stmt): For the
character load case, if get_stridx on MEM_REF's operand doesn't
look usable, retry with get_addr_stridx.

* gcc.dg/strlenopt-38.c: New test.

From-SVN: r255835

SFN: don't drop markers for skipping var-tracking

Although debug markers are more useful when bind stmts are placed
among them, there is value in keeping them even when VTA limits are
exceeded.

for  gcc/ChangeLog

PR debug/83422
* var-tracking.c (vt_debug_insns_local): Do not drop markers.
(variable_tracking_main_1): Keep markers even when VTA fails.

for  gcc/testsuite/ChangeLog

PR debug/83422
* gcc.dg/pr83422.c: New.

From-SVN: r255834

[SFN] start rtl block with label, then markers

Emitting markers before labels turned out to not be worth the trouble.
The markers outside BBs confuse the ebb scheduler, and they don't add
any useful information.  I'll arrange for markers to be moved past
labels, even in gimple, but for now this will fix the two remaining
known problems on ia64.

for  gcc/ChangeLog

PR bootstrap/83396
* cfgexpand.c (expand_gimple_basic_block): Expand label first,
even if there are markers before it.
* cfgrtl.c (rtl_verify_bb_layout): Reject DEBUG_INSNs outside BBs.

From-SVN: r255833

re PR testsuite/83454 (FAIL: gcc.dg/tree-ssa/cswtch-4.c and cswtch-5.c)

PR testsuite/83454
* gcc.dg/tree-ssa/cswtch-4.c: Require nonpic effective target.
* gcc.dg/tree-ssa/cswtch-5.c: Likewise.

From-SVN: r255832

read-rtl.c (parse_reg_note_name): Replace Yoda conditions with typical order conditions.

* read-rtl.c (parse_reg_note_name): Replace Yoda conditions with
typical order conditions.
* sel-sched.c (extract_new_fences_from): Likewise.
* config/visium/constraints.md (J, K, L): Likewise.
* config/visium/predicates.md (const_shift_operand): Likewise.
* config/visium/visium.c (visium_legitimize_address,
visium_legitimize_reload_address): Likewise.
* config/m68k/m68k.c (output_reg_adjust, emit_reg_adjust): Likewise.
* config/arm/arm.c (arm_block_move_unaligned_straight): Likewise.
* config/avr/constraints.md (Y01, Ym1, Y02, Ym2): Likewise.
* config/avr/avr-log.c (avr_vdump, avr_log_set_avr_log,
SET_DUMP_DETAIL): Likewise.
* config/avr/predicates.md (const_8_16_24_operand): Likewise.
* config/avr/avr.c (STR_PREFIX_P, avr_popcount_each_byte,
avr_is_casesi_sequence, avr_casei_sequence_check_operands,
avr_set_core_architecture, avr_set_current_function,
avr_legitimize_reload_address, avr_asm_len, avr_print_operand,
output_movqi, output_movsisf, avr_out_plus, avr_out_bitop,
avr_out_fract, avr_adjust_insn_length, avr_encode_section_info,
avr_2word_insn_p, output_reload_in_const, avr_has_nibble_0xf,
avr_map_decompose, avr_fold_builtin): Likewise.
* config/avr/driver-avr.c (avr_devicespecs_file): Likewise.
* config/avr/gen-avr-mmcu-specs.c (str_prefix_p, print_mcu): Likewise.
* config/i386/i386.c (ix86_parse_stringop_strategy_string): Likewise.
* config/m32c/m32c-pragma.c (m32c_pragma_memregs): Likewise.
* config/m32c/m32c.c (m32c_conditional_register_usage,
m32c_address_cost): Likewise.
* config/m32c/predicates.md (shiftcount_operand,
longshiftcount_operand): Likewise.
* config/iq2000/iq2000.c (iq2000_expand_prologue): Likewise.
* config/nios2/nios2.c (nios2_handle_custom_fpu_insn_option,
can_use_cdx_ldstw): Likewise.
* config/nios2/nios2.h (CDX_REG_P): Likewise.
* config/cr16/cr16.h (RETURN_ADDR_RTX, REGNO_MODE_OK_FOR_BASE_P):
Likewise.
* config/cr16/cr16.md (*mov<mode>_double): Likewise.
* config/cr16/cr16.c (cr16_create_dwarf_for_multi_push): Likewise.
* config/h8300/h8300.c (h8300_rtx_costs, get_shift_alg): Likewise.
* config/vax/constraints.md (U06, U08, U16, CN6, S08, S16): Likewise.
* config/vax/vax.c (adjacent_operands_p): Likewise.
* config/ft32/constraints.md (L, b, KA): Likewise.
* config/ft32/ft32.c (ft32_load_immediate, ft32_expand_prologue):
Likewise.
* cfgexpand.c (expand_stack_alignment): Likewise.
* gcse.c (insert_expr_in_table): Likewise.
* print-rtl.c (rtx_writer::print_rtx_operand_codes_E_and_V): Likewise.
* cgraphunit.c (cgraph_node::expand): Likewise.
* ira-build.c (setup_min_max_allocno_live_range_point): Likewise.
* emit-rtl.c (add_insn): Likewise.
* input.c (dump_location_info): Likewise.
* passes.c (NEXT_PASS): Likewise.
* read-rtl-function.c (parse_note_insn_name,
function_reader::read_rtx_operand_r, function_reader::parse_mem_expr):
Likewise.
* sched-rgn.c (sched_rgn_init): Likewise.
* diagnostic-show-locus.c (layout::show_ruler): Likewise.
* combine.c (find_split_point, simplify_if_then_else, force_to_mode,
if_then_else_cond, simplify_shift_const_1, simplify_comparison): Likewise.
* explow.c (eliminate_constant_term): Likewise.
* final.c (leaf_renumber_regs_insn): Likewise.
* cfgrtl.c (print_rtl_with_bb): Likewise.
* genhooks.c (emit_init_macros): Likewise.
* poly-int.h (maybe_ne, maybe_le, maybe_lt): Likewise.
* tree-data-ref.c (conflict_fn): Likewise.
* selftest.c (assert_streq): Likewise.
* expr.c (store_constructor_field, expand_expr_real_1): Likewise.
* fold-const.c (fold_range_test, extract_muldiv_1, fold_truth_andor,
fold_binary_loc, multiple_of_p): Likewise.
* reload.c (push_reload, find_equiv_reg): Likewise.
* et-forest.c (et_nca, et_below): Likewise.
* dbxout.c (dbxout_symbol_location): Likewise.
* reorg.c (relax_delay_slots): Likewise.
* dojump.c (do_compare_rtx_and_jump): Likewise.
* gengtype-parse.c (type): Likewise.
* simplify-rtx.c (simplify_gen_ternary, simplify_gen_relational,
simplify_const_relational_operation): Likewise.
* reload1.c (do_output_reload): Likewise.
* dumpfile.c (get_dump_file_info_by_switch): Likewise.
* gengtype.c (type_for_name): Likewise.
* gimple-ssa-sprintf.c (format_directive): Likewise.
ada/
* gcc-interface/trans.c (Loop_Statement_to_gnu): Replace Yoda
conditions with typical order conditions.
* gcc-interface/misc.c (gnat_get_array_descr_info,
default_pass_by_ref): Likewise.
* gcc-interface/decl.c (gnat_to_gnu_entity): Likewise.
* adaint.c (__gnat_tmp_name): Likewise.
c-family/
* known-headers.cc (get_stdlib_header_for_name): Replace Yoda
conditions with typical order conditions.
c/
* c-typeck.c (comptypes_internal, function_types_compatible_p,
perform_integral_promotions, digest_init): Replace Yoda conditions
with typical order conditions.
* c-decl.c (check_bitfield_type_and_width): Likewise.
cp/
* name-lookup.c (get_std_name_hint): Replace Yoda conditions with
typical order conditions.
* class.c (check_bitfield_decl): Likewise.
* pt.c (convert_template_argument): Likewise.
* decl.c (duplicate_decls): Likewise.
* typeck.c (commonparms): Likewise.
fortran/
* scanner.c (preprocessor_line): Replace Yoda conditions with typical
order conditions.
* dependency.c (check_section_vs_section): Likewise.
* trans-array.c (gfc_conv_expr_descriptor): Likewise.
jit/
* jit-playback.c (get_type, playback::compile_to_file::copy_file,
playback::context::acquire_mutex): Replace Yoda conditions with
typical order conditions.
* libgccjit.c (gcc_jit_context_new_struct_type,
gcc_jit_struct_set_fields, gcc_jit_context_new_union_type,
gcc_jit_context_new_function, gcc_jit_timer_pop): Likewise.
* jit-builtins.c (matches_builtin): Likewise.
* jit-recording.c (recording::compound_type::set_fields,
recording::fields::write_reproducer, recording::rvalue::set_scope,
recording::function::validate): Likewise.
* jit-logging.c (logger::decref): Likewise.

From-SVN: r255831

[arm] PR target/82975: Guard against reg_renumber being NULL in arm.h

In this bug we ICE when checking REGNO_OK_FOR_INDEX_P on arm during pre-IRA scheduling.
This is because REGNO_OK_FOR_INDEX_P ends up checking the reg_renumber array.
Before IRA reg_renumber is NULL and thus we segfault.

The fix is to guard the use of reg_renumber in the logic in TEST_REGNO in arm.h.
On aarch64, for example, we also guard against the reg_renumber == NULL case.
This fixes the ICE. I also remove the part of the comment that muses on when reg_renumber
is available as with this patch it should now be safe to use at any point.

Bootstrapped and tested on arm-none-linux-gnueabihf.

PR target/82975
* config/arm/arm.h (TEST_REGNO): Check reg_renumber is set before
accessing it.  Adjust comment.

* gcc.dg/pr82975.c: New test.

From-SVN: r255830

re PR middle-end/81914 (gcc 7.1 generates branch for code which was branchless in earlier gcc version)

PR middle-end/81914
* predict.c (zero_one_minusone): New function.
(apply_return_prediction): Avoid return prediction for functions
returning only -1, 0 and 1 values, unless they only return -1 and 0
or 0 and 1.

From-SVN: r255829

[ARC][COMMITTED] Clean up build warnings.

2017-12-19  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (legitimate_scaled_address_p): Clean
fall-through warning.
(arc_compute_frame_size): Remove unused variables.
(arc_print_operand): Fix fprintif format.
(arc_can_follow_jump): Clean fall-through warning.

From-SVN: r255827

Fix a file name.

From-SVN: r255826

Fix PR numbers.

From-SVN: r255825

re PR c++/83490 (ICE in find_call_stack_args, at dce.c:392)

PR c++/83490
* config/i386/i386.c (init_cumulative_args): Don't check TYPE_EMPTY_P
on an error node.

* g++.dg/abi/pr83490.C: New test.

From-SVN: r255824

commit_message

From-SVN: r255823

Add a tetst for PR middle-end/83471

PR middle-end/83471
* g++.dg/torture/pr83471.C: New test.

From-SVN: r255821

[PATCH] -fdump-tree, -save-temps=obj & subdirs

https://gcc.gnu.org/ml/gcc-patches/2017-12/msg00377.html
* opts.c (finish_options): Don't prefix dump_base_name if it
already contains directories.

From-SVN: r255820

Fix another failing test-case.

2017-12-19  Martin Liska  <mliska@suse.cz>

* gcc.dg/tree-prof/switch-case-1.c: Scan IPA profile dump
file instead of expand in order to not mix it with expanded
tree decision tree for the switch statement.

From-SVN: r255819

Fix off by one error in loop-unroll.c (PR rtl-optimization/82675).

2017-12-19  Martin Liska  <mliska@suse.cz>

PR rtl-optimization/82675
* loop-unroll.c (unroll_loop_constant_iterations): Allocate one
more element in sbitmap.

From-SVN: r255818

re PR c++/83482 (internal compiler error: Segmentation fault)

PR tree-optimization/83482
* g++.dg/torture/pr83482.C: New test.

From-SVN: r255817

gimplify.c (gimplify_expr): Use error_operand_p.

2017-12-19  Paolo Carlini  <paolo.carlini@oracle.com>

* gimplify.c (gimplify_expr): Use error_operand_p.

From-SVN: r255816

cp-tree.h: Fix typo in comment.

2017-12-19  Paolo Carlini  <paolo.carlini@oracle.com>

* cp-tree.h: Fix typo in comment.

From-SVN: r255815

Added testcase for PR 83329

2017-12-19  Martin Jambor  <mjambor@suse.cz>

PR tree-optimization/83329
* gcc.dg/tree-ssa/pr83329.c: New test

From-SVN: r255814

re PR c++/79490 (ICE when using -fsyntax-only with CATCH unit testing framework header)

2017-12-19  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/79490
* g++.dg/other/fsyntax-only2.C: New.

From-SVN: r255812

re PR ada/66205 (gnatbind generates invalid code when finalization is enabled in restricted runtime)

PR ada/66205
* bindgen.adb (Gen_AdaFinal): Revert previous change.

From-SVN: r255811

RTEMS/PowerPC: Remove 64-bit soft-float multilib

gcc/
PR target/83387
* config/rs6000/t-rtems (MULTILIB_REQUIRED): Remove 64-bit soft-float
multilib.

From-SVN: r255809

SPARC: Make sure that jump is to a label in errata workaround

In some cases the jump could be to a return instruction and in those cases the
next_active_insn() function tries to follow an invalid pointer which leads to a
crash. This error did not manifest when using a 32-bit version of GCC which is
why I did not detect it before.

gcc/
* config/sparc/sparc.c (sparc_do_work_around_errata): Make sure
the jump is to a label.

From-SVN: r255807

re PR tree-optimization/83444 (missing strlen optimization on a member array of a local struct)

PR tree-optimization/83444
* tree-ssa-strlen.c (strlen_check_and_optimize_stmt): Optimize
character loads.

* gcc.dg/strlenopt-38.c: New test.

From-SVN: r255806

re PR ipa/82801 (Internal compiler error with Eigen and __attribute__((always_inline, flatten)))

PR ipa/82801
PR ipa/83346
* ipa-inline.c (flatten_remove_node_hook): New function.
(ipa_inline): Keep only nodes with flatten attribute at the end of
the array in the order from ipa_reverse_postorder, only walk that
portion of array for flattening, if there is more than one such
node, temporarily register a removal hook and ignore removed nodes.

* g++.dg/ipa/pr82801.C: New test.

From-SVN: r255805

re PR tree-optimization/80631 (Compiling with -O3 -mavx2 gives wrong code)

PR tree-optimization/80631
* tree-vect-loop.c (vect_create_epilog_for_reduction): Compare
induc_code against MAX_EXPR or MIN_EXPR instead of reduc_fn against
IFN_REDUC_MAX or IFN_REDUC_MIN.

From-SVN: r255804

tree-ssa-dom.c (record_equivalences_from_phis): Do not record symbolic equivalences from backedges in the CFG.

* tree-ssa-dom.c (record_equivalences_from_phis): Do not
record symbolic equivalences from backedges in the CFG.

From-SVN: r255803

Revert
2017-11-19  Jeff Law  <law@redhat.com>

* tree-ssa-dom.c (record_equivalences_from_phis): Fix handling
of degenerates resulting from ignoring an edge.

From-SVN: r255802

Daily bump.

From-SVN: r255795

PR middle-end/83373 - False positive reported by -Wstringop-overflow

PR middle-end/83373 - False positive reported by -Wstringop-overflow
PR tree-optimization/78450 - strlen(s) return value can be assumed to be less than the size of s

gcc/ChangeLog:

PR middle-end/83373
PR tree-optimization/78450
* tree-ssa-strlen.c (maybe_set_strlen_range): New function.
(handle_builtin_strlen): Call it.

gcc/testsuite/ChangeLog:

PR middle-end/83373
PR tree-optimization/78450
* gcc.dg/pr83373.c: New test.
* gcc.dg/strlenopt-36.c: New test.
* gcc.dg/strlenopt-37.c: New test.

From-SVN: r255790

re PR libstdc++/80761 (std::set<T,C,A>::insert_return_type uses wrong iterator type)

2017-11-20  François Dumont  <fdumont@gcc.gnu.org>

PR libstdc++/80761
* include/debug/map.h
(std::__debug::map<>::insert_return_type): Define using
_Node_insert_return.
* include/debug/set.h (std::__debug::set<>::insert_return_type):
Likewise.
* include/debug/unordered_map:
(std::__debug::unordered_map<>::insert_return_type): Likewise.
* include/debug/unordered_set:
(std::__debug::unordered_set<>::insert_return_type): Likewise.

From-SVN: r255789

re PR c++/83116 (Statement with no effect causes wrong code of static object constexpr method)

PR c++/83116
* constexpr.c (cxx_eval_call_expression): Only look into
constexpr_call_table if ctx->strict.

* g++.dg/cpp1y/constexpr-83116.C: New test.

From-SVN: r255788

rtlanal: dead_or_set_regno_p should handle CLOBBER (PR83424)

In PR83424 combine's move_deaths puts a REG_DEAD note in the wrong place
because dead_or_set_regno_p does not account for CLOBBER insns.  This
fixes it.

PR rtl-optimization/83424
* rtlanal.c (dead_or_set_regno_p): Handle CLOBBER just like SET.

gcc/testsuite/
PR rtl-optimization/83424
* gcc.dg/pr83424.c: New testsuite.

From-SVN: r255787

testsuite: add coverage for diagnostics relating to inlining (PR tree-optimization/83336)

In theory, the diagnostics subsystem can print context information on
code inlining when diagnostics are emitted by the middle-end, describing
the chain of inlined callsites that led to a particular warning,
but PR tree-optimization/83336 describes various issues with this.

An underlying issue is that we have very little automated testing for
this code: gcc.dg/tm/pr52141.c has a test, but in general, prune.exp
filters out the various "inlined from" lines.

The following patch adds test coverage for it for C and C++ via a new
testsuite plugin, which emits a warning from the middle-end; the test
cases use dg-regexp to verify that the "inlined from" lines are
emitted correctly, with the correct function names and source locations.

Doing so requires a change to prune.exp: the dg-regexp lines have to
be handled *before* the "inlined from" lines are stripped.

gcc/testsuite/ChangeLog:
PR tree-optimization/83336
* g++.dg/cpp0x/missing-initializer_list-include.C: Update for
changes to prune.exp's handling of dg-regexp.
* g++.dg/plugin/diagnostic-test-inlining-1.C: New test case.
* g++.dg/plugin/plugin.exp (plugin_test_list): Add it, via
gcc.dg's plugin/diagnostic_plugin_test_inlining.c.
* gcc.dg/plugin/diagnostic-test-inlining-1.c: New test case.
* gcc.dg/plugin/diagnostic-test-inlining-2.c: Likewise.
* gcc.dg/plugin/diagnostic-test-inlining-3.c: Likewise.
* gcc.dg/plugin/diagnostic-test-inlining-4.c: Likewise.
* gcc.dg/plugin/diagnostic_plugin_test_inlining.c: New test
plugin.
* gcc.dg/plugin/plugin.exp (plugin_test_list): Add them.
* lib/prune.exp (prune_gcc_output): Move call to handle-dg-regexps
to before the various text stripping regsup invocations,
in particular, to before the stripping of "inlined from".

From-SVN: r255786

re PR libstdc++/68430 (std::is_constructible<T>::value == true for unconstructible type T)

PR libstdc++/68430

* testsuite/20_util/is_constructible/68430.cc: New.

From-SVN: r255785

re PR middle-end/83460 (FAIL: g++.dg/pr79095-4.C)

PR middle-end/83460
* g++.dg/pr79095-4.C: Remove compromised test.

From-SVN: r255784

re PR c++/83300 (Segmentation fault with template and __attribute__((vector_size (sizeof(int) * N)));)

PR c++/83300
* decl2.c (save_template_attributes): Add flags argument, if
not ATTR_FLAG_TYPE_IN_PLACE, *decl_p is a type and we want to
modify TYPE_ATTRIBUTES, add them on type attribute variant.

* g++.dg/ext/vector33.C: New test.

From-SVN: r255783

c-warn.c (warn_logical_operator): Return early if -Wlogical-op is not in effect.

* c-warn.c (warn_logical_operator): Return early if -Wlogical-op is
not in effect.

From-SVN: r255782

re PR middle-end/83463 (ICE: tree check: expected integer_type or enumeral_type or boolean_type or real_type or fixed_point_type, have pointer_type in builtin_memr ef, at gimple-ssa-warn-restrict.c:297)

PR middle-end/83463
* gimple-ssa-warn-restrict.c (builtin_memref::builtin_memref):
Check if TYPE is INTEGRAL_TYPE_P before accessing its min/max
values.

* gcc.dg/pr83463.c: New test.

From-SVN: r255781

[PR c++/59930] template friend injection

https://gcc.gnu.org/ml/gcc-patches/2017-12/msg01168.html
PR c++/59930
* name-lookup.c (name_lookup::search_unqualified): Don't search
parent namespace when looking for hidden things.
* pt.c (tsubst_friend_class): Always push to friend scope, drop
unneeded self-friend check. Inject new hidden friend into correct
scope.

PR c++/59930
* g++.dg/parse/pr81247-c.C: Adjust.
* g++.dg/template/pr59930-[123].C: New.

From-SVN: r255780

[ARC] Update (u)maddsidi patterns.

The accumulator registers are freely used by the compiler. However,
there are a number of instructions which are having an intrinsic use
of these registers. Update patterns to inform the compiler which ones.

gcc/
2017-09-19  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.md (maddsidi4, maddsidi4_split): Update pattern.
(umaddsidi4,umaddsidi4): Likewise.

gcc/testsuite
2017-09-19  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/tumaddsidi4.c: New test.

From-SVN: r255779

[ARC] Update legitimate constant hook.

Make sure we check the constants in all cases.

gcc/
2017-10-14  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (arc_legitimate_constant_p): Always check all
constants.

testsuite/
2017-10-14  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/tls-1.c: New test.

From-SVN: r255778

S/390: PR83420: Improve hotpatch option parsing.

With the attached patch we get rid of the following build failure:

/home/andreas/build/../gcc/gcc/config/s390/s390.c: In function ‘void
s390_option_override()’:
/home/andreas/build/../gcc/gcc/config/s390/s390.c:15361:16: error: ‘char*
strncpy(char*, const char*, size_t)’ specified bound 256 equals destination
size [-Werror=stringop-truncation]
        strncpy (s, opt->arg, 256);
        ~~~~~~~~^~~~~~~~~~~~~~~~~~

gcc/ChangeLog:

2017-12-18  Andreas Krebbel  <krebbel@linux.vnet.ibm.com>

PR target/83420
* config/s390/s390.c (s390_option_override): Avoid strncpy.

From-SVN: r255777

re PR ipa/81877 (Incorrect results with lto and -fipa-cp and -fipa-cp-clone)

2017-12-18  Richard Biener  <rguenther@suse.de>

PR tree-optimization/81877
* tree-ssa-loop-im.c (ref_indep_loop_p): Remove safelen parameters.
(outermost_indep_loop): Adjust.
(ref_indep_loop_p_1): Likewise.  Remove safelen handling again.
(can_sm_ref_p): Adjust.

* g++.dg/torture/pr81877.C: New testcase.
* g++.dg/vect/pr70729.cc: XFAIL.
* g++.dg/vect/pr70729-nest.cc: XFAIL.

From-SVN: r255776

re PR tree-optimization/77291 (False positive for -Warray-bounds)

2017-12-18  Richard Biener  <rguenther@suse.de>

PR middle-end/77291
* tree.c (array_at_struct_end_p): Return true if the underlying
object has space for at least one element in excess of what
the array domain specifies.

* gcc.dg/Warray-bounds-26.c: New testcase.

From-SVN: r255775

extend.texi (x86 Function Attributes): Reformat nocf_check example to avoid overfull hbox.

2017-12-17  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/extend.texi (x86 Function Attributes): Reformat nocf_check
example to avoid overfull hbox.
* doc/invoke.texi (Option Summary): Add missing @gol.
(C++ Dialect Options): Reformat -Wnoexcept-type example to avoid
overfull hbox.

From-SVN: r255774

invoke.texi (Option Summary): Add -mverbose-cost-dump to AArch64 and ARM lists, plus missing -mflip-thumb for ARM.

2017-12-17  Sandra Loosemore  <sandra@codesourcery.com>
    Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

gcc/
* doc/invoke.texi (Option Summary): Add -mverbose-cost-dump
to AArch64 and ARM lists, plus missing -mflip-thumb for ARM.
(AArch64 Options): Document -mverbose-cost-dump.
(ARM Options): Likewise, plus -mflip-thumb.

Co-Authored-By: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
From-SVN: r255773

PR bootstrap/83446 - Bootstrap failed on i686

gcc/testsuite/ChangeLog:

PR bootstrap/83446
* c-c++-common/Warray-bounds-3.c: Adjust.
* gcc.dg/Warray-bounds-25.c: New test.

gcc/ChangeLog:

PR bootstrap/83446
* gimple-ssa-warn-restrict.c
(builtin_memref::offset_out_of_bounds): Correct the handling of
anti-ranges.

From-SVN: r255772

Daily bump.

From-SVN: r255771

invoke.texi (ARC Options): Add missing -mlra entry.

2017-12-17  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/invoke.texi (ARC Options): Add missing -mlra entry.

From-SVN: r255768

c-attribs.c (common_handle_aligned_attribute): Avoid issuing an error for attribute warn_if_not_aligned.

gcc/c-family/ChangeLog:

* c-attribs.c (common_handle_aligned_attribute): Avoid issuing
an error for attribute warn_if_not_aligned.

From-SVN: r255767

Revert hunk with debug code that was comitted by mistake

From-SVN: r255766

guality.h (guality_check): Cast %lli arguments inf fprintf statements to long long int.

* gcc.dg/guality/guality.h (guality_check): Cast %lli arguments
inf fprintf statements to long long int.

From-SVN: r255765

pa.c (pa_som_asm_init_sections): Fix comment.

* config/pa/pa.c (pa_som_asm_init_sections): Fix comment.

From-SVN: r255764

Support -std=f2018

The Fortran committee has decided to rename the upcoming Fortran 2015
standard to Fortran 2018.  This is not a reflection of a three year
delay in the process, but rather they are following other standards in
adopting the year of publication for the name. For more details see
N2144.

This patch renames GFC_STD_F2015 to GFC_STD_F2018, and makes it a
separate flag rather than an alias for GFC_STD_GNU. Also, it adds a
-std=f2018 argument, and documents it.

Regtested on x86_64-pc-linux-gnu.

gcc/fortran/ChangeLog:

2017-12-17  Janne Blomqvist  <jb@gcc.gnu.org>

* decl.c (gfc_match_implicit_none): Use GFC_STD_F2018 instead of
GFC_STD_F2015.
* error.c (gfc_notify_std): Add GFC_STD_F2018{_DEL,_OBS} to
switch.
* gfortran.texi: Document -std=f2018.
* interface.c (compare_parameter): Fix comment.
* invoke.texi: Document -std=f2018.
* lang.opt: Add -std=f2018 argumnet.
* libgfortran.h (GFC_STD_F2015): Rename to GFC_STD_F0218, use
separate flag bit.
(GFC_STD_F2018_DEL): New macro.
(GFC_STD_F2018_OBS): Likewise.
* match.c (gfc_match_stopcode): Use GFC_STD_F2018.
* options.c (set_default_std_flags): Add F2018 flags to defaults.
(gfc_handle_option): Set options for -std=f2018.

gcc/testsuite/ChangeLog:

2017-12-17  Janne Blomqvist  <jb@gcc.gnu.org>

* gfortran.dg/error_stop_3.f90: Update -std= option, fix comments.
* gfortran.dg/error_stop_4.f90: Update error message.
* gfortran.dg/implicit_14.f90: Likewise.
* gfortran.dg/spellcheck-procedure_2.f90: Don't warn for F2018
features.

From-SVN: r255761

Correct imul (r64) latency for modern Intel CPUs

Since Sandybridge the 64bit multiplication latency is three cycles, not
four. So update the costs to reflect reality.

* x86-tune-costs.h (skylake_cost, core_cost): Decrease r64 multiply
latencies.

* gcc.target/i386/wmul-3.c: New test.

From-SVN: r255760

invoke.texi: Fix some typos.

2017-12-16  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* doc/invoke.texi: Fix some typos.

From-SVN: r255759

Daily bump.

From-SVN: r255758

PR tree-optimization/78918 - missing -Wrestrict on memcpy copying over self

gcc/c-family/ChangeLog:

PR tree-optimization/78918
* c-common.c (check_function_restrict): Avoid checking built-ins.
* c.opt (-Wrestrict): Include in -Wall.

gcc/ChangeLog:

PR tree-optimization/78918
* Makefile.in (OBJS): Add gimple-ssa-warn-restrict.o.
* builtins.c (check_sizes): Rename...
(check_access): ...to this.  Rename function arguments for clarity.
(check_memop_sizes): Adjust names.
(expand_builtin_memchr, expand_builtin_memcpy): Same.
(expand_builtin_memmove, expand_builtin_mempcpy): Same.
(expand_builtin_strcat, expand_builtin_stpncpy): Same.
(check_strncat_sizes, expand_builtin_strncat): Same.
(expand_builtin_strncpy, expand_builtin_memset): Same.
(expand_builtin_bzero, expand_builtin_memcmp): Same.
(expand_builtin_memory_chk, maybe_emit_chk_warning): Same.
(maybe_emit_sprintf_chk_warning): Same.
(expand_builtin_strcpy): Adjust.
(expand_builtin_stpcpy): Same.
(expand_builtin_with_bounds): Detect out-of-bounds accesses
in pointer-checking forms of memcpy, memmove, and mempcpy.
(gcall_to_tree_minimal, max_object_size): Define new functions.
* builtins.h (max_object_size): Declare.
* calls.c (alloc_max_size): Call max_object_size instead of
hardcoding ssizetype limit.
(get_size_range): Handle new argument.
* calls.h (get_size_range): Add a new argument.
* cfgexpand.c (expand_call_stmt): Propagate no-warning bit.
* doc/invoke.texi (-Wrestrict): Adjust, add example.
* gimple-fold.c (gimple_fold_builtin_memory_op): Detect overlapping
operations.
(gimple_fold_builtin_memory_chk): Same.
(gimple_fold_builtin_stxcpy_chk): New function.
* gimple-ssa-warn-restrict.c: New source.
* gimple-ssa-warn-restrict.h: New header.
* gimple.c (gimple_build_call_from_tree): Propagate location.
* passes.def (pass_warn_restrict): Add new pass.
* tree-pass.h (make_pass_warn_restrict): Declare.
* tree-ssa-strlen.c (handle_builtin_strcpy): Detect overlapping
operations.
(handle_builtin_strcat): Same.
(strlen_optimize_stmt): Rename...
(strlen_check_and_optimize_stmt): ...to this.  Handle strncat,
stpncpy, strncpy, and their checking forms.

gcc/testsuite/ChangeLog:

PR tree-optimization/78918
* c-c++-common/Warray-bounds.c: New test.
* c-c++-common/Warray-bounds-2.c: New test.
* c-c++-common/Warray-bounds-3.c: New test.
* c-c++-common/Warray-bounds-4.c: New test.
* c-c++-common/Warray-bounds-5.c: New test.
* c-c++-common/Wrestrict-2.c: New test.
* c-c++-common/Wrestrict.c: New test.
* c-c++-common/Wrestrict.s: New test.
* c-c++-common/Wsizeof-pointer-memaccess1.c: Adjust
* c-c++-common/Wsizeof-pointer-memaccess2.c: Same.
* g++.dg/torture/Wsizeof-pointer-memaccess1.C: Same.
* g++.dg/torture/Wsizeof-pointer-memaccess2.C: Same.
* gcc.dg/range.h: New header.
* gcc.dg/memcpy-6.c: New test.
* gcc.dg/pr69172.c: Adjust.
* gcc.dg/pr79223.c: Same.
* gcc.dg/pr81345.c: Adjust.
* gcc.dg/Wobjsize-1.c: Same.
* gcc.dg/Wrestrict-2.c: New test.
* gcc.dg/Wrestrict.c: New test.
* gcc.dg/Wsizeof-pointer-memaccess1.c: Adjust.
* gcc.dg/builtin-stpncpy.c: Same.
* gcc.dg/builtin-stringop-chk-1.c: Same.
* gcc.target/i386/chkp-stropt-17.c: New test.
* gcc.dg/torture/Wsizeof-pointer-memaccess1.c: Adjust.

From-SVN: r255755

PR tree-optimization/83239 - False positive from -Wstringop-overflow

PR tree-optimization/83239 - False positive from -Wstringop-overflow
on simple std::vector code

libstdc++/CHangeLog:
* include/bits/vector.tcc (vector::_M_default_append): Assert
        invariant to generate better code.

gcc/testsuite/ChangeLog:
* g++.dg/pr83239.C: New test case.

From-SVN: r255753

re PR libfortran/81937 (stack-buffer-overflow on memcpy in libgfortran/io/unix.c on character(kind=4))

2017-12-16  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR libgfortran/81937
        * io/list_read.c (next_char_internal): Don't attempt to read
        from the internal unit stream if no bytes are left. Decrement
        bytes_left in the right place.

From-SVN: r255750

re PR target/82767 (gcc.target/i386/pr71321.c scan-assembler-times fail)

PR testsuite/82767
 * gcc.target/i386/pr71321.c: Test with -mtune=generic

From-SVN: r255749