sourcebuild.texi (target attributes): Document new "inf" effective target keyword.

* doc/sourcebuild.texi (target attributes): Document new "inf"
effective target keyword.

From-SVN: r265815

S/390: Increase register move costs for CC_REGS

This increases the move costs for moves involving the CC
register which saves us some instructions in SPEC CPU2006.

gcc/ChangeLog:

2018-11-05  Robin Dapp  <rdapp@linux.ibm.com>

* config/s390/s390.c (s390_register_move_cost): Increase costs for
        moves involving the CC reg.

From-SVN: r265814

S/390: Make tests expect column numbers in RTL output

RTL output now includes column numbers in addition to line numbers,
like this:

  "gcc/testsuite/gcc.target/s390/md/andc-splitter-1.c":16:1

This confuses some S/390 tests.

gcc/testsuite/ChangeLog:

2018-11-05  Ilya Leoshkevich  <iii@linux.ibm.com>

* gcc.target/s390/md/andc-splitter-1.c: Add colon to
expectations.
* gcc.target/s390/md/andc-splitter-2.c: Likewise.
* gcc.target/s390/md/setmem_long-1.c: Likewise.

From-SVN: r265813

re PR tree-optimization/87873 (ICE: verify_gimple failed (error: incompatible types in PHI argument 0))

2018-11-05  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87873
* tree-ssa-loop-manip.h (split_loop_exit_edge): Add copy_constants_p
argument.
* tree-ssa-loop-manip.c (split_loop_exit_edge): Likewise.
* tree-vect-loop.c (vect_transform_loop): When splitting the
loop exit also create forwarder PHIs for constants.
* tree-vect-loop-manip.c (slpeel_duplicate_current_defs_from_edges):
Handle constant to_arg, add extra checking we match up the correct
PHIs.

* gcc.dg/pr87873.c: New testcase.

From-SVN: r265812

Do not use %zu format in libcpp.

2018-11-05  Martin Liska  <mliska@suse.cz>

* symtab.c (ht_dump_statistics): Replace %zu with %lu format.

From-SVN: r265811

Fix printf call in symtab.c.

2018-11-05  Martin Liska  <mliska@suse.cz>

* symtab.c (ht_dump_statistics): Fix format and
pass missing argument.

From-SVN: r265810

S/390: Add loc patterns for QImode and HImode

This adds QImode and HImode support to load on condition.

gcc/ChangeLog:

2018-11-05  Robin Dapp  <rdapp@linux.ibm.com>

* config/s390/s390.md: QImode and HImode for load on condition.

gcc/testsuite/ChangeLog:

2018-11-05  Robin Dapp  <rdapp@linux.ibm.com>

* gcc.target/s390/ifcvt-one-insn-bool.c: New test.
* gcc.target/s390/ifcvt-one-insn-char.c: New test.

From-SVN: r265808

S/390: Allow immediates in loc expander

This allows immediate operands for load on condition.

gcc/ChangeLog:

2018-11-05  Robin Dapp  <rdapp@linux.ibm.com>

* config/s390/predicates.md: Fix typo.
* config/s390/s390.md: Allow immediates for load on condition.

gcc/testsuite/ChangeLog:

2018-11-05  Robin Dapp  <rdapp@linux.ibm.com>

* gcc.dg/loop-8.c: On s390, always run the test with -march=zEC12.

From-SVN: r265804

Come up with SIZE_AMOUNT and use it in memory statistics and sort stats.

2018-11-05  Martin Liska  <mliska@suse.cz>

* alloc-pool.h (struct pool_usage): Use SIZE_AMOUNT.
* bitmap.h (struct bitmap_usage): Likewise.
* ggc-common.c (SCALE): Remove.
(LABEL): Likewise.
(struct ggc_usage): Use SIZE_AMOUNT. And update
compare method.
* ggc-page.c (SCALE): Remove.
(STAT_LABEL): Remove.
(ggc_print_statistics): Use SIZE_AMOUNT.
* gimple.h (SCALE): Remove.
(LABEL): Likewise.
* input.c (ONE_K): Remove.
(ONE_M): Likewise.
(SCALE): Likewise.
(STAT_LABEL): Likewise.
(FORMAT_AMOUNT): Likewise.
(dump_line_table_statistics): Use SIZE_AMOUNT.
* mem-stats.h (struct mem_usage): Likewise.
* rtl.c (dump_rtx_statistics): Likewise.
(rtx_alloc_counts): Change type to size_t.
(rtx_alloc_sizes): Likewise.
(rtx_count_cmp): New.
(dump_rtx_statistics): Sort first based on counts.
* tree.c (tree_nodes_cmp): New.
(tree_codes_cmp): New.
(dump_tree_statistics): Sort first based on counts.
* system.h (ONE_K): New.
(ONE_M): Likewise.
(SIZE_SCALE): Likewise.
(SIZE_LABEL): Likewise.
(SIZE_AMOUNT): Likewise.
* tree-cfg.c (dump_cfg_stats): Use SIZE_AMOUNT.
* tree-dfa.c (dump_dfa_stats): Likewise.
* tree-phinodes.c (phinodes_print_statistics): Likewise.
* tree-ssanames.c (ssanames_print_statistics): Likewise.
* tree.c (dump_tree_statistics): Likewise.
* vec.c (struct vec_usage): Likewise.
* trans-mem.c (tm_mangle): Enlarge buffer in order to not
trigger a -Werror=format-overflow with
--enable-gather-detailed-stats.

From-SVN: r265800

Fix vector memory statistics.

2018-11-05  Martin Liska  <mliska@suse.cz>

* mem-stats.h (mem_alloc_description::release_instance_overhead):
Return T *.
* vec.c (struct vec_usage): Register m_element_size.
(vec_prefix::register_overhead): New arguments: elements and
element_size.
(vec_prefix::release_overhead): Subtract elements.
* vec.h (struct vec_prefix): Change signature.
(va_heap::reserve): Pass proper arguments.
(va_heap::release): Likewise.

From-SVN: r265799

Fix GNU coding style.

2018-11-05  Martin Liska  <mliska@suse.cz>

* mem-stats.h (mem_alloc_description::get_list): Fix GNU coding
style.
* vec.c: Likewise.

From-SVN: r265798

Fix string pool statistics.

2018-11-05  Martin Liska  <mliska@suse.cz>

* symtab.c (ht_dump_statistics): Make dump conditional
based on alloc_subobject.

From-SVN: r265797

Fix build on sparc64-linux-gnu.

2018-11-05  Martin Liska  <mliska@suse.cz>

PR sanitizer/87860
* sanitizer_common/sanitizer_linux.cc:  Cherry-pick upstream
r346129.

From-SVN: r265796

tree-scalar-evolution.h (final_value_replacement_loop): Update prototype.

2018-11-05  Richard Biener  <rguenther@suse.de>

* tree-scalar-evolution.h (final_value_replacement_loop): Update
prototype.
* tree-scalar-evolution.c (final_value_replacement_loop): Return
whether anything was done.
(scev_const_prop): Remove constant propagation part, fold
remains into ...
* tree-ssa-loop.c (pass_scev_cprop::execute): ... here.
(pass_data_scev_cprop): TODO_cleanup_cfg is now done
conditionally.

* gcc.dg/pr41488.c: Scan ivcanon dump instead of sccp one.
* gcc.dg/tree-ssa/scev-7.c: Likewise.

From-SVN: r265795

re PR tree-optimization/87859 (store-merging miscompilation of mesa)

PR tree-optimization/87859
* gimple-ssa-store-merging.c (struct merged_store_group): Add
only_constants and first_nonmergeable_order members.
(merged_store_group::merged_store_group): Initialize them.
(merged_store_group::do_merge): Clear only_constants member if
adding something other than INTEGER_CST store.
(imm_store_chain_info::coalesce_immediate_stores): Don't merge
stores with order >= first_nonmergeable_order.  Use
merged_store->only_constants instead of always recomputing it.
Set merged_store->first_nonmergeable_order if we've skipped any
stores.  Attempt to merge overlapping INTEGER_CST stores that
we would otherwise skip.

* gcc.dg/store_merging_24.c: New test.
* gcc.dg/store_merging_25.c: New test.

From-SVN: r265794

re PR sanitizer/87837 (-O2 -fsanitize=signed-integer-overflow misses overflows on x86-64)

PR sanitizer/87837
* match.pd (X + Y < X): Don't optimize if TYPE_OVERFLOW_SANITIZED.

* c-c++-common/ubsan/pr87837.c: New test.

From-SVN: r265793

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

2018-11-05  Xuepeng Guo  <xuepeng.guo@intel.com>

PR target/87853
* config/i386/emmintrin.h (__v16qs): New to cope with option
-funsigned-char.
(_mm_cmpeq_epi8): Replace __v16qi with __v16qs.
(_mm_cmplt_epi8): Likewise.
(_mm_cmpgt_epi8): Likewise.

From-SVN: r265792

re PR middle-end/87852 (FAIL: gcc.c-torture/execute/pr53465.c)

2018-11-05  Richard Biener  <rguenther@suse.de>

PR rtl-optimization/87852
* fwprop.c (use_killed_between): Only consider single-defs of the
use whose definition statement dominates the use.

From-SVN: r265791

Implement UDL changes from P0732R2.

Implementing the UDL changes was pretty straightforward; I simplified
cp_parser_userdef_string_literal using the releasing_vec type from mangle.c.

While looking at this, I realized that the string UDL template taking a
character pack that we implemented for C++14 didn't actually make it into
C++14, so I've added a pedwarn for it and no longer suggest it in the
diagnostic about an invalid UDL template.

* cp-tree.h (struct releasing_vec): Move from mangle.c.
Add get_ref method.
* parser.c (cp_parser_userdef_string_literal): Use it.  Handle
passing the string to a single template parameter of class type.
(cp_parser_template_declaration_after_parameters): Allow it.
Pedwarn about the character pack template that was proposed but not
accepted for C++14, and don't suggest it.

From-SVN: r265790

Implement P0732R2, class types in non-type template parameters.

There is one significant piece of this that is not implemented yet: the
reliance on defaulted operator<=>, which someone else has been working on.
So, for the moment those lines are commented out of the testcases.

One tricky bit was treating template parameters of classtype as const
lvalues without making their decltype const; for this I used a
VIEW_CONVERT_EXPR wrapper, which previously could only appear in templates
as location wrappers.

The user-defined literal parts of P0732R2 are in the next patch.

gcc/cp/
* error.c (dump_simple_decl): Look through a template parm object.
* mangle.c (write_template_arg): Likewise.
(mangle_template_parm_object): New.
* pt.c (template_parm_object_p, get_template_parm_object): New.
(invalid_tparm_referent_p): Factor from convert_nontype_argument.
(convert_nontype_argument, invalid_nontype_parm_type_p): Handle
class-type template arguments.
* tree.c (lvalue_kind): Likewise.
gcc/c-family/
* c-cppbuiltin.c (c_cpp_builtins): Add
__cpp_nontype_template_parameter_class.
libiberty/
* cp-demangle.c (d_dump, d_make_comp, d_count_templates_scopes)
(d_print_comp_inner): Handle DEMANGLE_COMPONENT_TPARM_OBJ.
(d_special_name): Handle TA.
(d_expresion_1): Fix demangling of brace-enclosed initializer list.
include/
* demangle.h (enum demangle_component_type): Add
DEMANGLE_COMPONENT_TPARM_OBJ.

From-SVN: r265789

Fix various latent issues revealed by P0732 work.

The initialized_type hunk fixes handling of void AGGR_INIT_EXPRs that call a
non-constructor; an AGGR_INIT_EXPR can have void type if its initialization
semantics are more complicated than just expanding the call.

The cxx_eval_vec_init_1 hunk corrects AGGR_INIT_EXPRs that were
nonsensically built to initialize an object of void type.  And the
build_aggr_init_expr hunk makes sure we don't do that again.

The ocp_convert and cxx_eval_outermost_constant_expr hunks deal with making
sure that a constant CONSTRUCTOR has the right type.

* cvt.c (ocp_convert): Don't wrap a CONSTRUCTOR in a NOP_EXPR.
* constexpr.c (initialized_type): Fix AGGR_INIT_EXPR handling.
(cxx_eval_vec_init_1): Correct type of AGGR_INIT_EXPR.
(cxx_eval_outermost_constant_expr): Make sure a CONSTRUCTOR has the
right type.  Don't wrap a CONSTRUCTOR if one was passed in.
* tree.c (build_aggr_init_expr): Check for void.

From-SVN: r265788

PR c++/60503 - wrong lambda attribute syntax.

This patch fixes two issues with lambda attribute handling: First, it was in
the wrong place in the grammar.  Second, it was treating attributes as
applying to the whole declaration rather than to the function type, as
specified by the standard.

* parser.c (cp_parser_lambda_declarator_opt): Fix attribute
handling.

From-SVN: r265787

Remove options that are not disabled with -Os (PR web/87829).

2018-11-05  Martin Liska  <mliska@suse.cz>

PR web/87829
* doc/invoke.texi: Remove options that are
not disabled with -Os.

From-SVN: r265786

Verify that last argument of __builtin_expect_with_probability is a real cst (PR c/87811).

2018-11-05  Martin Liska  <mliska@suse.cz>

PR c/87811
* predict.c (expr_expected_value_1): Verify
that last argument is a real constants and emit
error.
2018-11-05  Martin Liska  <mliska@suse.cz>

PR c/87811
* gcc.dg/pr87811.c: New test.
* gcc.dg/pr87811-2.c: Likewise.
* gcc.dg/pr87811-3.c: Likewise.
2018-11-05  Martin Liska  <mliska@suse.cz>

PR c/87811
* doc/extend.texi: Update constrain about the last argument
of __builtin_expect_with_probability.

From-SVN: r265785

Fix setting of hotness in non-LTO mode (PR gcov-profile/77698).

2018-11-05  Martin Liska  <mliska@suse.cz>

PR gcov-profile/77698
* ipa-profile.c (ipa_profile): Adjust hotness threshold
only in LTO mode.
2018-11-05  Martin Liska  <mliska@suse.cz>

PR gcov-profile/77698
* gcc.dg/tree-prof/pr77698.c: New test.

From-SVN: r265784

Daily bump.

From-SVN: r265783

Merge dmd upstream 6243fa6d2

This introduces a new header that pulls in system includes for use only
in the DMD front-end part of the compiler, fixing up uses of problematic
functions that are prevalent throughout the code.

Commits merged from dmd.

    Fix build of the D frontend on the Hurd and KFreeBSD.
    Initial patch from Matthias Klose.
    https://github.com/dlang/dmd/pull/8893

    Don't care about D/C++ compatibility in C++ port.
    Fixes build error in https://gcc.gnu.org/PR87788
    https://github.com/dlang/dmd/pull/8895

    Allow compiling front-end headers with strict warnings.
    https://github.com/dlang/dmd/pull/8909

    Add root/system.h header for wrapping system includes.
    Fixes https://gcc.gnu.org/PR87865
    https://github.com/dlang/dmd/pull/8910

    Move checkedint to dmd/root.
    https://github.com/dlang/dmd/pull/8912

    Use rmem instead of libc for malloc() and strdup().
    https://github.com/dlang/dmd/pull/8913

    Use align(8) for alignment of UnionExp, fixing several BUS errors
    due to alignment issues on SPARC.
    https://github.com/dlang/dmd/pull/8914

    Don't pass NULL pointer as format parameter to errorSupplemental.
    https://github.com/dlang/dmd/pull/8916

gcc/d/ChangeLog:

2018-11-05  Iain Buclaw  <ibuclaw@gdcproject.org>

PR d/87865
* d-system.h: New file.

From-SVN: r265780

re PR tree-optimization/86572 (unsafe strlen folding of const arguments with non-const offset)

gcc:
2018-11-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR tree-optimization/86572
        * builtins.c (c_strlen): Handle negative offsets in a safe way.

testsuite:
2018-11-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR tree-optimization/86572
        * gcc.dg/pr86572.c: New test.

From-SVN: r265778

re PR tree-optimization/87672 (81512c36 causes ICE in bootstrap stage 3 using "-D_FORTIFY_SOURCE=2"   (invalid operand in unary operation, incorrect sharing of tree nodes, verify_gimple failed))

gcc:
2018-11-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR tree-optimization/87672
        * gimple-fold.c (gimple_fold_builtin_stxcpy_chk): Gimplify.
        * tree-ssa-strlen.c (handle_builtin_strcat): Adjust object size.

testsuite:
2018-11-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>

        PR tree-optimization/87672
        * gcc.dg/pr87672.c: New test.

From-SVN: r265777

re PR c++/58372 (internal compiler error: ix86_compute_frame_layout)

PR middle-end/58372
* cfgexpand.c (pass_expand::execute): Move the call to
finish_eh_generation in front of the call to expand_stack_alignment.

testsuite/ChangeLog:

PR middle-end/58372
* g++.target/i386/pr58372.C: New test.

From-SVN: r265776

Enable support for next generation AMD Zen CPU, via -march=znver2.

gcc/ChangeLog:
* common/config/i386/i386-common.c (processor_alias_table): Add znver2 entry.
* config.gcc (i[34567]86-*-linux* | ...): Add znver2.
(case ${target}): Add znver2.
* config/i386/driver-i386.c: (host_detect_local_cpu): Let
-march=native recognize znver2 processors.
* config/i386/i386-c.c (ix86_target_macros_internal): Add znver2.
* config/i386/i386.c (m_znver2): New definition.
(m_ZNVER): New definition.
(m_AMD_MULTIPLE): Includes m_znver2.
(processor_cost_table): Add znver2 entry.
(processor_target_table): Add znver2 entry.
(get_builtin_code_for_version): Set priority for
PROCESSOR_ZNVER2.
(processor_model): Add M_AMDFAM17H_ZNVER2.
(arch_names_table): Ditto.
(ix86_reassociation_width): Include znver2.
* config/i386/i386.h (TARGET_znver2): New definition.
(struct ix86_size_cost): Add TARGET_ZNVER2.
(enum processor_type): Add PROCESSOR_ZNVER2.
* config/i386/i386.md (define_attr "cpu"): Add znver2.
* config/i386/x86-tune-costs.h: (processor_costs) Add znver2 costs.
* config/i386/x86-tune-sched.c: (ix86_issue_rate): Add znver2.
(ix86_adjust_cost): Add znver2.
* config/i386/x86-tune.def:  Replace m_ZNVER1 by m_ZNVER
* gcc/doc/extend.texi: Add details about znver2.
* gcc/doc/invoke.texi: Add details about znver2.

libgcc/ChangeLog
* config/i386/cpuinfo.c: (get_amd_cpu): Add znver2.
* config/i386/cpuinfo.h(processor_subtypes): Ditto.

From-SVN: r265775

MAINTAINERS (Write After Approval): Add myself.

2018-11-04  Stafford Horne  <shorne@gmail.com>

* MAINTAINERS (Write After Approval): Add myself.

From-SVN: r265774

Daily bump.

From-SVN: r265773

re PR target/87079 (nios2 optimization for size - case of regression relatively to 5.3.0)

2018-11-03  Sandra Loosemore  <sandra@codesourcery.com>

PR target/87079

gcc/
* config/nios2/nios2.c (nios2_rtx_costs): Recognize <mul>sidi3
pattern.

gcc/testsuite/
* gcc.target/nios2/pr87079-1.c: New.
* gcc.target/nios2/pr87079-2.c: New.

From-SVN: r265770

Daily bump.

From-SVN: r265766

re PR fortran/46020 (Improve error string for BIND(C) diagnostic for len>1 character return type)

2018-11-02  Thomas Koenig  <tkoenig@gcc.gnu.org>

    PR fortran/46020
    * decl.c (verify_bind_c_sym): Remove unnecessary space
    in error message.

From-SVN: r265757

Move gcc.dg/compat dg-skip-if to *_main.c files

* gcc.dg/compat/pr83487-1_y.c: Move dg-skip-if ...
* gcc.dg/compat/pr83487-1_main.c: ... here.
* gcc.dg/compat/struct-by-value-10_main.c,
gcc.dg/compat/struct-by-value-10_x.c,
gcc.dg/compat/struct-by-value-11_main.c,
gcc.dg/compat/struct-by-value-11_x.c,
gcc.dg/compat/struct-by-value-12_main.c,
gcc.dg/compat/struct-by-value-12_x.c,
gcc.dg/compat/struct-by-value-13_main.c,
gcc.dg/compat/struct-by-value-13_x.c,
gcc.dg/compat/struct-by-value-14_main.c,
gcc.dg/compat/struct-by-value-14_x.c,
gcc.dg/compat/struct-by-value-15_main.c,
gcc.dg/compat/struct-by-value-15_x.c,
gcc.dg/compat/struct-by-value-17_main.c,
gcc.dg/compat/struct-by-value-17_x.c,
gcc.dg/compat/struct-by-value-18_main.c,
gcc.dg/compat/struct-by-value-18_x.c,
gcc.dg/compat/struct-by-value-2_main.c,
gcc.dg/compat/struct-by-value-2_x.c,
gcc.dg/compat/struct-by-value-22_main.c,
gcc.dg/compat/struct-by-value-22_x.c,
gcc.dg/compat/struct-by-value-3_main.c,
gcc.dg/compat/struct-by-value-3_x.c,
gcc.dg/compat/struct-by-value-4_main.c,
gcc.dg/compat/struct-by-value-4_x.c,
gcc.dg/compat/struct-by-value-5b_main.c,
gcc.dg/compat/struct-by-value-5b_x.c,
gcc.dg/compat/struct-by-value-6b_main.c,
gcc.dg/compat/struct-by-value-6b_x.c,
gcc.dg/compat/struct-by-value-6b_main.c,
gcc.dg/compat/struct-by-value-7b_x.c,
gcc.dg/compat/struct-by-value-7b_main.c,
gcc.dg/compat/struct-by-value-8_main.c,
gcc.dg/compat/struct-by-value-8_x.c,
gcc.dg/compat/struct-by-value-9_main.c,
gcc.dg/compat/struct-by-value-9_x.c,
gcc.dg/compat/struct-return-2_main.c,
gcc.dg/compat/struct-return-2_x.c: Likewise.

From-SVN: r265755

rs6000-string.c (expand_strncmp_gpr_sequence): Pay attention to TARGET_AVOID_XFORM and BYTES_BIG_ENDIAN.

2018-11-02  Aaron Sawdey  <acsawdey@linux.ibm.com>

* config/rs6000/rs6000-string.c (expand_strncmp_gpr_sequence): Pay
attention to TARGET_AVOID_XFORM and BYTES_BIG_ENDIAN.

From-SVN: r265751

Although there's no fundamental reason why shrink wrapping and speculation tracking are incompatible...

Although there's no fundamental reason why shrink wrapping and
speculation tracking are incompatible, a phase-ordering requirement
(we need to do speculation tracking before the final basic block
clean-up) means that the shrink wrapping pass can undo some of the
changes the speculation tracking pass makes.  The result is that the
tracking, while still safe is less comprehensive than we really want.

So to keep things simple, and because the tracking code is quite
expensive anyway, it seems best to just disable that pass when we are
tracking speculative execution.

 * config/aarch64/aarch64.c (aarch64_override_options): Disable
 shrink-wrapping when -mtrack-speculation.

From-SVN: r265747

tree-ssa-coalesce.c (struct coalesce_list): Add obstack member.

2018-11-02  Richard Biener  <rguenther@suse.de>

* tree-ssa-coalesce.c (struct coalesce_list): Add obstack member.
(pop_cost_one_pair): Do not free pair.
(pop_best_coalesce): Likewise.
(create_coalesce_list): Initialize obstack.
(delete_coalesce_list): Free obstack.
(find_coalesce_pair): Obstack-allocate coalesce pairs.
(add_cost_one_coalesce): Likewise.
(struct live_track): Remove bitmap pointer indirections.
(new_live_track): Adjust.
(delete_live_track): Likewise.
(live_track_remove_partition): Likewise.
(live_track_add_partition): Likewise.
(live_track_live_p): Likewise.
(live_track_process_def): Likewise.
(live_track_clear_base_vars): Likewise.

From-SVN: r265745

[C++ PATCH] refactor duplicate_decls

https://gcc.gnu.org/ml/gcc-patches/2018-11/msg00085.html
gcc/cp/
* decl.c (duplicate_decls): Refactor checks.
gcc/testsuite/
* g++.dg/lookup/crash6.C: Adjust error
* g++.dg/parse/crash38.C: Likewise.

From-SVN: r265744

Remove SPARC HAVE_AS_REGISTER_PSEUDO_OP

* configure.ac (gcc_cv_as_sparc_register_op): Remove.
* configure: Regenerate.
* config.in: Regenerate.
* config/sparc/sparc.c (sparc_output_scratch_registers): Remove
HAVE_AS_REGISTER_PSEUDO_OP guard.
* config/sparc/sparc.h (ASM_ARCH64_SPEC): Remove
!HAVE_AS_REGISTER_PSEUDO_OP support.
(ASM_DECLARE_REGISTER_GLOBAL): Remove HAVE_AS_REGISTER_PSEUDO_OP
guard.

From-SVN: r265743

Fix libgphobos.spec in the wrong place with --enable-version-specific-runtime-libs

libphobos/ChangeLog:

2018-11-01  Iain Buclaw  <ibuclaw@gdcproject.org>

PR d/87827
* Makefile.in: Rebuild.
* configure: Rebuild.
* configure.ac: Properly set MULTISUBDIR.
* d_rules.am: Set toolexecdir and toolexeclibdir.
* libdruntime/Makefile.in: Rebuild.
* m4/druntime.m4 (DRUNTIME_INSTALL_DIRECTORIES): Add
--enable-version-specific-runtime-libs.
* src/Makefile.in: Rebuild.
* testsuite/Makefile.in: Rebuild.

From-SVN: r265742

20010904-2.c: Fix last change.

* gcc.c-torture/execute/20010904-2.c: Fix last change.
* gcc.dg/Wattributes-10.c: Likewise.

From-SVN: r265741

dwarf2out.c (build_abbrev_table): Guard lookup_external_ref call with is_type_die.

2018-11-02  Richard Biener  <rguenther@suse.de>

* dwarf2out.c (build_abbrev_table): Guard lookup_external_ref call
with is_type_die.

From-SVN: r265740

re PR tree-optimization/87776 (Compile time hog during RPO VN)

2018-11-02  Richard Biener  <rguenther@suse.de>

PR tree-optimization/87776
* tree-ssa-sccvn.c (do_rpo_vn): Do not mark backedges
executable when iterating but running into rpo-vn-max-loop-depth
and not eliding the iteration.

* gcc.dg/torture/pr87776.c: New testcase.

From-SVN: r265739

Daily bump.

From-SVN: r265738

tree.c (free_lang_data_in_decl): Free TREE_TYPE and alignment of TYPE_DECL.

* tree.c (free_lang_data_in_decl): Free TREE_TYPE and alignment of
TYPE_DECL.

From-SVN: r265735

Implement P0846R0, ADL and function templates.

* decl.c (grokfndecl): Allow FUNCTION_DECL in assert.
* lex.c (unqualified_fn_lookup_error): Handle TEMPLATE_ID_EXPR.
* parser.c (cp_parser_postfix_expression): Do ADL for a template-name.
(cp_parser_template_id): Give errors if parsing the template argument
list didn't go well.  Allow FUNCTION_DECL in assert.
(cp_parser_template_name): Consider a name to refer to a template if
it is an unqualified-id followed by a <.  Don't return the identifier
if the decl is a function and dependent.
* pt.c (tsubst_copy) <case OVERLOAD>: Remove assert.

* g++.dg/addr_builtin-1.C: Adjust dg-error.
* g++.dg/cpp2a/fn-template1.C: New test.
* g++.dg/cpp2a/fn-template10.C: New test.
* g++.dg/cpp2a/fn-template11.C: New test.
* g++.dg/cpp2a/fn-template12.C: New test.
* g++.dg/cpp2a/fn-template13.C: New test.
* g++.dg/cpp2a/fn-template14.C: New test.
* g++.dg/cpp2a/fn-template15.C: New test.
* g++.dg/cpp2a/fn-template16.C: New test.
* g++.dg/cpp2a/fn-template2.C: New test.
* g++.dg/cpp2a/fn-template3.C: New test.
* g++.dg/cpp2a/fn-template4.C: New test.
* g++.dg/cpp2a/fn-template5.C: New test.
* g++.dg/cpp2a/fn-template6.C: New test.
* g++.dg/cpp2a/fn-template7.C: New test.
* g++.dg/cpp2a/fn-template8.C: New test.
* g++.dg/cpp2a/fn-template9.C: New test.
* g++.dg/parse/fn-template1.C: New test.
* g++.dg/parse/fn-template2.C: New test.
* g++.dg/parse/template19.C: Adjust dg-error.
* g++.dg/template/pr61745.C: Add target to dg-error.

From-SVN: r265734

rs6000-protos.h (rs6000_address_for_fpconvert): Remove prototype.

2018-11-01  Aaron Sawdey  <acsawdey@linux.ibm.com>

* config/rs6000/rs6000-protos.h (rs6000_address_for_fpconvert): Remove
prototype.
* config/rs6000/rs6000.c (rs6000_force_indexed_or_indirect_mem):
Combine with rs6000_address_for_fpconvert.
(rs6000_address_for_fpconvert): Combine with
rs6000_force_indexed_or_indirect_mem.
(rs6000_expand_vector_init): Change function call from
rs6000_address_for_fpconvert to rs6000_force_indexed_or_indirect_mem.
* config/rs6000/rs6000.md (floatsi<mode>2_lfiwax): Change call from
rs6000_address_for_fpconvert to rs6000_force_indexed_or_indirect_mem.
(floatsi<mode>2_lfiwax_mem): Ditto.
(floatunssi<mode>2_lfiwzx): Ditto.
(floatunssi<mode>2_lfiwzx_mem): Ditto.
(float<QHI:mode><FP_ISA3:mode>2): Ditto.
(floatuns<QHI:mode><FP_ISA3:mode>2): Ditto.
(fix_trunc<mode>si2_stfiwx): Ditto.
(fixuns_trunc<mode>si2_stfiwx): Ditto.
(float_<mode>si2_hw): Ditto.
(floatuns_<mode>si2_hw): Ditto.
* config/rs6000/vsx.md (*vsx_extract_si): Ditto.
(vsx_splat_<mode>): Ditto.

From-SVN: r265733

re PR fortran/54613 ([F08] Add FINDLOC plus support MAXLOC/MINLOC with KIND=/BACK=)

2017-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

    PR fortran/54613
    * gfortran.dg/findloc_1.f90: Actually commit.
    * gfortran.dg/findloc_2.f90: Actually commit.
    * gfortran.dg/findloc_3.f90: Actually commit.
    * gfortran.dg/findloc_4.f90: Actually commit.
    * gfortran.dg/findloc_5.f90: Actually commit.
    * gfortran.dg/findloc_6.f90: Actually commit.

From-SVN: r265732

re PR fortran/87782 (runtime error: load of value 1818451807, which is not a valid value for type 'expr_t')

2018-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/87782
* frontend-passes.c (constant_string_length): If there is a
substring with a length which cannot be reduced to a constant,
return NULL.

From-SVN: r265730

re PR fortran/40196 ([F03] [F08] Type parameter inquiry (str%len, a%kind) and Complex parts (z%re, z%im))

2018-11-01  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/40196
* dependency.c (are_identical_variables): Return false if the
inquiry refs are not the same.
(gfc_ref_needs_temporary_p): Break on an inquiry ref.
* dump_parse_tree.c (show_ref): Show the inquiry ref type.
* expr.c (gfc_free_ref_list): Break on an inquiry ref.
(gfc_copy_ref): Copy the inquiry ref types.
(find_inquiry_ref): New function.
(simplify_const_ref, simplify_ref_chain): Call it. Add new arg
to simplify_ref_chain.
(gfc_simplify_expr): Use the new arg in call to
simplify_ref_chain.
(gfc_get_full_arrayspec_from_expr, gfc_is_coarray): Break on
inquiry ref.
(gfc_traverse_expr): Return true for inquiry ref.
* frontend-passes.c (gfc_expr_walker): Break on inquiry ref.
* gfortran.h : Add enums and union member in gfc_ref to
implement inquiry refs.
* intrinsic.c : Fix white nois.
* match.c (gfc_match_assignment): A constant lavlue is an
error.
* module.c : Add DECL_MIO_NAME for inquiry_type and the mstring
for inquiry_types.
(mio_ref): Handle inquiry refs.
* primary.c (is_inquiry_ref): New function.
(gfc_match_varspec): Handle inquiry refs calling new function.
(gfc_variable_attr): Detect inquiry ref for disambiguation
with components.
(caf_variable_attr): Treat inquiry and substring refs in the
same way.
* resolve.c (find_array_spec): ditto.
(gfc_resolve_substring_charlen): If there is neither a charlen
ref not an inquiry ref, return.
(resolve_ref): Handle inqiry refs as appropriate.
(resolve_allocate_expr): Entities with an inquiry ref cannot be
allocated.
* simplify.c (simplify_bound, simplify_cobound): Punt on
inquiry refs.
* trans-array.c (get_array_ctor_var_strlen): Break on inquiry
ref.
*trans-expr.c (conv_inquiry): New function.
(gfc_conv_variable): Retain the last typespec to pass to
conv_inquiry on detecting an inquiry ref.

2018-11-01  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/40196
* gfortran.dg/inquiry_part_ref_1.f08: New test.
* gfortran.dg/inquiry_part_ref_2.f90: New test.
* gfortran.dg/inquiry_part_ref_3.f90: New test.

From-SVN: r265729

20010904-1.c: Align 2 if pdp11.

* gcc.c-torture/execute/20010904-1.c: Align 2 if pdp11.
* gcc.c-torture/execute/20010904-2.c: Ditto.
* c-c++-common/builtin-arith-overflow-2.c: Skip if pdp11.
* gcc.dg/Walloc-size-larger-than-4.c: Ditto.
* gcc.dg/Walloc-size-larger-than-5.c: Ditto.
* gcc.dg/Walloc-size-larger-than-6.c: Ditto.
* gcc.dg/Walloc-size-larger-than-7.c: Ditto.
* gcc.dg/Walloca-14.c: Ditto.
* gcc.dg/Wlarger-than3.c: Ditto.
* gcc.dg/compat/pr83487-1_y.c: Ditto.
* gcc.dg/compat/struct-by-value-2_x.c: Ditto.
* gcc.dg/compat/struct-by-value-3_x.c: Ditto.
* gcc.dg/compat/struct-by-value-4_x.c: Ditto.
* gcc.dg/compat/struct-by-value-5b_x.c: Ditto.
* gcc.dg/compat/struct-by-value-6b_x.c: Ditto.
* gcc.dg/compat/struct-by-value-7b_x.c: Ditto.
* gcc.dg/compat/struct-by-value-8_x.c: Ditto.
* gcc.dg/compat/struct-by-value-9_x.c: Ditto.
* gcc.dg/compat/struct-by-value-10_x.c: Ditto.
* gcc.dg/compat/struct-by-value-11_x.c: Ditto.
* gcc.dg/compat/struct-by-value-12_x.c: Ditto.
* gcc.dg/compat/struct-by-value-13_x.c: Ditto.
* gcc.dg/compat/struct-by-value-14_x.c: Ditto.
* gcc.dg/compat/struct-by-value-15_x.c: Ditto.
* gcc.dg/compat/struct-by-value-16_x.c: Ditto.
* gcc.dg/compat/struct-by-value-17_x.c: Ditto.
* gcc.dg/compat/struct-by-value-18_x.c: Ditto.
* gcc.dg/compat/struct-by-value-22_x.c: Ditto.
* gcc.dg/compat/struct-return-2_x.c: Ditto.
* gcc.dg/falign-labels-1.c: Ditto.
* gcc.dg/long_branch.c: Ditto.
* gcc.dg/nextafter-1.c: Ditto.
* gcc.dg/pr35045.c: Ditto.
* gcc.dg/pr48616.c: Ditto.
* gcc.dg/pr84100.c: Ditto.
* gcc.dg/tree-ssa/builtin-sprintf-9.c: Ditto.
* gcc.dg/tree-ssa/builtin-sprintf-warn-10.c: Ditto.
* gcc.dg/tree-ssa/builtin-sprintf.c: Ditto.
* gcc.dg/Wattributes-10.c: Expect error if pdp11.
* gcc.dg/attr-alloc_size-11.c: Don't XFAIL if pdp11.
* gcc.dg/builtin-inf-1.c: Add pdp11 to warnings about INF.
* gcc.dg/builtins-1.c: Ditto.

From-SVN: r265728

re PR fortran/46020 (Improve error string for BIND(C) diagnostic for len>1 character return type)

2018-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

    PR fortran/46020
    * gfortran.dg/bind_c_18.f90: Adjust error message.

From-SVN: r265727

t-pdp11 (LIB2ADD): Add divmod.c.

* config/pdp11/t-pdp11 (LIB2ADD): Add divmod.c.
(HOST_LIBGCC2_CFLAGS): Change to optimize for size.

From-SVN: r265726

* g++.dg/cpp1z/constexpr-lambda22.C: New test,

From-SVN: r265725

Avoid stray [] in ia64 assembler test.

Noted in <https://gcc.gnu.org/ml/gcc-patches/2018-11/msg00034.html>.
Fix by adding extra newline in test input is analogous to many such
newlines added in cases where argument to a newly added AC_LANG_SOURCE
had a preprocessor directive on its last line.

* configure.ac (gcc_cv_as_ia64_ltoffx_ldxmov_relocs): Add newline
at end of assembler input text.
* configure: Regenerate.

From-SVN: r265724

cunroll-13.c: Add missing cast from literal to int32.

2018-11-01  Jozef Lawrynowicz  <jozef.l@mittosystems.com>

* gcc.dg/tree-ssa/cunroll-13.c: Add missing cast from literal to int32.

From-SVN: r265723

tree.c (free_lang_data_in_decl): Free TREE_TYPE and alignment of TYPE_DECL.

* tree.c (free_lang_data_in_decl): Free TREE_TYPE and alignment of
TYPE_DECL.

* g++.dg/lto/pr84805_1.C: Update template.
* g++.dg/lto/pr84805_0.C: Update template.
* g++.dg/lto/pr84805_1b.C: Update template.
* g++.dg/lto/pr84805_0b.C: Update template.

From-SVN: r265722

re PR tree-optimization/87826 (ubsan: gimple-ssa-store-merging.c:281)

PR tree-optimization/87826
* gimple-ssa-store-merging.c (do_shift_rotate): Punt if count is
negative or larger or equal to type's precision.

* gcc.dg/pr87826.c: New test.

From-SVN: r265720

re PR fortran/46020 (Improve error string for BIND(C) diagnostic for len>1 character return type)

2018-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/46020
* decl.c (verify_bind_c_sym): Improve error message.

From-SVN: r265719

re PR fortran/46020 (Improve error string for BIND(C) diagnostic for len>1 character return type)

2018-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

    PR fortran/46020
    * gfortran.dg/bind_c_usage_32.f90: Really commit.

From-SVN: r265717

re PR fortran/46020 (Improve error string for BIND(C) diagnostic for len>1 character return type)

2018-11-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

    PR fortran/46020
    * gfortran.dg/bind_c_usage_7.f90: Adjust error message.
    * gfortran.dg/bind_c_usage_32.f90: New test.

From-SVN: r265716

re PR fortran/54613 ([F08] Add FINDLOC plus support MAXLOC/MINLOC with KIND=/BACK=)

2018-10-28  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/54613
* gfortran.texi (File format of unformatted sequential files):
Replace random comma with period.
* intrinsic.texi (Intrinsic Procedures): Add FINDLOC to menu.
(FINDLOC): Document.
(MAXLOC): Add refrence to FINDLOC.
(MINLOC): Likewise.

From-SVN: r265715

[ABI PATCH] static anonymous unions of function scope

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02076.html
* cp-tree.h (struct lang_function): Delete x_local_names field.
(struct lang_decl_base): Rename u2sel to spare.
(struct lang_decl_min): Remove lang_decl_u2 union.  Keep access
field.
(LANG_DECL_U2_CHECK): Delete.
(DECL_DISCRIMINATOR_P): Require function scope.
(DECL_DISCRIMINATOR): Adjust.
(DECL_DISCRIMINATOR_SET_P): Delete.
(DECL_CAPTURED_VARIABLE, DECL_ACCESS, THUnK_VIRTUAL_OFFSET): Adjust.
(local_classes): Don't declare.
(determine_local_discriminator): Declare.
* decl.c (push_local_name): Delete.
(local_entities, determina_local_discrminator): New.
(duplicate_decls): Copy DECL_ACCESS. Fix formatting.
(cp_finish_decl): Use determine_local_discriminator.
(save_function_data): Drop x_local_names.
(finish_function): Drop local_names.
* decl2.c (finish_anon_union): Use determine_local_disciminator.
* mangle.c (write_unnamed_type_name): Use
discriminator_for_local_entity.
(local_class_index): Delete.
(discriminator_for_local_entity): Reimplement.
(write_local_name): Adjust discriminator code.
* name-lookup.c (do_pushtag): Call determine_local_discrimiator.
* semantics.c (finish_omp_threadprivate): Drop DECL_DISCRIMINATOR
handling.
* class.c (local_classes): Delete.
(init_class_processing): Don't init it.

* g++.dg/abi/anon5.C: New.

From-SVN: r265714

re PR d/87824 (x86_64-linux multilib issues)

PR d/87824
* lang.opt (Wpsabi): New option.

* gdc.dg/simd.d: Add -Wno-psabi.
* gdc.dg/compilable.d: Likewise.

From-SVN: r265713

Make __PRETTY_FUNCTION__-like functions mergeable string csts (PR c++/64266).

2018-11-01  Martin Liska  <mliska@suse.cz>
    Jason Merrill  <jason@redhat.com>

PR c++/64266
PR bootstrap/70422
PR ipa/81277
* cp-tree.h (DECL_FNAME_P): New macro.
* decl.c (cp_make_fname_decl): Set DECL_DECLARED_CONSTEXPR_P,
        DECL_VALUE_EXPR, DECL_INITIALIZED_BY_CONSTANT_EXPRESSION_P.
(cp_finish_decl):
* lambda.c (is_capture_proxy): Use DECL_FNAME_P.
* pt.c (tsubst_expr): Handle DECL_PRETTY_FUNCTION_P.
2018-11-01  Martin Liska  <mliska@suse.cz>
    Jason Merrill  <jason@redhat.com>

PR c++/64266
PR bootstrap/70422
PR ipa/81277
* g++.dg/cpp0x/constexpr-__func__2.C: Make it a compilation
test.
* g++.old-deja/g++.ext/pretty4.C: Remove as the run-time
assumptions are not longer valid.

Co-Authored-By: Jason Merrill <jason@redhat.com>
From-SVN: r265711

go-gcc.cc (Gcc_backend::write_global_definitions): Don't call gimplify_function_tree.

* go-gcc.cc (Gcc_backend::write_global_definitions): Don't call
gimplify_function_tree.  Instead call allocate_struct_function if
necessary.

From-SVN: r265710

Fix typo in gcc/ChangeLog

From-SVN: r265709

gOlogy: skip dbranch at -Og

Delayed slot filling moves insns without any regard to variable
location notes, causing the location information in them to become
incorrect.

Fixing that appears to be quite difficult, but filling delay slots is
hardly an essential optimization to run at -Og, so if the user wants
to privilege debuggability, skip delay slot filling.

for  gcc/ChangeLog

* opts.c (default_options_table): Do not enable
OPT_fdelayed_branch at -Og.
* doc/invoke.texi (-fdelayed-branch): Document it.

From-SVN: r265708

Daily bump.

From-SVN: r265707

Add visibility to libfunc constructors

* optabs-libfuncs.c (build_libfunc_function_visibility):
New, split out from...
(build_libfunc_function): ... here.
(init_one_libfunc_visibility): New, split out from ...
(init_one_libfunc): ... here.

From-SVN: r265704

aarch64: Remove early clobber from ATOMIC_LDOP scratch

* config/aarch64/atomics.md (aarch64_atomic_<ATOMIC_LDOP><ALLI>_lse):
The scratch register need not be early-clobber.  Document the reason
why we cannot use ST<OP>.

From-SVN: r265703

Fix profiledbootstrap when D is enabled

gcc/d/ChangeLog:

2018-10-31  Iain Buclaw  <ibuclaw@gdcproject.org>

PR d/87825
* Make-lang.in (d/idgen) Link with BUILD_LINKERFLAGS.
(d/impcvgen): Likewise.

From-SVN: r265702

re PR bootstrap/82856 (--enable-maintainter-mode broken by incompatiblity of gcc's required automake and modern Perl)

PR bootstrap/82856

    libgo: update to autoconf 2.69 and automake 1.15.1

    Initial patch from Joseph Myers.

    Reviewed-on: https://go-review.googlesource.com/c/146417

From-SVN: r265701

re PR fortran/20520 (allocatable arrays used uninitialized without a warning)

2018-10-31  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/20520
* gfortran.dg/allocatable_uninitialized_1.f90: New test.

From-SVN: r265698

[C++ PATCH] missed testcase

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02065.html
* g++.dg/lookup/friend21.C: New.

From-SVN: r265697

[PATCH] command line macros

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02062.html
* c-opts.c (c_finish_options): Force command line macro
location.  Refactor to avoid repeating main debug hook.
(push_command_line_include): Clarify comment.

From-SVN: r265696

Update GCC to autoconf 2.69, automake 1.15.1 (PR bootstrap/82856).

This patch updates GCC to use autoconf 2.69 and automake 1.15.1.
(That's not the latest automake version, but it's the one used by
binutils-gdb, with which consistency is desirable, and in any case
seems a useful incremental update that should make a future update to
1.16.1 easier.)

The changes are generally similar to the binutils-gdb ones, and are
copied from there where shared files and directories are involved
(there are some further changes to such shared directories, however,
which I'd expect to apply to binutils-gdb once this patch is in GCC).
Largely, obsolete AC_PREREQ calls are removed, while many
AC_LANG_SOURCE calls are added to avoid warnings from aclocal and
autoconf.  Multilib support is no longer included in core automake,
meaning that multilib.am needs copying from automake's contrib
directory into the GCC source tree.  Autoconf 2.69 has Go support, so
local copies of that support are removed.  I hope the D support will
soon be submitted to upstream autoconf so the local copy of that can
be removed in a future update.  Changes to how automake generates
runtest calls mean quotes are removed from RUNTEST definitions in five
lib*/testsuite/Makefile.am files (libatomic, libgomp, libitm,
libphobos, libvtv; some others have RUNTEST definitions without
quotes, which are still OK); libgo and libphobos also get
-Wno-override added to AM_INIT_AUTOMAKE so those overrides of RUNTEST
do not generate automake warnings.

Note that the regeneration did not include regeneration of
fixincludes/config.h.in (attempting such regeneration resulted in all
the USED_FOR_TARGET conditionals disappearing; and I don't see
anything in the fixincludes/ directory that would result in such
conditionals being generated, unlike in the gcc/ directory).  Also
note that libvtv/testsuite/other-tests/Makefile.in was not
regenerated; that directory is not listed as a subdirectory for which
Makefile.in gets regenerated by calling "automake" in libvtv/, so I'm
not sure how it's meant to be regenerated.

While I mostly fixed warnings should running aclocal / automake /
autoconf, there were various such warnings from automake in the
libgfortran, libgo, libgomp, liboffloadmic, libsanitizer, libphobos
directories that I did not fix, preferring to leave those to the
relevant subsystem maintainers.  Specifically, most of those warnings
were of the following form (example from libgfortran):

Makefile.am:48: warning: source file 'caf/single.c' is in a subdirectory,
Makefile.am:48: but option 'subdir-objects' is disabled
automake: warning: possible forward-incompatibility.
automake: At least a source file is in a subdirectory, but the 'subdir-objects'
automake: automake option hasn't been enabled.  For now, the corresponding output
automake: object file(s) will be placed in the top-level directory.  However,
automake: this behaviour will change in future Automake versions: they
will
automake: unconditionally cause object files to be placed in the same subdirectory
automake: of the corresponding sources.
automake: You are advised to start using 'subdir-objects' option throughout your
automake: project, to avoid future incompatibilities.

I think it's best for the relevant maintainers to add subdir-objects
and do any other associated Makefile.am changes needed.  In some cases
the paths in the warnings involved ../; I don't know if that adds any
extra complications to the use of subdir-objects.

I've tested this with native, cross and Canadian cross builds.  The
risk of any OS-specific issues should I hope be rather lower than if a
libtool upgrade were included (we *should* do such an upgrade at some
point, but it's more complicated - it involves identifying all our
local libtool changes to see if any aren't included in the upstream
version we update to, and reverting an upstream libtool patch that's
inappropriate for use in GCC); I think it would be better to get this
update into GCC so that people can test in different configurations
and we can fix any issues found, rather than to try to get more and
more testing done before it goes in.

top level:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* multilib.am: New file.  From automake.

Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* libtool.m4: Use AC_LANG_SOURCE.
* configure.ac: Remove AC_PREREQ, use AC_LANG_SOURCE.
* ar-lib: New file.
* test-driver: New file.
* configure: Re-generate.

config:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* math.m4, tls.m4: Use AC_LANG_SOURCE.

Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* override.m4 (_GCC_AUTOCONF_VERSION): Bump from 2.64 to 2.69.

fixincludes:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* aclocal.m4, configure: Regenerate.

gcc:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.  Use single
line for second argument of AC_DEFINE_UNQUOTED.
* doc/install.texi (Tools/packages necessary for modifying GCC):
Update to autoconf 2.69 and automake 1.15.1.
* aclocal.m4, config.in, configure: Regenerate.

gnattools:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* configure: Regenerate.

gotools:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* config/go.m4: Remove file.
* Makefile.am (ACLOCAL_AMFLAGS): Do not use -I ./config.
* configure.ac:  Remove AC_PREREQ.  Do not include config/go.m4.
* Makefile.in, aclocal.m4, configure: Regenerate.

intl:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* configure.ac: Add AC_USE_SYSTEM_EXTENSIONS, remove AC_PREREQ.
* configure: Re-generate.
* config.h.in: Re-generate.
* aclocal.m4: Re-generate.

libada:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* configure: Regenerate.

libatomic:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* acinclude.m4: Use AC_LANG_SOURCE.
* configure.ac: Remove AC_PREREQ.
* testsuite/Makefile.am (RUNTEST): Remove quotes.
* Makefile.in, aclocal.m4, configure, testsuite/Makefile.in:
Regenerate.

libbacktrace:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.
* Makefile.in, aclocal.m4, config.h.in, configure: Regenerate.

libcc1:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, configure: Regenerate.

libcpp:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.
* aclocal.m4, config.in, configure: Regenerate.

libdecnumber:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* configure.ac: Remove AC_PREREQ.
* configure: Re-generate.
* aclocal.m4.

libffi:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
(AUTOMAKE_OPTIONS): Add info-in-builddir.
(CLEANFILES): Remove doc/libffi.info.
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, configure, fficonfig.h.in,
include/Makefile.in, man/Makefile.in, testsuite/Makefile.in:
Regenerate.

libgcc:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.
* configure: Regenerate.

libgfortran:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, config.h.in, configure: Regenerate.

libgo [logically part of this change but omitted from the commit]:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* config/go.m4: Remove file.
* config/libtool.m4: Use AC_LANG_SOURCE.
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.  Use
-Wno-override in AM_INIT_AUTOMAKE call.
* Makefile.in, aclocal.m4, configure, testsuite/Makefile.in:
Regenerate.

libgomp:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am
(AUTOMAKE_OPTIONS): Add info-in-builddir.
(CLEANFILES): Remove libgomp.info.
* configure.ac: Remove AC_PREREQ.
* testsuite/Makefile.am (RUNTEST): Remove quotes.
* Makefile.in, aclocal.m4, configure, testsuite/Makefile.in:
Regenerate.

libhsail-rt:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, configure: Regenerate.

libiberty:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* configure.ac: Remove AC_PREREQ.
* configure: Re-generate.
* config.in: Re-generate.

libitm:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
(AUTOMAKE_OPTIONS): Add info-in-builddir.
(CLEANFILES): Remove libitm.info.
* configure.ac: Remove AC_PREREQ.
* testsuite/Makefile.am (RUNTEST): Remove quotes.
* Makefile.in, aclocal.m4, configure, testsuite/Makefile.in:
Regenerate.

libobjc:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.
* aclocal.m4, config.h.in, configure: Regenerate.

liboffloadmic:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.
* plugin/Makefile.am: Include multilib.am.
* plugin/configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, configure, plugin/Makefile.in,
plugin/aclocal.m4, plugin/configure: Regenerate.

libphobos:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.  Use -Wno-override in
AM_INIT_AUTOMAKE call.
* m4/autoconf.m4: Add extra argument to AC_LANG_DEFINE call.
* m4/druntime/os.m4: Use AC_LANG_SOURCE.
* testsuite/Makefile.am (RUNTEST): Remove quotes.
* Makefile.in, aclocal.m4, configure, libdruntime/Makefile.in,
src/Makefile.in, testsuite/Makefile.in: Regenerate.

libquadmath:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
(AUTOMAKE_OPTIONS): Remove 1.8.  Add info-in-builddir.
(all-local): Define outside conditional code.
(CLEANFILES): Remove libquadmath.info.
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, config.h.in, configure: Regenerate.

libsanitizer:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.
* Makefile.in, aclocal.m4, asan/Makefile.in, configure,
interception/Makefile.in, libbacktrace/Makefile.in,
lsan/Makefile.in, sanitizer_common/Makefile.in, tsan/Makefile.in,
ubsan/Makefile.in: Regenerate.

libssp:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
(AUTOMAKE_OPTIONS): Remove 1.9.5.
* configure.ac: Remove AC_PREREQ.  Quote argument to
AC_RUN_IFELSE.
* Makefile.in, aclocal.m4, configure: Regenerate.

libstdc++-v3:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.
* Makefile.in, aclocal.m4, configure, doc/Makefile.in,
include/Makefile.in, libsupc++/Makefile.in, po/Makefile.in,
python/Makefile.in, src/Makefile.in, src/c++11/Makefile.in,
src/c++17/Makefile.in, src/c++98/Makefile.in,
src/filesystem/Makefile.in, testsuite/Makefile.in: Regenerate.

libvtv:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.
* configure.ac: Remove AC_PREREQ.
* testsuite/Makefile.am (RUNTEST): Remove quotes.
* Makefile.in, aclocal.m4, configure, testsuite/Makefile.in:
Regenerate.

lto-plugin:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* configure.ac: Remove AC_PREREQ.  Use AC_LANG_SOURCE.
* Makefile.in, aclocal.m4, config.h.in, configure: Regenerate.

zlib:
2018-10-31  Joseph Myers  <joseph@codesourcery.com>

PR bootstrap/82856
* Makefile.am: Include multilib.am.

Merge from binutils-gdb:
2018-06-19  Simon Marchi  <simon.marchi@ericsson.com>

* configure.ac: Modernize AC_INIT call, remove AC_PREREQ.
* Makefile.am (AUTOMAKE_OPTIONS): Remove 1.8, cygnus, add foreign.
* Makefile.in: Re-generate.
* aclocal.m4: Re-generate.
* configure: Re-generate.

From-SVN: r265695

tree-ssa-loop-ivcanon.c (try_peel_loop): Fix typo and remove dead initialization.

* tree-ssa-loop-ivcanon.c (try_peel_loop): Fix typo and remove dead
initialization.

From-SVN: r265693

[6/6] Preprocessor forced macro location

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02044.html
libcpp/
* internal.h (struct cpp_reader): Rename forced_token_location_p
to forced_token_location and drop its pointerness.
* include/cpplib.h (cpp_force_token_locations): Take location, not
pointer to one.
* init.c (cpp_create_reader): Adjust.
* lex.c (cpp_read_main_file):

gcc/c-family/
* c-opts.c (c_finish_options): Adjust cpp_force_token_locations call.

gcc/fortran/
* cpp.c (gfc_cpp_init): Adjust cpp_force_token_locations call.

From-SVN: r265692

Add myself to MAINTAINERS

ChangeLog:

2018-10-31  Michael Ploujnikov  <michael.ploujnikov@oracle.com>

* MAINTAINERS (Write After Approval): Add myself.

From-SVN: r265691

[5/6] Preprocessor include

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02041.html
* directives.c (do_include_common): Commonize cleanup path.
(_cpp_pop_buffer): Fix leak.

From-SVN: r265690

[4/7] Preprocessor location-kind predicates

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02040.html
* include/line-map.h (IS_ORDINARY_LOC, IS_MACRO_LOC): New
predicates.
(IS_ADHOC_LOC): Move earlier.
(MAP_ORDINARY_P): Use IS_ORDINARY_LOC.
* line-map.c (linemap_location_from_macro_expansion_p): Use
IS_MACRO_LOC.

From-SVN: r265689

[3/7] Preprocessor macro loc

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02039.html
* include/cpplib.h (cpp_macro_definition_location): Make inline.
* macro.c (warn_of_redefinition): Fix comments, examine macro
type, use C++ for.
(cpp_macro_definition_location): Don't define here.

From-SVN: r265688

[2/7] Preprocessor node access

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02038.html
* include/cpplib.h (HT_NODE): Don't cast NODE.
(NODE_LEN, NODE_NAME): Use HT_NODE.

From-SVN: r265687

Provide extension hint for aarch64 target (PR driver/83193).

2018-10-31  Martin Liska  <mliska@suse.cz>

PR driver/83193
* common/config/aarch64/aarch64-common.c (aarch64_parse_extension):
Add new argument invalid_extension.
(aarch64_get_all_extension_candidates): New function.
(aarch64_rewrite_selected_cpu): Add NULL to function call.
* config/aarch64/aarch64-protos.h (aarch64_parse_extension): Add
new argument.
(aarch64_get_all_extension_candidates): New function.
* config/aarch64/aarch64.c (aarch64_parse_arch): Add new
argument invalid_extension.
(aarch64_parse_cpu): Likewise.
(aarch64_print_hint_for_extensions): New function.
(aarch64_validate_mcpu): Provide hint about invalid extension.
(aarch64_validate_march): Likewise.
(aarch64_handle_attr_arch): Pass new argument.
(aarch64_handle_attr_cpu): Provide hint about invalid extension.
(aarch64_handle_attr_isa_flags): Likewise.
2018-10-31  Martin Liska  <mliska@suse.cz>

PR driver/83193
* gcc.target/aarch64/spellcheck_7.c: New test.
* gcc.target/aarch64/spellcheck_8.c: New test.
* gcc.target/aarch64/spellcheck_9.c: New test.

From-SVN: r265686

[1/7] Preprocessor cleanup

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02037.html
* directives.c (DIRECTIVE_TABLE): Drop historical frequency
comments.
* files.c (_cpp_stack_file): Fix indentation.

From-SVN: r265685

More testing for std::pair layout change

* testsuite/20_util/pair/87822.cc: Test deeper nesting.

From-SVN: r265680

[C++ PATCH] Simplify overloads

https://gcc.gnu.org/ml/gcc-patches/2018-10/msg02026.html
gcc/cp/
* cp-tree.h (OVL_DEDUP_P): New.
* name-lookup.c (name_lookup::add_overload): Check OVL_DEDUP_P.
(get_class_binding_direct): Likwise.
* tree.c (ovl_make): Propagate OVL_DEDUP_P.
(ovl_copy): Copy it.
(ovl_insert): Do not keep using-decls ordered.
(lookup_maybe_add): Adjust comment.

gcc/testsuite/
* g++.dg/lookup/using60.C: New.

From-SVN: r265679

PR libstdc++/87822 fix layout change for nested std::pair

The introduction of the empty __pair_base base class for PR 86751
changed the layout of std::pair<std::pair<...>, ...>. The outer pair and
its first member both have a base class of the same type, which cannot
exist at the same address. This causes the first member to be at a
non-zero offset.

The solution is to make the base class depend on the template
parameters, so that each pair type has a different base class type,
which allows the base classes of the outer pair and its first member to
have the same address.

PR libstdc++/87822
* include/bits/stl_pair.h (__pair_base): Change to class template.
(pair): Make base class type depend on template parameters.
* testsuite/20_util/pair/87822.cc: New test.

From-SVN: r265678

re PR middle-end/70359 (Code size increase for x86/ARM/others compared to gcc-5.3.0)

2018-10-31  Richard Biener  <rguenther@suse.de>

PR middle-end/70359
PR middle-end/86270
* tree-outof-ssa.c (insert_backedge_copies): Restrict
copy generation to useful cases.  Place the copy before
the definition of the backedge value when possible.

* gcc.target/i386/pr70359.c: New testcase.
* gcc.target/i386/pr86270.c: Likewise.

From-SVN: r265677

[ARC] Handle store cacheline hazard.

Handle store cacheline hazard for A700 cpus by inserting two NOP_S
between ST ST LD or their logical equivalent (like ST ST NOP_S NOP_S
J_L.D LD)

gcc/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

        * config/arc/arc-arch.h (ARC_TUNE_ARC7XX): New tune value.
        * config/arc/arc.c (arc_active_insn): New function.
        (check_store_cacheline_hazard): Likewise.
        (workaround_arc_anomaly): Use check_store_cacheline_hazard.
        (arc_override_options): Disable delay slot scheduler for older
        A7.
        (arc_store_addr_hazard_p): New implementation, old one renamed to
        ...
        (arc_store_addr_hazard_internal_p): Renamed.
        (arc_reorg): Don't combine into brcc instructions which are part
        of hardware hazard solution.
        * config/arc/arc.md (attr tune): Consider new arc7xx tune value.
        (tune_arc700): Likewise.
        * config/arc/arc.opt (arc7xx): New tune value.
        * config/arc/arc700.md: Improve A7 scheduler.

From-SVN: r265676

[ARC] Add BI/BIH instruction support.

Use BI/BIH instruction to implement casesi pattern. Only ARC V2.

gcc/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

        * config/arc/arc.c (arc_override_options): Remove
        TARGET_COMPACT_CASESI.
        * config/arc/arc.h (ASM_OUTPUT_ADDR_DIFF_ELT): Update.
        (CASE_VECTOR_MODE): Likewise.
        (CASE_VECTOR_PC_RELATIVE): Likewise.
        (CASE_VECTOR_SHORTEN_MODE): Likewise.
        (CASE_VECTOR_SHORTEN_MODE1): Delete.
        (ADDR_VEC_ALIGN): Update.
        (ASM_OUTPUT_CASE_LABEL): Undefine.
        (ASM_OUTPUT_BEFORE_CASE_LABEL): Undefine.
        (TARGET_BI_BIH): Define.
        (DEFAULT_BRANCH_INDEX): Likewise.
        * config/arc/arc.md (casesi): Rework to accept BI/BIH
        instructions, remove compact_casesi use case.
        (casesi_compact_jump): Remove.
        (casesi_dispatch): New pattern.
        * config/arc/arc.opt: Add mbranch-index option. Deprecate
        compact_casesi option.
        * doc/invoke.texi: Document mbranch-index option.

gcc/testsuite
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

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

From-SVN: r265675

[ARC] Cleanup TLS implementation and add a number of tests.

gcc/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (arc_get_tp): Remove function.
(arc_emit_call_tls_get_addr): Likewise.
(arc_call_tls_get_addr): New function.
(arc_legitimize_tls_address): Make use of arc_call_tls_get_addr.
* config/arc/arc.md (tls_load_tp_soft): Remove.
(tls_gd_get_addr): Likewise.

testsuite/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/tls-gd.c: New file.
* gcc.target/arc/tls-ie.c: Likewise.
* gcc.target/arc/tls-ld.c: Likewise.
* gcc.target/arc/tls-le.c: Likewise.
* gcc.target/arc/tls-1.c: Update test.

From-SVN: r265674

Update LOCAL_PATCHES after libsanitizer merge.

2018-10-31  Martin Liska  <mliska@suse.cz>

* LOCAL_PATCHES: Update to installed revisions.

From-SVN: r265673

[ARC] Remove non standard funcions calls.

Replace all custom "library" calls with compiler known patterns.

gcc/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.md (mulsi3): Remove call to mulsi_600_lib.
(mulsi3_600_lib): Remove pattern.
(umulsi3_highpart_600_lib_le): Likewise.
(umulsi3_highpart): Remove call to umulsi3_highpart_600_lib_le.
(umulsidi3): Remove call to umulsidi3_600_lib.
(umulsidi3_600_lib): Remove pattern.
(peephole2): Remove peephole using the above deprecated patterns.

testsuite/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/mulsi3_highpart-2.c: Update test.

libgcc/
xxxx-xx-xx  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/lib1funcs.S (_muldi3): New function.
* config/arc/t-arc (LIB1ASMFUNCS): Add _muldi3.

From-SVN: r265672

Update test-suite expected output after rewording in libsanitizer.

2018-10-31  Martin Liska  <mliska@suse.cz>

* c-c++-common/ubsan/pr63839.c: Update to new sanitizer message.
* c-c++-common/ubsan/unreachable-1.c (main): Likewise.
* c-c++-common/ubsan/unreachable-2.c: Likewise.
* c-c++-common/ubsan/unreachable-4.c (main): Likewise.

From-SVN: r265670

New local GCC patch for CAN_SANITIZE_UB ifdef.

2018-10-31  Martin Liska  <mliska@suse.cz>

* ubsan/ubsan_platform.h: Add ifndef as we define it with
-DCAN_SANITIZE_UB CFLAGS.

From-SVN: r265669