Test whitespace handling in std::complex extraction

* testsuite/26_numerics/complex/inserters_extractors/char/dr2714.cc:
Add tests using noskipws.

From-SVN: r255632

gigi.h (pad_type_has_rm_size): Declare.

* gcc-interface/gigi.h (pad_type_has_rm_size): Declare.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Variable>: Do not build
a padding type for the alignment before validating the size.
Flip conditional construct and add a comment.
* gcc-interface/trans.c (Attribute_to_gnu) <Attr_Size>: Make sure to
apply the exception for padded objects to the type of the object.
* gcc-interface/utils.c (hash_pad_type): New static function.
(lookup_and_insert_pad_type): Rename into...
(canonicalize_pad_type): ...this.  Call hash_pad_type, do only one
lookup with insertion and always return the canonical type.
(maybe_pad_type): Adjust to above changes.  Set debug type later.
(pad_type_has_rm_size): New predicate.
(set_reverse_storage_order_on_pad_type): Adjust to above changes.

From-SVN: r255631

Improve std::complex test and move to sub-directory

* testsuite/26_numerics/complex/dr2714.cc: Move to ...
* testsuite/26_numerics/complex/inserters_extractors/char/dr2714.cc:
... Here. Remove duplicate header and dg-options. Check first invalid
character gets putback. Remove wchar_t test.

From-SVN: r255630

re PR middle-end/83415 (ICE during gimplification of assignment to read-only vector)

2017-12-14  Richard Biener  <rguenther@suse.de>

PR c/83415
c/
* c-fold.c (c_fully_fold_internal): Treat VIEW_CONVERT_EXPR
like REALPART_EXPR for the behavior of whether its operand
is an lvalue.

* gcc.dg/pr83415.c: New testcase.

From-SVN: r255629

re PR tree-optimization/83418 (ICE in extract_range_for_var_from_comparison_expr, at vr-values.c:448)

2017-12-14  Richard Biener  <rguenther@suse.de>

PR tree-optimization/83418
* vr-values.c (vr_values::extract_range_for_var_from_comparison_expr):
Instead of asserting we don't get unfolded comparisons deal with
them.

* gcc.dg/torture/pr83418.c: New testcase.

From-SVN: r255628

re PR bootstrap/83396 (Bootstrap failures with Statement Frontiers)

PR bootstrap/83396
* var-tracking.c (vt_initialize): Ignore non-DEBUG_INSNs outside of
basic blocks.  Assert debug bind insns don't appear outside of bbs,
don't reset them.  Assert insns without BLOCK_FOR_INSN are outside of
bb.  Simplify.

* gcc.dg/pr83396.c: New test.

From-SVN: r255627

re PR tree-optimization/83198 (ICE in format_floating, at gimple-ssa-sprintf.c:1900)

PR tree-optimization/83198
* gimple-ssa-sprintf.c (format_floating): Set type solely based on
dir.modifier, regardless of TREE_TYPE (arg).  Assume non-REAL_CST
value if arg is a REAL_CST with incompatible type.

* gcc.dg/pr83198.c: New test.
* gcc.dg/tree-ssa/pr83198.c: New test.

From-SVN: r255626

[PATCH PR81228][AARCH64]Fix ICE by adding LTGT in vec_cmp<mode><v_int_equiv>

This patch is a follow up to the existing discussions on
https://gcc.gnu.org/ml/gcc-patches/2017-07/msg01904.html
Bin had earlier submitted this patch to fix the ICE that
occurs because of the missing LTGT in aarch64-simd.md.
That discussion opened up a new bug report PR81647 for
an inconsistent behavior.

As discussed earlier on the gcc-patches discussion and on
the bug report, PR81647 was occurring because of how UNEQ
was handled in aarch64-simd.md rather than LTGT.
Since __builtin_islessgreater is guaranteed to not give an
FP exception but LTGT might, __builtin_islessgreater gets
converted to ~UNEQ very early on in fold_builtin_unordered_cmp.
Thus I will post a separate patch for correcting how UNEQ and
other unordered comparisons are handled in aarch64-simd.md.

This patch is only adding the missing LTGT to plug the ICE.

Testing done: Checked for regressions on bootstrapped
aarch64-none-linux-gnu and added a new compile time test case
that gives out LTGT to make sure it doesn't ICE

*** gcc/ChangeLog ***

2017-12-14  Sudakshina Das  <sudi.das@arm.com>
    Bin Cheng  <bin.cheng@arm.com>

PR target/81228
* config/aarch64/aarch64.c (aarch64_select_cc_mode): Move LTGT
to CCFPEmode.
* config/aarch64/aarch64-simd.md (vec_cmp<mode><v_int_equiv>): Add
LTGT.

*** gcc/testsuite/ChangeLog ***

2017-12-14  Sudakshina Das  <sudi.das@arm.com>

PR target/81228
* gcc.dg/pr81228.c: New.

Co-Authored-By: Bin Cheng <bin.cheng@arm.com>
From-SVN: r255625

[AArch64] Specify fp16 support for Cortex-A55 and Cortex-A75

The Cortex-A55 and Cortex-A75 processors support the fp16 extension.
We already specify them as such in the arm port.
This patch makes aarch64 consistent on this front.

Bootstrapped and tested on aarch64-none-linux-gnu.
Manually checked that compiling with aarch64-none-linux-gnu-gcc -mcpu=cortex-a55 -dM -E - < /dev/null
shows __ARM_FEATURE_FP16_VECTOR_ARITHMETIC and __ARM_FEATURE_FP16_SCALAR_ARITHMETIC being specified
as expected whereas they were not before this patch.

* config/aarch64/aarch64-cores.def (cortex-a55, cortex-a75,
cortex-a75.cortex-a55): Specify AARCH64_FL_F16 in the arch features.

From-SVN: r255624

re PR c++/79650 (ICE on invalid c++ code with label arithmetic in convert_nontype_argument (pt.c:6515))

PR c++/79650
* pt.c (convert_nontype_argument): Diagnose
reduced_constant_expression_p expressions that aren't INTEGER_CST.

* g++.dg/template/pr79650.C: New test.

From-SVN: r255623

re PR libfortran/78549 (Very slow formatted internal file output)

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

        PR libgfortran/78549
        * io/inquire.c (inquire_via_unit): Adjust test for existence for
        pre-connected internal units.
        * io/transfer.c (finalize_transfer): When done with a transfer
        to internal units, free the format buffer and close the stream.
        (st_read_done): Delete freeing the stream, now handled using
        sclose in finalize_transfer. (st_write_done): Likewise.
        * io/unit.c (get_unit): Return NULL for special reserved unit
        numbers, signifying not accessible to the user.
        (init_units): Insert the two special internal units into the
        unit treap. This makes these unit structures available without
        further allocations for later use by internal unit I/O. These
        units are automatically deleted by normal program termination.
        * io/unix.c (mem_close): Add a guard check to protect from double free.

From-SVN: r255621

Daily bump.

From-SVN: r255620

poly_int: add poly-int.h

This patch adds a new "poly_int" class to represent polynomial integers
of the form:

  C0 + C1*X1 + C2*X2 ... + Cn*Xn

It also adds poly_int-based typedefs for offsets and sizes of various
precisions.  In these typedefs, the Ci coefficients are compile-time
constants and the Xi indeterminates are run-time invariants.  The number
of coefficients is controlled by the target and is initially 1 for all
ports.

Most routines can handle general coefficient counts, but for now a few
are specific to one or two coefficients.  Support for other coefficient
counts can be added when needed.

The patch also adds a new macro, IN_TARGET_CODE, that can be
set to indicate that a TU contains target-specific rather than
target-independent code.  When this macro is set and the number of
coefficients is 1, the poly-int.h classes define a conversion operator
to a constant.  This allows most existing target code to work without
modification.  The main exceptions are:

- values passed through ..., which need an explicit conversion to a
  constant

- ?: expression in which one arm ends up being a polynomial and the
  other remains a constant.  In these cases it would be valid to convert
  the constant to a polynomial and the polynomial to a constant, so a
  cast is needed to break the ambiguity.

The patch also adds a new target hook to return the estimated
value of a polynomial for costing purposes.

The patch also adds operator<< on wide_ints (it was already defined
for offset_int and widest_int).  I think this was originally excluded
because >> is ambiguous for wide_int, but << is useful for converting
bytes to bits, etc., so is worth defining on its own.  The patch also
adds operator% and operator/ for offset_int and widest_int, since those
types are always signed.  These changes allow the poly_int interface to
be more predictable.

I'd originally tried adding the tests as selftests, but that ended up
bloating cc1 by at least a third.  It also took a while to build them
at -O2.  The patch therefore uses plugin tests instead, where we can
force the tests to be built at -O0.  They still run in negligible time
when built that way.

2017-12-14  Richard Sandiford  <richard.sandiford@linaro.org>
    Alan Hayward  <alan.hayward@arm.com>
    David Sherwood  <david.sherwood@arm.com>

gcc/
* poly-int.h: New file.
* poly-int-types.h: Likewise.
* coretypes.h: Include them.
(POLY_INT_CONVERSION): Define.
* target.def (estimated_poly_value): New hook.
* doc/tm.texi.in (TARGET_ESTIMATED_POLY_VALUE): New hook.
* doc/tm.texi: Regenerate.
* doc/poly-int.texi: New file.
* doc/gccint.texi: Include it.
* doc/rtl.texi: Describe restrictions on subreg modes.
* Makefile.in (TEXI_GCCINT_FILES): Add poly-int.texi.
* genmodes.c (NUM_POLY_INT_COEFFS): Provide a default definition.
(emit_insn_modes_h): Emit a definition of NUM_POLY_INT_COEFFS.
* targhooks.h (default_estimated_poly_value): Declare.
* targhooks.c (default_estimated_poly_value): New function.
* target.h (estimated_poly_value): Likewise.
* wide-int.h (WI_UNARY_RESULT): Use wi::binary_traits.
(wi::unary_traits): Delete.
(wi::binary_traits::signed_shift_result_type): Define for
offset_int << HOST_WIDE_INT, etc.
(generic_wide_int::operator <<=): Define for all types and use
wi::lshift instead of <<.
(wi::hwi_with_prec): Add a default constructor.
(wi::ints_for): New class.
(operator <<): Define for all wide-int types.
(operator /): New function.
(operator %): Likewise.
* selftest.h (ASSERT_KNOWN_EQ, ASSERT_KNOWN_EQ_AT, ASSERT_MAYBE_NE)
(ASSERT_MAYBE_NE_AT): New macros.

gcc/testsuite/
* gcc.dg/plugin/poly-int-tests.h,
gcc.dg/plugin/poly-int-test-1.c,
gcc.dg/plugin/poly-int-01_plugin.c,
gcc.dg/plugin/poly-int-02_plugin.c,
gcc.dg/plugin/poly-int-03_plugin.c,
gcc.dg/plugin/poly-int-04_plugin.c,
gcc.dg/plugin/poly-int-05_plugin.c,
gcc.dg/plugin/poly-int-06_plugin.c,
gcc.dg/plugin/poly-int-07_plugin.c: New tests.
* gcc.dg/plugin/plugin.exp: Run them.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r255617

re PR middle-end/78468 (libgomp.c/reduction-10.c and many more FAIL)

PR middle-end/78468
* emit-rtl.c (init_emit): Remove ??? comment.
* explow.c (get_dynamic_stack_size): Take known alignment of stack
pointer + STACK_DYNAMIC_OFFSET into account in lieu of STACK_BOUNDARY.
* config/sparc/sparc.h (INIT_EXPANDERS): In 32-bit mode, lower the
alignment of 3 virtual registers to BITS_PER_WORD.

* config/sparc/sparc.c (sparc_compute_frame_size): Simplify.

Co-Authored-By: Dominik Vogt <vogt@linux.vnet.ibm.com>
From-SVN: r255616

ppc-auxv.h (PPC_FEATURE2_HTM_NO_SUSPEND): New define.

gcc/
* config/rs6000/ppc-auxv.h (PPC_FEATURE2_HTM_NO_SUSPEND): New define.
* config/rs6000/rs6000.c (cpu_supports_info): Use it.

gcc/testsuite/
* gcc.target/powerpc/cpu-builtin-1.c (htm-no-suspend): Add test.

From-SVN: r255614

re PR c++/81061 (ICE modifying read-only variable)

/cp
2017-12-13  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/81061
* typeck.c (cp_build_modify_expr): Upon cxx_readonly_error
immediately return error_mark_node.

/testsuite
2017-12-13  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/81061
* g++.dg/other/const5.C: New.

From-SVN: r255613

[SFN] don't eliminate regs in markers

Eliminate regs in debug bind insns, but not in markers.

for  gcc/ChangeLog

PR bootstrap/83396
* reload1.c (eliminate_regs_in_insn): Skip debug markers.

From-SVN: r255612

tree-cfg.c (verify_gimple_in_cfg): Verify no non-label stmts with the exception of debug begin stmt markers appear...

* tree-cfg.c (verify_gimple_in_cfg): Verify no non-label stmts
with the exception of debug begin stmt markers appear before
labels.

From-SVN: r255611

re PR bootstrap/83396 (Bootstrap failures with Statement Frontiers)

PR bootstrap/83396
* final.c (rest_of_handle_final): Call variable_tracking_main only
if !flag_var_tracking.

From-SVN: r255610

re PR bootstrap/83396 (Bootstrap failures with Statement Frontiers)

PR bootstrap/83396
PR debug/83391
* tree-cfgcleanup.c (remove_forwarder_block): Keep after
labels debug stmts that can only appear after labels.

* gcc.dg/torture/pr83396.c: New test.
* g++.dg/torture/pr83391.C: New test.

Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r255609

PR libstdc++/59568 fix error handling for std::complex stream extraction

PR libstdc++/59568
* include/std/complex (operator>>): Implement proposed resolution to
LWG 2714. Use putback if and only if a character has been successfully
extracted but isn't a delimiter. Use ctype::widen and traits::eq when
testing if extracted characters match delimiters.
* testsuite/26_numerics/complex/dr2714.cc: New test.

From-SVN: r255608

sel-sched: fix sel_rank_for_schedule for qsort (PR 82398)

PR rtl-optimization/82398
* sel-sched.c (sel_rank_for_schedule): Fix check for zero
EXPR_USEFULNESS in priority comparison.

From-SVN: r255607

combine: Fix PR83393

In move_deaths we move a REG_DEAD note if the instruction combination
has extended the lifetime of a register so that the existing note is
no longer valid.  We find that note using reg_stat, but what that finds
can refer to a later insn.  If so, we cannot use the cached value.  This
patch implements that.

PR rtl-optimization/83393
* combine.c (move_deaths): If reg_stat points to a too new insn in
last_death, do not use it: find the proper insn instead.

gcc/testsuite/
PR rtl-optimization/83393
* gcc.dg/pr83393.c: New testcase.

From-SVN: r255606

[PR C++/15272] lookups with ambiguating dependent base

https://gcc.gnu.org/ml/gcc-patches/2017-12/msg00766.html
PR c++/15272
* pt.c (tsubst_baselink): Don't repeat the lookup for
non-dependent baselinks.

PR c++/15272
* g++.dg/template/pr71826.C: Adjust for 15272 fix.

From-SVN: r255605

* gcc-interface/trans.c (Subprogram_Body_to_gnu): Initialize locus.

From-SVN: r255603

trans.c (process_freeze_entity): Be prepared for a package without body.

* gcc-interface/trans.c (process_freeze_entity): Be prepared for a
package without body.

From-SVN: r255601

decl.c (gnat_to_gnu_entity): Robustify test for types descendant of System.Address.

* gcc-interface/decl.c (gnat_to_gnu_entity): Robustify test for types
descendant of System.Address.
(gnat_to_gnu_subprog_type): Likewise.

From-SVN: r255599

Use C version of multi3 for RVE support.

libgcc/
* config/riscv/t-elf: Use multi3.c instead of multi3.S.
* config/riscv/multi3.c: New file.
* config/riscv/multi3.S: Remove.

From-SVN: r255598

Daily bump.

From-SVN: r255596

re PR tree-optimization/83298 (wrong code at -O1, -O2 and -O3 on x86_64-linux-gnu)

PR tree-optimization/83298
PR tree-optimization/83362
PR tree-optimization/83383
* gimple-ssa-evrp-analyze.h (class evrp_range_analyzer): Make
push_value_range a public interface.  Add new argument to
record_ranges_from_stmt.
* gimple-ssa-evrp-analyze.c
(evrp_range_analyzer::record_ranges_from_stmt): Add new argument.
Update comments.  Handle recording temporary equivalences.
* tree-ssa-dom.c (dom_opt_opt_walker::before_dom_children): Add
new argument to call to evrp_range_analyzer::record_ranges_from_stmt.
* gimple-ssa-evrp.c (evrp_dom_walker::before_dom_children): Likewise.
* tree-ssa-threadedge.c: Include alloc-pool.h, vr-values.h and
gimple-ssa-evrp-analyze.h.
(record_temporary_equivalences_from_phis): Add new argument.  When
the PHI arg is an SSA_NAME, set the result's range to the range
of the PHI arg.
(record_temporary_equivalences_from_stmts_at_dest): Record ranges
from statements too.
(thread_through_normal_block): Accept new argument, evrp_range_analyzer.
Pass it down to children as needed.
(thread_outgoing_edges): Likewise.
(thread_across_edge): Likewise.   Push/pop range state as needed.
* tree-ssa-threadedge.h (thread_outgoing_edges): Update prototype.

PR tree-optimization/83298
PR tree-optimization/83362
PR tree-optimization/83383
* gcc.c-torture/execute/pr83298.c: New test.
* gcc.c-torture/execute/pr83362.c New test.
* gcc.c-torture/execute/pr83383.c New test.

From-SVN: r255593

re PR target/82190 (Possibly latent miscompilation issue on ppc64le-linux-gnu for memcpy-bi.c with -fweb -fno-optimize-strlen)

2017-12-12  Aaron Sawdey  <acsawdey@linux.vnet.ibm.com>

PR target/82190
* config/rs6000/rs6000-string.c (expand_block_compare,
expand_strn_compare): Fix set_mem_size() calls.

From-SVN: r255592

Remove type_dependent_init_p.

* decl.c (value_dependent_init_p): Check the type of a CONSTRUCTOR.
(type_dependent_init_p): Remove.

From-SVN: r255591

PR c++/82115 - ICE with variable initialized with its own address.

* cp-tree.h (struct lang_decl_base): Add dependent_init_p.
(DECL_DEPENDENT_INIT_P, SET_DECL_DEPENDENT_INIT_P): New.
* decl.c (cp_finish_decl): Set it.
(duplicate_decls): Copy it.
* pt.c (tsubst_decl): Clear it.
(value_dependent_expression_p): Revert earlier change.  Check it.

From-SVN: r255590

i386.c (PTA_SKYLAKE_AVX512): Add PTA_CLWB.

* config/i386/i386.c (PTA_SKYLAKE_AVX512): Add PTA_CLWB.
(PTA_CANNONLAKE): Remove PTA_CLWB.

From-SVN: r255589

re PR target/83332 (new test case gfortran.dg/vect/pr81303.f fails (r255499))

2017-12-12  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

PR target/83332
* config/rs6000/vector.md (vcondv2dfv2di): New define_expand.
(vcondv2div2df): Likewise.
(vconduv2dfv2di): Likewise.

From-SVN: r255588

intrinsic.texi (MAXLOC): Remove double description of return value.

2017-12-12  Thomas Koenig  <tkoenig@gcc.gnu.org>

* intrinsic.texi (MAXLOC): Remove double description
of return value.
(MINLOC): Likewise.

From-SVN: r255587

builtins.def (BUILT_IN_NEXTAFTER, [...]): Use ATTR_MATHFN_ERRNO instead of ATTR_MATHFN_FPROUNDING_ERRNO.

* builtins.def (BUILT_IN_NEXTAFTER, BUILT_IN_NEXTAFTERF,
BUILT_IN_NEXTAFTERL, BUILT_IN_NEXTTOWARD, BUILT_IN_NEXTTOWARDF,
BUILT_IN_NEXTTOWARDL): Use ATTR_MATHFN_ERRNO instead of
ATTR_MATHFN_FPROUNDING_ERRNO.

From-SVN: r255586

PR libstdc++/83395 fix invocable traits for INVOKE<cv void>

PR libstdc++/83395
* include/std/type_traits (__is_invocable_impl): Remove partial
specialization for INVOKE<void> and restore is_void<R> check in
primary template.
(__is_nt_invocable_impl): Likewise.
* testsuite/20_util/is_invocable/83395.cc: New test.
* testsuite/20_util/is_nothrow_invocable/83395.cc: New test.

From-SVN: r255584

Updated email in MAINTAINERS file.

From-SVN: r255583

go-lang.c (TARGET_AIX): Define if not defined.

* go-lang.c (TARGET_AIX): Define if not defined.
(go_langhook_init): Set nil_check_size_threshold to -1 on AIX.

Co-Authored-By: Ian Lance Taylor <iant@golang.org>
From-SVN: r255582

configure.ac: Remove logic adding gdb to noconfigsdirs for or1k.

2017-12-12  Stafford Horne  <shorne@gmail.com>

* configure.ac: Remove logic adding gdb to noconfigsdirs for or1k.
* configure: Regenerate.

From-SVN: r255581

Fix g++.old-deja/g++.pt/const2.C on Solaris

* lib/gcc-dg.exp (process-message): Avoid additional whitespace in
$expmsg.

From-SVN: r255580

re PR tree-optimization/83385 ([graphite] Wrong code w/ -O1 -floop-nest-optimize)

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

PR tree-optimization/83385
* graphite-scop-detection.c (get_order, order): Remove.
(bb_to_rpo): New global.
(cmp_pbbs): Adjust.
(build_scops): Sort pbbs in RPO order.

* gcc.dg/graphite/pr83385.c: New testcase.

From-SVN: r255579

[Patch combine] Don't create vector mode ZERO_EXTEND from subregs

The code in simplify set to handle transforming the paradoxical subreg
expression:

  (set FOO (subreg:M (mem:N BAR) 0))

in to:

  (set FOO (zero_extend:M (mem:N BAR)))

Does not consider the case where M is a vector mode, allowing it to
construct (for example):

  (zero_extend:V4SI (mem:SI))

For one, this has the wrong semantics - but fortunately we fail long
before then in expand_compound_operation.

We need to explicitly reject vector modes from this transformation.

gcc/

* combine.c (simplify_set): Do not transform subregs to zero_extends
if the destination mode is a vector mode.

From-SVN: r255578

re PR c/82679 (Uses of typedefs of arrays of _Atomic-qualified types are rejected)

PR c/82679
* c-decl.c (grokdeclarator): Check declspecs insted of atomicp.

* gcc.dg/c11-atomic-5.c: New test.

From-SVN: r255577

re PR rtl-optimization/83363 (wrong code at -O2 -fno-forward-propagate)

PR rtl-optimization/83363
PR rtl-optimization/83364
* gcc.dg/pr83363.c: New test.
* gcc.dg/pr83364.c: New test.

From-SVN: r255576

re PR tree-optimization/83359 (ICE in expand_LOOP_DIST_ALIAS, at internal-fn.c:2362)

PR tree-optimization/83359
* tree-cfg.h (fold_loop_internal_call): Declare.
* tree-vectorizer.c (fold_loop_internal_call): Moved to ...
* tree-cfg.c (fold_loop_internal_call): ... here.  No longer static.
(find_loop_dist_alias): New function.
(move_sese_region_to_fn): If any dloop->orig_loop_num value is
updated, also adjust any corresponding LOOP_DIST_ALIAS internal
calls.

* gcc.dg/graphite/pr83359.c: New test.

From-SVN: r255575

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

PR tree-optimization/80631
* tree-vect-loop.c (get_initial_def_for_reduction): Fix comment typo.
(vect_create_epilog_for_reduction): Add INDUC_VAL and INDUC_CODE
arguments, for INTEGER_INDUC_COND_REDUCTION use INDUC_VAL instead of
hardcoding zero as the value if COND_EXPR is never true.  For
INTEGER_INDUC_COND_REDUCTION don't emit the final COND_EXPR if
INDUC_VAL is equal to INITIAL_DEF, and use INDUC_CODE instead of
hardcoding MAX_EXPR as the reduction operation.
(is_nonwrapping_integer_induction): Allow negative step.
(vectorizable_reduction): Compute INDUC_VAL and INDUC_CODE for
vect_create_epilog_for_reduction, if no value is suitable, don't
use INTEGER_INDUC_COND_REDUCTION for now.  Formatting fixes.

* gcc.dg/vect/pr80631-1.c: New test.
* gcc.dg/vect/pr80631-2.c: New test.
* gcc.dg/vect/pr65947-13.c: Expect integer induc cond reduction
vectorization.

From-SVN: r255574

re PR middle-end/81889 (bogus warnings with -Wmaybe-uninitialized -O3)

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

PR tree-optimization/81889
* tree-ssa-loop-niter.c (infer_loop_bounds_from_signedness): Use
range info from the non-wrapping IV instead of just the range
of the type.

* gfortran.dg/pr81889.f90: New testcase.
* gcc.dg/tree-ssa/pr64183.c: Adjust.

From-SVN: r255573

Enable VAES support [2/5]

gcc/
* config.gcc: Add vaesintrin.h.
* config/i386/i386-builtin-types.def (V64QI_FTYPE_V64QI_V64QI): New type.
* config/i386/i386-builtin.def (__builtin_ia32_vaesdec_v16qi,
__builtin_ia32_vaesdec_v32qi, __builtin_ia32_vaesdec_v64qi): New builtins.
* config/i386/i386.c (ix86_expand_args_builtin): Handle new type.
* config/i386/immintrin.h: Include vaesintrin.h.
* config/i386/sse.md (vaesdec_<mode>): New pattern.
* config/i386/vaesintrin.h (_mm256_aesdec_epi128, _mm512_aesdec_epi128,
_mm_aesdec_epi128): New intrinsics.

gcc/testsuite/
* gcc.target/i386/avx512-check.h: Handle bit_VAES.
* gcc.target/i386/avx512f-aesdec-2.c: New test.
* gcc.target/i386/avx512fvl-vaes-1.c: Ditto.
* gcc.target/i386/avx512vl-aesdec-2.c: Ditto.
* gcc.target/i386/i386.exp (check_effective_target_avx512vaes): New.

From-SVN: r255572

Enable VAES support [1/5]

gcc/
* common/config/i386/i386-common.c (OPTION_MASK_ISA_VAES_SET,
OPTION_MASK_ISA_VAES_UNSET): New.
(ix86_handle_option): Handle -mvaes.
* config/i386/cpuid.h: Define bit_VAES.
* config/i386/driver-i386.c (host_detect_local_cpu): Detect -mvaes.
* config/i386/i386-c.c (__VAES__): New.
* config/i386/i386.c (ix86_target_string): Add -mvaes.
(ix86_valid_target_attribute_inner_p): Ditto.
* config/i386/i386.h (TARGET_VAES, TARGET_VAES_P): New.
* config/i386/i386.opt: Add -mvaes.
* doc/invoke.texi: Ditto.

From-SVN: r255571

[IEPM] Introduce debug hook for inline entry point markers

The inline_entry hook will be given a definition in a later patch.

for  gcc/ChangeLog

* debug.h (gcc_debug_hooks): Add inline_entry.
* dbxout.c (dbx_debug_hooks, xcoff_debug_hooks): Likewise.
* debug.c (do_nothing_debug_hooks): Likewise.
* vmsdbgout.c (vmsdbg_debug_hooks): Likewise.
* dwarf2out.c (dwarf2_debug_hooks): Likewise.
(dwarf2_lineno_debug_hooks): Likewise.

From-SVN: r255570

[SFN] Introduce -gstatement-frontiers option, enable debug markers

Introduce a command line option to enable statement frontiers, enabled
by default in optimized builds with DWARF2+ debug information.

This patch depends on an earlier patch that completed the
infrastructure for debug markers, and on another patch that turns -g
into a negatable option prefix.

for  gcc/ChangeLog

* common.opt (gstatement-frontiers): New, setting
debug_nonbind_markers_p.
* rtl.h (MAY_HAVE_DEBUG_MARKER_INSNS): Activate.
* toplev.c (process_options): Autodetect value for debug statement
frontiers option.
* tree.h (MAY_HAVE_DEBUG_MARKER_STMTS): Activate.
* doc/invoke.texi (gstatement-frontiers, gno-statement-frontiers): New.

From-SVN: r255569

[SFN] introduce statement frontier notes, still disabled

This patch completes the infrastructure for the introduction of
statement frontiers in C-family languages.

It brings in all the code remaining code needed to introduce and
transform begin stmt trees, gimple stmts, insns and notes, and
ultimately use them to generate the is_stmt column in DWARF2+ line
number tables/programs, however none of it is activated: the option
that would do so will be introduced in a subsequent patch.

This patch depends on an earlier patch with not-quite-boilerplate
changes towards SFN.

for  gcc/c-family/ChangeLog

* c-semantics.c (pop_stmt_list): Move begin stmt marker into
subsequent statement list.

for  gcc/c/ChangeLog

* c-objc-common.h (LANG_HOOKS_EMITS_BEGIN_STMT): Redefine as true.
* c-parser.c (add_debug_begin_stmt): New.
(c_parser_declaration_or_fndef): Call it.
(c_parser_compound_statement_nostart): Likewise.
(c_parser_statement_after_labels): Likewise.
* c-typeck (c_finish_stmt_expr): Skip begin stmts markers.

for  gcc/cp/ChangeLog

* constexpr.c (check_constexpr_ctor_body_1): Skip begin stmt
markers.
(constexpr_fn_retval): Likewise.
(potential_constant_expression_1): Likewise.
(cxx_eval_statement_list): Check that a begin stmt marker is
not used as the value of a statement list.
(cxx_eval_constant_expression): Return begin stmt markers
unchanged.
* cp-array-notation.c (stmt_location): New.
(cp_expand_cond_array_notations): Use it.
* cp-objcp-common.h (LANG_HOOKS_EMITS_BEGIN_STMT): Redefine as true.
* parser.c (add_debug_begin_stmt): New.
(cp_parser_statement): Call it.
* pt.c (tsubst_copy): Handle begin stmt markers.

for  gcc/ChangeLog

* cfgexpand.c (expand_gimple_basic_block): Handle begin stmt
markers.  Integrate source bind into debug stmt expand loop.
(pass_expand::execute): Check debug marker limit.  Avoid deep
TER and expand debug locations for debug bind insns only.
* cse.c (insn_live_p): Keep nonbind markers and debug bindings
followed by them.
* df-scan.c (df_insn_delete): Accept out-of-block debug insn.
* final.c (reemit_insn_block_notes): Take current block from
nonbind markers.  Declare note where it's first set.
(final_scan_insn): Handle begin stmt notes.  Emit is_stmt according to
begin stmt markers if enabled.
(notice_source_line): Handle nonbind markers.  Fail if their
location is unknown or that of builtins.
(rest_of_handle_final): Convert begin stmt markers to notes if
var-tracking didn't run.
(rest_of_clean_state): Skip begin stmt markers.
* gimple-pretty-print.c (dump_gimple_debug): Handle begin stmt
markers.
* function.c (allocate_struct_function): Set begin_stmt_markers.
* function.h (struct function): Add debug_marker_count counter
and debug_nonbind_markers flag.
* gimple-iterator.c (gsi_remove): Adjust debug_marker_count.
* gimple-low.c (lower_function_body): Adjust
debug_nonbind_markers.
(lower_stmt): Drop or skip gimple debug stmts.
(lower_try_catch): Skip debug stmts.
* gimple.c (gimple_build_debug_begin_stmt): New.
(gimple_copy): Increment debug_marker_count if copying one.
* gimple.h (gimple_build_debug_begin_stmt): Declare.
* gimplify.c (rexpr_location): New.
(rexpr_has_location): New.
(warn_switch_unreachable_r): Handle gimple debug stmts.
(shortcut_cond_r): Call expr_location.
(find_goto): New.
(find_goto_label): New.
(shortcut_cond_expr): Call expr_has_location, expr_location, and
find_goto_label.
(gimplify_cond_expr): Call find_goto_label, expr_has_location, and
expr_location.
(gimplify_expr): Handle begin stmt markers.  Reject debug expr decls.
* langhooks-def.h (LANG_HOOKS_EMITS_BEGIN_STMT): New.  Add to...
(LANG_HOOKS_INITIALIZER): ... this.
* langhooks.h (struct lang_hooks): Add emits_begin_stmt.
* lra-contraints.c (inherit_reload_reg): Tolerate between-blocks
debug insns.
(update_ebb_live_info): Skip debug insn markers.
* lra.c (debug_insn_static_data): Rename to...
(debug_bind_static_data): ... this.
(debug_marker_static_data): New.
(lra_set_insn_recog_data): Select one of the above depending
on debug insn kind.
(lra_update_isn_regno_info): Don't assume debug insns have
freqs.
(push_insns): Skip debug insns.
* lto-streamer-in.c (input_function): Drop debug stmts
depending on active options.  Adjust debug_nonbind_markers.
* params.def (PARAM_MAX_DEBUG_MARKER_COUNT): New.
* print-rtl.c (rtx_writer::print_rtx_operand_code_0): Handle
begin stmt marker notes.
(print_insn): Likewise.
* recog.c (extract_insn): Recognize rtl for debug markers.
* rtl.def (DEBUG_MARKER): New.
* tree-inline.c: Include params.h.
(remap_gimple_stmt): Handle nonbind markers.
(maybe_move_debug_stmts_to_successors): Likewise.
(copy_debug_stmt): Likewise.
* tree-iterator.c (append_to_statement_list_1): Append begin stmt
markers regardless of no side effects.
(tsi_link_before): Don't update container's side effects when adding
a begin stmt marker.
(tsi_link_after): Likewise.
(expr_first): Skip begin stmt markers.
(expr_last): Likewise.
* tree-pretty-print (dump_generic_node): Handle begin stmt markers.
* tree-ssa-threadedge.c (propagate_threaded_block_debug_info):
Disregard nonbind markers.
* tree.c (make_node_stat): Don't set side effects for begin stmt
markers.
(build1_stat): Likewise.
* tree.def (DEBUG_BEGIN_STMT): New.
* tree.h (GOTO_DESTINATION): Require a GOTO_EXPR.
* var-tracking.c (delete_debug_insns): Renamed to...
(delete_vta_debug_insns): ... this.
(reemit_marker_as_note): New.
(vt_initialize): Reemit markers.
(delete_vta_debug_insns): Likewise.
(vt_debug_insns_local): Reemit or delete markers.
(variable_tracking_main_1): Likewise.
* doc/generic.texi (DEBUG_BEGIN_STMT): Document.
* doc/gimple.texi (gimple_debug_begin_stmt_p): New.
(gimple_debug_nonbind_marker_p): New.
(gimple_build_debug_bind): Adjust.
(gimple_build_debug_begin_stmt): New.
* doc/invoke.texi (max-debug-marker-count): New param.
* doc/rtl.texi (debug_implicit_ptr, entry_value): New.
(debug_parameter_ref, debug_marker): New.
(NOTE_INSN_BEGIN_STMT): New.
(DEBUG_INSN): Describe begin stmt markers.

From-SVN: r255568

[SFN] stabilize find_bb_boundaries

If find_bb_boundaries is given a block with zero or one nondebug insn
beside debug insns, it shouldn't purge dead edges, because without
debug insns we wouldn't purge them at that point.  Doing so may change
the order in which edges are processed, and ultimately lead to
different transformations to the CFG and then to different
optimizations.

We shouldn't, however, retain debug insns after control flow insns, so
if we find debug insns after a single insn that happens to be a
control flow insn, do the debug insn cleanups, but still refrain from
purging dead edges at that point.

for  gcc/ChangeLog

* cfgbuild.c (find_bb_boundaries): Don't purge dead edges if,
without debug insns, we wouldn't, but clean up debug insns
after a control flow insn nevertheless.

From-SVN: r255567

[SFN] not-quite-boilerplate changes in preparation to introduce nonbind markers

This patch adjusts numerous parts of the compiler that would
malfunction should they find debug markers at points where they may be
introduced.  The changes purport to allow the compiler to pass
bootstrap-debug-lean (-fcompare-debug in stage3) at various
optimization levels, as well as bootstrap-debug-lib (-fcompare-debug
for target libraries), even after the compiler is changed so that
debug markers are introduced in code streams at spots where earlier
debug stmts, insns and notes wouldn't normally appear.

This patch depends on an earlier SFN boilerplate patch, and on another
SFN patch that introduces new RTL insn-walking functions.

for  gcc/ChangeLog

* cfgcleanup.c (delete_unreachable_blocks): Use alternate
block removal order if MAY_HAVE_DEBUG_BIND_INSNS.
* cfgexpand.c (label_rtx_for_bb): Skip debug insns.
* cfgrtl.c (try_redirect_by_replacing_jump): Skip debug insns.
(rtl_tidy_fallthru_edge): Likewise.
(rtl_verify_fallthru): Likewise.
(rtl_verify_bb_layout): Likewise.
(skip_insns_after_block): Likewise.
(duplicate_insn_chain): Use DEBUG_BIND_INSN_P.
* dwarf2out.c: Include print-rtl.h.
(dwarf2out_next_real_insn): New.
(dwarf2out_var_location): Call it.  Disregard begin stmt markers.
Dump debug binds in asm comments.
* gimple-iterator.c (gimple_find_edge_insert_loc): Skip debug stmts.
* gimple-iterator.h (gsi_start_bb_nondebug): Remove; adjust
callers to use gsi_start_nondebug_bb instead.
(gsi_after_labels): Skip gimple debug stmts.
(gsi_start_nondebug): New.
* gimple-loop-interchange.c (find_deps_in_bb_for_stmt): Adjust.
(proper_loop_form_for_interchange): Adjust.
* gimple-low.c (gimple_seq_may_fallthru): Take last nondebug stmt.
* gimple.h (gimple_seq_last_nondebug_stmt): New.
* gimplify.c (last_stmt_in_scope): Skip debug stmts.
(collect_fallthrough_labels): Likewise.
(should_warn_for_implicit_fallthrough): Likewise.
(warn_implicit_fallthrough_r): Likewise.
(expand_FALLTHROUGH_r): Likewise.
* graphite-isl-ast-to-gimple.c (gsi_insert_earliest): Adjust.
(graphite_copy_stmts_from_block): Skip nonbind markers.
* haifa-sched.c (sched_extend_bb): Skip debug insns.
* ipa-icf-gimple.c (func_checker::compare_bb): Adjust.
* jump.c (clean_barriers): Skip debug insns.
* omp-expand.c (expand_parallel_call): Skip debug insns.
(expand_task_call): Likewise.
(remove_exit_barrier): Likewise.
(expand_omp_taskreg): Likewise.
(expand_omp_for_init_counts): Likewise.
(expand_omp_for_generic): Likewise.
(expand_omp_for_static_nochunk): Likewise.
(expand_omp_for_static_chunk): Likewise.
(expand_omp_simd): Likewise.
(expand_omp_taskloop_for_outer): Likewise.
(expand_omp_taskloop_for_inner): Likewise.
(expand_oacc_for): Likewise.
(expand_omp_sections): Likewise.
(expand_omp_single): Likewise.
(expand_omp_synch): Likewise.
(expand_omp_atomic_load): Likewise.
(expand_omp_atomic_store): Likewise.
(expand_omp_atomic_fetch_op): Likewise.
(expand_omp_atomic_pipeline): Likewise.
(expand_omp_atomic_mutex): Likewise.
(expand_omp_target): Likewise.
(grid_expand_omp_for_loop): Likewise.
(grid_expand_target_grid_body): Likewise.
(build_omp_regions_1): Likewise.
* omp-low.c (check_combined_parallel): Skip debug stmts.
* postreload.c (fixup_debug_insns): Skip nonbind debug insns.
* regcprop.c (find_oldest_value_reg): Ensure REGNO is not a pseudo.
* sese.c (sese_trivially_empty_bb_p): Call is_gimple_debug in
test.
* tree-cfg.c (make_blobs_1): Skip debug stmts.
(make_edges): Likewise.
(cleanup_dead_labels): Likewise.
(gimple_can_merge_blocks_p): Likewise.
(stmt_starts_bb_p): Likewise.
(gimple_block_label): Likewise.
(gimple_redirect_edge_and_branch): Likewise.
* tree-cfgcleanup.c (remove_forwarder_block): Rearrange skipping
of debug stmts.
(execute_cleanup_cfg_post_optimizing): Dump enumerated decls with
TDF_SLIM.
* tree-pretty-print (print_declaration): Omit initializer in slim
dumps.
* tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Mark begin stmt
markers.
(eliminate_unnecessary_stmts): Stabilize block removal order.
* tree-ssa-tail-merge.c (find_duplicate): Skip debug stmts.
* var-tracking.c (get_first_insn): New.
(vt_emit_notes): Call it.
(vt_initialize): Walk any insns before the first BB.
(delete_debug_insns): Likewise.

From-SVN: r255566

[SFN] boilerplate changes in preparation to introduce nonbind markers

This patch introduces a number of new macros and functions that will
be used to distinguish between different kinds of debug stmts, insns
and notes, namely, preexisting debug bind ones and to-be-introduced
nonbind markers.

In a seemingly mechanical way, it adjusts several uses of the macros
and functions, so that they refer to narrower categories when
appropriate.

These changes, by themselves, should not have any visible effect in
the compiler behavior, since the upcoming debug markers are never
created with this patch alone.

for  gcc/ChangeLog

* gimple.h (enum gimple_debug_subcode): Add
GIMPLE_DEBUG_BEGIN_STMT.
(gimple_debug_begin_stmt_p): New.
(gimple_debug_nonbind_marker_p): New.
* tree.h (MAY_HAVE_DEBUG_MARKER_STMTS): New.
(MAY_HAVE_DEBUG_BIND_STMTS): Renamed from....
(MAY_HAVE_DEBUG_STMTS): ... this.  Check both.
* insn-notes.def (BEGIN_STMT): New.
* rtl.h (MAY_HAVE_DEBUG_MARKER_INSNS): New.
(MAY_HAVE_DEBUG_BIND_INSNS): Renamed from....
(MAY_HAVE_DEBUG_INSNS): ... this.  Check both.
(NOTE_MARKER_LOCATION, NOTE_MARKER_P): New.
(DEBUG_BIND_INSN_P, DEBUG_MARKER_INSN_P): New.
(INSN_DEBUG_MARKER_KIND): New.
(GEN_RTX_DEBUG_MARKER_BEGIN_STMT_PAT): New.
(INSN_VAR_LOCATION): Check for VAR_LOCATION.
(INSN_VAR_LOCATION_PTR): New.
* cfgexpand.c (expand_debug_locations): Handle debug bind insns
only.
(expand_gimple_basic_block): Likewise.  Emit debug temps for TER
deps only if debug bind insns are enabled.
(pass_expand::execute): Avoid deep TER and expand
debug locations for debug bind insns only.
* cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Narrow
debug stmts special handling down to debug bind stmts.
* combine.c (try_combine): Narrow debug insns special handling
down to debug bind insns.
* cse.c (delete_trivially_dead_insns): Handle debug bindings.
Narrow debug insns preexisting special handling down to debug
bind insns.
* dce.c (rest_of_handle_ud_dce): Narrow debug insns special
handling down to debug bind insns.
* function.c (instantiate_virtual_regs): Skip debug markers,
adjust handling of debug binds.
* gimple-ssa-backprop.c (backprop::prepare_change): Try debug
temp insertion iff MAY_HAVE_DEBUG_BIND_STMTS.
* haifa-sched.c (schedule_insn): Narrow special handling of debug
insns to debug bind insns.
* ipa-param-manipulation.c (ipa_modify_call_arguments): Narrow
special handling of debug stmts to debug bind stmts.
* ipa-split.c (split_function): Likewise.
* ira.c (combine_and_move_insns): Adjust debug bind insns only.
* loop-unroll.c (apply_opt_in_copies): Adjust tests on bind
debug insns.
* reg-stack.c (convert_regs_1): Use DEBUG_BIND_INSN_P.
* regrename.c (build_def_use): Likewise.
* regcprop.c (copyprop_hardreg_forward_1): Likewise.
(pass_cprop_hardreg): Narrow special casing of debug insns to
debug bind insns.
* regstat.c (regstat_init_n_sets_and_refs): Likewise.
* reload1.c (reload): Likewise.
* sese.c (sese_insert_phis_for_liveouts): Narrow special
casing of debug stmts to debug bind stmts.
* shrink-wrap.c (move_insn_for_shrink_wrap): Likewise.
* ssa-iterators.h (num_imm_uses): Likewise.
* tree-cfg.c (gimple_merge_blocks): Narrow special casing of
debug stmts to debug bind stmts.
* tree-inline.c (tree_function_versioning): Narrow special casing
of debug stmts to debug bind stmts.
* tree-loop-distribution.c (generate_loops_for_partition):
Narrow special casing of debug stmts to debug bind stmts.
* tree-sra.c (analyze_access_subtree): Narrow special casing
of debug stmts to debug bind stmts.
* tree-ssa-dce.c (remove_dead_stmt): Narrow special casing of debug
stmts to debug bind stmts.
* tree-ssa-loop-ivopt.c (remove_unused_ivs): Narrow special
casing of debug stmts to debug bind stmts.
* tree-ssa-reassoc.c (reassoc_remove_stmt): Likewise.
* tree-ssa-tail-merge.c (tail_merge_optimize): Narrow special
casing of debug stmts to debug bind stmts.
* tree-ssa-threadedge.c (propagate_threaded_block_debug_info):
Likewise.
* tree-ssa.c (flush_pending_stmts): Narrow special casing of
debug stmts to debug bind stmts.
(gimple_replace_ssa_lhs): Likewise.
(insert_debug_temp_for_var_def): Likewise.
(insert_debug_temps_for_defs): Likewise.
(reset_debug_uses): Likewise.
* tree-ssanames.c (release_ssa_name_fn): Likewise.
* tree-vect-loop-manip.c (adjust_debug_stmts_now): Likewise.
(adjust_debug_stmts): Likewise.
(adjust_phi_and_debug_stmts): Likewise.
(vect_do_peeling): Likewise.
* tree-vect-loop.c (vect_transform_loop): Likewise.
* valtrack.c (propagate_for_debug): Use BIND_DEBUG_INSN_P.
* var-tracking.c (adjust_mems): Narrow special casing of debug
insns to debug bind insns.
(dv_onepart_p, dataflow_set_clar_at_call, use_type): Likewise.
(compute_bb_dataflow, vt_find_locations): Likewise.
(vt_expand_loc, emit_notes_for_changes): Likewise.
(vt_init_cfa_base): Likewise.
(vt_emit_notes): Likewise.
(vt_initialize): Likewise.
(vt_finalize): Likewise.

From-SVN: r255565

[SFN] adjust RTL insn-walking API

This patch removes unused RTL functions, introduces alternate ones for
use in a later SFN patch, and regroups other related functions so that
they appear in a more consistent order.

for  gcc/ChangeLog

* emit-rtl.c (next_nondebug_insn, prev_nondebug_insn): Reorder.
(next_nonnote_nondebug_insn, prev_nonnote_nondebug_insn): Reorder.
(next_nonnote_nondebug_insn_bb): New.
(prev_nonnote_nondebug_insn_bb): New.
(prev_nonnote_insn_bb, next_nonnote_insn_bb): Remove.
* rtl.h (prev_nonnote_insn_bb, next_nonnote_insn_bb): Remove decls.
(prev_nonnote_nondebug_insn_bb): Declare.
(next_nonnote_nondebug_insn_bb): Declare.
* cfgbuild.c (find_bb_boundaries): Adjust to skip debug insns.
* cfgrtl.c (get_last_bb_insn): Likewise.
* lra.c (push_insns): Likewise.

From-SVN: r255564

Expensive selftests: torture testing for fix-it boundary conditions (PR c/82050)

This patch adds selftest coverage for the fix for PR c/82050.

The selftest iterates over various "interesting" column and line-width
values to try to shake out bugs in the fix-it printing routines, a kind
of "torture" selftest.

Unfortunately this selftest is noticably slower than the other selftests;
adding it to diagnostic-show-locus.c led to:
  -fself-test: 40218 pass(es) in 0.172000 seconds
slowing down to:
  -fself-test: 97315 pass(es) in 6.109000 seconds
for an unoptimized build (e.g. when hacking with --disable-bootstrap).

Given that this affects the compile-edit-test cycle of the "gcc"
subdirectory, this felt like an unacceptable amount of overhead to add.

I attempted to optimize the test by reducing the amount of coverage, but
the test seems useful, and there seems to be a valid role for "torture"
selftests.

Hence this patch adds a:
  gcc.dg/plugin/expensive_selftests_plugin.c
with the responsibility for running "expensive" selftests, and adds the
expensive test there.  The patch moves a small amount of code from
selftest::run_tests into a helper class so that the plugin can print
a useful summary line (to reassure us that the tests are actually being
run).

With that, the compile-edit-test cycle of the "gcc" subdir is unaffected;
the plugin takes:
  expensive_selftests_plugin: 26641 pass(es) in 3.127000 seconds
which seems reasonable within the much longer time taken by "make check"
(I optimized some of the overhead away, hence the reduction from 6 seconds
above down to 3 seconds).

gcc/ChangeLog:
PR c/82050
* selftest-run-tests.c (selftest::run_tests): Move start/finish code
to...
* selftest.c (selftest::test_runner::test_runner): New ctor.
(selftest::test_runner::~test_runner): New dtor.
* selftest.h (class selftest::test_runner): New class.

gcc/testsuite/ChangeLog:
PR c/82050
* gcc.dg/plugin/expensive-selftests-1.c: New file.
* gcc.dg/plugin/expensive_selftests_plugin.c: New file.
* gcc.dg/plugin/plugin.exp (plugin_test_list): Add the above.

From-SVN: r255563

Daily bump.

From-SVN: r255562

Fix the new pr83361.c testcase

gcc/testsuite/
* gcc.dg/pr83361.c: Add -Wno-div-by-zero to dg-options.

From-SVN: r255559

Remove a stray line.

From-SVN: r255558

PR c/81544 - attribute noreturn and warn_unused_result on the same function

PR c/81544 - attribute noreturn and warn_unused_result on the same function
accepted

Add tests missed in r255469.

From-SVN: r255557

altivec-12.c (main): Add tests for vec_avg.

gcc/testsuite/ChangeLog:

2017-12-07  Carl Love  <cel@us.ibm.com>

* gcc.target/powerpc/altivec-12.c (main):  Add tests for vec_avg.
* gcc.target/powerpc/builtins-1.c: Renamed builtins-1.h.
* gcc.target/powerpc/builtins-1.h: Add tests for vec_and,
vec_andc, vec_cpsgn, vec_mergeh,vec_mergel, vec_nor, vec_or,
vec_sel, vec_xor, vec_all_eq, vec_all_ge, vec_all_gt, vec_all_le,
vec_all_lt, vec_all_ne, vec_any_eq,vec_any_ge, vec_any_gt,vec_any_le,
vec_any_lt, vec_any_ne, vec_cntlz, vec_round, vec_rsqrt, vec_rsqrte,
vec_splat, vec_div, vec_mul,vec_ctf, vec_cts, vec_ctu.
* gcc.target/powerpc/builtins-1-be.c: New file, big endian
test file for tests in builtins-1.h.
* gcc.target/powerpc/builtins-1-le.c: New file, little endian
test file for tests in builtins-1.h.
* gcc.target/powerpc/builtins-2.c (main): Add tests for vec_ctf,
vec_cts, vec_ctu.
* gcc.target/powerpc/builtins-3-runnable.c (test_result_sp): Fix printf
conversion character.
  * gcc.target/powerpc/fold-vec-minmax-char.c (test1_min, test2_min,
test4_min, test1_max, test2_max): Add tests for vec_max and vec_min.
  * gcc.target/powerpc/fold-vec-minmax-int.c (test4_min, test5_min,
test7_min, test8_min): Add tests for vec_min.
  * gcc.target/powerpc/fold-vec-minmax-longlong.c (test7_max, test8_max,
test9_max, test10_max): Add tests for and vec_max.
  * gcc.target/powerpc/fold-vec-minmax-short.c (test4_min, test5_min,
tes7_min, test8_min): Add tests for and vec_min.
  * gcc.target/powerpc/p8vector-builtin-8.c (foo): Add test for vec_adds,
vec_subs, vec_sum2s, vec_sum4s.
  * gcc.target/powerpc/vsx-splats-p8.c: New file to test vec_splats.
* gcc.target/vsx-vector-5.c (arg1_tests_d, arg1_tests_d): Add
tests for vec_ceil, vec_floor, vec_nearbyint, vec_sqrt.
* gcc.target/vsx-vector-7.c (foo): Add tests for vec_sel.
* gcc.target/vsx-vector-abss.c: New file to test vec_abss.

From-SVN: r255556

altivec.h (vec_extract_fp32_from_shorth, [...]): Add #defines.

gcc/ChangeLog:

2017-12-11  Carl Love  <cel@us.ibm.com>

* config/rs6000/altivec.h (vec_extract_fp32_from_shorth,
vec_extract_fp32_from_shortl]): Add #defines.
* config/rs6000/rs6000-builtin.def (VSLDOI_2DI): Add macro expansion.
* config/rs6000/rs6000-c.c (ALTIVEC_BUILTIN_VEC_UNPACKH,
ALTIVEC_BUILTIN_VEC_UNPACKL, ALTIVEC_BUILTIN_VEC_AND,
ALTIVEC_BUILTIN_VEC_SLD, ALTIVEC_BUILTIN_VEC_SRL,
ALTIVEC_BUILTIN_VEC_SRO, ALTIVEC_BUILTIN_VEC_SLD,
ALTIVEC_BUILTIN_VEC_SLL): Add expansions.
* doc/extend.texi: Add documentation for the added builtins.

gcc/testsuite/ChangeLog:

2017-12-11  Carl Love  <cel@us.ibm.com>
* gcc.target/powerpc/altivec-7.c: Renamed altivec-7.h.
* gcc.target/powerpc/altivec-7.h (main): Add testcases for vec_unpackl.
Add dg-final tests for the instructions generated.
* gcc.target/powerpc/altivec-7-be.c: New file to test on big endian.
* gcc.target/powerpc/altivec-7-le.c: New file to test on little endian.
* gcc.target/powerpc/altivec-13.c (foo): Add vec_sld, vec_srl,
 vec_sro testcases. Add dg-final tests for the instructions generated.
* gcc.target/powerpc/builtins-3-p8.c (test_vsi_packs_vui,
test_vsi_packs_vsi, test_vsi_packs_vssi, test_vsi_packs_vusi,
test_vsi_packsu-vssi, test_vsi_packsu-vusi, test_vsi_packsu-vsll,
test_vsi_packsu-vull, test_vsi_packsu-vsi, test_vsi_packsu-vui): Add
testcases. Add dg-final tests for new instructions.
* gcc.target/powerpc/p8vector-builtin-2.c (vbschar_eq, vbchar_eq,
vuchar_eq, vbint_eq, vsint_eq, viint_eq, vuint_eq, vbool_eq, vbint_ne,
vsint_ne, vuint_ne, vbool_ne, vsign_ne, vuns_ne, vbshort_ne): Add
tests.
Add dg-final instruction tests.
* gcc.target/powerpc/vsx-vector-6.c: Renamed vsx-vector-6.h.
* gcc.target/powerpc/vsx-vector-6.h (vec_andc,vec_nmsub, vec_nmadd,
vec_or, vec_nor, vec_andc, vec_or, vec_andc, vec_msums): Add tests.
Add dg-final tests for the generated instructions.
* gcc.target/powerpc/builtins-3.c (test_sll_vsc_vsc_vsuc,
test_sll_vuc_vuc, test_sll_vsi_vsi_vuc, test_sll_vui_vui_vuc,
test_sll_vbll_vull, test_sll_vbll_vbll_vus, test_sll_vp_vp_vuc,
test_sll_vssi_vssi_vuc, test_sll_vusi_vusi_vuc, test_slo_vsc_vsc_vsc,
test_slo_vuc_vuc_vsc, test_slo_vsi_vsi_vsc, test_slo_vsi_vsi_vuc,
test_slo_vui_vui_vsc, test_slo_vui_vui_vuc, test_slo_vsll_slo_vsll_vsc,
test_slo_vsll_slo_vsll_vuc, test_slo_vull_slo_vull_vsc,
test_slo_vull_slo_vull_vuc, test_slo_vp_vp_vsc, test_slo_vp_vp_vuc,
test_slo_vssi_vssi_vsc, test_slo_vssi_vssi_vuc, test_slo_vusi_vusi_vsc,
test_slo_vusi_vusi_vuc, test_slo_vusi_vusi_vuc, test_slo_vf_vf_vsc,
test_slo_vf_vf_vuc, test_cmpb_float): Add tests.

From-SVN: r255555

[PR80693] drop value of parallel SETs dropped by combine

When combine drops a REG_UNUSED SET in a parallel, we have to clear
cached values, so that, even if the REGs remain used (e.g. because
they were referenced in the used SET_SRC), we will not use properties
of the dropped modified value as if they applied to the preserved
original one.

We fail to adjust REG_N_SETS.

for  gcc/ChangeLog

PR rtl-optimization/80693
PR rtl-optimization/81019
PR rtl-optimization/81020
* combine.c (distribute_notes): Reset any REG_UNUSED REGs that
are not mentioned in i3.  Place the REG_UNUSED note on i2,
possibly modified to REG_DEAD, if it did not originate in i3.

for  gcc/testsuite/ChangeLog

PR rtl-optimization/80693
PR rtl-optimization/81019
PR rtl-optimization/81020
* gcc.dg/pr80693.c: New.
* gcc.dg/pr81019.c: New.

From-SVN: r255554

recog.c (store_data_bypass_p_1): New function.

* recog.c (store_data_bypass_p_1): New function.
(store_data_bypass_p): Handle USE in a PARALLEL like CLOBBER.  Use
store_data_bypass_p_1 to avoid code duplication.  Formatting fixes.

From-SVN: r255553

ifcvt: Call fixup_partitions (PR83361)

After converting a conditional branch to an unconditional trap to a
conditional trap, if the original trap is still reachable from another
path, it may be that it is in a hot basic block  and only reachable from
cold blocks.  Fix that up.

PR rtl-optimization/83361
* ifcvt.c (if_convert): Call fixup_partitions.

gcc/testsuite/
PR rtl-optimization/83361
* gcc.dg/pr83361.c: New testcase.

From-SVN: r255550

rs6000.c (rs6000_gimple_fold_builtin): Add support for early folding of splat_u{8,16,32}.

[gcc]

2017-12-11  Will Schmidt  <will_schmidt@vnet.ibm.com>

* config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add support for
early folding of splat_u{8,16,32}.

[testsuite]

2017-12-11  Will Schmidt  <will_schmidt@vnet.ibm.com>

* gcc.target/powerpc/fold-vec-splat-misc-invalid.c: New.

From-SVN: r255549

intrinsic.texi (MAXLOC): Update documentation for character arrays and KIND argument.

2017-12-11  Thomas Koenig  <tkoenig@gcc.gnu.org>

* intrinsic.texi (MAXLOC): Update documentation for
character arrays and KIND argument.
(MINLOC): Likewise.

From-SVN: r255546

aarch64.c (aarch64_print_operand): Don't start output_operand_lossage first argument with capital letter.

* config/aarch64/aarch64.c (aarch64_print_operand): Don't start
output_operand_lossage first argument with capital letter.
(aarch64_override_options): Don't start error and sorry first argument
with capital letter.

From-SVN: r255544

re PR fortran/83379 (Intermittent failure of min/maxval_char*.f90 test cases)

2017-12-11  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83379
* gfortran.dg/minval_char_1.f90: Correct logic for mask
calculation.
* gfortran.dg/minval_char_2.f90: Likewise.
* gfortran.dg/maxval_char_1.f90: Likewise.
* gfortran.dg/maxval_char_2.f90: Likewise.

From-SVN: r255543

re PR fortran/45689 ([F03] Missing transformational intrinsic in the trans_func_f2003 list)

2017-12-11  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/45689
* simplify.c (min_max_choose): Add prototype.
(gfc_count): Format correctly.
(simplify_transformation): Pass array argument to init_result_expr.
(gfc_simplify_minval_maxval): Remove.
(gfc_min): New function.
(gfc_simplify_minval): Call simplify_transformation.
(gfc_max): New function.
(gfc_simplify_maxval): Call simplify_transformation.

2017-12-11  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/45689
* gfortran.dg/minval_parameter_1.f90: New test.
* gfortran.dg/maxval_parameter_1.f90: New test.

From-SVN: r255542

Fix stack overflow with autofdo (PR83355)

g++.dg/bprob* is failing currently with autofdo.

Running in gdb shows that there is a very deep recursion in get_index_by_decl until it
overflows the stack.

gcc/:
2017-12-11  Andi Kleen  <ak@linux.intel.com>

PR gcov-profile/83355
* auto-profile.c (string_table::get_index_by_decl): Don't
recurse when abstract origin points to itself.

From-SVN: r255540

re PR tree-optimization/83320 (Mismatched free() / delete / delete []  in gimple-loop-interchange.cc:948)

PR tree-optimization/83320
* gimple-loop-interchange.cc (free_data_refs_with_aux): Use delete.
(prune_datarefs_not_in_loop): Ditto.

From-SVN: r255539

Daily bump.

From-SVN: r255537

* ChangeLog: Fix whitespace some more.

From-SVN: r255534

ChangeLog: Fix whitespace.

* ChangeLog: Fix whitespace.
* testsuite/ChangeLog: Ditto.

From-SVN: r255533

re PR fortran/53478 (gfortran segfaults when module name clashes with C binding name of procedure)

2017-12-10  Dominique d'Humieres  <dominiq@lps.ens.fr>

PR fortran/53478
* gfortran.h (gfc_find_case_gsymbol): New prototype.
* symbol.c (gfc_find_case_gsymbol): New procedure, case
insensistive version of gfc_find_gsymbol.
* resolve.c (resolve_common_blocks): Use it.
Replace %s with %qs where needed.

* gfortran.dg/binding_label_tests_4.f03: Update dg-error.
* gfortran.dg/binding_label_tests_6.f03: Likewise.
* gfortran.dg/binding_label_tests_7.f03: Likewise.
* gfortran.dg/binding_label_tests_8.f03: Likewise.
* gfortran.dg/binding_label_tests_10_main.f03: Likewise.
* gfortran.dg/binding_label_tests_11_main.f03: Likewise.
* gfortran.dg/binding_label_tests_13_main.f03: Likewise.
* gfortran.dg/test_common_binding_labels_3_main.f03: Likewise.
* gfortran.dg/binding_label_tests_29.f90: New test.

From-SVN: r255530

* doc/install.texi (Specific): Tweak link to mkssoftware.com.

From-SVN: r255529

re PR tree-optimization/83337 (ICE at -O3 x86_64-linux-gnu: in interpret_rhs_expr, at tree-scalar-evolution.c:1775)

PR tree-optimization/83337
* gimple-loop-interchange.cc (compute_access_stride): Handle bitfield DRs
properly.

* gcc.dg/tree-ssa/loop-interchange-14.c: New test.
* gcc.dg/tree-ssa/loop-interchange-15.c: New test.

From-SVN: r255528

Daily bump.

From-SVN: r255527

re PR fortran/82934 (Segfault on assumed character length in allocate)

2017-12-09  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/82934
PR fortran/83318
* match.c (gfc_match_allocate): Enforce F2008:C631.

2017-12-09  Steven G. Kargl  <kargl@gcc.gnu.org>

PR fortran/82934
PR fortran/83318
* gfortran.dg/allocate_assumed_charlen_2.f90: new test.

From-SVN: r255524

re PR lto/83338 (SPEC CPU2017 510.parest_r ICE)

PR tree-optimization/83338
* tree-vect-stmts.c (vectorizable_operation): Handle POINTER_DIFF_EXPR
vectorization as MINUS_EXPR with a subsequent VIEW_CONVERT_EXPR from
vector of unsigned integers to vector of signed integers.

* gcc.dg/vect/pr83338.c: New test.

From-SVN: r255523

re PR fortran/83316 (ICE: minval/maxval and characters)

2017-12-09  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83316
* arith.c (gfc_character2character): New function.
* arith.h: Add prototype.
* simplify.c (gfc_convert_constant): Handle character type.

2017-12-09  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83316
* gfortran.dg/minval_char_5.f90: New test.

From-SVN: r255522

Add .type and .size directives to riscv libgcc functions.

libgcc/
* config/riscv/div.S: Use FUNC_* macros.
* config/riscv/muldi3.S, config/riscv/multi3.S: Likewise
* config/riscv/save-restore.S: Likewise.
* config/riscv/riscv-asm.h: New.

From-SVN: r255521

Daily bump.

From-SVN: r255520

[multiple changes]

2017-12-08  Vladimir Makarov  <vmakarov@redhat.com>

PR rtl-optimization/83317
* lra-constraints.c (process_address_1): Add insn code check.

2017-12-08  Jakub Jelinek  <jakub@redhat.com>

PR rtl-optimization/83317
* gcc.target/i386/pr83317.c: New test.

From-SVN: r255517

re PR rtl-optimization/81595 (ICE: verify_flow_info failed (error: multiple hot/cold transitions found))

PR rtl-optimization/81595
* gcc.c-torture/compile/pr81595.c: New test.

From-SVN: r255516

re PR tree-optimization/83323 (186.crafty miscompares)

Fix PR83323

* gimple-loop-jam (unroll_jam_possible_p): Correct test for
head-controlled loops and loop BBs.
* common.opt (funroll-and-jam): Remove, instead ...
(floop-unroll-and-jam): ... reuse this option.
* opts.c (default_options_table): Use OPT_floop_unroll_and_jam.
* doc/invoke.texi (-funroll-and-jam): Move docu to ...
(-floop-unroll-and-jam): ... this option.

testsuite/
* gcc.dg/pr83323.c: New test.
* gcc.dg/unroll-and-jam.c: Use -floop-unroll-and-jam.

From-SVN: r255514

ipa-polymorphic-call.c (noncall_stmt_may_be_vtbl_ptr_store): Fix a comment typo, get_base_ref_and_offset -> get_ref_base_and_extent.

* ipa-polymorphic-call.c (noncall_stmt_may_be_vtbl_ptr_store): Fix
a comment typo, get_base_ref_and_offset -> get_ref_base_and_extent.
* ipa-prop.c (stmt_may_be_vtbl_ptr_store): Likewise.

From-SVN: r255513

re PR middle-end/81782 (Yet another -Wmaybe-uninitialized false positive with empty array)

2017-12-08  Richard Biener  <rguenther@suse.de>

PR middle-end/81782
* tree-ssa-uninit.c (warn_uninitialized_vars): Properly
handle accesses outside of zero-sized vars.

* gcc.dg/uninit-pr81782.c: New testcase.

From-SVN: r255512

Disable hwcap on Solaris in g++.dg/opt/pr83252.C

* g++.dg/opt/pr83252.C: Add -mclear-hwcap on *-*-solaris*.

From-SVN: r255511

Prevent SRA from removing type changing assignment

2017-12-08  Martin Jambor  <mjambor@suse.cz>

PR tree-optimization/83141
* tree-sra.c (contains_vce_or_bfcref_p): Move up in the file, also
test for MEM_REFs implicitely changing types with padding.  Remove
inline keyword.
(build_accesses_from_assign): Added contains_vce_or_bfcref_p checks.

testsuite/
* gcc.dg/tree-ssa/pr83141.c: New test.
* gcc.dg/guality/pr54970.c: XFAIL tests querying a[0].

From-SVN: r255510

arc.c (arc_attribute_table): Add exclusions to the comment.

* config/arc/arc.c (arc_attribute_table): Add exclusions to
the comment.
* config/avr/avr.c (avr_attribute_table): Likewise.
* config/msp430/msp430.c (msp430_attribute_table): Likewise.
* config/rl78/rl78.c (rl78_attribute_table): Likewise.
* config/nds32/nds32.c (nds32_attribute_table): Likewise.
* config/darwin.h (SUBTARGET_ATTRIBUTE_TABLE): Initialize new member
of struct attribute_spec.
* config/i386/cygming.h (SUBTARGET_ATTRIBUTE_TABLE): Likewise.
ada/
* gcc-interface/utils.c (gnat_internal_attribute_table): Add
exclusions to the comment.
brig/
* brig-lang.c (brig_attribute_table): Fix up comment.

From-SVN: r255509

re PR target/82960 (spu_machine_dependent_reorg does not handle jump_table_data insn)

gcc/
PR target/82960
* config/spu/spu.c (pad_bb): Only check INSN_CODE when INSN_P is true.

From-SVN: r255508

re PR tree-optimization/83609 (ICE in read_complex_part at gcc/expr.c:3202)

* profile-count.c (profile_count::from_gcov_type): Move from
profile-count.h; handle overflow.
* profile-count. (profile_count::from_gcov_type): Move offline.

PR middle-end/83609
* gcc.c-torture/compile/pr83069.c: New testcase.

From-SVN: r255507

combine: Fix PR83304

In PR83304 two insns are combined, where the I2 uses a register that
has a REG_DEAD note on an insn after I2 but before I3.  In such a case
move_deaths should move that death note.  But move_deaths only looks
at the reg_stat[regno].last_death insn, and that field can be zeroed
out (previously, use_crosses_set_p would prevent the combination in
this case).

If the last_death field is zero it means "unknown", not "no death", so
we have to find if there is a REG_DEAD note.

PR rtl-optimization/83304
* combine.c (move_deaths): If we do not know where a register died,
search for it.

From-SVN: r255506

gimple-loop-interchange.cc (tree_loop_interchange::interchange): Provide -fopt-info-loop feedback when we interchange in a nest.

2017-12-08  Richard Biener  <rguenther@suse.de>

* gimple-loop-interchange.cc (tree_loop_interchange::interchange):
Provide -fopt-info-loop feedback when we interchange in a nest.

From-SVN: r255505

[arm] PR target/83206: Make native driver select fp-capable armv6 cores

A quirk in the historical naming of some ARMv6 products means that the
main CPU name implies the presence or otherwise of the floating point unit.
This causes problems when using -mfpu=auto with -mcpu=native: the driver is
picking a CPU that does not support a floating-point unit, even though
one may well exist.

This patch addresses this by selecting the FP-capable names so that FP
instructions will be generated if the other options suggest this is
permitted.

Note that a more complete fix is really needed here to look up the
FP/simd capabilities and append the appropriate capability extensions.
This will be the subject of some follow-up patches.

* config/arm/driver-arm.c (arm_cpu_table): Use fp-capable product names
for armv6 ARM CPU IDs.

From-SVN: r255504