varasm.c (output_constructor_regular_field): Check TYPE_SIZE_UNIT of the init value.

* varasm.c (output_constructor_regular_field): Check TYPE_SIZE_UNIT of
the init value.

* c-common.c (complete_flexible_array_elts): New helper function.
* c-common.h (complete_flexible_array_elts): Declare.

* c-decl.c (finish_decl): Call complete_flexible_array_elts.

* decl.c (check_initializer): Call cp_complete_array_type.

From-SVN: r264147

rs6000.c (rs6000_gimple_fold_builtin): Add support for early gimple folding of vec_splat().

[gcc]

2018-09-06  Will Schmidt  <will_schmidt@vnet.ibm.com>

* config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add support for
early gimple folding of vec_splat().
* tree-vect-generic.c: Remove static from tree_vec_extract() definition.
* gimple-fold.h: Add an extern define for tree_vec_extract().

From-SVN: r264146

rs6000.c (fold_mergehl_helper): Add types_compatible_p wrappers around TREE_TYPE comparisons.

[gcc]

2018-08-20  Will Schmidt  <will_schmidt@vnet.ibm.com>

* config/rs6000/rs6000.c (fold_mergehl_helper): Add types_compatible_p
wrappers around TREE_TYPE comparisons.

From-SVN: r264145

S/390: Prohibit SYMBOL_REF in UNSPECV_CAS

Inhibit constant propagation inlining SYMBOL_REF loads into
UNSPECV_CAS.  Even though reload can later undo it, the resulting
code will be less efficient.

gcc/ChangeLog:

2018-09-06  Ilya Leoshkevich  <iii@linux.ibm.com>

PR target/80080
* config/s390/predicates.md: Add nonsym_memory_operand.
* config/s390/s390.c (s390_legitimize_cs_operand): If operand
contains a SYMBOL_REF, load it into an intermediate pseudo.
(s390_emit_compare_and_swap): Legitimize operand.
* config/s390/s390.md: Use the new nonsym_memory_operand
with UNSPECV_CAS patterns.

gcc/testsuite/ChangeLog:

2018-09-06  Ilya Leoshkevich  <iii@linux.ibm.com>

PR target/80080
* gcc.target/s390/pr80080-3.c: New test.
* gcc.target/s390/s390.exp: Make sure the new test passes
on all optimization levels.

From-SVN: r264143

S/390: Register pass_s390_early_mach statically

The dump file used to come at the end of the sorted dump file list,
because the pass was registered dynamically. This did not reflect the
order in which passes are executed. Static registration fixes this:

* foo4.c.277r.split2
* foo4.c.281r.early_mach
* foo4.c.282r.pro_and_epilogue

gcc/ChangeLog:

2018-09-06  Ilya Leoshkevich  <iii@linux.ibm.com>

PR target/80080
* config/s390/s390-passes.def: New file.
* config/s390/s390-protos.h (class rtl_opt_pass): Add forward
declaration.
(make_pass_s390_early_mach): Add declaration.
* config/s390/s390.c (make_pass_s390_early_mach):
(s390_option_override): Remove dynamic registration.
* config/s390/t-s390: Add s390-passes.def.

From-SVN: r264142

S/390: Remove UNSPEC_LTREL_BASE

It was needed only for g5/g6 machines, which are now gone.

gcc/ChangeLog:

2018-09-06  Ilya Leoshkevich  <iii@linux.ibm.com>

* config/s390/s390.c (s390_decompose_constant_pool_ref):
Remove UNSPEC_LTREL_BASE check.
(annotate_constant_pool_refs): Likewise.
(find_constant_pool_ref): Likewise.
(find_ltrel_base): Removed.
(replace_ltrel_base): Removed.
(s390_mainpool_finish): Remove replace_ltrel_base call.
(s390_chunkify_start): Remove pending LTREL_BASE logic.
(s390_chunkify_finish): Remove replace_ltrel_base call.
* config/s390/s390.md: Remove UNSPEC_LTREL_BASE.

From-SVN: r264141

PR c++/86982, -Wreturn-local-addr and std::move and std::forward.

* typeck.c (maybe_warn_about_returning_address_of_local): Handle calls
to std::move or std::forward.
(is_std_forward_p): New function.

* g++.dg/warn/Wreturn-local-addr-5.C: New test.

From-SVN: r264139

Daily bump.

From-SVN: r264138

re PR target/86779 (Cris port needs updating for CVE-2017-5753)

PR target/86779
* config/cris/cris.c (TARGET_HAVE_SPECULATION_SAFE_VALUE): Redefine
to speculation_safe_value_not_needed.

From-SVN: r264134

[nvptx] Basic -misa support for nvptx

gcc/
* config/nvptx/nvptx-opts.h: New file.
* config/nvptx/nvptx.c (nvptx_file_start): Print the correct .target.
* config/nvptx/nvptx.h: Include "nvptx-opts.h".
(ASM_SPEC): Define.
(TARGET_SM35): New macro.
* config/nvptx/nvptx.md (atomic_fetch_<logic><mode>): Enable with the
correct predicate.
* config/nvptx/nvptx.opt (ptx_isa, sm_30, sm_35): New enum and its
values.
(misa=): New option.
* doc/invoke.texi (Nvidia PTX Options): Document -misa.

gcc/testsuite/
* gcc.target/nvptx/atomic_fetch-1.c: New test.
* gcc.target/nvptx/atomic_fetch-1.c: New test.

Co-Authored-By: Bernd Schmidt <bernds_cb1@t-online.de>
From-SVN: r264133

PR c++/87109, wrong overload with ref-qualifiers.

* call.c (build_user_type_conversion_1): Use NULL instead of 0.  Bail
out if performing the maybe-rvalue overload resolution and a conversion
function is getting called.

* g++.dg/cpp0x/ref-qual19.C: New test.

From-SVN: r264132

i386.md (truncdfsf2): Remove expander.

* config/i386/i386.md (truncdfsf2): Remove expander.
(truncdfsf2_with_temp): Ditto.
(truncxf<mode>2): Ditto.
(*truncdfsf_fast_mixed): Remove insn pattern.
(*truncdfsf_fast_i387): Ditto.
(*truncdfsf_mixed): Ditto.
(*truncdfsf_i387): Ditto.
(*truncdfsf2_i387_1): Ditto.
(*truncxfsf2_mixed): Ditto.
(*truncxfdf2_mixed): Ditto.
(*truncxf<mode>2_i387_noop): Ditto. Update callers
to call gen_truncxf<mode>2 instead.
(*truncxf<mode>2_i387): Remove.
(reg->reg splitters): Remove splitter pattern.
(reg->mem splitters): Ditto.

(truncdfsf2): New insn pattern.
(truncxf<mode>2): Ditto.

From-SVN: r264130

PR52665 do not let .ident confuse assembler scan tests

A branch with a name matching scan-assembler pattern triggers
inappropriate FAIL.

E.g. branch fixups-testsuite and
- gcc.target/i386/pr65871-?.c (scan-assembler-not "test")
- gcc.target/i386/pr41442.c (scan-assembler-times "test|cmp" 2)
etc.

This is a recurring problem as can be seen by some -fno-ident additions
by commits from e.g. Michael Meissner over the years: builtins-58.c,
powerpc/pr46728-?.c

The patch below adds -fno-ident if a testcase contains one of
scan-assembler, scan-assembler-not or scan-assembler-times.

Regression tested on x86_64-unknown-linux on a fixups-testsuite branch
where it fixes several false FAILs without regressions.

gcc/testsuite/ChangeLog

2016-06-18  Bernhard Reutner-Fischer  <aldot@gcc.gnu.org>

        PR testsuite/52665
        * lib/gcc-dg.exp (gcc-dg-test-1): Iterate over _required_options.
        * lib/target-supports.exp (scan-assembler_required_options,
        scan-assembler-not_required_options,
        scan-assembler-times_required_options): Add -fno-ident.
        * lib/scanasm.exp (scan-assembler-times): Fix error message.
        * c-c++-common/ident-0a.c: New test.
        * c-c++-common/ident-0b.c: New test.
        * c-c++-common/ident-1a.c: New test.
        * c-c++-common/ident-1b.c: New test.
        * c-c++-common/ident-2a.c: New test.
        * c-c++-common/ident-2b.c: New test.

From-SVN: r264128

* libsupc++/cxxabi.h (__cxa_demangle): Clarify doxygen comment.

From-SVN: r264127

Optimise sqrt reciprocal multiplications

This patch aims to optimise sequences involving uses of 1.0 / sqrt (a) under -freciprocal-math and -funsafe-math-optimizations.
In particular consider:

x = 1.0 / sqrt (a);
r1 = x * x;  // same as 1.0 / a
r2 = a * x; // same as sqrt (a)

If x, r1 and r2 are all used further on in the code, this can be transformed into:
tmp1 = 1.0 / a
tmp2 = sqrt (a)
tmp3 = tmp1 * tmp2
x = tmp3
r1 = tmp1
r2 = tmp2

A bit convoluted, but this saves us one multiplication and, more importantly, the sqrt and division are now independent.
This also allows optimisation of a subset of these expressions.
For example:
x = 1.0 / sqrt (a)
r1 = x * x

can be transformed to r1 = 1.0 / a, eliminating the sqrt if x is not used anywhere else.
And similarly:
x = 1.0 / sqrt (a)
r1 = a * x

can be transformed to sqrt (a) eliminating the division.

For the testcase:
double res, res2, tmp;
void
foo (double a, double b)
{
  tmp = 1.0 / __builtin_sqrt (a);
  res = tmp * tmp;
  res2 = a * tmp;
}

We now generate for aarch64 with -Ofast:
foo:
        fmov    d2, 1.0e+0
        adrp    x2, res2
        fsqrt   d1, d0
        adrp    x1, res
        fdiv    d0, d2, d0
        adrp    x0, tmp
        str     d1, [x2, #:lo12:res2]
        fmul    d1, d1, d0
        str     d0, [x1, #:lo12:res]
        str     d1, [x0, #:lo12:tmp]
        ret

where before it generated:
foo:
        fsqrt   d2, d0
        fmov    d1, 1.0e+0
        adrp    x1, res2
        adrp    x2, tmp
        adrp    x0, res
        fdiv    d1, d1, d2
        fmul    d0, d1, d0
        fmul    d2, d1, d1
        str     d1, [x2, #:lo12:tmp]
        str     d0, [x1, #:lo12:res2]
        str     d2, [x0, #:lo12:res]
        ret

As you can see, the new sequence has one fewer multiply and the fsqrt and fdiv are independent.

* tree-ssa-math-opts.c (is_mult_by): New function.
(is_square_of): Use the above.
(optimize_recip_sqrt): New function.
(pass_cse_reciprocals::execute): Use the above.

* gcc.dg/recip_sqrt_mult_1.c: New test.
* gcc.dg/recip_sqrt_mult_2.c: Likewise.
* gcc.dg/recip_sqrt_mult_3.c: Likewise.
* gcc.dg/recip_sqrt_mult_4.c: Likewise.
* gcc.dg/recip_sqrt_mult_5.c: Likewise.
* g++.dg/recip_sqrt_mult_1.C: Likewise.
* g++.dg/recip_sqrt_mult_2.C: Likewise.

From-SVN: r264126

re PR bootstrap/87134 (SEGV in cc1 caused by r263875)

2018-09-05  Richard Biener  <rguenther@suse.de>

PR bootstrap/87134
* tree-ssa-sccvn.c (rpo_elim::eliminate_push_avail): Make sure
to zero-init the emplaced vec.

From-SVN: r264125

Group switch cases in switch lowering (PR tree-optimization/87205).

2018-09-05  Martin Liska  <mliska@suse.cz>

PR tree-optimization/87205
* tree-switch-conversion.c (pass_lower_switch::execute):
Group cases for switch statements.
2018-09-05  Martin Liska  <mliska@suse.cz>

PR tree-optimization/87205
* gcc.dg/tree-ssa/pr87205-2.c: New test.
* gcc.dg/tree-ssa/pr87205.c: New test.

From-SVN: r264124

re PR tree-optimization/87217 (ICE in in check_loop_closed_ssa_def, at tree-ssa-loop-manip.c:709 when compiling SPEC2000 starting with r264069)

2018-09-05  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87217
* tree-ssa-sccvn.c (vuse_valueize): New.
(vn_reference_lookup_pieces): Use it.
(vn_reference_lookup): Likewise.

* gfortran.dg/pr87217.f: New testcase.

From-SVN: r264121

* io/async.h: Use __gthread_mutex_t, not pthread_mutex_t.

From-SVN: r264120

PR c++/87137] GCC-8 Fix

https://gcc.gnu.org/ml/gcc-patches/2018-08/msg01966.html
PR c++/87137
* stor-layout.c (place_field): Scan forwards to check last
bitfield when ms_bitfield_placement is in effect.
gcc/testsuite/
* g++.dg/abi/pr87137.C: New.

From-SVN: r264119

[PR c++/87185] ICE in prune-lambdas

https://gcc.gnu.org/ml/gcc-patches/2018-09/msg00249.html
cp/
PR c++/87185
* lambda.c (prune_lambda_captures): Protect against const_vars.get
returning NULL.
testsuite/
PR c++/87185
* g++.dg/pr87185.C: New.

From-SVN: r264118

re PR bootstrap/87225 (tree-vect-stmts.c:3748 error: converting to 'bool' from 'std::nullptr_t' requires direct-initialization [-fpermissive])

2018-09-05  Richard Biener  <rguenther@suse.de>

PR bootstrap/87225
* tree-vect-stmts.c (vectorizable_simd_clone_call): Fix bogus
return.

From-SVN: r264116

[aarch64] Avoid tag collisions for loads falkor

This is a rewrite of the tag collision avoidance patch that Kugan had
written as a machine reorg pass back in February.

The falkor hardware prefetching system uses a combination of the
source, destination and offset to decide which prefetcher unit to
train with the load.  This is great when loads in a loop are
sequential but sub-optimal if there are unrelated loads in a loop that
tag to the same prefetcher unit.

This pass attempts to rename the desination register of such colliding
loads using routines available in regrename.c so that their tags do
not collide.  This shows some performance gains with mcf and xalancbmk
(~5% each) and will be tweaked further.  The pass is placed near the
fag end of the pass list so that subsequent passes don't inadvertantly
end up undoing the renames.

2018-07-02  Siddhesh Poyarekar  <siddhesh@sourceware.org>
            Kugan Vivekanandarajah  <kugan.vivekanandarajah@linaro.org>

* config/aarch64/falkor-tag-collision-avoidance.c: New file.
* config.gcc (extra_objs): Build it.
* config/aarch64/t-aarch64 (falkor-tag-collision-avoidance.o):
Likewise.
* config/aarch64/aarch64-passes.def
(pass_tag_collision_avoidance): New pass.
* config/aarch64/aarch64.c (qdf24xx_tunings): Add
AARCH64_EXTRA_TUNE_RENAME_LOAD_REGS to tuning_flags.
(aarch64_classify_address): Remove static qualifier.
(aarch64_address_info, aarch64_address_type): Move to...
* config/aarch64/aarch64-protos.h: ... here.
(make_pass_tag_collision_avoidance): New function.
* config/aarch64/aarch64-tuning-flags.def (rename_load_regs):
New tuning flag.

Co-Authored-By: Kugan Vivekanandarajah <kuganv@linaro.org>
From-SVN: r264115

Fix typo in ChangeLog.

From-SVN: r264114

Fix a scan in test for Darwin target (PR testsuite/87216).

2018-09-05  Martin Liska  <mliska@suse.cz>

PR testsuite/87216
* gcc.dg/tree-prof/pr59521-3.c: Update scanned pattern
to support Dawring names.

From-SVN: r264113

GCOV: Print one decimal place in human readable mode.

2018-09-05  Martin Liska  <mliska@suse.cz>

* doc/gcov.texi: Update documentation of humar
readable mode.
* gcov.c (format_count): Print one decimal place, it provides
more fine number of situations like '1G' vs. '1.4G'.
2018-09-05  Martin Liska  <mliska@suse.cz>

* g++.dg/gcov/loop.C: Update test to support new format.

From-SVN: r264112

Ignore properly -mdirect-move (PR target/87164).

2018-09-05  Martin Liska  <mliska@suse.cz>

PR target/87164
* config/rs6000/rs6000.opt: Mark the option as Deprecated.
* optc-gen.awk: Allow 'Var' for Deprecated options in order
to generate a MASK value.

From-SVN: r264111

cp-tree.h (treat_lvalue_as_rvalue_p): Declare.

* cp-tree.h (treat_lvalue_as_rvalue_p): Declare.
* except.c (build_throw): Use it.  Use CP_TYPE_VOLATILE_P.
* typeck.c (treat_lvalue_as_rvalue_p): No longer static.  Add PARM_OK
parameter.
(maybe_warn_pessimizing_move): Adjust treat_lvalue_as_rvalue_p call.
(check_return_expr): Likewise.

From-SVN: r264101

Daily bump.

From-SVN: r264100

DWARF: Allow hard frame pointer even if frame pointer isn't used

r251028

commit cd557ff63f388ad27c376d0a225e74d3594a6f9d
Author: hjl <hjl@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Thu Aug 10 15:29:05 2017 +0000

    i386: Don't use frame pointer without stack access

    When there is no stack access, there is no need to use frame pointer
    even if -fno-omit-frame-pointer is used and caller's frame pointer is
    unchanged.

frame pointer may not be available even if -fno-omit-frame-pointer is
used.  When this happened, arg pointer may be eliminated by hard frame
pointer.  Since hard frame pointer is encoded with DW_OP_fbreg which
uses the DW_AT_frame_base attribute, not hard frame pointer directly,
we should allow hard frame pointer when generating DWARF info even if
frame pointer isn't used.

gcc/

PR debug/86593
* dwarf2out.c (based_loc_descr): Allow hard frame pointer even
if frame pointer isn't used.
(compute_frame_pointer_to_fb_displacement): Likewise.

gcc/testsuite/

PR debug/86593
* g++.dg/pr86593.C: New test.

From-SVN: r264096

re PR target/87198 (ICE in extract_insn, at recog.c:2304)

PR target/87198
* common/config/i386/i386-common.c (OPTION_MASK_ISA_XSAVEOPT_SET,
OPTION_MASK_ISA_XSAVES_SET, OPTION_MASK_ISA_XSAVEC_SET): Use
OPTION_MASK_ISA_XSAVE_SET instead of OPTION_MASK_ISA_XSAVE.
(OPTION_MASK_ISA_XSAVE_UNSET): Add OPTION_MASK_ISA_XSAVES_UNSET
and OPTION_MASK_ISA_XSAVEC_UNSET.

* gcc.target/i386/pr87198.c: New test.

From-SVN: r264088

gcc: xtensa: fix NAND code in xtensa_expand_atomic

NAND is ~(a1 & a2), but xtensa_expand_atomic does ~a1 & a2.
That fixes libatomic tests atomic-op-{1,2}.

gcc/
2018-09-04  Max Filippov  <jcmvbkbc@gmail.com>

* config/xtensa/xtensa.c (xtensa_expand_atomic): Reorder AND and
XOR operations in NAND case.

From-SVN: r264087

Don't xfail gcc.target/i386/addr-sel-1.c (PR target/86744)

PR target/86744
* gcc.target/i386/addr-sel-1.c: Don't xfail "b\\+1" scan.

From-SVN: r264086

wide-int-range.cc (wide_int_range_convert): New.

* wide-int-range.cc (wide_int_range_convert): New.
* wide-int-range.h (wide_int_range_convert): New.
* tree-vrp.c (extract_range_from_unary_expr): Abstract wide int
code into wide_int_range_convert.
(extract_range_into_wide_ints): Do not munge anti range constants
into the entire domain.  Just return the range back.

From-SVN: r264085

genmatch: put reporting on a cold path

2018-09-04  Martin Liska  <mliska@suse.cz>

* genmatch.c (output_line_directive): Add new argument
fnargs.
(dt_simplify::gen_1): Encapsulate dump within __builtin_expect.

From-SVN: r264084

Add whitespace before warning option added in previous commmit

* doc/invoke.texi (Option Summary): Add whitespace.

From-SVN: r264083

Add -Waligned-new to Option Summary

* doc/invoke.texi (Option Summary): Add -Waligned-new.

From-SVN: r264080

re PR tree-optimization/87211 (gcc ICE at O2: in set_ssa_val_to, at tree-ssa-sccvn.c:3628)

2018-09-04  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87211
* tree-ssa-sccvn.c (visit_phi): When value-numbering to a
backedge value we're supposed to treat as VARYING also number
the PHI to VARYING in case it got a different value-number already.

* gcc.dg/torture/pr87211.c: New testcase.

From-SVN: r264079

tree-vrp.c (vrp_can_optimize_bit_op): Remove.

* tree-vrp.c (vrp_can_optimize_bit_op): Remove.
(extract_range_from_binary_expr_1): Do not call
vrp_can_optimize_bit_op.
* wide-int-range.cc (wide_int_range_can_optimize_bit_op): Make
static.
(wide_int_range_get_mask_and_bounds): New.
(wide_int_range_optimize_bit_op): New.
(wide_int_range_bit_ior): Call wide_int_range_optimize_bit_op.
(wide_int_range_bit_and): Same.
* wide-int-range.h (wide_int_range_can_optimize_bit_op): Remove.
(wide_int_range_optimize_bit_op): New.
(wide_int_range_get_mask_and_bounds): New.

From-SVN: r264078

re PR tree-optimization/87176 (wrong code at -Os and above on x86-64-linux-gnu)

2018-09-04  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87176
* tree-ssa-sccvn.c (visit_phi): Remove redundant allsame
variable.  When value-numbering a virtual PHI node make sure
to not value-number to the backedge value.

* gcc.dg/torture/pr87176.c: New testcase.
* gcc.dg/torture/ssa-fre-1.c: Likewise.

From-SVN: r264077

Update C Extensions docs for support in latest C++

* doc/extend.texi (Long Long, Hex Floats): Document support for
long long and hex floats in more recent versions of ISO C++.

From-SVN: r264076

* MAINTAINERS: Update my email address.

From-SVN: r264075

Daily bump.

From-SVN: r264074

simplify.c (gfc_simplify_modulo): Re-arrange code to test whether 'P' is zero and issue an error if it is.

2018-09-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

* simplify.c (gfc_simplify_modulo): Re-arrange code to test whether
'P' is zero and issue an error if it is.
* gfortran.dg/modulo_check: New test.

From-SVN: r264070

re PR tree-optimization/87177 (wrong code at -O2 and -O3 on x86-64-linux-gnu)

2018-09-03  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87177
* tree-ssa-sccvn.c (vuse_ssa_val): Revert previous change, keep
cleanup.

* gcc.dg/torture/pr87177.c: New testcase.
* gcc.dg/torture/pr87177-2.c: Likewise.

From-SVN: r264069

bb-reorder: convert to gcc_stablesort

* bb-reorder.c (edge_order): Convert to C-qsort-style
tri-state comparator.
(reorder_basic_blocks_simple): Change std::stable_sort to gcc_stablesort.

From-SVN: r264068

tree-loop-distribution: convert to gcc_stablesort

* tree-loop-distribution.c (offset_cmp): Convert to C-qsort-style
tri-state comparator.
(fuse_memset_builtins): Change std::stable_sort to gcc_stablesort.

From-SVN: r264067

introduce gcc_stablesort

* sort.cc (struct sort_ctx): New field 'nlim'.  Use it...
(mergesort): ... here as maximum count for using netsort.
(gcc_qsort): Set nlim to 3 if stable sort is requested.
(gcc_stablesort): New.
* system.h (gcc_stablesort): Declare.

From-SVN: r264066

qsort_chk: call from gcc_qsort instead of wrapping it

* sort.cc (gcc_qsort) [CHECKING_P]: Call qsort_chk.
* system.h (qsort): Always redirect to gcc_qsort.  Update comment.
* vec.c (qsort_chk): Do not call gcc_qsort.  Update comment.

From-SVN: r264065

rs6000: lxsdux / stxsdux do not exist

Our md files refer to {l,st}xsd%U<n>x, but no {l,st}xsdux insns exist.
This patch removes the update forms.  All these use  constraint "Z"
which does not allow update form, so there is no practical difference.

* config/rs6000/rs6000.md (*mov<mode>_hardfloat32): Remove %U from the
lxsdx and stxsdx alternatives.
(*mov<mode>_hardfloat64): Ditto.
* config/rs6000/vsx.md (*vsx_extract_<mode>_store): Ditto.

From-SVN: r264064

PR libstdc++/78179 run long double tests separately

Split the long double testing into a separate file, so that we can XFAIL
targets where the long double precision doesn't meet the expected
tolerances. The float and double tests are still expefted to PASS for
all targets.

PR libstdc++/78179
* testsuite/26_numerics/headers/cmath/hypot-long-double.cc: New test
that runs the long double part of hypot.cc.
* testsuite/26_numerics/headers/cmath/hypot.cc: Disable long double
tests unless TEST_HYPOT_LONG_DOUBLE is defined.

From-SVN: r264063

re PR tree-optimization/87200 (ICE in set_ssa_val_to, at tree-ssa-sccvn.c:3629)

2018-09-03  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87200
* tree-ssa-sccvn.c (vn_nary_build_or_lookup_1): Valueize a
simplify result.

* gcc.dg/torture/pr87200.c: New testcase.

From-SVN: r264062

Fix vector::_Temporary_value::_M_ptr

The pointer argument to allocator_traits::construct and
allocator_traits::destroy should be a raw pointer, not the allocator's
pointer type. _Temporary_value::_M_ptr was returning the wrong type.

* include/bits/stl_vector.h (vector::_Temporary_value::_M_ptr):
Return raw pointer not allocator's pointer type.
(vector::_Temporary_value::_M_val): Use _M_ptr.

From-SVN: r264061

PR libstdc++/87194 fix range insertion into maps and sets

Since C++11 range insertion and construction of maps and sets from a
pair of iterators only requires that the iterator's value_type is
convertible to the container's value_type (previously it had to be the
same).

This fixes the implementation to meet that relaxed requirement, by
defining a pair of overloads that either insert or emplace, depending on
the iterator's value_type. Instead of adding yet another overload of
_M_insert_unique and _M_insert_equal, the overloads taking iterators are
renamed to _M_insert_range_unique and _M_insert_range_equal.

PR libstdc++/87194
* include/bits/stl_map.h
(map::map(initializer_list<value_type>, const Compare&, const Alloc&))
(map::map(initializer_list<value_type>, const Alloc&))
(map::map(InputIterator, InputIterator, const Alloc&))
(map::map(InputIterator, InputIterator))
(map::map(InputIterator, InputIterator, const Compare&, const Alloc&))
(map::insert(InputIterator, InputIterator)):
Call _M_insert_range_unique instead of _M_insert_unique.
* include/bits/stl_multimap.h
(multimap::multimap(initializer_list<value_type>, const C&, const A&))
(multimap::multimap(initializer_list<value_type>, const A&))
(multimap::multimap(InputIterator, InputIterator, const A&))
(multimap::multimap(InputIterator, InputIterator))
(multimap::multimap(InputIterator, InputIterator, const C&, const A&))
(multimap::insert(InputIterator, InputIterator)): Call
_M_insert_range_equal instead of _M_insert_equal.
* include/bits/stl_multiset.h
(multiset::multiset(InputIterator, InputIterator))
(multiset::multiset(InputIterator, InputIterator, const C&, const A&))
(multiset::multiset(initializer_list<value_type>, const C&, const A&))
(multiset::multiset(initializer_list<value_type>, const A&))
(multiset::multiset(InputIterator, InputIterator, const A&))
(multiset::insert(InputIterator, InputIterator)): Call
_M_insert_range_equal instead of _M_insert_equal.
* include/bits/stl_set.h
(set::set(InputIterator, InputIterator))
(set::set(InputIterator, InputIterator, const Compare&, const Alloc&))
(set::set(initializer_list<value_type>, const Compare&, const Alloc&))
(set::set(initializer_list<value_type>, const Alloc&))
(set::set(InputIterator, InputIterator, const Alloc&))
(set::insert(InputIterator, InputIterator)):
Call _M_insert_range_unique instead of _M_insert_unique.
* include/bits/stl_tree.h
[__cplusplus >= 201103L] (_Rb_tree::__same_value_type): New alias
template for SFINAE constraints.
[__cplusplus >= 201103L] (_Rb_tree::_M_insert_range_unique): Pair of
constrained overloads that either insert or emplace, depending on
iterator's value_type.
[__cplusplus >= 201103L] (_Rb_tree::_M_insert_range_equal): Likewise.
[__cplusplus < 201103L] (_Rb_tree::_M_insert_range_unique)
(_Rb_tree::_M_insert_range_equal): New functions replacing range
versions of _M_insert_unique and _M_insert_equal.
(_Rb_tree::_M_insert_unique(_InputIterator, _InputIterator))
(_Rb_tree::_M_insert_equal(_InputIterator, _InputIterator)): Remove.
* testsuite/23_containers/map/modifiers/insert/87194.cc: New test.
* testsuite/23_containers/multimap/modifiers/insert/87194.cc: New test.
* testsuite/23_containers/multiset/modifiers/insert/87194.cc: New test.
* testsuite/23_containers/set/modifiers/insert/87194.cc: New test.

From-SVN: r264060

PR libstdc++/78595 implement insertion into maps in terms of emplace

C++14 simplified the specification of the generic insert function
templates to be equivalent to calling emplace (or emplace_hint).
Defining them in terms of emplace takes care of the problems described
in PR 78595, ensuring a single conversion to value_type is done at the
right time.

PR libstdc++/78595
* include/bits/stl_map.h (map::insert(_Pair&&))
(map::insert(const_iterator, _Pair&&)): Do emplace instead of insert.
* include/bits/stl_multimap.h (multimap::insert(_Pair&&))
(multimap::insert(const_iterator, _Pair&&)): Likewise.
* include/bits/unordered_map.h (unordered_map::insert(_Pair&&))
(unordered_map::insert(const_iterator, _Pair&&))
(unordered_multimap::insert(_Pair&&))
(unordered_multimap::insert(const_iterator, _Pair&&)): Likewise.
* testsuite/23_containers/map/modifiers/insert/78595.cc: New test.
* testsuite/23_containers/multimap/modifiers/insert/78595.cc: New test.
* testsuite/23_containers/unordered_map/modifiers/78595.cc: New test.
* testsuite/23_containers/unordered_multimap/modifiers/78595.cc: New
test.

From-SVN: r264059

Fix thinko (PR tree-optimization/87201).

2018-09-03  Martin Liska  <mliska@suse.cz>

PR tree-optimization/87201
* tree-switch-conversion.c (switch_decision_tree::balance_case_nodes):
Fix parenthesis in an expression.

From-SVN: r264058

re PR tree-optimization/87197 (ICE in is_gimple_reg_type at gimple-expr.h:75 since r264021)

2018-09-03  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87197
* tree-ssa-sccvn.c (vn_nary_build_or_lookup_1): Mark the new def
visited.  CSE the VN_INFO hashtable lookup.

* gcc.dg/torture/pr87197.c: New testcase.

PR tree-optimization/87169
* tree-ssa-sccvn.c (do_rpo_vn): When marking loops for not
iterating make sure there's no extra backedges from irreducible
regions feeding the header.  Mark the destination block
executable.

* gcc.dg/torture/pr87169.c: New testcase.

From-SVN: r264057

fixincludes: vxworks: remove unnecessary parentheses in ioctl wrapper macro

The rationale for the fixinclude ioctl macro wrapper is, as far as I can
tell (https://gcc.gnu.org/ml/gcc-patches/2012-09/msg01619.html)

  Fix 2: Add hack for ioctl() on VxWorks.

  ioctl() is supposed to be variadic, but VxWorks only has a three
  argument version with the third argument of type int.  This messes up
  when the third argument is not implicitly convertible to int.  This
  adds a macro which wraps around ioctl() and explicitly casts the third
  argument to an int.  This way, the most common use case of ioctl (with
  a const char * for the third argument) will compile in C++, where
  pointers must be explicitly casted to int.

However, we have existing C++ code that calls the ioctl function via

  ::ioctl(foo, bar, baz)

and obviously this breaks when it gets expanded to

  ::(ioctl)(foo, bar, (int)(baz))

Since the GNU C preprocessor already prevents recursive expansion of
function-like macros, the parentheses around ioctl are unnecessary.

Incidentally, there is also a macro sioIoctl() in the vxworks sioLib.h
header that expands to

  ((pSioChan)->pDrvFuncs->ioctl (pSioChan, cmd, arg))

which also breaks when that gets further expanded to

  ((pSioChan)->pDrvFuncs->(ioctl) (pSioChan, cmd, (int)(arg)))

This patch partly fixes that issue as well, but the third argument to
the pDrvFuncs->ioctl method should be void*, so the cast to (int) is
slightly annoying. Internally, we've simply patched the sioIoctl macro:

  (((pSioChan)->pDrvFuncs->ioctl) (pSioChan, cmd, arg))

From-SVN: r264056

Come up with TARGET_GET_VALID_OPTION_VALUES option hook (PR driver/83193).

2018-09-03  Martin Liska  <mliska@suse.cz>

        PR driver/83193
* common/common-target.def: Add TARGET_GET_VALID_OPTION_VALUES.
* common/common-targhooks.c (default_get_valid_option_values):
        New function.
* common/common-targhooks.h (default_get_valid_option_values):
        Likewise.
* common/config/i386/i386-common.c: Move processor_target_table
        from i386.c.
(ix86_get_valid_option_values): New function.
(TARGET_GET_VALID_OPTION_VALUES): New macro.
* config/i386/i386.c (struct ptt): Move to i386-common.c.
(PTA_*): Move all defined masks into i386-common.c.
(ix86_function_specific_restore): Use new processor_cost_table.
* config/i386/i386.h (struct ptt): Moved from i386.c.
(struct pta): Likewise.
* doc/tm.texi: Document new TARGET_GET_VALID_OPTION_VALUES.
* doc/tm.texi.in: Likewise.
* opt-suggestions.c (option_proposer::suggest_option):
        Pass prefix to build_option_suggestions.
(option_proposer::get_completions): Likewise.
(option_proposer::build_option_suggestions): Use the new target
        hook.
* opts.c (struct option_help_tuple): New struct.
(print_filtered_help): Use the new target hook.
2018-09-03  Martin Liska  <mliska@suse.cz>

        PR driver/83193
* gcc.dg/completion-4.c: New test.

From-SVN: r264052

re PR c++/84980 ([concepts] ICE with missing typename in concept)

/cp
2018-09-03  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/84980
* constraint.cc (finish_shorthand_constraint): Early return if the
constraint is erroneous.

/testsuite
2018-09-03  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/84980
* g++.dg/concepts/pr84980.C: New.

From-SVN: r264051

Make __builtin_expect effective in switch statements (PR middle-end/PR59521).

2018-09-03  Martin Liska  <mliska@suse.cz>

  PR middle-end/59521
* predict.c (set_even_probabilities): Add likely_edges
        argument and handle cases where we have precisely one
        likely edge.
(combine_predictions_for_bb): Catch also likely_edges.
(tree_predict_by_opcode): Handle gswitch statements.
* tree-cfg.h (find_case_label_for_value): New declaration.
(find_taken_edge_switch_expr): Likewise.
* tree-switch-conversion.c (switch_decision_tree::balance_case_nodes):
        Find pivot in decision tree based on probabily, not by number of
        nodes.
2018-09-03  Martin Liska  <mliska@suse.cz>

  PR middle-end/59521
* c-c++-common/pr59521-1.c: New test.
* c-c++-common/pr59521-2.c: New test.
* gcc.dg/tree-prof/pr59521-3.c: New test.

From-SVN: r264050

Daily bump.

From-SVN: r264049

io.h: Change declaration of vlist type to gfc_full_array_i4 to eliminate warning for...

2018-09-01  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

* io/io.h: Change declaration of vlist type to
gfc_full_array_i4 to eliminate warning for mismatched type.
* io/format.c ((parse_format_list): Use gfc_full_array_i4.
* io/io.h: Use gfc_full_array_i4.

From-SVN: r264043

c-common.c (braced_list_to_string): Remove eval parameter.

* c-common.c (braced_list_to_string): Remove eval parameter.
Add some more checks.  Always create zero-terminated STRING_CST.
* c-common.h (braced_list_to_string): Adjust prototype.

* c-decl.c (finish_decl): Call braced_list_to_string here ...
* c-parser.c (c_parser_declaration_or_fndef): ... instead of here.

* decl.c (eval_check_narrowing): Remove.
(check_initializer): Move call to braced_list_to_string from here ...
* typeck2.c (store_init_value): ... to here.
(digest_init_r): Remove handing of signed/unsigned char strings.

* c-c++-common/array-init.c: New test.
* g++.dg/init/string2.C: Remove xfail.

From-SVN: r264042

2018-09-02  François Dumont  <fdumont@gcc.gnu.org>

* include/debug/safe_iterator.h
(_Safe_iterator<_It, _Seq, _Cat>::_Self): New.
(_Safe_iterator<_It, _Seq, std::random_access_iterator_tag>::_Self):
New.
(_Safe_iterator<_It, _Seq, std::random_access_iterator_tag>
::_OtherSelf): New.
(_GLIBCXX_DEBUG_VERIFY_OPERANDS, _GLIBCXX_DEBUG_VERIFY_EQ_OPERANDS)
(_GLIBCXX_DEBUG_VERIFY_REL_OPERANDS)
(_GLIBCXX_DEBUG_VERIFY_DIST_OPERANDS): Define macros.
(_Safe_iterator<_It, _Seq, std::random_access_iterator_tag>
::operator+(difference_type)): Use latters, inline as friend.
(_Safe_iterator<_It, _Seq, std::random_access_iterator_tag>
::operator-(difference_type)): Likewise.
(operator==(const _Safe_iterator<>&, const _Safe_iterator<>&)):
Likewise.
(operator!=(const _Safe_iterator<>&, const _Safe_iterator<>&)):
Likewise.
(operator<(const _Safe_iterator<>&, const _Safe_iterator<>&)): Likewise.
(operator<=(const _Safe_iterator<>&, const _Safe_iterator<>&)):
Likewise.
(operator>(const _Safe_iterator<>&, const _Safe_iterator<>&)): Likewise.
(operator>=(const _Safe_iterator<>&, const _Safe_iterator<>&)):
Likewise.
(operator-(const _Safe_iterator<>&, const _Safe_iterator<>&)): Likewise.
(operator+(difference_type, const _Safe_iterator<>&)): Likewise.
(operator-(const _Safe_iterator<>&, difference_type)): Likewise.
* include/debug/safe_iterator.tcc
(_Safe_iterator<>::_M_can_advance(difference_type)): Take parameter by
copy.
* include/debug/safe_local_iterator.h
(_Safe_local_iterator<_It, _Seq>::_Self): New.
(_Safe_local_iterator<_It, _Seq>::_OtherSelf): New.
(_GLIBCXX_DEBUG_VERIFY_OPERANDS): Define macro.
(operator==(const _Safe_local_iterator<>&,
const _Safe_local_iterator<>&)): Use latter, inline as friend.
(operator!=(const _Safe_local_iterator<>&,
const _Safe_local_iterator<>&)): Likewise.
* testsuite/util/testsuite_containers.h: Include utility.
(struct forward_members_unordered<_Tp, bool>): Remove 2nd template
parameter.
(forward_members_unordered<>::forward_members_unordered(value_type&)):
Add using namespace std::rel_ops.
Add iterator_concept_checks on local_iterator and const_local_iterator.
Add asserts on comparison between const_local_iterator and
local_iterator.
(struct forward_members_unordered<_Tp, false>): Remove partial
specialization.
* testsuite/23_containers/forward_list/types/1.cc: New.
* testsuite/23_containers/list/types/1.cc: New.

From-SVN: r264039

* doc/standards.texi (Standards): Update Objective-C reference.

From-SVN: r264038

Daily bump.

From-SVN: r264037

* doc/install.texi (Prerequisites): Update link for MPC.

From-SVN: r264033

generic.texi (OpenMP): Adjust link to openmp.org.

* doc/generic.texi (OpenMP): Adjust link to openmp.org.
* doc/invoke.texi (C Dialect Options): Ditto.

From-SVN: r264032

profile_mode.xml: Update three ieeexplore.ieee.org references.

* doc/xml/manual/profile_mode.xml: Update three ieeexplore.ieee.org
references.

From-SVN: r264031

re PR tree-optimization/87074 (Unroll and jam bug: O3 result differ from O2)

Fix PR87074

PR tree-optimization/87074
* gimple-loop-jam.c (unroll_jam_possible_p): Check loop exit
PHIs for outer-loop uses.

testsuite/
* gcc.dg/pr87074.c: New test.

From-SVN: r264029

* doc/install.texi (Prerequisites): Adjust link mpfr.org.

From-SVN: r264028

Daily bump.

From-SVN: r264026

10.cc: Add dg-require-fileio.

2018-08-31  Sandra Loosemore  <sandra@codesourcery.com>

libstdc++-v3/

* testsuite/21_strings/basic_string/inserters_extractors/wchar_t/10.cc:
Add dg-require-fileio.
* testsuite/21_strings/basic_string/inserters_extractors/wchar_t/11.cc:
Likewise.
* testsuite/21_strings/basic_string/inserters_extractors/wchar_t/5.cc:
Likewise.
* testsuite/21_strings/basic_string_view/inserters/wchar_t/2.cc:
Likewise.
* testsuite/25_algorithms/copy/streambuf_iterators/wchar_t/4.cc:
Likewise.
* testsuite/25_algorithms/find/istreambuf_iterators/wchar_t/2.cc:
Likewise.
* testsuite/27_io/basic_filebuf/close/wchar_t/12790-1.cc: Likewise.
* testsuite/27_io/basic_filebuf/close/wchar_t/12790-2.cc: Likewise.
* testsuite/27_io/basic_filebuf/close/wchar_t/12790-3.cc: Likewise.
* testsuite/27_io/basic_filebuf/close/wchar_t/12790-4.cc: Likewise.
* testsuite/27_io/basic_filebuf/seekoff/wchar_t/11543.cc: Likewise.
* testsuite/27_io/basic_filebuf/seekoff/wchar_t/12790-1.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekoff/wchar_t/12790-2.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekoff/wchar_t/12790-3.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekoff/wchar_t/12790-4.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekpos/wchar_t/12790-1.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekpos/wchar_t/12790-2.cc:
Likewise.
* testsuite/27_io/basic_filebuf/seekpos/wchar_t/12790-4.cc:
Likewise.
* testsuite/27_io/basic_filebuf/underflow/wchar_t/2.cc: Likewise.
* testsuite/27_io/basic_filebuf/underflow/wchar_t/3.cc: Likewise.
* testsuite/27_io/basic_ifstream/cons/wchar_t/1.cc: Likewise.
* testsuite/27_io/basic_ifstream/open/wchar_t/1.cc: Likewise.
* testsuite/27_io/basic_istream/extractors_character/wchar_t/4.cc:
Likewise.
* testsuite/27_io/basic_istream/extractors_other/wchar_t/2.cc:
Likewise.
* testsuite/27_io/basic_istream/get/wchar_t/2.cc: Likewise.
* testsuite/27_io/basic_istream/getline/wchar_t/5.cc: Likewise.
* testsuite/27_io/basic_istream/ignore/wchar_t/2.cc: Likewise.
* testsuite/27_io/basic_istream/ignore/wchar_t/3.cc: Likewise.
* testsuite/27_io/basic_istream/seekg/wchar_t/sstream.cc: Likewise.
* testsuite/27_io/basic_istream/tellg/wchar_t/sstream.cc: Likewise.
* testsuite/27_io/basic_ofstream/cons/wchar_t/1.cc: Likewise.
* testsuite/27_io/basic_ofstream/open/wchar_t/1.cc: Likewise.
* testsuite/27_io/basic_ostream/inserters_other/wchar_t/1.cc:
Likewise.
* testsuite/27_io/objects/wchar_t/10.cc: Likewise.
* testsuite/27_io/objects/wchar_t/12048-1.cc: Likewise.
* testsuite/27_io/objects/wchar_t/12048-2.cc: Likewise.
* testsuite/27_io/objects/wchar_t/12048-3.cc: Likewise.
* testsuite/27_io/objects/wchar_t/12048-4.cc: Likewise.
* testsuite/27_io/objects/wchar_t/12048-5.cc: Likewise.
* testsuite/experimental/string_view/inserters/wchar_t/2.cc:
Likewise.
* testsuite/ext/stdio_sync_filebuf/wchar_t/1.cc: Likewise.

From-SVN: r264022

re PR tree-optimization/87168 (ICE on valid code at -Os and above on x86_64-linux-gnu: verify_ssa failed)

2018-08-31  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87168
* tree-ssa-sccvn.c (SSA_VAL): Add visited output parameter.
(rpo_elim::eliminate_avail): When OP was not visited it must
be available.

* gcc.dg/torture/pr87168.c: New testcase.

From-SVN: r264021

tree-vrp: add "const" qualifier to various value_range pointers

gcc/ChangeLog:
* tree-vrp.c (copy_value_range): Convert param "from" from
"value_range *" to "const value_range *".
(range_is_null): Likewise for param "vr".
(range_int_cst_p): Likewise.
(range_int_cst_singleton_p): Likewise.
(symbolic_range_p): Likewise.
(value_ranges_intersect_p): Likewise for both params.
(value_range_nonnegative_p): Likewise for param "vr".
(value_range_constant_singleton): Likewise.
(vrp_set_zero_nonzero_bits): Likewise for param "ar".
(extract_range_into_wide_ints): Likewise for param "vr".
(extract_range_from_multiplicative_op): Likewise for params "vr0"
and "vr1".
(vrp_can_optimize_bit_op): Likewise.
(extract_range_from_binary_expr_1): Likewise for params "vr0_" and
"vr1_".
(extract_range_from_unary_expr): Likewise.
(debug_value_range): Likewise for param "vr".
(value_range::dump): Add "const" qualifier.
(vrp_prop::check_array_ref): Convert local "vr" from
"value_range *" to "const value_range *".
(vrp_prop::check_mem_ref): Likewise.
(vrp_prop::visit_stmt): Likewise for local "old_vr".
(vrp_intersect_ranges_1): Likewise for param "vr_1".
(vrp_intersect_ranges): Likewise.
(simplify_stmt_for_jump_threading): Likewise for local "vr".
(vrp_prop::vrp_finalize): Likewise.
* tree-vrp.h (value_range::dump): Add "const" qualifier.
(vrp_intersect_ranges): Add "const" qualifier to params as above.
(extract_range_from_unary_expr): Likewise.
(value_range_constant_singleton): Likewise.
(symbolic_range_p): Likewise.
(copy_value_range): Likewise.
(extract_range_from_binary_expr_1): Likewise.
(range_int_cst_p): Likewise.
(vrp_set_zero_nonzero_bits): Likewise.
(range_int_cst_singleton_p): Likewise.

From-SVN: r264020

[AArch64] Implement new intrinsics vabsd_s64 and vnegd_s64.

gcc/
2018-08-31  Vlad Lazar  <vlad.lazar@arm.com>

* config/aarch64/arm_neon.h (vabsd_s64): New.
(vnegd_s64): Likewise.

gcc/testsuite/
2018-08-31  Vlad Lazar  <vlad.lazar@arm.com>

* gcc.target/aarch64/scalar_intrinsics.c (test_vnegd_s64): New.
* gcc.target/aarch64/vneg_s.c (RUN_TEST_SCALAR): New.
(test_vnegd_s64): Likewise.
* gcc.target/aarch64/vnegd_64.c: New.
* gcc.target/aarch64/vabsd_64.c: New.
* gcc.tartget/aarch64/vabs_intrinsic_3.c: New

From-SVN: r264019

[PR c++/87155] Anonymous namespace and

https://gcc.gnu.org/ml/gcc-patches/2018-08/msg02031.html
PR c++/87155
PR c++/84707
cp/
* name-lookup.c (name_lookup::search_namespace): Don't look at
inlines when searching for NULL names.
testsuite/
* g++.dg/cpp0x/pr87155.C: New.
* g++.dg/cpp0x/inline-ns10.C: Adjust.

From-SVN: r264016

[PATCH] Fix thinko in estimate_local_effects in IPA-CP

2018-08-31  Martin Jambor  <mjambor@suse.cz>

* ipa-cp.c (estimate_local_effects): Replace wrong MAX with MIN.

From-SVN: r264015

IPA ICF: make type cache a static field sem_item.

2018-08-31  Martin Liska  <mliska@suse.cz>

* ipa-icf.c (sem_item::add_type): Use
sem_item::m_type_hash_cache.
* ipa-icf.h: Move the cache from sem_item_optimizer
to sem_item.

From-SVN: r264014

[C++ PATCH] Remove K&R declaration hack.

https://gcc.gnu.org/ml/gcc-patches/2018-08/msg01885.html
gcc/
* doc/extend.texi (Backwards Compatibility): Remove implicit
extern C leeway of () being (...).
gcc/cp/
* decl.c (decls_match): Remove SYSTEM_IMPLICIT_EXTERN_C matching
of return types and parms.
* parser.c (cp_parser_parameter_declaration_clause): Likewise,
'()' always means '(void)'.

From-SVN: r264013

[ipa-inline][obvious] Fix typos in comment

* ipa-inline.c (can_inline_edge_by_limits_p): Fix typos in comment.

From-SVN: r264012

rs6000: Make lround<mode>di2 depend on TARGET_FPRND (PR86684)

TARGET_FPRND should be on for everything ISA 2.04 and later, and
TARGET_VSX implies ISA 2.06 or later; but it is possible to disable
TARGET_FPRND (separately via -mno-fprnd, but also implicitly)
currently, and then things fall down.  This patch makes things not
fall down.

PR target/86684
PR target/87149
* config/rs6000/rs6000.md (lround<mode>di2): Gate on TARGET_FPRND.

From-SVN: r264011

re PR middle-end/87138 (wrong code with TI multiplication with -O -march=k8 -fno-tree-fre -mavx512bw)

PR middle-end/87138
* expmed.c (expand_mult_const): Use immed_wide_int_const instead of
gen_int_mode.  Formatting fixes.

* gcc.target/i386/avx512bw-pr87138.c: New test.

From-SVN: r264009

re PR fortran/86328 (Runtime segfault reading an allocatable class(*) object in allocate statements)

2018-08-31  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/86328
PR fortran/86760
* trans-array.c (gfc_conv_scalarized_array_ref): Do not fix
info->descriptor but pass it directly to gfc_build_array_ref.
(gfc_conv_array_ref): Likewise for se->expr.
* trans.c (gfc_build_array_ref): If 'decl' is a COMPONENT_REF
obtain the span field directly from it.

2018-08-31  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/86328
PR fortran/86760
* gfortran.dg/pr86328.f90 : New test.
in comment 12 of the PR.
* gfortran.dg/pr86760.f90 : New test.

From-SVN: r264008

Fix PR testsuite/86519 reference.

From-SVN: r264007

Daily bump.

From-SVN: r264006

noexcept30.C: Make dependence on -fdelete-null-pointer-checks explicit.

2018-08-30  Sandra Loosemore  <sandra@codesourcery.com>

gcc/testsuite/
* g++.dg/cpp0x/noexcept30.C: Make dependence on
-fdelete-null-pointer-checks explicit.
* g++.dg/cpp1y/constexpr-82218.C: Likewise.
* g++.dg/expr/pmf-3.C: Likewise.
* g++.dg/ext/attr-returns-nonnull.C: Likewise.
* g++.dg/tree-ssa/lifetime-dse1.C: Likewise.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* gcc.dg/addr_equal-1.c: Likewise.
* gcc.dg/ipa/pr85734.c: Likewise.
* gcc.dg/ipa/propmalloc-1.c: Likewise.
* gcc.dg/ipa/propmalloc-2.c: Likewise.
* gcc.dg/ipa/propmalloc-3.c: Likewise.
* gcc.dg/tree-ssa/pr78154.c: Likewise.
* gcc.dg/tree-ssa/pr83648.c: Likewise.

libstdc++-v3/
* testsuite/experimental/propagate_const/observers/1.cc: Make
dependence on -fdelete-null-pointer-checks explicit.

From-SVN: r264001

PR testsuite/87158 - FAIL gcc.c-torture/execute/memchr-1.c on big endian targets

gcc/testsuite/ChangeLog:
* gcc.c-torture/execute/memchr-1.c: Correct big-endian expectations.

From-SVN: r264000

target.def (custom_function_descriptors): Improve documentation.

2018-08-30  Sandra Loosemore  <sandra@codesourcery.com>

gcc/
* target.def (custom_function_descriptors): Improve documentation.
* doc/tm.texi.in (Trampolines): Expand discussion of function
descriptors and move TARGET_CUSTOM_FUNCTION_DESCRIPTORS to the
beginning of the section.
* doc/tm.texi: Regenerated.

From-SVN: r263999

cfg.h (class auto_edge_flag): Spell out the template-id of the base class in the initializer list.

gcc:

2018-08-30  Jose E. Marchesi  <jose.marchesi@oracle.com>

* cfg.h (class auto_edge_flag): Spell out the template-id of the
base class in the initializer list.  This is a workaround for
building with older GCC.
(class auto_bb_flag): Likewise.

From-SVN: r263997

altivec.md (altivec_eq<mode>): Remove star.

2018-08-30  Aaron Sawdey  <acsawdey@linux.ibm.com>

* config/rs6000/altivec.md (altivec_eq<mode>): Remove star.
(altivec_vcmpequ<VI_char>_p): Remove star.
* config/rs6000/rs6000-string.c (do_load_for_compare): Support
vector load modes.
(expand_strncmp_vec_sequence): New function.
(emit_final_str_compare_vec): New function.
(expand_strn_compare): Add support for vector strncmp.
* config/rs6000/rs6000.opt (-mstring-compare-inline-limit): Change
length specification to bytes.
* config/rs6000/vsx.md (vsx_ld_elemrev_v16qi_internal): Remove star.
(vcmpnezb_p): New pattern.
* doc/invoke.texi (RS/6000 and PowerPC Options): Update documentation
for option -mstring-compare-inline-limit.

From-SVN: r263991

x86: Move AESNI generation to Skylake and Goldmont

The instruction set first appeared with Westmere, but not all processors
in that and the next few generations have the instructions. According to
Wikipedia[1], the first generation in which all SKUs have AES
instructions are Skylake and Goldmont. I can't find any Skylake,
Kabylake, Kabylake-R or Cannon Lake currently listed at
https://ark.intel.com that says "Intel® AES New Instructions" "No".

[1] https://en.wikipedia.org/wiki/AES_instruction_set

2018-08-30  Thiago Macieira  <thiago.macieira@intel.com>

* config/i386/i386.c (PTA_WESTMERE): Remove PTA_AES.
(PTA_SKYLAKE): Add PTA_AES.
(PTA_GOLDMONT): Likewise.

From-SVN: r263989

re PR lto/86517 (relocation R_X86_64_32 against `.rodata.str1.1' can not be used when making a shared object with LTO)

PR lto/86517
* lto-opts.c (lto_write_options): Always stream PIC/PIE mode.
* lto-wrapper.c (merge_and_complain): Fix merging of PIC/PIE.

From-SVN: r263988

[libcpp] fix some line map comments

https://gcc.gnu.org/ml/gcc-patches/2018-08/msg01978.html
* include/line-map.h (enum lc_reason): Comment each member
separately.
(struct line_maps): Fix reallocator comment.

From-SVN: r263987

Avoid undefined shifts in ceil2 operations

For values where the result cannot be represented the shift width would
be equal to the width of the type, which is undefined. Perform two
well-defined shifts instead of one possible undefined shift.

* include/bits/hashtable_policy.h (__clp2): Fix calculation for LLP64
targets where sizeof(size_t) > sizeof(long). Avoid undefined shifts
of the number of bits in the type.
* include/std/bit (__ceil2): Avoid undefined shifts.
* testsuite/26_numerics/bit/bit.pow.two/ceil2.cc: Test values with
the most signifiant bit set.

From-SVN: r263986

lto-streamer-out.c (DFS::DFS_write_tree_body): Do not follow TYPE_STUB_DECL.

* lto-streamer-out.c (DFS::DFS_write_tree_body): Do not follow
TYPE_STUB_DECL.
(hash_tree): Do not visit TYPE_STUB_DECL.
* tree-streamer-out.c (write_ts_type_common_tree_pointers): Do not
stream TYPE_STUB_DECL.
* tree-streamer-in.c (lto_input_ts_type_common_tree_pointers): Likewise.
* ipa-utils.h (type_with_linkage_p): Do not rely on TYPE_STUB_DECL
after free_lang_data.
(type_in_anonymous_namespace_p): Likewise.

From-SVN: r263985

Move strcmpopt_6.c from gcc.dg to gcc.target/aarch64 and gcc.target/i386.

From-SVN: r263983

gimplefe: expose MULT_HIGHPART_EXPR

* gimple-parser.c (c_parser_gimple_binary_expression): Accept infix
"__MULT_HIGHPART" for MULT_HIGHPART_EXPR.

From-SVN: r263982

sreal.h (SREAL_PART_BITS): Change to 31; remove seemingly unnecessary comment that it has to be even number.

* sreal.h (SREAL_PART_BITS): Change to 31; remove seemingly unnecessary
comment that it has to be even number.
(class sreal): Change m_sig type to int32_t.
* sreal.c (sreal::dump, sreal::to_int, opreator+, operator-): Use
int64_t for temporary calculations.
(sreal_verify_basics): Drop one bit from minimum and maximum.

From-SVN: r263981