Daily bump.

From-SVN: r247709

store-motion.c (remove_reachable_equiv_notes): Reformat long lines.

* store-motion.c (remove_reachable_equiv_notes): Reformat long
lines.  Use for (;;).

From-SVN: r247705

rs6000.c (rs6000_vect_nonmem): New static var.

[gcc]

2017-05-05  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

* config/rs6000/rs6000.c (rs6000_vect_nonmem): New static var.
(rs6000_init_cost): Initialize rs6000_vect_nonmem.
(rs6000_add_stmt_cost): Update rs6000_vect_nonmem.
(rs6000_finish_cost): Avoid vectorizing simple copy loops with
VF=2 that require versioning.

[gcc/testsuite]

2017-05-05  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

* gcc.target/powerpc/versioned-copy-loop.c: New file.

From-SVN: r247671

Convert CARET_LINE_MARGIN to const int

gcc/ChangeLog:
* diagnostic.h (CARET_LINE_MARGIN): Convert from macro to const
int.

From-SVN: r247666

Convert diagnostic_override_option_index from macro to inline function

gcc/ChangeLog:
* diagnostic.h (diagnostic_override_option_index): Convert from
macro to inline function.

From-SVN: r247665

Get rid of macros for diagnostic_report_current_module

diagnostic.h has a couple of macros (diagnostic_last_module_changed
and diagnostic_set_last_module) which are only used within
diagnostic_report_current_module.

This patch eliminates the macros in favor of static functions within
diagnostic.c.

No functional change intended.

gcc/ChangeLog:
* diagnostic.c (last_module_changed_p): New function.
(set_last_module): New function.
(diagnostic_report_current_module): Convert macro usage to
the above functions.
* diagnostic.h (diagnostic_context::last_module): Strengthen
from const line_map * to const line_map_ordinary *.
(diagnostic_last_module_changed): Delete macro.
(diagnostic_set_last_module): Delete macro.

From-SVN: r247664

Eliminate report_diagnostic macro

This patch eliminates the report_diagnostic macro, manually
expanding it in all sites in the code.

No functional change intended.

gcc/c-family/ChangeLog:
* c-common.c (c_cpp_error): Replace report_diagnostic
with diagnostic_report_diagnostic.

gcc/c/ChangeLog:
* c-decl.c (warn_defaults_to): Replace report_diagnostic
with diagnostic_report_diagnostic.
* c-errors.c (pedwarn_c99): Likewise.
(pedwarn_c90): Likewise.

gcc/cp/ChangeLog:
* error.c (pedwarn_cxx98): Replace report_diagnostic
with diagnostic_report_diagnostic.

gcc/ChangeLog:
* diagnostic.c (diagnostic_impl): Replace report_diagnostic
with diagnostic_report_diagnostic.
(diagnostic_n_impl_richloc): Likewise.
* diagnostic.h (report_diagnostic): Delete macro.
* rtl-error.c (diagnostic_for_asm): Replace report_diagnostic
with diagnostic_report_diagnostic.
* substring-locations.c (format_warning_va): Likewise.

gcc/fortran/ChangeLog:
* cpp.c (cb_cpp_error): Replace report_diagnostic
with diagnostic_report_diagnostic.
* error.c (gfc_warning): Likewise.
(gfc_warning_now_at): Likewise.
(gfc_warning_now): Likewise.
(gfc_warning_internal): Likewise.
(gfc_error_now): Likewise.
(gfc_fatal_error): Likewise.
(gfc_error_opt): Likewise.
(gfc_internal_error): Likewise.

From-SVN: r247663

diagnostic.c: add print_option_information

This patch simplifies diagnostic_report_diagnostic by moving
option-printing to a new subroutine.

Doing so required a slight rewrite.  In both the old and new
code, context->option_name returns a malloc-ed string.
The old behavior was to then use ACONCAT to manipulate the
format_spec, appending the option metadata.
ACONCAT calcs the buffer size, then uses alloca, and then copies the
data to the on-stack buffer.

Given the alloca, this needs rewriting when moving the printing to
a subroutine.  In the new version, the metadata is simply printed
using pp_* calls (so it's hitting the obstack within the
pretty_printer).

This means we can get rid of the save/restore of format_spec: I don't
believe anything else in the code modifies it.

It also seems inherently simpler; it seems odd to me to be
appending metadata to the formatting string, rather than simply
printing the metadata after the formatted string is printed
(the old code also assumed that no option name contained a '%').

No functional change intended.

gcc/ChangeLog:
* diagnostic.c (diagnostic_report_diagnostic): Eliminate
save/restor of format_spec.  Move option-printing code to...
(print_option_information): ...this new function, and
reimplement by simply printing to the pretty_printer,
rather than appending to the format string.

From-SVN: r247661

diagnostic_report_diagnostic: refactor pragma-handling

This patch simplifies diagnostic_report_diagnostic by moving the
pragma-handling logic into a subroutine.

No functional change intended.

gcc/ChangeLog:
* diagnostic.c (diagnostic_report_diagnostic): Split out pragma
handling logic into...
(update_effective_level_from_pragmas): ...this new function.

From-SVN: r247660

RISC-V: Add -mstrict-align option

The RISC-V user ISA permits misaligned accesses, but they may trap
and be emulated.  That emulation software needs to be compiled assuming
strict alignment.

Even when strict alignment is not required, set SLOW_UNALIGNED_ACCESS
based upon -mtune to avoid a performance pitfall.

gcc/ChangeLog:

2017-05-04  Andrew Waterman  <andrew@sifive.com>

* config/riscv/riscv.opt (mstrict-align): New option.
* config/riscv/riscv.h (STRICT_ALIGNMENT): Use it.  Update comment.
(SLOW_UNALIGNED_ACCESS): Define.
(riscv_slow_unaligned_access): Declare.
* config/riscv/riscv.c (riscv_tune_info): Add slow_unaligned_access
field.
(riscv_slow_unaligned_access): New variable.
(rocket_tune_info): Set slow_unaligned_access to true.
(optimize_size_tune_info): Set slow_unaligned_access to false.
(riscv_cpu_info_table): Add entry for optimize_size_tune_info.
(riscv_valid_lo_sum_p): Use TARGET_STRICT_ALIGN.
(riscv_option_override): Set riscv_slow_unaligned_access.
        * doc/invoke.texi: Add -mstrict-align to RISC-V.

From-SVN: r247659

RISC-V: Unify indention in riscv.md

This contains only whitespace changes.

gcc/ChangeLog

2017-05-04  Kito Cheng  <kito.cheng@gmail.com>

* config/riscv/riscv.md: Unify indentation.

From-SVN: r247658

re PR target/79038 (Improve PowerPC ISA 3.0 conversion between integers and hardware _Float128)

[gcc]
2017-05-05  Michael Meissner  <meissner@linux.vnet.ibm.com>

PR target/79038
PR target/79202
PR target/79203
* config/rs6000/rs6000.md (u code attribute): Add FIX and
UNSIGNED_FIX.
(extendsi<mode>2): Add support for doing sign extension via
VUPKHSW and XXPERMDI if the value is in Altivec registers and we
don't have ISA 3.0 instructions.
(extendsi<mode>2 splitter): Likewise.
(fix_trunc<mode>si2): If we are at ISA 2.07 (VSX small integer),
generate the normal insns since SImode can now go in vector
registers.  Disallow the special UNSPECs needed for previous
machines to hide SImode being used.  Add new insns
fctiw{,w}_<mode>_smallint if SImode can go in vector registers.
(fix_trunc<mode>si2_stfiwx): Likewise.
(fix_trunc<mode>si2_internal): Likewise.
(fixuns_trunc<mode>si2): Likewise.
(fixuns_trunc<mode>si2_stfiwx): Likewise.
(fctiw<u>z_<mode>_smallint): Likewise.
(fctiw<u>z_<mode>_mem): New combiner pattern to prevent conversion
of floating point to 32-bit integer from doing a direct move to
the GPR registers to do a store.
(fctiwz_<mode>): Break long line.

[gcc/testsuite]
2017-05-05  Michael Meissner  <meissner@linux.vnet.ibm.com>

PR target/79038
PR target/79202
PR target/79203
* gcc.target/powerpc/ppc-round3.c: New test.
* gcc.target/powerpc/ppc-round2.c: Update expected code.

From-SVN: r247657

cp-tree.h (IDENTIFIER_GLOBAL_VALUE): Use get_namespace_value.

* cp-tree.h (IDENTIFIER_GLOBAL_VALUE): Use get_namespace_value.
(SET_IDENTIFIER_GLOBAL_VALUE): Use set_global_value.
(IDENTIFIER_NAMESPACE_VALUE): Delete.
* name-lookup.h (namespace_binding, set_namespace_binding):
Replace
with ...
(get_namespace_value, set_global_value): ... these.
(get_global_value_if_present, is_typename_at_global_scope):
Delete.
* decl.c (poplevel): Use get_namespace_value.
(grokdeclarator): Use IDENTIFIER_GLOBAL_VALUE.
* class.c (build_vtbl_initializer): Stash library decl in
static var. Use IDENTIFIER_GLOBAL_VALUE.
* except.c (do_get_exception_ptr, do_begin_catch, do_end_catch)
do_allocate_exception, do_free_exception, build_throw): Likewise.
* init.c (throw_bad_array_new_length): Likewise.
* rtti.c (throw_bad_cast, throw_bad_typeid): Likewise.
* name-lookup.c (arg_assoc_namespace, pushdecl_maybe_friend_1)
check_for_our_of_scope_variable, push_overloaded_decl_1): Use
get_namespace_value.
(set_namespace_binding_1): Rename to
(set_namespace_binding): ... here.
(set_global_value): New.
(lookup_name_innermost_nonclass_level_1, push_namespace): Use
get_namespace_value.
* pt.c (listify): Use get_namespace_value.
((--This line, and those below, will be ignored--

M    cp/name-lookup.c
M    cp/name-lookup.h
M    cp/ChangeLog
M    cp/except.c
M    cp/class.c
M    cp/pt.c
M    cp/init.c
M    cp/cp-tree.h
M    cp/decl.c
M    cp/rtti.c

From-SVN: r247654

Makefile.in (GTFILES): Add tree-ssa-loop-ivopts.c.

* Makefile.in (GTFILES): Add tree-ssa-loop-ivopts.c.
* tree-ssa-loop-ivopts.c (compute_max_addr_offset): Delete.
(addr_list, addr_offset_valid_p): New.
(split_address_groups): Check offset validity with above function.
(gt-tree-ssa-loop-ivopts.h): Include header file.

From-SVN: r247653

builtins.c (check_sizes, [...]): Fix typos introduced during merge conflict resolution.

gcc/ChangeLog:
* builtins.c (check_sizes, check_strncat_sizes): Fix typos
introduced during merge conflict resolution.
(expand_builtin_strncat): Same.

gcc/testsuite/ChangeLog:
* gcc.dg/pr78138.c: Adjust text of expected diagnostics.
* gcc.dg/pr79214.c: Same.
* gcc.dg/pr79222.c: Same.
* gcc.dg/pr79223.c: Same.
* gcc.dg/tree-ssa/builtins-folding-gimple-ub.c: Expect warnings.

From-SVN: r247652

* config.gcc (arm*-*-*): Add missing 'fi'.

From-SVN: r247650

invoke.texi (-fopt-info): Explicitly say order of options included in -fopt-info does not matter.

2017-05-05  Steve Ellcey  <sellcey@cavium.com>

* doc/invoke.texi (-fopt-info): Explicitly say order of options
included in -fopt-info does not matter.
* doc/optinfo.texi (-fopt-info): Fix description of default
behavour. Explicitly say order of options included in -fopt-info
does not matter.

From-SVN: r247648

[ARM] Allow combination of aprofile and rmprofile multilibs

2017-05-05  Thomas Preud'homme  <thomas.preudhomme@arm.com>

    gcc/
    * config.gcc: Allow combinations of aprofile and rmprofile values for
    --with-multilib-list.
    * config/arm/t-multilib: New file.
    * config/arm/t-aprofile: Remove initialization of MULTILIB_*
    variables.  Remove setting of ISA and floating-point ABI in
    MULTILIB_OPTIONS and MULTILIB_DIRNAMES.  Set architecture and FPU in
    MULTI_ARCH_OPTS_A and MULTI_ARCH_DIRS_A rather than MULTILIB_OPTIONS
    and MULTILIB_DIRNAMES respectively.  Add comment to introduce all
    matches.  Add architecture matches for marvel-pj4 and generic-armv7-a
    CPU options.
    * config/arm/t-rmprofile: Likewise except for the matches changes.
    * doc/install.texi (--with-multilib-list): Document the combination of
    aprofile and rmprofile values and warn about pitfalls in doing that.

From-SVN: r247646

call.c (make_temporary_var_for_ref_to_temp): Push decl into current scope.

* call.c (make_temporary_var_for_ref_to_temp): Push decl into
current scope.
* lex.c (unqualified_name_lookup_error): Likewise.

From-SVN: r247645

class.c (alter_class): Use retrofit_lang_decl directly.

* class.c (alter_class): Use retrofit_lang_decl directly.
* decl.c (push_local_name, dupliate_decls): Likewise.
* semantics.c (omp_privatize_field): Likewise.

From-SVN: r247644

Float to int moves currently generate inefficient code due to hacks used in the movsi and movdi patterns.

Float to int moves currently generate inefficient code due to
hacks used in the movsi and movdi patterns.  The 'r = w' variant
uses '*' which tells the register allocator to ignore it.
As a result the float to int moves typically spill to the stack,
which is extremely inefficient.

    gcc/
* config/aarch64/aarch64.md (movsi_aarch64): Remove '*' from r=w.
(movdi_aarch64): Likewise.

From-SVN: r247643

re PR tree-optimization/80632 (error: invalid PHI argument with -O2)

PR tree-optimization/80632
* tree-switch-conversion.c (struct switch_conv_info): Add target_vop
field.
(build_arrays): Initialize it for virtual phis.
(fix_phi_nodes): Use it for virtual phis.

* gcc.dg/pr80632.c: New test.

From-SVN: r247642

re PR tree-optimization/80558 (VRP not handling x & -2 well)

PR tree-optimization/80558
* tree-vrp.c (extract_range_from_binary_expr_1): Optimize
[x, y] op z into [x op, y op z] for op & or | if conditions
are met.

* gcc.dg/tree-ssa/vrp115.c: New test.

From-SVN: r247641

[ARM] PR71607: Fix ICE when loading constant

2017-05-05  Andre Vieira  <andre.simoesdiasvieira@arm.com>
            Prakhar Bahuguna  <prakhar.bahuguna@arm.com>

    gcc/
    PR target/71607
    * config/arm/arm.md (use_literal_pool): Remove.
    (64-bit immediate split): No longer takes cost into consideration
    if arm_disable_literal_pool is enabled.
    * config/arm/arm.c (arm_tls_referenced_p): Add diagnostic if TLS is
    used when arm_disable_literal_pool is enabled.
    (arm_max_const_double_inline_cost): Remove use of
    arm_disable_literal_pool.
    (push_minipool_fix): Add assert.
    (arm_reorg): Add return if arm_disable_literal_pool is enabled.
    * config/arm/vfp.md (no_literal_pool_df_immediate): New.
    (no_literal_pool_sf_immediate): New.

2017-05-05  Andre Vieira  <andre.simoesdiasvieira@arm.com>
        Thomas Preud'homme  <thomas.preudhomme@arm.com>
        Prakhar Bahuguna  <prakhar.bahuguna@arm.com>

    gcc/testsuite/
    PR target/71607
    * gcc.target/arm/thumb2-slow-flash-data.c: Renamed to ...
    * gcc.target/arm/thumb2-slow-flash-data-1.c: ... this.
    * gcc.target/arm/thumb2-slow-flash-data-2.c: New.
    * gcc.target/arm/thumb2-slow-flash-data-3.c: New.
    * gcc.target/arm/thumb2-slow-flash-data-4.c: New.
    * gcc.target/arm/thumb2-slow-flash-data-5.c: New.
    * gcc.target/arm/tls-disable-literal-pool.c: New.

Co-Authored-By: Prakhar Bahuguna <prakhar.bahuguna@arm.com>
Co-Authored-By: Thomas Preud'homme <thomas.preudhomme@arm.com>
From-SVN: r247640

Kill walk_namespaces.

* cp-tree.h (walk_namespaces_fn, walk_namespaces): Delete.
* decl.c (walk_namespaces_r, walk_namespaces): Delete.

From-SVN: r247638

re PR tree-optimization/80613 (ICE in is_gimple_reg_type with -O2)

2017-05-05  Prathamesh Kulkarni  <prathamesh.kulkarni@linaro.org>

PR tree-optimization/80613
* tree-ssa-dce.c (propagate_necessity): Remove cases for
BUILT_IN_STRDUP and BUILT_IN_STRNDUP.

testsuite/
* gcc.dg/tree-ssa/pr79697.c (k): Remove.

From-SVN: r247635

tree-ssa-pre.c (get_or_alloc_expr_for): Simplify.

2017-05-05  Richard Biener  <rguenther@suse.de>

* tree-ssa-pre.c (get_or_alloc_expr_for): Simplify.

From-SVN: r247634

Kill per-namespace static_decls.

* cp-tree.h (static_decls): Declare.
(wrapup_globals_for_namespace)
diagnose_inline_vars_for_namespace): Replace with ...
(wrapup_namespace_globals): ... this.
* decl.c (static_decls): Define.
(wrapup_globals_for_namespace)
diagnose_inline_vars_for_namespace): Replace with ...
(wrapup_namespace_globals): ... this.
(cxx_init_decl_processing): Initialize static_decls.
* decl2.c (c_parse_final_cleanups): Adjust.
* name-lookup.h (cp_binding_level): Remove static_decls member.
* name-lookup.c (add_decl_to_level): Adjust.
(begin_scope): Adjust.
((--This line, and those below, will be ignored--

M    cp/cp-tree.h
M    cp/name-lookup.c
M    cp/name-lookup.h
M    cp/ChangeLog
M    cp/decl2.c
M    cp/decl.c

From-SVN: r247633

avr.md [...]: Remove any occurence of this flag from insn conditions due to removal from r247495.

* config/avr/avr.md [flag_strict_overflow]: Remove any occurence
of this flag from insn conditions due to removal from r247495.

From-SVN: r247632

Code scheduling for Cortex-A53 isn't as good as it could be.

Code scheduling for Cortex-A53 isn't as good as it could be.  It turns out
code runs faster overall if we place loads and stores with a dependency
closer together.  To achieve this effect, this patch adds a bypass between
cortex_a53_load1 and cortex_a53_load*/cortex_a53_store* if the result of an
earlier load is used in an address calculation.  This significantly improved
benchmark scores in a proprietary benchmark suite.

    gcc/
* config/arm/aarch-common.c (arm_early_load_addr_dep_ptr):
New function.
(arm_early_store_addr_dep_ptr): Likewise.
* config/arm/aarch-common-protos.h
(arm_early_load_addr_dep_ptr): Add prototype.
(arm_early_store_addr_dep_ptr): Likewise.
* config/arm/cortex-a53.md: Add new bypasses.

From-SVN: r247631

re PR c++/71577 (ICE on invalid C++11 code (with extra struct initializer): in digest_init_r, at cp/typeck2.c:1117)

/cp
2017-05-05  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/71577
* decl.c (reshape_init): Unconditionally return error_mark_node
upon error about too many initializers.

/testsuite
2017-05-05  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/71577
* g++.dg/cpp0x/pr71577.C: New.

From-SVN: r247630

tree.c (next_type_uid): Change type to unsigned.

* tree.c (next_type_uid): Change type to unsigned.
(type_hash_canon): Decrement back next_type_uid if
freeing a type node with the highest TYPE_UID.  For INTEGER_TYPEs
also ggc_free TYPE_MIN_VALUE, TYPE_MAX_VALUE and TYPE_CACHED_VALUES
if possible.

From-SVN: r247628

Daily bump.

From-SVN: r247627

builtins.c: Fix a trivial typo in a comment.

gcc/ChangeLog:
* builtins.c: Fix a trivial typo in a comment.

From-SVN: r247623

PR libstdc++/54924 - Warn for std::string constructor with wrong size

PR libstdc++/54924 - Warn for std::string constructor with wrong size
PR middle-end/79234 - warn on past the end reads by library functions

gcc/ChangeLog:

PR middle-end/79234
* builtins.c (check_sizes): Adjust to handle reading past the end.
Avoid printing excessive upper bound of ranges.  Use %E to print
tree nodes instead of converting them to %wu.
(expand_builtin_memchr): New function.
(compute_dest_size): Rename...
(compute_objsize): ...to this.
(expand_builtin_memcpy): Adjust.
(expand_builtin_mempcpy): Adjust.
(expand_builtin_strcat): Adjust.
(expand_builtin_strcpy): Adjust.
(check_strncat_sizes): Adjust.
(expand_builtin_strncat): Adjust.
(expand_builtin_strncpy): Adjust and simplify.
(expand_builtin_memset): Adjust.
(expand_builtin_bzero): Adjust.
(expand_builtin_memcmp): Adjust.
(expand_builtin): Handle memcmp.
(maybe_emit_chk_warning): Check strncat just once.

gcc/testsuite/ChangeLog:

PR middle-end/79234
* gcc.dg/builtin-stringop-chk-8.c: New test.
* gcc.dg/builtin-stringop-chk-1.c: Adjust.
* gcc.dg/builtin-stringop-chk-4.c: Same.
* gcc.dg/builtin-strncat-chk-1.c: Same.
* g++.dg/ext/strncpy-chk1.C: Same.
* g++.dg/torture/Wsizeof-pointer-memaccess1.C: Same.
* gcc.dg/out-of-bounds-1.c: Same.
* gcc.dg/pr78138.c: Same.
* gcc.dg/torture/Wsizeof-pointer-memaccess1.c: Same.
* gfortran.dg/mvbits_7.f90: Same.

From-SVN: r247622

Fix a typo.

From-SVN: r247619

PR preprocessor/79214 - -Wno-system-header defeats strncat buffer overflow warnings

PR preprocessor/79214 -  -Wno-system-header defeats strncat buffer overflow warnings
PR middle-end/79222 - missing -Wstringop-overflow= on a stpcpy overflow
PR middle-end/79223 - missing -Wstringop-overflow on a memmove overflow

gcc/ChangeLog:

PR preprocessor/79214
PR middle-end/79222
PR middle-end/79223
* builtins.c (check_sizes): Add inlinining context and issue
warnings even when -Wno-system-headers is set.
(check_strncat_sizes): Same.
(expand_builtin_strncat): Same.
(expand_builtin_memmove): New function.
(expand_builtin_stpncpy): Same.
(expand_builtin): Handle memmove and stpncpy.

gcc/testsuite/ChangeLog:

PR preprocessor/79214
PR middle-end/79222
PR middle-end/79223
* gcc.dg/pr79214.c: New test.
* gcc.dg/pr79214.h: New test header.
* gcc.dg/pr79222.c: New test.
* gcc.dg/pr79223.c: New test.
* gcc.dg/pr78138.c: Adjust.
* gfortran.dg/unconstrained_commons.f: Same.

From-SVN: r247618

* constraint.cc (diagnose_check_constraint): Fix %E thinko.

From-SVN: r247614

tree-ssa-loop-ivopts.c (struct cost_pair): Remove field inv_expr which is not used any more.

* tree-ssa-loop-ivopts.c (struct cost_pair): Remove field inv_expr
which is not used any more.

From-SVN: r247612

Many supported cores use the AUTOPREFETCHER_WEAK setting which tries to order...

Many supported cores use the AUTOPREFETCHER_WEAK setting which tries
to order loads and stores to improve streaming performance.  Since significant
gains were reported in http://patchwork.ozlabs.org/patch/534469/ it seems
like a good idea to enable this setting too for -mcpu=generic.  Since the
weak model only keeps the order if it doesn't make the schedule worse, it
should not impact performance adversely on cores that don't show a gain.

    gcc/
        * config/aarch64/aarch64.c (generic_tunings): Update prefetch model.

From-SVN: r247610

Set jump alignment to 4 for Cortex cores as it reduces codesize by 0.4% on...

Set jump alignment to 4 for Cortex cores as it reduces codesize by 0.4% on
average with no obvious performance difference.  See original discussion of
the overheads of various alignments:
https://gcc.gnu.org/ml/gcc-patches/2016-06/msg02075.html.

    gcc/
* config/aarch64/aarch64.c (cortexa35_tunings): Set jump alignment to 4.
(cortexa53_tunings): Likewise.
(cortexa57_tunings): Likewise.
(cortexa72_tunings): Likewise.
(cortexa73_tunings): Likewise.

From-SVN: r247609

With -mcpu=generic the loop alignment is currently 4.

With -mcpu=generic the loop alignment is currently 4.  All but one of the
supported cores use 8 or higher.  Since using 8 provides performance gains
on several cores, it is best to use that by default.  As discussed in [1],
the jump alignment has no effect on performance, yet has a relatively high
codesize cost [2], so setting it to 4 is best.  This gives a 0.2% overall
codesize improvement as well as performance gains in several benchmarks.

    gcc/
* config/aarch64/aarch64.c (generic_tunings): Set jump alignment to 4.
Set loop alignment to 8.

[1] https://gcc.gnu.org/ml/gcc-patches/2017-04/msg00574.html
[2] https://gcc.gnu.org/ml/gcc-patches/2016-06/msg02075.html

From-SVN: r247608

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/c-family/ChangeLog:

PR translation/80280
* c-attribs.c (handle_alias_ifunc_attribute): Quote a %D directive.
(handle_weakref_attribute): Same.

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/ChangeLog:

PR translation/80280
* config/i386/i386.c (ix86_function_versions): Quote a %D directive.

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/cp/ChangeLog:

PR translation/80280
* call.c (print_z_candidate): Add missing quoting to %D and other
like directives.
(build_op_call_1): Same.
* constraint.cc (diagnose_check_constraint): Same.
* mangle.c (mangle_decl): Same.
* name-lookup.c (cp_binding_level_debug): Same.
(set_decl_namespace): Same.
* parser.c (cp_parser_tx_qualifier_opt): Same.
* pt.c (print_candidates_1): Same.
(check_template_variable): Same.
(tsubst_default_argument): Same.
(most_specialized_partial_spec): Same.
* semantics.c (omp_reduction_lookup): Same.
* tree.c (check_abi_tag_redeclaration): Same.
* typeck.c (comptypes): Same.
* typeck2.c (abstract_virtuals_error_sfinae): Same.

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/ChangeLog:

PR translation/80280
* builtins.c (expand_builtin_object_size): Add missing quoting to
%D and like directives.
* hsa-gen.c (hsa_type_for_scalar_tree_type): Same.
(hsa_type_for_tree_type): Same.
(verify_function_arguments): Same.
* symtab.c (symbol_table::change_decl_assembler_name): Same.
* varasm.c (get_section): Same.
(mark_weak): Same.

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/objc/ChangeLog:

PR translation/80280
* objc-gnu-runtime-abi-01.c (objc_eh_runtime_type): Add missing
quoting to directives.
* objc-next-runtime-abi-01.c (objc_eh_runtime_type): Ditto.
* objc-next-runtime-abi-02.c (next_runtime_02_eh_type): Ditto.

PR translation/80280 - Missing closing quote (%>) c/semantics.c and c/c-typeck.c

gcc/testsuite/ChangeLog:

PR translation/80280
* g++.dg/abi/Wabi-2-3.C: Adjust.
* g++.dg/abi/Wabi-3-2.C: Ditto.
* g++.dg/lookup/using17.C: Ditto.
* gcc.dg/format/gcc_diag-1.c (foo): Ditto.

From-SVN: r247607

All cores which add a cpu_addrcost_table use a non-zero value for HI and TI...

All cores which add a cpu_addrcost_table use a non-zero value for
HI and TI mode shifts (a non-zero value for general indexing also
applies to all shifts).  Given this, it makes no sense to use a
different setting in generic_addrcost_table.  So change it so that
all supported cores, including -mcpu=generic, now generate the same:

int f(short *p, short *q, long x) { return p[x] + q[x]; }

        lsl     x2, x2, 1
        ldrsh   w3, [x0, x2]
        ldrsh   w0, [x1, x2]
        add     w0, w3, w0
        ret

    gcc/
* config/aarch64/aarch64.c (generic_addrcost_table):
Change HI/TI mode setting.

From-SVN: r247606

[PR 80622] Treat const pools as initialized in SRA

2017-05-04  Martin Jambor  <mjambor@suse.cz>

PR tree-optimization/80622
* tree-sra.c (comes_initialized_p): New function.
(build_accesses_from_assign): Only set write lazily when
comes_initialized_p is false.
(analyze_access_subtree): Use comes_initialized_p.
(propagate_subaccesses_across_link): Assert !comes_initialized_p
instead of testing for PARM_DECL.

testsuite/
* gcc.dg/tree-ssa/pr80622.c: New test.

From-SVN: r247604

[AArch64] Accept more addressing modes for PRFM

* config/aarch64/aarch64.md (prefetch); Adjust predicate and
constraint on operand 0 to allow more general addressing modes.
Adjust output template.
* config/aarch64/aarch64.c (aarch64_address_valid_for_prefetch_p):
New function.
* config/aarch64/aarch64-protos.h
(aarch64_address_valid_for_prefetch_p): Declare prototype.
* config/aarch64/constraints.md (Dp): New address constraint.
* config/aarch64/predicates.md (aarch64_prefetch_operand): New
predicate.

* gcc.target/aarch64/prfm_imm_offset_1.c: New test.

From-SVN: r247603

Update .po files.

* be.po, da.po, el.po, fi.po, hr.po, id.po, ja.po, nl.po, ru.po,
sr.po, sv.po, tr.po, uk.po, vi.po, zh_CN.po, zh_TW.po: Update.

From-SVN: r247601

More global trees.

* cp-tree.h (enum cp_tree_index): Add CPTI_GLOBAL,
CPTI_GLOBAL_TYPE, CPTI_GLOBAL_IDENTIFIER, CPTI_ANON_IDENTIFIER,
CPTI_INIT_LIST_IDENTIFIER.
(global_namespace, global_type_node, global_identifier)
anon_identifier, init_list_identifier): New.
* decl.c (global_type_node, global_scope_name): Delete.
(initialize_predefined_identifiers): Add new identifiers.
(cxx_init_decl_processing): Adjust.
* name-lookup.h (global_namespace, global_type_node): Delete.
* name-lookup.c (global_namespace, anonymous_namespace_name)
get_anonymous_namespace_name): Delete.
(namespace_scope_ht_size, begin_scope, pushtag_1)
push_namespace): Adjust,
* call.c (type_has_extended_temps): Use init_list_identifier.
* pt.c (listify): Likewise.
(((--This line, and those below, will be ignored--

M    cp/name-lookup.c
M    cp/name-lookup.h
M    cp/ChangeLog
M    cp/call.c
M    cp/pt.c
M    cp/cp-tree.h
M    cp/decl.c

From-SVN: r247599

ipa-cp.c (perform_estimation_of_a_value): Drop base_time parameter; update use of estimate_ipcp_clone_size_and_time.

* ipa-cp.c (perform_estimation_of_a_value): Drop base_time parameter;
update use of estimate_ipcp_clone_size_and_time.
(estimate_local_effects): Update use of
estimate_ipcp_clone_size_and_time and perform_estimation_of_a_value.
* ipa-inline.h (estimate_ipcp_clone_size_and_time): Update prototype.
* ipa-inline-analysis.c (estimate_ipcp_clone_size_and_time):
Return nonspecialized time.

* gcc.dg/ipa/ipcp-2.c: Decrease ipcp threshold
* gcc.dg/tree-ssa/ipa-cp-1.c: Likewise.

From-SVN: r247597

tree-ssa-alias.c (get_continuation_for_phi): Improve looking for the last VUSE which def dominates the PHI.

2017-05-04  Richard Biener  <rguenther@suse.de>

* tree-ssa-alias.c (get_continuation_for_phi): Improve looking
for the last VUSE which def dominates the PHI.  Directly call
maybe_skip_until.
(get_continuation_for_phi_1): Remove.

* gcc.dg/tree-ssa/ssa-fre-58.c: New testcase.

From-SVN: r247596

name-lookup.c: Reorder functions to make merging from modules branch simpler.

* name-lookup.c: Reorder functions to make merging from modules
branch simpler.

From-SVN: r247592

Cap niter_for_unrolled_loop to upper bound

For the reasons explained in PR77536, niter_for_unrolled_loop assumes 5
iterations in the absence of profiling information, although it doesn't
increase beyond the estimate for the original loop.  This left a hole in
which the new estimate could be less than the old one but still greater
than the limit imposed by CEIL (nb_iterations_upper_bound, unroll factor).

2017-05-04  Richard Sandiford  <richard.sandiford@linaro.org>

gcc/
* tree-ssa-loop-manip.c (niter_for_unrolled_loop): Add commentary
to explain the use of truncating division.  Cap the number of
iterations to the maximum given by nb_iterations_upper_bound,
if defined.

gcc/testsuite/
* gcc.dg/vect/vect-profile-1.c: New test.

From-SVN: r247591

Fix previous commit

From-SVN: r247589

configure.ac (--enable-mingw-wildcard): Add new configurable feature.

2017-05-04  Thomas Preud'homme  <thomas.preudhomme@arm.com>

* configure.ac (--enable-mingw-wildcard): Add new configurable feature.
* configure: Regenerate.
* config.in: Regenerate.
* config/i386/driver-mingw32.c: new file.
* config/i386/x-mingw32: Add rule to build driver-mingw32.o.
* config.host: Link driver-mingw32.o on MinGW host.
* doc/install.texi: Document new --enable-mingw-wildcard configure
option.

From-SVN: r247588

Remove bogus top-level ChangeLog commit (sorry!)

From-SVN: r247587

re PR tree-optimization/80612 (ICE in get_range_info, at tree-ssanames.c:375)

PR tree-optimization/80612
* calls.c (get_size_range): Check for INTEGRAL_TYPE_P.

* gcc.dg/torture/pr80612.c: New test.

From-SVN: r247586

[ARM] Enable Purecode for ARMv8-M Baseline

This patch adds support for purecode to ARMv8-M Baseline, in addition to
the existing support for ARMv7-M and ARMv8-M Mainline.

2017-05-04  Prakhar Bahuguna  <prakhar.bahuguna@arm.com>
            Andre Simoes Dias Vieira  <andre.simoesdiasvieira@arm.com>

    gcc/
    * config/arm/arm.md (movsi): Change TARGET_32BIT to TARGET_HAVE_MOVT.
    (movt splitter): Likewise.
    * config/arm/arm.c (arm_option_check_internal): Change arm_arch_thumb2
    to TARGET_HAVE_MOVT, and merge with -mslow-flash-data check.
    (const_ok_for_arm): Change else to else if (TARGET_THUMB2) and add else
    block for Thumb-1 with MOVT.
    (thumb2_legitimate_address_p): Move code block ...
    (can_avoid_literal_pool_for_label_p): ... into this new function.
    (thumb1_legitimate_address_p): Add check for TARGET_HAVE_MOVT and
    literal pool.
    (thumb_legitimate_constant_p): Add conditional on TARGET_HAVE_MOVT
    * doc/invoke.texi (-mpure-code): Change "ARMv7-M targets" for
    "M-profile targets with the MOVT instruction".

    gcc/testsuite/
    * gcc.target/arm/pure-code/pure-code.exp: Add conditional for
    check_effective_target_arm_thumb1_movt_ok.

Co-Authored-By: Andre Vieira <andre.simoesdiasvieira@arm.com>
From-SVN: r247585

[ARM] Rename FPSCR builtins to correct names

The GCC documentation in section 6.60.8 ARM Floating Point Status and
Control Intrinsics states that the FPSCR register can be read and
written to using the intrinsics __builtin_arm_get_fpscr and
__builtin_arm_set_fpscr. However, these are misnamed within GCC itself
and these intrinsic names are not recognised.

This patch corrects the intrinsic names to match the documentation, and
adds tests to verify these intrinsics generate the correct
instructions.

2017-05-04  Prakhar Bahuguna  <prakhar.bahuguna@arm.com>

    gcc/
    * gcc/config/arm/arm-builtins.c (arm_init_builtins): Rename
    __builtin_arm_ldfscr to __builtin_arm_get_fpscr, and rename
    __builtin_arm_stfscr to __builtin_arm_set_fpscr.

    gcc/testsuite/
    * gcc.target/arm/fpscr.c: New file.

From-SVN: r247584

Remove an unused variable.

2017-05-04  Martin Liska  <mliska@suse.cz>

* tree-vrp.c (simplify_cond_using_ranges_2): Remove unused
variable cond_code.

From-SVN: r247583

Require c99_runtime for pr78622.c

2017-05-04  Thomas Preud'homme  <thomas.preudhomme@arm.com>

gcc/testsuite/
    * gcc.c-torture/execute/pr78622.c: Require c99_runtime effective
    target.

From-SVN: r247582

tree.c (array_at_struct_end_p): Handle arrays at struct end with flexarrays more conservatively.

2017-05-04  Richard Biener  <rguenther@suse.de>

* tree.c (array_at_struct_end_p): Handle arrays at struct
end with flexarrays more conservatively.  Refactor and treat
arrays of arrays or aggregates more strict.  Fix
VIEW_CONVERT_EXPR handling.  Remove allow_compref argument.
* tree.c (array_at_struct_end_p): Adjust prototype.
* emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
* gimple-fold.c (get_range_strlen): Likewise.
* tree-chkp.c (chkp_may_narrow_to_field): Likewise.

From-SVN: r247581

Replace absolute line numbers in g++.dg/warn

2017-05-04  Tom de Vries  <tom@codesourcery.com>

PR testsuite/80557
* g++.dg/warn/Wstrict-aliasing-float-ref-int-obj.C: Replace absolute
line numbers.
* g++.dg/warn/miss-format-1.C: Same.
* g++.dg/warn/overflow-warn-1.C: Same.
* g++.dg/warn/overflow-warn-3.C: Same.
* g++.dg/warn/overflow-warn-4.C: Same.

From-SVN: r247580

Replace absolute line numbers in gcc.target/i386

2017-05-04  Tom de Vries  <tom@codesourcery.com>

PR testsuite/80557
* gcc.target/i386/pr57655.c: Replace absolute line numbers.
* gcc.target/i386/pr68657.c: Same.
* gcc.target/i386/pr69255-1.c: Same.
* gcc.target/i386/pr69255-2.c: Same.
* gcc.target/i386/pr69255-3.c: Same.

From-SVN: r247579

re PR tree-optimization/31130 (VRP no longer derives range for division after negation)

2017-05-04  Richard Biener  <rguenther@suse.de>

PR tree-optimization/31130
* tree-vrp.c (needs_overflow_infinity): Remove as always returning
false.
(supports_overflow_infinity): Likewise.
(is_negative_overflow_infinity): Likewise.
(is_positive_overflow_infinity): Likewise.
(is_overflow_infinity): Likewise.
(stmt_overflow_infinity): Likewise.
(overflow_infinity_range_p): Likewise.
(usable_range_p): Remove as always returning true.
(make_overflow_infinity): Remove.
(negative_overflow_infinity): Likewise.
(positive_overflow_infinity): Likewise.
(avoid_overflow_infinity): Likewise.
(set_value_range): Adjust accordingly.
(set_value_range_to_nonnegative): Likewise, remove now unused
overflow_infinity arg.
(vrp_operand_equal_p): Adjust.
(update_value_range): Likewise.
(range_int_cst_singleton_p): Likewise.
(operand_less_p): Likewise.
(compare_values_warnv): Likewise.
(extract_range_for_var_from_comparison_expr): Likewise.
(vrp_int_const_binop): Likewise.
(zero_nonzero_bits_from_vr): Likewise.
(extract_range_from_multiplicative_op_1): Likewise.
(extract_range_from_binary_expr_1): Likewise.
(extract_range_from_unary_expr): Likewise.
(extract_range_from_comparison): Likewise.
(extract_range_basic): Likewise.
(adjust_range_with_scev): Likewise.
(compare_ranges): Likewise.
(compare_range_with_value): Likewise.
(dump_value_range): Likewise.
(test_for_singularity): Likewise, remove strict_overflow_p parameter
never used.
(simplify_cond_using_ranges): Adjust.

* gcc.dg/Wstrict-overflow-12.c: XFAIL.
* gcc.dg/Wstrict-overflow-13.c: Likewise.
* gcc.dg/Wstrict-overflow-21.c: Likewise.
* gcc.dg/pr52904.c: Remove XFAIL.
* gcc.dg/tree-ssa/vrp114.c: New testcase.

From-SVN: r247578

Removed accidentally committed conflict marker.

From-SVN: r247577

Minor BRIG/HSAIL frontend updates and bug fixes:

* brig-builtins.def: Added a builtin for class_f64.
* builtin-types.def: Added a builtin type needed by class_f64.
* brigfrontend/brig-code-entry-handler.cc
 (brig_code_entry_handler::build_address_operand): Fix a bug
 with reg+offset addressing on 32b segments. In large mode,
 the offset is treated as 32bits unless it's global, readonly or
 kernarg address space.
* rt/workitems.c: Removed a leftover comment.
* rt/arithmetic.c (__hsail_class_f32, __hsail_class_f64): Fix the
 check for signaling/non-signalling NaN. Add class_f64 default
 implementation.

From-SVN: r247576

Daily bump.

From-SVN: r247575

* de.po: Update.

From-SVN: r247570

invoke.texi: Note that -faligned-new is on by default for C++17.

* doc/invoke.texi: Note that -faligned-new is on by default
for C++17.

From-SVN: r247564

timevar.def: Add TV_CONSTEXPR.

* timevar.def: Add TV_CONSTEXPR.

* constexpr.c (cxx_eval_outermost_constant_expr): Use TV_CONSTEXPR.

From-SVN: r247563

Fix typo in common.opt

gcc/ChangeLog:
* common.opt (fdiagnostics-parseable-fixits): Fix typo.

From-SVN: r247562

Remove ipa_update_after_lto_read

2017-05-03  Martin Jambor  <mjambor@suse.cz>

* ipa-prop.c (ipa_update_after_lto_read): Removed.
* ipa-prop.h (ipa_update_after_lto_read): Remove declaration.
* ipa-cp.c (ipcp_propagate_stage): Do not call
ipa_update_after_lto_read.
* ipa-inline.c (ipa_inline): Likewise.

From-SVN: r247559

Use call_summary in ipa-prop and ipa-cp

2017-05-03  Martin Jambor  <mjambor@suse.cz>

* ipa-prop.h (ipa_edge_args): Make a class.  Mark with for_user GTY
tag.  Added a default constructor and a destructor.
(ipa_edge_args_sum_t): New class;
(ipa_edge_args_sum): Declare.
(ipa_edge_args_vector): Remove declaration.
(IPA_EDGE_REF): Use ipa_edge_args_sum.
(ipa_free_edge_args_substructures): Remove declaration.
(ipa_check_create_edge_args): Use ipa_edge_args_sum.
(ipa_edge_args_info_available_for_edge_p): Likewise.
* ipa-prop.c (ipa_edge_args_vector): Removed.
(edge_removal_hook_holder): Likewise.
(edge_duplication_hook_holder): Likewise.
(ipa_edge_args_sum): New variable.
(ipa_propagate_indirect_call_infos): Test ipa_edge_args_sum instead of
ipa_edge_args_vector.
(ipa_free_edge_args_substructures): Likewise.
(ipa_free_all_edge_args): Free ipa_edge_args_sum instead of
ipa_edge_args_vector.
(ipa_edge_removal_hook): Turned into method
ipa_edge_args_sum_t::remove.
(ipa_edge_duplication_hook): Turned into method
ipa_edge_args_sum_t::duplicate.
(ipa_register_cgraph_hooks): Create ipa_edge_args_sum instead of
registering edge hooks.
(ipa_unregister_cgraph_hooks): Do not unregister edge hooks.
* ipa-inline-analysis.c (estimate_function_body_sizes): Test
ipa_edge_args_sum instead of ipa_edge_args_vector.
* ipa-profile.c (ipa_profile): Likewise.

From-SVN: r247558

call_summary to keep info about cgraph_edges

2017-05-03  Martin Jambor  <mjambor@suse.cz>

* symbol-summary.h (function_summary): New method exists.
(function_summary::symtab_removal): Deallocate through release.
(call_summary): New class.
(gt_ggc_mx): New overload.
(gt_pch_nx): Likewise.
(gt_pch_nx): Likewise.

From-SVN: r247557

re PR tree-optimization/78496 (Missed opportunities for jump threading)

PR tree-optimization/78496
* tree-vrp.c (simplify_cond_using_ranges_1): Renamed
from simplify_cond_using_ranges.  Split off code to walk
backwards through casts into ...
(simplify_cond_using_ranges_2): New function.
(simplify_stmt_using_ranges): Call simplify_cond_using_ranges_1.
(execute_vrp): After identifying jump threads, call
simplify_cond_using_ranges_2.

PR tree-optimization/78496
* gcc.dg/tree-ssa/ssa-thread-15.c: New test.

From-SVN: r247556

re PR ipa/80609 (crash_signal in reset_inline_summary on ia64 bootstrap)

PR bootstrap/80609
* ipa-inline.h (inline_summary): Add ctor.
(create_ggc): Do not use ggc_cleared_alloc.

From-SVN: r247555

config-list.mk (am33_2.0-linux): Remove from list of targets to build.

2007-05-03  Jeff Law  <law@redhat.com>

* config-list.mk (am33_2.0-linux): Remove from list of targets
to build.

From-SVN: r247553

Support escaping special characters in specs

2017-05-03  Jeff Downs  <heydowns@somuchpressure.net>
    Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>

* gcc.c (handle_braces): Support escaping in switch matching
text.
* doc/invoke.texi (Spec Files): Document it.
Remove superfluous @code markup in items.

Co-Authored-By: Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
From-SVN: r247552

pr79671_0.C (foo): Fix asm constraints.

* g++.dg/lto/pr79671_0.C (foo): Fix asm constraints.

From-SVN: r247550

New fix-it printer

The existing fix-it printer can lead to difficult-to-read output
when fix-it hints are near each other.  For example, in a recent
patch to add fix-it hints to the C++ frontend's -Wold-style-cast,
e.g. for:

  foo *f = (foo *)ptr->field;
                       ^~~~~

the fix-it hints:
 replace the open paren with "const_cast<"
 replace the close paren with "> ("
 insert ")" after the "ptr->field"

would be printed in this odd-looking way:

  foo *f = (foo *)ptr->field;
                       ^~~~~
           -
           const_cast<
                 -
                 > (        )

class rich_location consolidates adjacent fix-it hints, which helps
somewhat, but the underlying problem is that the existing printer
simply walks through the list of hints printing them, starting newlines
as necessary.

This patch reimplements fix-it printing by introducing a planning
stage: a new class line_corrections "plans" how to print the
fix-it hints affecting a line, generating a vec of "correction"
instances.  Hints that are sufficiently close to each other are
consolidated at this stage.

This leads to the much more reasonable output for the above case:

  foo *f = (foo *)ptr->field;
                       ^~~~~
           -----------------
           const_cast<foo *> (ptr->field);

where the 3 hints are consolidated into one "correction" at printing.

gcc/ChangeLog:
* diagnostic-show-locus.c (struct column_range): New struct.
(get_affected_columns): New function.
(get_printed_columns): New function.
(struct correction): New struct.
(correction::ensure_capacity): New function.
(correction::ensure_terminated): New function.
(struct line_corrections): New struct.
(line_corrections::~line_corrections): New dtor.
(line_corrections::add_hint): New function.
(layout::print_trailing_fixits): Reimplement in terms of the new
classes.
(selftest::test_overlapped_fixit_printing): New function.
(selftest::diagnostic_show_locus_c_tests): Call it.

From-SVN: r247548

cp-tree.h (enum cp_tree_index, [...]): Move earlier, along with #defines, to before name-lookup include.

* cp-tree.h (enum cp_tree_index, cp_global_trees): Move earlier,
along with #defines, to before name-lookup include.

From-SVN: r247547

Canonicalize canonical type hashing

Canonicalize canonical type hashing
gcc/
* tree.h (type_hash_canon_hash): Declare.
* tree.c (type_hash_list, attribute_hash_list): Move into
type_hash_canon_hash.
(build_type_attribute_qual_variant): Break out hash code calc into
type_hash_canon_hash.
(type_hash_canon_hash): New.  Generic type hash computation.
(build_range_type_1, build_array_type_1, build_function_type)
build_method_type_directly, build_offset_type, build_complex_type,
make_vector_type): Call it.
gcc/c-family/
* c-common.c (complete_array_type): Use type_hash_canon.
(--This line, and those below, will be ignored--

M    gcc/tree.c
M    gcc/tree.h
M    gcc/c-family/ChangeLog
M    gcc/c-family/c-common.c
M    gcc/ChangeLog

From-SVN: r247546

tree-vect-data-refs.c (vect_enhance_data_refs_alignment): When all DRs have unknown misaligned do not always peel when...

2017-05-03  Richard Biener  <rguenther@suse.de>

* tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
When all DRs have unknown misaligned do not always peel
when there is a store but apply the same costing model as if
there were only loads.

* gcc.dg/vect/costmodel/x86_64/costmodel-alignpeel.c: New testcase.

From-SVN: r247544

revert: re PR tree-optimization/80492 (Wrong code when unrolling a loop with inline asm and local regs)

2017-05-03  Richard Biener  <rguenther@suse.de>

Revert
PR tree-optimization/80492
* tree-ssa-alias.c (decl_refs_may_alias_p): Handle
compare_base_decls returning dont-know properly.

From-SVN: r247543

[ARM] Set mode for success result of atomic compare and swap

2017-05-03  Thomas Preud'homme  <thomas.preudhomme@arm.com>

gcc/
    * config/arm/iterators.md (CCSI): New mode iterator.
    (arch): New mode attribute.
    * config/arm/sync.md (atomic_compare_and_swap<mode>_1): Rename into ...
    (atomic_compare_and_swap<CCSI:arch><NARROW:mode>_1): This and ...
    (atomic_compare_and_swap<CCSI:arch><SIDI:mode>_1): This.  Use CCSI
    code iterator for success result mode.
    * config/arm/arm.c (arm_expand_compare_and_swap): Adapt code to use
    the corresponding new insn generators.

From-SVN: r247542

Revert r247509 2017-05-02 Bin Cheng <bin.cheng@arm.com>

Revert r247509
2017-05-02  Bin Cheng  <bin.cheng@arm.com>
* rtlanal.c (rtx_cost): Handle TRUNCATE between tieable modes.

From-SVN: r247541

Wrap tree-data-ref.h macro arguments

gcc/
2016-05-03  Richard Sandiford  <richard.sandiford@linaro.org>

* tree-data-ref.h (SUB_CONFLICTS_IN_A): Wrap SUB argument in brackets.
(SUB_CONFLICTS_IN_B, SUB_LAST_CONFLICT, SUB_DISTANCE): Likewise.
(DDR_A): Wrap DDR argument in brackets.
(DDR_B, DDR_AFFINE_P, DDR_ARE_DEPENDENT, DDR_SUBSCRIPTS): Likewise.
(DDR_LOOP_NEST, DDR_INNER_LOOP, DDR_SELF_REFERENCE): Likewise.
(DDR_REVERSED_P): Likewise.

From-SVN: r247539

re PR tree-optimization/79472 (x86-64: Switch table generation fails if default case has different code)

PR tree-optimization/79472
* tree-switch-conversion.c (struct switch_conv_info): Add
contiguous_range and default_case_nonstandard fields.
(collect_switch_conv_info): Compute contiguous_range and
default_case_nonstandard fields, don't clear final_bb if
contiguous_range and only the default case doesn't have the required
structure.
(check_all_empty_except_final): Set default_case_nonstandard instead
of failing if contiguous_range and the default case doesn't have empty
block.
(check_final_bb): Add SWTCH argument, don't fail if contiguous_range
and only the default case doesn't have the required constants.  Skip
virtual phis.
(gather_default_values): Skip virtual phis.  Allow non-NULL CASE_LOW
if default_case_nonstandard.
(build_constructors): Build constant 1 just once.  Assert that default
values aren't inserted in between cases if contiguous_range.  Skip
virtual phis.
(build_arrays): Skip virtual phis.
(prune_bbs): Add DEFAULT_BB argument, don't remove that bb.
(fix_phi_nodes): Don't add e2f phi arg if default_case_nonstandard.
Handle virtual phis.
(gen_inbound_check): Handle default_case_nonstandard case.
(process_switch): Adjust check_final_bb caller.  Call
gather_default_values with the first non-default case instead of
default case if default_case_nonstandard.

* gcc.dg/tree-ssa/vrp40.c: Add -fno-tree-switch-conversion to dg-options.
* gcc.dg/tree-ssa/vrp113.c: New test.
* gcc.dg/tree-ssa/cswtch-3.c: New test.
* gcc.dg/tree-ssa/cswtch-4.c: New test.
* gcc.dg/tree-ssa/cswtch-5.c: New test.

From-SVN: r247538

Replace absolute line numbers in c-c++-common

2017-05-03  Tom de Vries  <tom@codesourcery.com>

PR testsuite/80557
* c-c++-common/Wshift-negative-value-1.c: Replace absolute line numbers.
* c-c++-common/Wshift-negative-value-2.c: Same.
* c-c++-common/Wshift-negative-value-3.c: Same.
* c-c++-common/Wshift-negative-value-4.c: Same.
* c-c++-common/cilk-plus/AN/pr57541.c: Same.
* c-c++-common/cpp/pr60400.c: Same.
* c-c++-common/fmax-errors.c: Same.
* c-c++-common/goacc/data-2.c: Same.
* c-c++-common/goacc/host_data-2.c: Same.
* c-c++-common/gomp/simd4.c: Same.
* c-c++-common/pr28656.c: Same.
* c-c++-common/pr43395.c: Same.
* c-c++-common/torture/pr57945.c: Same.

From-SVN: r247537

Add quotes to numerical comment arg of dg directive

2017-05-03  Tom de Vries  <tom@codesourcery.com>

* c-c++-common/goacc/data-default-1.c: Add quotes to numerical comment
arg of dg directive.
* c-c++-common/goacc/routine-3.c: Same.
* c-c++-common/goacc/routine-4.c: Same.

From-SVN: r247536

Daily bump.

From-SVN: r247533

ggc-page.c (move_ptes_to_front): Replace unsigned >0 with i-- check.

* ggc-page.c (move_ptes_to_front): Replace unsigned >0 with i--
check.  Fix formatting.

From-SVN: r247529

ipa-inline-analysis.c (estimate_node_size_and_time): Allow roundoff errors when comparing specialized and unspecialized times.

* ipa-inline-analysis.c (estimate_node_size_and_time): Allow roundoff
errors when comparing specialized and unspecialized times.

From-SVN: r247528

pt.c (is_auto_or_concept): Remove.

2017-05-02  Paolo Carlini  <paolo.carlini@oracle.com>

* pt.c (is_auto_or_concept): Remove.
(type_uses_auto_or_concept): Remove, unused.
(find_parameter_packs_r, extract_autos_r, is_auto_r): Adjust.
* parser.c (tree_type_is_auto_or_concept): Remove, unused.
* cp-tree.h (is_auto_or_concept): Remove.

From-SVN: r247526

* es.po: Update.

From-SVN: r247524

Support fix-it hints that add new lines

Previously fix-it hints couldn't contain newlines.  This is
due to the need to print something user-readable for them
within diagnostic-show-locus, and for handling them within
edit-context for printing diffs and regenerating content.

This patch enables limited support for fix-it hints with newlines,
for suggesting adding new lines.
Such a fix-it hint must have exactly one newline character, at the
end of the content.  It must be an insertion at the beginning of
a line (so that e.g. fix-its that split a pre-existing line are
still rejected).

They are printed by diagnostic-show-locus with a '+' in the
left-hand margin, like this:

test.c:42:4: note: suggest adding 'break;' here
+      break;
     case 'b':
     ^~~~~~~~~

and the printer injects "spans" if the insertion location is not
near the primary range of the diagnostic e.g.:

test.c:4:2: note: unrecognized 'putchar'; suggest including '<stdio.h>'
test.c:1:1:
+#include <stdio.h>

test.c:4:2:
  putchar (ch);
  ^~~~~~~

gcc/ChangeLog:
* diagnostic-show-locus.c
(layout::should_print_annotation_line_p): Make private.
(layout::print_annotation_line): Make private.
(layout::annotation_line_showed_range_p): Make private.
(layout::show_ruler): Make private.
(layout::print_source_line): Make private.  Pass in line and
line_width, rather than calling location_get_source_line.  Drop
returned value.
(layout::print_leading_fixits): New method.
(layout::print_any_fixits): Rename to...
(layout::print_trailing_fixits): ...this, and make private.
Don't print newline fixits.
(diagnostic_show_locus): Move logic for printing one row into...
(layout::print_line): ...this new function.  Move the
location_get_source_line call and error-handling from
print_source_line to here.  Call print_leading_fixits, and rename
print_any_fixits to print_trailing_fixits.
(selftest::test_fixit_insert_containing_newline): Update now that
newlines are partially supported.
(selftest::test_fixit_insert_containing_newline_2): New test.
(selftest::test_fixit_replace_containing_newline): Update comments.
(selftest::diagnostic_show_locus_c_tests): Call the new test.
* edit-context.c (class added_line): New class.
(class edited_line): Describe newline handling in comment.
(edited_line::actually_edited_p): New method.
(edited_line::print_content): Delete redundant decl.
(edited_line::m_predecessors): New field.
(edited_file::print_content): Call edited_line::print_content.
(edited_file::print_diff): Update to support newlines.
(edited_file::print_diff_hunk): Likewise.
(edited_file::print_run_of_changed_lines): New function.
(edited_file::print_diff_line): Convert to...
(print_diff_line): ...this.
(edited_file::get_effective_line_count): New function.
(edited_line::edited_line): Initialize new field m_predecessors.
(edited_line::~edited_line): Clean up m_predecessors.
(edited_line::apply_fixit): Handle newlines.
(edited_line::get_effective_line_count): New function.
(edited_line::print_content): New function.
(edited_line::print_diff_lines): New function.
(selftest::test_applying_fixits_insert_containing_newline): New
test.
(selftest::test_applying_fixits_replace_containing_newline): New
test.
(selftest::insert_line): New function.
(selftest::test_applying_fixits_multiple_lines): Add example of
inserting a line.
(selftest::edit_context_c_tests): Call the new tests.

gcc/testsuite/ChangeLog:
* gcc.dg/plugin/diagnostic-test-show-locus-bw.c
(test_fixit_insert_newline): New function.
* gcc.dg/plugin/diagnostic-test-show-locus-color.c
(test_fixit_insert_newline): New function.
* gcc.dg/plugin/diagnostic-test-show-locus-generate-patch.c
(test_fixit_insert_newline): New function.
* gcc.dg/plugin/diagnostic-test-show-locus-parseable-fixits.c
(test_fixit_insert_newline): New function.
* gcc.dg/plugin/diagnostic_plugin_test_show_locus.c
(test_show_locus): Handle test_fixit_insert_newline.

libcpp/ChangeLog:
* include/line-map.h (class rich_location): Update description of
newline handling.
(class fixit_hint): Likewise.
(fixit_hint::ends_with_newline_p): New decl.
* line-map.c (rich_location::maybe_add_fixit): Support newlines
in fix-it hints that are insertions of single lines at the start
of a line.  Don't consolidate into such fix-it hints.
(fixit_hint::ends_with_newline_p): New method.

From-SVN: r247522

* fr.po: Update.

From-SVN: r247520

tree-ssa-loop-ivopts.c (get_scaled_computation_cost_at): Delete parameter cand.

* tree-ssa-loop-ivopts.c (get_scaled_computation_cost_at): Delete
parameter cand.  Update dump information.
(get_computation_cost): Update uses.

From-SVN: r247519

tree-ssa-loop-ivopts.c (get_computation_aff_1): New.

* tree-ssa-loop-ivopts.c (get_computation_aff_1): New.
(get_computation_aff): Reorder parameters.  Use get_computation_aff_1.
(get_computation_at, rewrite_use_address): Update use of
get_computation_aff.

From-SVN: r247518

tree-ssa-loop-ivopts.c (get_computation_at): Reorder parameters.

* tree-ssa-loop-ivopts.c (get_computation_at): Reorder parameters.
(get_computation): Delete.
(get_computation_cost): Implement like get_computation_cost_at.
Use get_computation_at.
(get_computation_cost_at): Delete.
(rewrite_use_nonlinear_expr): Use get_computation_at.
(rewrite_use_compare, remove_unused_ivs): Ditto.

From-SVN: r247517