pa.h (BIGGEST_ALIGNMENT): Adjust comment.

* config/pa/pa.h (BIGGEST_ALIGNMENT): Adjust comment.
(MALLOC_ABI_ALIGNMENT): Define.

From-SVN: r241678

Daily bump.

From-SVN: r241670

Daily bump.

From-SVN: r241638

Daily bump.

From-SVN: r241607

Daily bump.

From-SVN: r241543

Daily bump.

From-SVN: r241500

backport: re PR fortran/78021 (Wrong result with optimization on character constant)

2016-10-24  Thomas Koenig  <tkoenig@gcc.gnu.org>

Backport from trunk
PR fortran/78021
* gfc_compare_functions:  Strings with different lengths in
argument lists compare unequal.

2016-10-24  Thomas Koenig  <tkoenig@gcc.gnu.org>

Backport from trunk
PR fortran/78021
* gfortran.dg/string_length_3.f90:  New test.

From-SVN: r241498

re PR fortran/71895 (ICE in gfc_compare_derived_types, at fortran/interface.c:520)

2016-10-24  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/71895
* interface.c (gfc_compare_derived_types):  Convert gcc_assert()
to a gfc_internal_error() to prevent an ICE.

2016-10-24  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/71895
* gfortran.dg/pr71895.f90: New test.

From-SVN: r241496

re PR fortran/61420 ([OOP] type-bound procedure returning a procedure pointer fails to compile)

2016-10-24  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/61420
PR fortran/78013
* resolve.c (resolve_variable): Obtain the typespec for a
variable expression, when the variable is a function result
that is a procedure pointer.

2016-10-24  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/61420
PR fortran/78013
* gfortran.dg/proc_ptr_49.f90: New test.

From-SVN: r241474

backport: re PR tree-optimization/77916 (ICE in verify_gimple_in_cfg: invalid (pointer) operands to plus/minus)

[gcc]

2016-10-23  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

Backport from mainline
2016-10-17  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR tree-optimization/77916
* gimple-ssa-strength-reduction.c (create_add_on_incoming_edge):
Don't allow a MINUS_EXPR for pointer arithmetic for either known
or unknown strides.
(record_increment): Increments of -1 for unknown strides just use
a multiply initializer like other negative values.
(analyze_increments): Remove stopgap solution for -1 increment
applied to pointer arithmetic.
(insert_initializers): Requirement of initializer for -1 should be
based on pointer-typedness of the candidate basis.

[gcc/testsuite]

2016-10-23  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

Backport from mainline
2016-10-17  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR tree-optimization/77916
* gcc.dg/torture/pr77916.c: New.

From-SVN: r241462

Daily bump.

From-SVN: r241458

Daily bump.

From-SVN: r241442

Daily bump.

From-SVN: r241434

Daily bump.

From-SVN: r241389

PR78052 Define std::allocator<void>::{construct,destroy}

PR libstdc++/78052
* include/bits/allocator.h (allocator<void>::construct)
(allocator<void>::destroy): Define.
* testsuite/20_util/allocator/void.cc: New test.

From-SVN: r241370

Daily bump.

From-SVN: r241357

re PR target/77991 (ICE on x32 in plus_constant, at explow.c:87)

PR target/77991
* config/i386/i386.c (legitimize_tls_address)
<case TLS_MODEL_INITIAL_EXEC>: For TARGET_64BIT || TARGET_ANY_GNU_TLS
convert dest to Pmode if different than Pmode.

testsuite/ChangeLog:

PR target/77991
* gcc.target/i386/pr77991.c: New test.

From-SVN: r241346

PR77990 fix unique_ptr for non-copyable deleters

PR libstdc++/77990
* include/bits/unique_ptr.h (unique_ptr::unique_ptr(pointer)): Set
pointer member after value-initialization of tuple.
* testsuite/20_util/unique_ptr/assign/48635_neg.cc: Adjust dg-errors.
* testsuite/20_util/unique_ptr/cons/77990.cc: New test.

From-SVN: r241337

Daily bump.

From-SVN: r241324

backport: tree-vrp.c (vrp_intersect_ranges_1): Allocate bitmap before copying.

2016-10-18  Kugan Vivekanandarajah  <kuganv@linaro.org>

Backport from mainline
2016-10-11  Kugan Vivekanandarajah  <kuganv@linaro.org>

* tree-vrp.c (vrp_intersect_ranges_1): Allocate bitmap before
  copying.

From-SVN: r241298

Daily bump.

From-SVN: r241290

backport: re PR fortran/77978 (stop codes misinterpreted in both f2003 and f2008)

2016-10-17  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/77978
* match.c (gfc_match_stopcode): Fix error reporting for several
deficiencies in matching stop-codes.

2016-10-17  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/77978
* gfortran.dg/pr77978_1.f90: New test.
* gfortran.dg/pr77978_2.f90: Ditto.
* gfortran.dg/pr77978_3.f90: Ditto.

From-SVN: r241286

Update docs on libstdc++ source-code layout

Backport from mainline:
2016-10-10  Jonathan Wakely  <jwakely@redhat.com>

* doc/xml/manual/appendix_contributing.xml (contrib.organization):
Replace <literallayout> with nested <variablelist> elements. Update
some more outdated text.
* doc/html/*: Regenerate.

Backport from mainline:
2016-10-10  Jonathan Wakely  <jwakely@redhat.com>

* doc/xml/manual/appendix_contributing.xml (contrib.organization):
Describe other subdirectories and add markup. Remove outdated
reference to check-script target.

From-SVN: r241270

Change test to use VERIFY not assert

Backport from mainline:
2016-10-13  Jonathan Wakely  <jwakely@redhat.com>

* testsuite/26_numerics/random/default_random_engine.cc: Use VERIFY
instead of assert.

From-SVN: r241269

Remove redundant const qualifiers from string_view

Backport from mainline:
2016-10-11  Jonathan Wakely  <jwakely@redhat.com>

* include/experimental/string_view (basic_string_view::_S_compare):
Remove redundant const from return type.
* include/std/string_view (basic_string_view:_S_compare): Likewise.

From-SVN: r241268

Use normal mode containers in searchers

Backport from mainline:
2016-10-13  Jonathan Wakely  <jwakely@redhat.com>

* include/experimental/functional (boyer_moore_searcher)
(__boyer_moore_map_base, __boyer_moore_array_base): Qualify containers
with _GLIBCXX_STD_C.

From-SVN: r241267

Add noexcept to std::function swap

Backport from mainline:
2016-08-22  Jonathan Wakely  <jwakely@redhat.com>

PR libstdc++/77322
* doc/xml/manual/intro.xml: Document DR 2062 change.
* include/std/functional (function::swap): Add noexcept.
(swap(function<Res(Args...)&, function<Res(Args...)&)): Likewise.
* testsuite/20_util/function/77322.cc: New test.

From-SVN: r241266

Fix invalid dg-do directive

Backport from mainline:
2016-08-17  Jonathan Wakely  <jwakely@redhat.com>

* testsuite/20_util/reference_wrapper/invoke-2.cc: Fix invalid dg-do
directive.

From-SVN: r241265

Use ::new to avoid finding overloaded operator new

Backport from mainline:
2016-08-06  Jonathan Wakely  <jwakely@redhat.com>

PR libstdc++/72820
* include/std/functional (_Function_base::_Base_manager::_M_clone):
Qualify new operator.
* testsuite/20_util/function/cons/72820.cc: New test.

From-SVN: r241264

Backport fixes to std::experimental::sample

PR libstdc++/77994
* include/experimental/algorithm (experimental::sample): Convert size
argument to iterator difference type. Fix invalid use of input
iterator. Defend against overloaded comma operator.

From-SVN: r241263

backport: re PR tree-optimization/77937 (ICE: in replace_one_candidate, at gimple-ssa-strength-reduction.c:3370)

[gcc]

2016-10-17  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

Backport from mainline
2016-10-12  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR tree-optimization/77937
* gimple-ssa-strength-reduction.c (analyze_increments): Use
POINTER_TYPE_P on the candidate type to determine whether
candidates in this chain require pointer arithmetic.

2016-10-13  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR tree-optimization/77937
* gimple-ssa-strength-reduction.c (analyze_increments): Set cost
to infinite when we have a pointer with an increment of -1.

[gcc/testsuite]

2016-10-17  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

Backport from mainline
2016-10-13  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR tree-optimization/77937
* gcc.dg/torture/pr77937-1.c: New.
* gcc.dg/torture/pr77937-2.c: New.

From-SVN: r241244

Do not merge BBs with a different EH landing pads (PR

Backported from mainline
2016-10-13  Martin Liska  <mliska@suse.cz>

PR tree-optimization/77943
* g++.dg/tree-ssa/pr77943.C: New test.
Backported from mainline
2016-10-13  Martin Liska  <mliska@suse.cz>

PR tree-optimization/77943
* tree-ssa-tail-merge.c (merge_stmts_p): Do not merge BBs with
a different EH landing pads.

From-SVN: r241240

system-darwin-ppc64.ads (Support_64_Bit_Divides): Delete.

* system-darwin-ppc64.ads (Support_64_Bit_Divides): Delete.
* system-linux-armeb.ads (Support_64_Bit_Divides): Likewise.
* system-linux-mips.ads (Support_64_Bit_Divides): Likewise.
* system-linux-mips64el.ads (Support_64_Bit_Divides): Likewise.
* system-linux-mipsel.ads (Support_64_Bit_Divides): Likewise.
* system-linux-sparcv9.ads (Support_64_Bit_Divides): Likewise.
* system-rtems.ads (Support_64_Bit_Divides): Likewise.

From-SVN: r241233

Daily bump.

From-SVN: r241224

sparc.c (sparc_expand_vector_init): Only accept literal constants in CONST_VECTORs.

* config/sparc/sparc.c (sparc_expand_vector_init): Only accept literal
constants in CONST_VECTORs.

From-SVN: r241220

backport: system-aix.ads (Always_Compatible_Rep): Change to False.

Backport from mainline
2016-10-16  Eric Botcazou  <ebotcazou@adacore.com>

* system-aix.ads (Always_Compatible_Rep): Change to False.
* system-aix64.ads (Always_Compatible_Rep): Likewise.
* system-hpux-ia64.ads (Always_Compatible_Rep): Likewise.
* system-hpux.ads (Always_Compatible_Rep): Likewise.
* system-linux-alpha.ads (Always_Compatible_Rep): Likewise.
* system-linux-hppa.ads (Always_Compatible_Rep): Likewise.
* system-linux-ia64.ads (Always_Compatible_Rep): Likewise.
* system-linux-mips.ads (Always_Compatible_Rep): Likewise.
* system-linux-mips64el.ads (Always_Compatible_Rep): Likewise.
* system-linux-mipsel.ads (Always_Compatible_Rep): Likewise.
* system-linux-s390.ads (Always_Compatible_Rep): Likewise.
* system-linux-s390x.ads (Always_Compatible_Rep): Likewise.
* system-linux-sh4.ads (Always_Compatible_Rep): Likewise.
* system-linux-sparc.ads (Always_Compatible_Rep): Likewise.
* system-linux-sparcv9.ads (Always_Compatible_Rep): Likewise.
* system-rtems.ads (Always_Compatible_Rep): Likewise.

From-SVN: r241215

Daily bump.

From-SVN: r241208

sparc.c (sparc_expand_vec_perm_bmask): Use a scratch register as destination of bmask.

* config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use a scratch
register as destination of bmask.
(vector_init_bshuffle): Likewise.
* config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
(bmaskdi_vis): Enable only in 64-bit mode.

From-SVN: r241207

Daily bump.

From-SVN: r241194

Daily bump.

From-SVN: r241144

Delete x86 deprecated pcommit instruction support

Delete x86 pcommit instruction support, which has been deprecated:

https://software.intel.com/en-us/blogs/2016/09/12/deprecate-pcommit-instruction

gcc/

2016-10-13  Andrew Senkevich  <andrew.senkevich@intel.com>

Backport from trunk
* common/config/i386/i386-common.c (OPTION_MASK_ISA_PCOMMIT_UNSET,
OPTION_MASK_ISA_PCOMMIT_SET): Deleted definitions.
(ix86_handle_option): Deleted handle of OPT_mpcommit.
* config.gcc: Deleted pcommitintrin.h
* config/i386/pcommitintrin.h: Deleted file.
* config/i386/cpuid.h (bit_PCOMMIT): Deleted.
* config/i386/driver-i386.c (host_detect_local_cpu): Deleted pcommit
detection.
* config/i386/i386-c.c (ix86_target_macros_internal): Deleted define
__PCOMMIT__.
* config/i386/i386.c (ix86_target_string): Deleted -mpcommit.
(PTA_PCOMMIT, IX86_BUILTIN_PCOMMIT, __builtin_ia32_pcommit): Deleted.
(ix86_option_override_internal): Deleted handle of option.
(ix86_valid_target_attribute_inner_p): Deleted pcommit.
* config/i386/i386.h (TARGET_PCOMMIT, TARGET_PCOMMIT_P): Deleted.
* config/i386/i386.md (unspecv): Deleted UNSPECV_PCOMMIT.
(pcommit): Deleted instruction.
* config/i386/i386.opt: Mention -mpcommit deprecation.
* config/i386/x86intrin.h: Deleted inclusion of pcommitintrin.h.

gcc/testsuite/

2016-10-13  Andrew Senkevich  <andrew.senkevich@intel.com>

Backport from trunk
* testsuite/gcc.target/i386/pcommit-1.c: Deleted test.
* testsuite/gcc.target/i386/sse-12.c: Deleted -pcommit option.
* testsuite/gcc.target/i386/sse-13.c: Ditto.
* testsuite/gcc.target/i386/sse-14.c: Ditto.
* testsuite/gcc.target/i386/sse-22.c: Ditto.
* testsuite/gcc.target/i386/sse-23.c: Ditto.
* testsuite/g++.dg/other/i386-2.C: Ditto.
* testsuite/g++.dg/other/i386-3.C: Ditto.

From-SVN: r241142

Daily bump.

From-SVN: r241079

Daily bump.

From-SVN: r241014

backport: system-linux-armeb.ads (Backend_Overflow_Checks): Change to True.

Backport from mainline
2016-10-11  Eric Botcazou  <ebotcazou@adacore.com>

* system-linux-armeb.ads (Backend_Overflow_Checks): Change to True.
* system-linux-mips.ads (Backend_Overflow_Checks): Likewise.
* system-linux-mips64el.ads (Backend_Overflow_Checks): Likewise.
* system-linux-mipsel.ads (Backend_Overflow_Checks): Likewise.
* system-linux-sparcv9.ads (Backend_Overflow_Checks): Likewise.
* system-rtems.ads: Likewise.

2016-10-11  Eric Botcazou  <ebotcazou@adacore.com>
    Tristan Gingold  <gingold@adacore.com>

* system-linux-ppc64.ads: Delete.
* system-linux-ppc.ads: Make 32-bit/64-bit neutral.
* gcc-interface/Makefile.in (PowerPC/Linux): Simplify.

From-SVN: r241006

Daily bump.

From-SVN: r240960

gen-avr-mmcu-texi.c (string.h): Include.

* config/avr/gen-avr-mmcu-texi.c (string.h): Include.

From-SVN: r240927

utils.c (convert): For a biased input type, convert the bias itself to the base type before adding it.

* gcc-interface/utils.c (convert): For a biased input type, convert
the bias itself to the base type before adding it.

From-SVN: r240912

Daily bump.

From-SVN: r240905

Daily bump.

From-SVN: r240894

Daily bump.

From-SVN: r240880

Daily bump.

From-SVN: r240847

re PR target/77874 (two problems with gcc.target/i386/avx-1.c)

PR target/77874
* config/i386/sse.md (<mask_codefor><code><mode>3<mask_name>):
Remove wrong assert.
(<floatsuffix>float<sseintvecmodelower><mode>2<mask_name><round_name>:
Use <round_constraint> as operand 1 constraint.

From-SVN: r240838

Daily bump.

From-SVN: r240821

re PR fortran/58991 (ICE with associate and character string constant)

2016-10-05  Steven G. Kargl  <kargls@gcc.gnu.org>

PR fortran/58991
PR fortran/58992
* resolve.c (resolve_assoc_var):  Fix CHARACTER type-spec for a
selector in ASSOCIATE.
(resolve_fl_variable): Skip checks for an ASSOCIATE variable.

2016-10-05  Steven G. Kargl  <kargls@gcc.gnu.org>

PR fortran/58991
PR fortran/58992
* gfortran.dg/associate_22.f90: New test.

From-SVN: r240817

PR 70564 disambiguate constructor for not_fn call wrapper

PR libstdc++/70564
* include/experimental/functional (_Not_fn): Add second parameter to
disambiguate copying from initialization by not_fn.
(not_fn): Add second argument to initialization.
* testsuite/experimental/functional/not_fn.cc: Copy call wrapper using
direct-initialization. Test abstract class.

From-SVN: r240790

77864 Fix noexcept conditions for map/set default constructors

PR libstdc++/77864
* include/bits/stl_map.h (map::map()): Use nothrow constructibility
of comparison function in conditional noexcept.
* include/bits/stl_multimap.h (multimap::multimap()): Likewise.
* include/bits/stl_multiset.h (multiset::multiset()): Likewise.
* include/bits/stl_set.h (set::set()): Likewise.
* testsuite/23_containers/map/cons/noexcept_default_construct.cc:
New test.
* testsuite/23_containers/multimap/cons/noexcept_default_construct.cc:
Likewise.
* testsuite/23_containers/multiset/cons/noexcept_default_construct.cc:
Likewise.
* testsuite/23_containers/set/cons/noexcept_default_construct.cc:
Likewise.

From-SVN: r240789

Daily bump.

From-SVN: r240759

Daily bump.

From-SVN: r240736

Fix libstdc++ versioned namespace build

PR libstdc++/68323
PR libstdc++/77794
* config/abi/pre/gnu-versioned-namespace.ver: Add export for
__cxa_thread_atexit.
* include/Makefile.am: Add <experimental/lfts_config.h>
* include/Makefile.in: Regenerate.
* src/filesystem/Makefile.in: Likewise.
* include.bits/basic_string.h: Fix nesting of versioned namespaces.
* include/bits/c++config: Declare versioned namespaces for literals.
* include/bits/regex.h (basic_regex, match_results): Add workarounds
for PR c++/59256.
* include/bits/uniform_int_dist.h: Fix nesting of versioned namespace.
* include/std/chrono: Likewise.
* include/std/complex: Likewise.
* include/experimental/fs_fwd.h: Declare versioned namespace.
* include/experimental/lfts_config.h: Declare versioned namespaces.
* include/experimental/algorithm: Include
<experimental/lfts_config.h>.
* include/experimental/any: Likewise.
* include/experimental/chrono: Likewise.
* include/experimental/functional: Likewise.
* include/experimental/optional: Likewise.
* include/experimental/ratio: Likewise.
* include/experimental/system_error: Likewise.
* include/experimental/tuple: Likewise.
* include/experimental/type_traits: Likewise.
* include/experimental/string_view: Likewise. Fix nesting of
versioned namespaces.
* include/experimental/string_view.tcc: Reopen inline namespace for
non-inline function definitions.
* testsuite/17_intro/using_namespace_std_exp_neg.cc: New test.
* testsuite/20_util/duration/literals/range.cc: Adjust dg-error line.
* testsuite/experimental/any/misc/any_cast_neg.cc: Likewise.
* testsuite/ext/profile/mutex_extensions_neg.cc: Likewise.

From-SVN: r240720

Daily bump.

From-SVN: r240703

Daily bump.

From-SVN: r240697

backport: re PR fortran/77391 (gfortran allows CHARACTER(LEN=:),PARAMETER:: STRING='constant'   buts does not report it as an extension)

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/77391
* resolve.c (deferred_requirements): New function to check F2008:C402.
(resolve_fl_variable,resolve_fl_parameter): Use it.

PR fortran/77391
* gfortran.dg/pr77391.f90: New test.

From-SVN: r240693

backport: re PR fortran/77380 (ICE in gfc_check_dependency, at fortran/dependency.c:1255)

2016-09-30  Steven G. Kargl <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/77380
* dependency.c (gfc_check_dependency): Do not assert with
-fcoarray=lib.

PR fortran/77380
* gfortran.dg/pr77380.f90: New test.

From-SVN: r240692

Daily bump.

From-SVN: r240689

backport: re PR fortran/61318 (Improve error diagnostic by pointing to the expression and not to declared-at of a USE-associated variable)

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/61318
* interface.c (compare_parameter): Use better locus for error message.

PR fortran/61318
* gfortran.dg/pr61318.f90: New test.

From-SVN: r240683

backport: re PR fortran/71859 (ICE on same variable/subroutine name (verify_gimple failed))

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/71859
* gfortran.dg/pr71859.f90: New test.
* gfortran.dg/intrinsic_numeric_arg.f: Update error message.
* gfortran.dg/coarray_collectives_1.f90: Ditto.

PR fortran/71859
* check.c(numeric_check): Prevent ICE.  Issue error for invalid
subroutine as an actual argument when numeric argument is expected.

From-SVN: r240682

backport: re PR fortran/69867 (ICE on initializing character in type with array of incompatible data)

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/69867
* decl.c (build_struct): Ensure that pointers point to something.

PR fortran/69962
* decl.c (gfc_set_constant_character_len):  if expr is not
constant issue an error instead of an ICE.

PR fortran/69962
* gfortran.dg/pr69962.f90: New test.

PR fortran/69867
* gfortran.dg/pr69867.f90: New test.

From-SVN: r240679

backport: re PR fortran/77351 (ICE in remove_trim, at frontend-passes.c:1145)

2016-09-30  Steven g. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/77351
* frontend-passes.c (remove_trim,combine_array_constructor): Check for
NULL pointer.

PR fortran/77351
* gfortran.dg/pr77351.f90: New test.

From-SVN: r240678

backport: re PR fortran/68566 (ICE on using unusable array in reshape (double free or corruption))

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/68566
* check.c (gfc_check_reshape): Check for constant expression.

PR fortran/68566
* gfortran.dg/pr68566.f90: new test.

From-SVN: r240671

backport: re PR fortran/71799 (ICE in DO loop code emission (gfc_resolve_iterator))

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/71799
* gfortran.dg/pr71799.f90: New test.

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/71799
* resolve.c(gfc_resolve_iterator): Failure of type conversion need
not ICE.

From-SVN: r240668

backport: re PR fortran/70006 (Duplicate errors "label not defined")

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/70006
* io.c (gfc_resolve_dt): Use correct locus.
* resolve.c (resolve_branch): Ditto.

2016-09-30  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/70006
* gfortran.dg/pr70006.f90: New test.

From-SVN: r240659

backport: re PR fortran/77707 (formatted direct access: nextrec off by one)

2016-09-29  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

Backport from trunk
PR libgfortran/77707
io/transfer.c (next_record): Flush before calculating next_record.
Correctly calculate.

* gfortran.dg/inquire_17.f90: New test.

From-SVN: r240645

Daily bump.

From-SVN: r240641

backport: re PR fortran/77420 (gfortran and equivalence produces internal compiler error)

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/77420
* trans-common.c:  Handle array elements in equivalence when
the lower and upper bounds of array spec are NULL.

PR fortran/77460
* simplify.c (simplify_transformation_to_scalar):  On error, result
may be NULL, simply return.

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk

PR fortran/77420
* gfortran.dg/pr77420_1.f90: New test.
* gfortran.dg/pr77420_2.f90: Ditto.
* gfortran.dg/pr77420_3.f90: New test. Requires ...
* gfortran.dg/pr77420_4.f90: this file.

PR fortran/77460
* gfortran.dg/pr77460.f90: New test.

From-SVN: r240640

backport: re PR fortran/77506 (F2008 Standard does not allow CHARACTER(LEN=*) in array constructor)

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/77506
* array.c (gfc_match_array_constructor): CHARACTER(len=*) cannot
appear in an array constructor.

PR fortran/77507
* intrinsic.c (add_functions):  Use correct keyword.

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/77507
  * ieee/ieee_arithmetic.F90 (IEEE_VALUE_4,IEEE_VALUE_8,IEEE_VALULE_10,
IEEE_VALUE_16):  Use correct keyword.

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/77506
* gfortran.dg/pr77506.f90: New test.

PR fortran/77507
* gfortran.dg/pr77507.f90: New test.

From-SVN: r240636

backport: re PR fortran/71067 (ICE on data initialization with insufficient value)

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/71067
* decl.c (match_data_constant): On error, set 'result' to NULL.

PR fortran/77260
* gcc/fortran/trans-decl.c (generate_local_decl): Suppress warning
for unused variable if symbol is entry point.

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/71067
* gfortran.dg/pr71067_1.f90: New test.
* gfortran.dg/pr71067_2.f90: Ditto.

PR fortran/77260
* gfortran.dg/pr77260_1.f90: New test.
* gfortran.dg/pr77260_2.f90: Ditto.

From-SVN: r240635

backport: re PR fortran/69514 (ICE with nested array constructor)

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>

Backport from trunk
PR fortran/69514
* array.c (gfc_match_array_constructor):  If type-spec is present,
walk the array constructor performing possible conversions for
numeric types.

2016-09-29  Steven G. Kargl  <kargl@gcc.gnu.org>
    Louis Krupp      <lkrupp@gcc.gnu.org>

Backport from trunk
PR fortran/69514
* gfortran.dg/pr69514_1.f90: New test.
* gfortran.dg/pr69514_2.f90: New test.

Co-Authored-By: Louis Krupp <lkrupp@gcc.gnu.org>
From-SVN: r240628

backport: re PR fortran/71730 (ICE when character length specification uses an undefined variable)

2016-09-28  Steven G. Kargl  <kargl@gcc.gnu.org>

backport from trunk
PR fortran/71730
* gfortran.dg/pr71730.f90: New test.
* gfortran.dg/bounds_check_strlen_2.f90: Fix invalid code.
* gfortran.dg/array_constructor_27.f03: Update dg-error message.
* gfortran.dg/array_constructor_26.f03: Ditto.

PR fortran/77612
* gfortran.dg/pr77612.f90: New test.

2016-09-28  Steven G. Kargl  <kargl@gcc.gnu.org>

backport from trunk
PR fortran/71730
* decl.c (char_len_param_value): Check return value of
gfc_reduce_init_expr().

PR fortran/77612
* decl.c (char_len_param_value): Check parent namespace for
seen_implicit_none.

From-SVN: r240608

Daily bump.

From-SVN: r240605

backport: re PR fortran/77694 (ICE in optimize_binop_array_assignment, at fortran/frontend-passes.c:1080)

2016-09-28  Steven G. Kargl  <kargl@gcc.gnu.org>

backport from trunk
PR fortran/77694
* frontend-passes.c (optimize_binop_array_assignment): Check pointer
for NULL.

2016-09-28  Steven G. Kargl  <kargl@gcc.gnu.org>

backport from trunk
PR fortran/77694
* gfortran.dg/pr77694.f90: New test.

From-SVN: r240604

backport: re PR rtl-optimization/71709 (powerpc64le: argument to strcpy() optimised out)

2016-09-28  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
    Alan Modra  <amodra@gmail.com>

Backport from mainline
2016-07-01  Alan Modra  <amodra@gmail.com>

PR rtl-optimization/71709
* ira-lives.c (find_call_crossed_cheap_reg): Exit loop on arg reg
being set, not referenced.

Co-Authored-By: Alan Modra <amodra@gmail.com>
From-SVN: r240599

Correct libgcc complex multiply excess precision handling (PR libgcc/77519).

libgcc complex multiply is meant to eliminate excess
precision from certain internal values by forcing them to memory in
exactly those cases where the type has excess precision.  But in
https://gcc.gnu.org/ml/gcc-patches/2014-09/msg01894.html I
accidentally inverted the logic so that values get forced to memory in
exactly the cases where it's not needed.  (This is a pessimization in
the no-excess-precision case, in principle could lead to bad results
depending on code generation in the excess-precision case.  Note: I do
not have a test demonstrating bad results.)

Bootstrapped with no regressions on x86_64-pc-linux-gnu.  Code size
went down on x86_64 as expected; old sizes:

   text    data     bss     dec     hex filename
    887       0       0     887     377 _muldc3.o
    810       0       0     810     32a _mulsc3.o
   2032       0       0    2032     7f0 _multc3.o
    983       0       0     983     3d7 _mulxc3.o

New sizes:

    847       0       0     847     34f _muldc3.o
    770       0       0     770     302 _mulsc3.o
   2032       0       0    2032     7f0 _multc3.o
    951       0       0     951     3b7 _mulxc3.o

Backport from mainline
2016-09-07  Joseph Myers  <joseph@codesourcery.com>

PR libgcc/77519
* libgcc2.c (NOTRUNC): Invert settings.

From-SVN: r240586

backport: re PR testsuite/77411 (object-size-9.c -fpic -m32 failure)

backport from mainline
2016-09-23  Bernd Edlinger  <bernd.edlinger@hotmail.de>
            Tom de Vries  <tom@codesourcery.com>

        PR testsuite/77411
        * c-c++-common/ubsan/object-size-9.c: Call __builtin_exit in C++.

From-SVN: r240575

libstdc++/77686 use may_alias for std::function storage

PR libstdc++/77686
* include/std/functional (_Any_data): Add may_alias attribute.

From-SVN: r240574

Daily bump.

From-SVN: r240561

Fix lifetime of mutex lock in debug iterator

* include/debug/safe_iterator.h (_Safe_iterator::operator++()): Fix
lifetime of lock.

From-SVN: r240549

backport: rs6000.c (rs6000_output_symbol): Don't modify VAR_DECL string.

        Backport from mainline
        2016-09-27  David Edelsohn  <dje.gcc@gmail.com>

        * config/rs6000/rs6000.c (rs6000_output_symbol): Don't modify
        VAR_DECL string.

From-SVN: r240544

re PR target/77478 (Incorrect code generated with -O3, m32, -msse2 and -ffast-math)

2016-09-27  Richard Biener  <rguenther@suse.de>

PR tree-optimization/77478
* tree-vect-loop-manip.c (vect_gen_niters_for_prolog_loop):
Fix alignment of SSA var used before the alignment prologue.

* gcc.dg/torture/pr77478.c: New testcase.

From-SVN: r240532

Backport fix for PR 65210

gcc/ChangeLog

2016-09-27  Senthil Kumar Selvaraj  <senthil_kumar.selvaraj@atmel.com>

Backport from trunk r227496

PR target/65210
* config/avr/avr.c (avr_eval_addr_attrib): Look for io_low
attribute as well.

gcc/testsuite/ChangeLog

2016-09-27  Senthil Kumar Selvaraj  <senthil_kumar.selvaraj@atmel.com>

Backport from trunk r227496

PR target/65210
* gcc.target/avr/pr65210.c: New test.

From-SVN: r240521

Revert part of r240397 that got accidentally committed

From-SVN: r240520

Daily bump.

From-SVN: r240512

Daily bump.

From-SVN: r240480

Daily bump.

From-SVN: r240468

Daily bump.

From-SVN: r240459

backport: avr.c (avr_rtx_costs_1): Handle DImode MULT.

Backport from 2016-09-22 trunk r240388.

2016-09-22  Senthil Kumar Selvaraj  <senthil_kumar.selvaraj@atmel.com>

* config/avr/avr.c (avr_rtx_costs_1): Handle DImode MULT.
(avr_address_cost): Replace 61 with MAX_LD_OFFSET(mode).

From-SVN: r240397

Daily bump.

From-SVN: r240384

Daily bump.

From-SVN: r240337

Fix JUMP_LABEL documentation

* doc/rtl.texi (JUMP_LABEL): Document RETURN and SIMPLE_RETURN values.

From-SVN: r240327

dump-noaddr.x: Remove debug statements.

* gcc.c-torture/unsorted/dump-noaddr.x: Remove debug statements.

From-SVN: r240320

backport: re PR target/77326 ([avr] Invalid optimization omits comparison)

gcc/
Backport from 2016-09-21 trunk r240306.
PR target/77326
* config/avr/avr.c (hard-reg-set.h): Move #include up in front
of rtl.h to that HARD_CONST is defined in rtl.h.
(avr_notice_update_cc) [CC_NONE]: If insn touches some regs
mentioned in cc_status, do CC_STATUS_INIT.

gcc/testsuite/
Backport from 2016-09-21 trunk r240306.
PR target/77326
* gcc.target/avr/torture/pr77326.c: New test.

From-SVN: r240315