Support extended aggregate jump function in ipa-cp

2019-11-14  Feng Xue  <fxue@os.amperecomputing.com>

PR ipa/91682
* ipa-prop.h (jump_func_type): New value IPA_JF_LOAD_AGG.
(ipa_load_agg_data, ipa_agg_value, ipa_agg_value_set): New structs.
(ipa_agg_jf_item): Add new field jftype and type, redefine field value.
(ipa_agg_jump_function): Remove member function equal_to.
(ipa_agg_jump_function_p): Remove typedef.
(ipa_copy_agg_values, ipa_release_agg_values): New functions.
* ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump
information for aggregate jump function.
(get_ssa_def_if_simple_copy): Add new parameter rhs_stmt to
record last definition statement.
(load_from_unmodified_param_or_agg): New function.
(ipa_known_agg_contents_list): Add new field type and value, remove
field constant.
(build_agg_jump_func_from_list): Rename parameter const_count to
value_count, build aggregate jump function from ipa_load_agg_data.
(analyze_agg_content_value): New function.
(extract_mem_content): Analyze memory store assignment to prepare
information for aggregate jump function generation.
(determine_known_aggregate_parts): Add new parameter fbi, remove
parameter aa_walk_budeget_p.
(update_jump_functions_after_inlining): Update aggregate jump function.
(ipa_find_agg_cst_for_param): Change type of parameter agg.
(try_make_edge_direct_simple_call): Add new parameter new_root.
(try_make_edge_direct_virtual_call): Add new parameter new_root and
new_root_info.
(update_indirect_edges_after_inlining): Pass new argument to
try_make_edge_direct_simple_call and try_make_edge_direct_virtual_call.
(ipa_write_jump_function): Write aggregate jump function to file.
(ipa_read_jump_function): Read aggregate jump function from file.
(ipa_agg_value::equal_to): Migrate from ipa_agg_jf_item::equal_to.
* ipa-cp.c (ipa_get_jf_arith_result): New function.
(ipa_agg_value_from_node): Likewise.
(ipa_agg_value_set_from_jfunc): Likewise.
(propagate_vals_across_arith_jfunc): Likewise.
(propagate_aggregate_lattice): Likewise.
(ipa_get_jf_pass_through_result): Call ipa_get_jf_arith_result.
(propagate_vals_across_pass_through): Call
propagate_vals_across_arith_jfunc.
(get_clone_agg_value): Move forward.
(propagate_aggs_across_jump_function): Handle value propagation for
aggregate jump function.
(agg_jmp_p_vec_for_t_vec): Remove.
(context_independent_aggregate_values): Replace vec<ipa_agg_jf_item>
with vec<ipa_agg_value>.
(copy_plats_to_inter, intersect_with_plats): Likewise.
(agg_replacements_to_vector, intersect_with_agg_replacements): Likewise.
(intersect_aggregate_with_edge): Likewise.
(find_aggregate_values_for_callers_subset): Likewise.
(cgraph_edge_brings_all_agg_vals_for_node): Likewise.
(estimate_local_effects): Replace vec<ipa_agg_jump_function> and
vec<ipa_agg_jump_function_p> with vec<ipa_agg_value_set>.
(gather_context_independent_values): Likewise.
(perform_estimation_of_a_value, decide_whether_version_node): Likewise.
* ipa-fnsummary.c (evaluate_conditions_for_known_args): Replace
vec<ipa_agg_jump_function_p> with vec<ipa_agg_value_set>.
(evaluate_properties_for_edge): Likewise.
(estimate_edge_devirt_benefit): Likewise.
(estimate_edge_size_and_time):  Likewise.
(estimate_calls_size_and_time): Likewise.
(ipa_call_context::ipa_call_context): Likewise.
(estimate_ipcp_clone_size_and_time):  Likewise.
* ipa-fnsummary.h (ipa_call_context): Replace
vec<ipa_agg_jump_function_p> with vec<ipa_agg_value_set>.
* ipa-inline-analysis.c (do_estimate_edge_time): Replace
vec<ipa_agg_jump_function_p> with vec<ipa_agg_value_set>.
(do_estimate_edge_size): Likewise.
(do_estimate_edge_hints): Likewise.

2019-11-14  Feng Xue  <fxue@os.amperecomputing.com>

        PR ipa/91682
        * gcc.dg/ipa/ipcp-agg-10.c: Change dg-scan string.
        * gcc.dg/ipa/ipcp-agg-11.c: New test.

From-SVN: r278193

re PR ipa/92421 (ICE in inline_small_functions, at ipa-inline.c:2001 since r277759)

PR ipa/92421
* g++.dg/torture/pr92421.C: Add -Wno-return-type to
dg-additional-options.  Avoid -Wwrite-string warnings, most of
-Wreturn-type warnings, define bf ctor.  Use struct instead of class
with public: at the start.

From-SVN: r278192

alias-decl-pr92206-3.C: Require effective target c++14 rather than c++11.

* g++.dg/cpp0x/alias-decl-pr92206-3.C: Require effective target c++14
rather than c++11.

From-SVN: r278191

ISO_Fortran_binding_17.c: Include ../../../libgfortran/ISO_Fortran_binding.h rather than...

* gfortran.dg/ISO_Fortran_binding_17.c: Include
../../../libgfortran/ISO_Fortran_binding.h rather than
ISO_Fortran_binding.h.

From-SVN: r278190

Daily bump.

From-SVN: r278189

ipa-cp.c (propagate_vr_across_jump_function): Propagate also across binary operations.

* ipa-cp.c (propagate_vr_across_jump_function): Propagate also across
binary operations.

From-SVN: r278185

ipa-profile.c (check_argument_count): Check properly that e_info is non-NULL; do not check descriptors.

* ipa-profile.c (check_argument_count): Check properly that e_info
is non-NULL; do not check descriptors.

From-SVN: r278184

ipa-inline-analysis.c (do_estimate_edge_time): Relax check for ipa profiles.

* ipa-inline-analysis.c (do_estimate_edge_time): Relax
check for ipa profiles.

From-SVN: r278179

re PR ipa/92421 (ICE in inline_small_functions, at ipa-inline.c:2001 since r277759)

PR c++/92421
* ipa-prop.c (update_indirect_edges_after_inlining):
Mark parameter as used.
* ipa-inline.c (recursive_inlining): Reset node cache
after inlining.
(inline_small_functions): Remove checking ifdef.
* ipa-inline-analysis.c (do_estimate_edge_time): Verify
cache consistency.
* g++.dg/torture/pr92421.C: New testcase.

From-SVN: r278159

re PR ipa/92498 (gcc.dg/tree-prof/crossmodule-indircall-1.c fails starting with r278100)

PR ipa/92498
* ipa-profile.c (check_argument_count): Do not ICE when descriptors
is NULL.
(ipa_profile): Fix reversed test.

From-SVN: r278157

ipa-cp.c (ignore_edge_p): Do not check caller flags.

* ipa-cp.c (ignore_edge_p): Do not check caller flags.
(propagate_constants_topo): Fix typo.

From-SVN: r278156

Move plain value_range things to value-range.[hc]*.

From-SVN: r278153

Sanitize the constant argument for rotr<mode>3

This was dormant for quite some time, but it started happening for me
on gcc.c-torture/compile/pr65153.c sometime after r276645 for -mabi=32 linux runs.

The pattern accepts any SMALL_OPERAND constant value while it asserts during the final
that the value is in the mode size range. I this case it happens that combine_and_move_insns
during ira makes a pattern with negative "shift count" which fails at final stage.

This simple fix just truncates the constant operand to mode size the same as shift patterns.

gcc/ChangeLog:

2019-11-13  Dragan Mladjenovic  <dmladjenovic@wavecomp.com>

* config/mips/mips.md (rotr<mode>3): Sanitize the constant argument
instead of asserting its value.

From-SVN: r278152

download_prerequisites: Use http instead of ftp for downloading.

2019-11-13  Janne Blomqvist  <jb@gcc.gnu.org>

        * download_prerequisites: Use http instead of ftp for downloading.

From-SVN: r278151

pr92090.c: Limit -mbig to powerpc64le-*-*.

        * gcc.target/powerpc/pr92090.c: Limit -mbig to powerpc64le-*-*.
        * gcc.target/powerpc/pr92090-2.c: Likewise.

From-SVN: r278150

libsupc++: Implement comparison algorithms for C++20

This is incomplete because std::strong_order doesn't support
floating-point types.

The partial_order and weak_order tests use VERIFY instead of
static_assert because of PR 92431.

* libsupc++/compare (strong_order, weak_order, partial_order)
(compare_strong_order_fallback, compare_weak_order_fallback)
(compare_partial_order_fallback): Define customization point objects
for C++20.
* testsuite/18_support/comparisons/algorithms/partial_order.cc: New
test.
* testsuite/18_support/comparisons/algorithms/strong_order.cc: New
test.
* testsuite/18_support/comparisons/algorithms/weak_order.cc: New test.

From-SVN: r278149

Rewrite value_range constructors to the value_range_kind is at the end, and defaults to VR_RANGE.

Rewrite value_range constructors to the value_range_kind is at the
end, and defaults to VR_RANGE.  Similarly for set() methods.

From-SVN: r278148

PR c++/89070 - bogus [[nodiscard]] warning in SFINAE.

This is a complaint that we issue a [[nodiscard]] warning even in SFINAE
contexts.  Here 'complain' is tf_decltype, but not tf_warning so I guess
we can fix it as below.

* cvt.c (convert_to_void): Guard maybe_warn_nodiscard calls with
tf_warning.

* g++.dg/cpp1z/nodiscard7.C: New test.

From-SVN: r278147

tree-dump.c (dequeue_and_dump): Print first tree operand for VIEW_CONVERT_EXPR.

2019-11-13  Ulrich Drepper  <drepper@redhat.com>

* tree-dump.c (dequeue_and_dump): Print first tree operand
for VIEW_CONVERT_EXPR.

From-SVN: r278146

Add C2x *_NORM_MAX constants to <float.h>.

C2x adds <float.h> constants FLT_NORM_MAX, DBL_NORM_MAX and
LDBL_NORM_MAX.  These are for the maximum "normalized" finite
floating-point number, where the given definition of normalized is
that all possible values with MANT_DIG significand digits (leading one
not zero) can be represented with that exponent.  The effect of that
definition is that these macros are the same as the corresponding MAX
macros for all formats except IBM long double, where the NORM_MAX
value has exponent 1 smaller than the MAX one so that all 106 digits
can be 1.

This patch adds those macros to GCC.  They are only defined for float,
double and long double; C2x does not include such macros for DFP
types, and while the integration of TS 18661-3 into C2x has not yet
occurred, the draft proposed text does not add them for the _FloatN /
_FloatNx types (where they would always be the same as the MAX
macros).

Bootstrapped with no regressions on x86_64-pc-linux-gnu.  Also tested
compilation of the new test for powerpc-linux-gnu to confirm the check
of LDBL_NORM_MAX in the IBM long double case does get properly
optimized out.

gcc:
* ginclude/float.c [__STDC_VERSION__ > 201710L] (FLT_NORM_MAX,
DBL_NORM_MAX, LDBL_NORM_MAX): Define.
* real.c (get_max_float): Add norm_max argument.
* real.h (get_max_float): Update prototype.
* builtins.c (fold_builtin_interclass_mathfn): Update calls to
get_max_float.

gcc/c-family:
* c-cppbuiltin.c (builtin_define_float_constants): Also define
NORM_MAX constants.  Update call to get_max_float.
(LAZY_HEX_FP_VALUES_CNT): Update value to include NORM_MAX
constants.

gcc/d:
* d-target.cc (define_float_constants): Update call to
get_max_float.

gcc/testsuite:
* gcc.dg/c11-float-3.c, gcc.dg/c2x-float-1.c: New tests.

From-SVN: r278145

Come up with selftests for dbgcnt.

2019-11-13  Martin Liska  <mliska@suse.cz>

* dbgcnt.c (test_sorted_dbg_counters): New.
(dbgcnt_c_tests): Likewise.
* selftest-run-tests.c (selftest::run_tests): Likewise.
* selftest.h (dbgcnt_c_tests): Likewise.

From-SVN: r278144

Add a few missing checks that IPA_NODE_REF is not NULL (PR 92454)

2019-11-13  Jan Hubicka  <hubicka@ucw.cz>
    Martin Jambor  <mjambor@suse.cz>

PR ipa/92454
* ipa-cp.c (spread_undeadness): Check that IPA_NODE_REF exists.
(identify_dead_nodes): Likewise.

testsuite/
* g++.dg/ipa/pr92454.C: New test.

From-SVN: r278142

Do not overuse push/pop_cfun in IPA ICF.

2019-11-13  Martin Liska  <mliska@suse.cz>

* ipa-icf.c (sem_function::equals_private): Do not overuse
push/pop_cfun functions.

From-SVN: r278141

Enhance syntax of -fdbg-cnt.

2019-11-13  Martin Liska  <mliska@suse.cz>

* common.opt: Document change of -fdbg-cnt option.
* dbgcnt.c (DEBUG_COUNTER): Remove.
(dbg_cnt_is_enabled): Remove.
(dbg_cnt): Work with new intervals.
(dbg_cnt_set_limit_by_index): Set to new
list of intervals.
(dbg_cnt_set_limit_by_name): Likewise.
(dbg_cnt_process_single_pair): Process new format.
(dbg_cnt_process_opt): Likewise.
(dbg_cnt_list_all_counters): Likewise.
* doc/invoke.texi: Document change of -fdbg-cnt option.
(cmp_tuples): New.
2019-11-13  Martin Liska  <mliska@suse.cz>

* gcc.dg/ipa/ipa-icf-39.c: Update -fdbg-cnt to the new format.
* gcc.dg/pr68766.c: Likewise.

From-SVN: r278140

ipa-inline.c (ipa_inline): Check that function is defined before flattening.

* ipa-inline.c (ipa_inline): Check that function is defined before
flattening.
* gcc.c-torture/compile/flatten.c: New testcase.

From-SVN: r278139

GCN Libgomp Plugin

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>
    Kwok Cheung Yeung  <kcy@codesourcery.com>
    Julian Brown  <julian@codesourcery.com>
    Tom de Vries  <tom@codesourcery.com>

libgomp/
* plugin/Makefrag.am: Add amdgcn plugin support.
* plugin/configfrag.ac: Likewise.
* plugin/plugin-gcn.c: New file.
* configure: Regenerate.
* Makefile.in: Regenerate.
* testsuite/Makefile.in: Regenerate.

Co-Authored-By: Julian Brown <julian@codesourcery.com>
Co-Authored-By: Kwok Cheung Yeung <kcy@codesourcery.com>
Co-Authored-By: Tom de Vries <tom@codesourcery.com>
From-SVN: r278138

Use a single worker for OpenACC on AMD GCN

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>
    Julian Brown  <julian@codesourcery.com>

gcc/
* config/gcn/gcn.c (gcn_goacc_validate_dims): Ensure
flag_worker_partitioning is not set.
(TARGET_GOACC_WORKER_PARTITIONING): Remove target hook definition.
* config/gcn/gcn.opt (macc-experimental-workers): Default to off.

Co-Authored-By: Julian Brown <julian@codesourcery.com>
From-SVN: r278137

Optimize GCN OpenMP malloc performance

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>

libgomp/
* config/gcn/team.c (gomp_gcn_enter_kernel): Set up the team arena
and use team_malloc variants.
(gomp_gcn_exit_kernel): Use team_free.
* libgomp.h (TEAM_ARENA_SIZE): Define.
(TEAM_ARENA_START): Define.
(TEAM_ARENA_FREE): Define.
(TEAM_ARENA_END): Define.
(team_malloc): New function.
(team_malloc_cleared): New function.
(team_free): New function.
* team.c (gomp_new_team): Initialize and use team_malloc.
(free_team): Use team_free.
(gomp_free_thread): Use team_free.
(gomp_pause_host): Use team_free.
* work.c (gomp_init_work_share): Use team_malloc.
(gomp_fini_work_share): Use team_free.

From-SVN: r278136

GCN libgomp port

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>
    Kwok Cheung Yeung  <kcy@codesourcery.com>
    Julian Brown  <julian@codesourcery.com>
    Tom de Vries  <tom@codesourcery.com>

include/
* gomp-constants.h (GOMP_DEVICE_GCN): Define.
(GOMP_VERSION_GCN): Define.

libgomp/
* Makefile.am (libgomp_la_SOURCES): Add oacc-target.c.
* Makefile.in: Regenerate.
* config.h.in (PLUGIN_GCN): Add new undef.
* config/accel/openacc.f90 (acc_device_gcn): New parameter.
* config/gcn/affinity-fmt.c: New file.
* config/gcn/bar.c: New file.
* config/gcn/bar.h: New file.
* config/gcn/doacross.h: New file.
* config/gcn/icv-device.c: New file.
* config/gcn/oacc-target.c: New file.
* config/gcn/simple-bar.h: New file.
* config/gcn/target.c: New file.
* config/gcn/task.c: New file.
* config/gcn/team.c: New file.
* config/gcn/time.c: New file.
* configure.ac: Add amdgcn*-*-*.
* configure: Regenerate.
* configure.tgt: Add amdgcn*-*-*.
* libgomp-plugin.h (offload_target_type): Add OFFLOAD_TARGET_TYPE_GCN.
* libgomp.h (gcn_thrs): Add amdgcn variant.
(set_gcn_thrs): Likewise.
(gomp_thread): Likewise.
* oacc-int.h (goacc_thread): Likewise.
* oacc-target.c: New file.
* openacc.f90 (acc_device_gcn): New parameter.
* openacc.h (acc_device_t): Add acc_device_gcn.
* team.c (gomp_free_pool_helper): Add amdgcn support.

Co-Authored-By: Julian Brown <julian@codesourcery.com>
Co-Authored-By: Kwok Cheung Yeung <kcy@codesourcery.com>
Co-Authored-By: Tom de Vries <tom@codesourcery.com>
From-SVN: r278135

Add device number to GOMP_OFFLOAD_openacc_async_construct

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>
    Julian Brown  <julian@codesourcery.com>

libgomp/
* libgomp-plugin.h (GOMP_OFFLOAD_openacc_async_construct): Add int
parameter.
* oacc-async.c (lookup_goacc_asyncqueue): Pass device number to the
queue constructor.
* oacc-host.c (host_openacc_async_construct): Add device parameter.
* plugin/plugin-nvptx.c (GOMP_OFFLOAD_openacc_async_construct): Add
device parameter.

Co-Authored-By: Julian Brown <julian@codesourcery.com>
From-SVN: r278134

GCN mkoffload

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>
    Kwok Cheung Yeung  <kcy@codesourcery.com>
    Julian Brown  <julian@codesourcery.com>
    Tom de Vries  <tom@codesourcery.com>

gcc/
* config/gcn/mkoffload.c: New file.
* config/gcn/offload.h: New file.

From-SVN: r278133

Move generic libgomp files from nvptx to accel

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>

libgomp/
* configure.tgt (nvptx*-*-*): Add "accel" directory.
* config/nvptx/libgomp-plugin.c: Move ...
* config/accel/libgomp-plugin.c: ... to here.
* config/nvptx/lock.c: Move ...
* config/accel/lock.c: ... to here.
* config/nvptx/mutex.c: Move ...
* config/accel/mutex.c: ... to here.
* config/nvptx/mutex.h: Move ...
* config/accel/mutex.h: ... to here.
* config/nvptx/oacc-async.c: Move ...
* config/accel/oacc-async.c: ... to here.
* config/nvptx/oacc-cuda.c: Move ...
* config/accel/oacc-cuda.c: ... to here.
* config/nvptx/oacc-host.c: Move ...
* config/accel/oacc-host.c: ... to here.
* config/nvptx/oacc-init.c: Move ...
* config/accel/oacc-init.c: ... to here.
* config/nvptx/oacc-mem.c: Move ...
* config/accel/oacc-mem.c: ... to here.
* config/nvptx/oacc-plugin.c: Move ...
* config/accel/oacc-plugin.c: ... to here.
* config/nvptx/omp-lock.h: Move ...
* config/accel/omp-lock.h: ... to here.
* config/nvptx/openacc.f90: Move ...
* config/accel/openacc.f90: ... to here.
* config/nvptx/pool.h: Move ...
* config/accel/pool.h: ... to here.
* config/nvptx/proc.c: Move ...
* config/accel/proc.c: ... to here.
* config/nvptx/ptrlock.c: Move ...
* config/accel/ptrlock.c: ... to here.
* config/nvptx/ptrlock.h: Move ...
* config/accel/ptrlock.h: ... to here.
* config/nvptx/sem.c: Move ...
* config/accel/sem.c: ... to here.
* config/nvptx/sem.h: Move ...
* config/accel/sem.h: ... to here.
* config/nvptx/thread-stacksize.h: Move ...
* config/accel/thread-stacksize.h: ... to here.

From-SVN: r278132

Move gcn-run heap into GPU memory.

2019-11-13  Andrew Stubbs  <ams@codesourcery.com>

gcc/
* config/gcn/gcn-run.c (heap_region): New global variable.
(struct hsa_runtime_fn_info): Add hsa_memory_assign_agent_fn.
(init_hsa_runtime_functions): Initialize hsa_memory_assign_agent.
(get_kernarg_region): Move contents to ....
(get_memory_region): .... here.
(get_heap_region): New function.
(init_device): Initialize the heap_region.
(device_malloc): Add region parameter.
(struct kernargs): Move heap ....
(heap): ... to global scope.
(main): Allocate heap separate to kernargs.

From-SVN: r278131

c-ada-spec.c (get_underlying_decl): Do not look through typedefs.

* c-ada-spec.c (get_underlying_decl): Do not look through typedefs.
(dump_forward_type): Do not generate a declaration for function types.
(dump_nested_type) <ARRAY_TYPE>: Do not generate a nested declaration
of the component type if it is declared in another file.

From-SVN: r278129

PR fortran/92470 Fixes for CFI_address

        libgfortran/
        PR fortran/92470
        * runtime/ISO_Fortran_binding.c (CFI_establish): Set lower_bound to 0
        also for CFI_attribute_other.

        gcc/testsuite/
        PR fortran/92470
        * gfortran.dg/ISO_Fortran_binding_1.c (establish_c): Add assert for
        lower_bound == 0.

From-SVN: r278128

ipa-prop.c (ipa_print_node_jump_functions, [...]): Print info about missing summaries.

* ipa-prop.c (ipa_print_node_jump_functions,
ipa_print_node_params): Print info about missing summaries.

From-SVN: r278127

Enable libsanitizer build on riscv64

* configure.tgt (riscv64-*-linux*): Enable build.

From-SVN: r278126

Account for the cost of generating loop masks

We didn't take the cost of generating loop masks into account, and so
tended to underestimate the cost of loops that need multiple masks.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* tree-vect-loop.c (vect_estimate_min_profitable_iters): Include
the cost of generating loop masks.

gcc/testsuite/
* gcc.target/aarch64/sve/mask_struct_store_3.c: Add
-fno-vect-cost-model.
* gcc.target/aarch64/sve/mask_struct_store_3_run.c: Likewise.
* gcc.target/aarch64/sve/peel_ind_2.c: Likewise.
* gcc.target/aarch64/sve/peel_ind_2_run.c: Likewise.
* gcc.target/aarch64/sve/peel_ind_3.c: Likewise.
* gcc.target/aarch64/sve/peel_ind_3_run.c: Likewise.

From-SVN: r278125

Avoid accounting for non-existent vector loop versioning

vect_analyze_loop_costing uses two profitability thresholds: a runtime
one and a static compile-time one.  The runtime one is simply the point
at which the vector loop is cheaper than the scalar loop, while the
static one also takes into account the cost of choosing between the
scalar and vector loops at runtime.  We compare this static cost against
the expected execution frequency to decide whether it's worth generating
any vector code at all.

However, we never reclaimed the cost of applying the runtime threshold
if it turned out that the vector code can always be used.  And we only
know whether that's true once we've calculated what the runtime
threshold would be.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* tree-vectorizer.h (vect_apply_runtime_profitability_check_p):
New function.
* tree-vect-loop-manip.c (vect_loop_versioning): Use it.
* tree-vect-loop.c (vect_analyze_loop_2): Likewise.
(vect_transform_loop): Likewise.
(vect_analyze_loop_costing): Don't take the cost of versioning
into account for the static profitability threshold if it turns
out that no versioning is needed.

From-SVN: r278124

ipa.c (cgraph_build_static_cdtor): Pass optimization_default_node and target_option_default_node to get...

* ipa.c (cgraph_build_static_cdtor): Pass optimization_default_node
and target_option_default_node to get -fprofile-generate ctors working
right with LTO.

From-SVN: r278123

Don't assign a cost to vectorizable_assignment

vectorizable_assignment handles true SSA-to-SSA copies (which hopefully
we don't see in practice) and no-op conversions that are required
to maintain correct gimple, such as changes between signed and
unsigned types.  These cases shouldn't generate any code and so
shouldn't count against either the scalar or vector costs.

Later patches test this, but it seemed worth splitting out.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* tree-vectorizer.h (vect_nop_conversion_p): Declare.
* tree-vect-stmts.c (vect_nop_conversion_p): New function.
(vectorizable_assignment): Don't add a cost for nop conversions.
* tree-vect-loop.c (vect_compute_single_scalar_iteration_cost):
Likewise.
* tree-vect-slp.c (vect_bb_slp_scalar_cost): Likewise.

From-SVN: r278122

Fix vectorizable_conversion costs

This patch makes two tweaks to vectorizable_conversion.  The first
is to use "modifier" to distinguish between promotion, demotion,
and neither promotion nor demotion, rather than using a code for
some cases and "modifier" for others.  The second is to take ncopies
into account for the promotion and demotion costs; previously we gave
multiple copies the same cost as a single copy.

Later patches test this, but it seemed worth splitting out.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* tree-vect-stmts.c (vect_model_promotion_demotion_cost): Take the
number of ncopies as an additional argument.
(vectorizable_conversion): Update call accordingly.  Use "modifier"
to check whether a conversion is between vectors with the same
numbers of units.

From-SVN: r278121

[AArch64] Use aarch64_sve_int_mode in SVE ACLE code

This is a like-for-like change at the moment, but is a prerequisite
for removing mode_for_int_vector.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
* config/aarch64/aarch64-sve-builtins-functions.h
(unary_count::expand): Use aarch64_sve_int_mode instead of
mode_for_int_vector.

From-SVN: r278120

[C++] Fix interaction between aka changes and DR1558 (PR92206)

One of the changes in r277281 was to make the typedef variant
handling in strip_typedefs pass the raw DECL_ORIGINAL_TYPE to the
recursive call, instead of applying TYPE_MAIN_VARIANT first.
This PR shows that that interacts badly with the implementation
of DR1558, because we then refuse to strip aliases with dependent
template parameters and trip:

  gcc_assert (!typedef_variant_p (result)
      || ((flags & STF_USER_VISIBLE)
  && !user_facing_original_type_p (result)));

Keeping the current behaviour but suppressing the ICE leads to a
duplicate error (the dg-bogus in the first test), so that didn't
seem like a good fix.

I assume keeping the alias should never actually be necessary for
DECL_ORIGINAL_TYPEs, because it will already have been checked
somewhere, even for implicit TYPE_DECLs.  This patch therefore
passes a flag to say that we can safely strip aliases with
dependent template parameters.

2019-11-13  Richard Sandiford  <richard.sandiford@arm.com>

gcc/cp/
PR c++/92206
* cp-tree.h (STF_STRIP_DEPENDENT): New constant.
* tree.c (strip_typedefs): Add STF_STRIP_DEPENDENT to the flags
when calling strip_typedefs recursively on a DECL_ORIGINAL_TYPE.
Don't apply the fix for DR1558 in that case; allow aliases with
dependent template parameters to be stripped instead.

gcc/testsuite/
PR c++/92206
* g++.dg/cpp0x/alias-decl-pr92206-1.C: New test.
* g++.dg/cpp0x/alias-decl-pr92206-2.C: Likewise.
* g++.dg/cpp0x/alias-decl-pr92206-3.C: Likewise.

From-SVN: r278119

Update comment in opts.c.

2019-11-13  Martin Liska  <mliska@suse.cz>

* opts.c: Update comment about OPT_LEVELS_2_PLUS_SPEED_ONLY.

From-SVN: r278118

Fix params.exp by parsing output of --help=params -Q.

2019-11-13  Martin Liska  <mliska@suse.cz>

* gcc.dg/params/params.exp: Restore test by parsing output
of --help=params -Q.

From-SVN: r278117

Remove leftover call to finalize_options_struct.

2019-11-13  Martin Liska  <mliska@suse.cz>

* tree-streamer-in.c (lto_input_ts_function_decl_tree_pointers):
Remove call to finalize_options_struct.

From-SVN: r278116

re PR target/92055 ([avr] Support 64-bit double)

PR target/92055
* config/avr/t-avr (avr-mcus): Do not depend on
$(srcdir)/config/avr/t-multilib.

From-SVN: r278115

Fortran] Use proper type for hidden is-present argument

        gcc/fortran/
        * trans-expr.c (gfc_conv_procedure_call): Fold hidden
        is-present argument to the right type.

From-SVN: r278114

re PR target/92473 (test pr92324-2.c fails on arm and aarch64)

2019-11-13  Richard Biener  <rguenther@suse.de>

PR tree-optimization/92473
* tree-vect-loop.c (vect_create_epilog_for_reduction): Perform
direct optab reduction in the correct type.

From-SVN: r278113

Add option -fweb for pr47763.c

This case is testing 'web' on ignore naked clobber.
-funroll-loops no longer implies -fweb for powerpc.
So, add -fweb to enable 'web' for this case.

gcc.testsuite/
2019-11-13  Jiufu Guo  <guojiufu@linux.ibm.com>

PR target/92465
* gcc.dg/pr47763.c: Add option -fweb.

From-SVN: r278112

Daily bump.

From-SVN: r278111

rs6000: Use ULL on big hexadecimal literal

C++98 does not have long long int, and does not use (unsigned) long
long int for hexadecimal literals.  So let's use an ULL suffix here,
which is still not strict C++98, but which works with more compilers.

* config/rs6000/rs6000.md (rs6000_set_fpscr_drn): Use ULL on big
hexadecimal literal.

From-SVN: r278107

[C++ PATCH] Merge some using-decl handling

https://gcc.gnu.org/ml/gcc-patches/2019-11/msg00971.html
gcc/cp/
* name-lookup.c (lookup_using_decl): New function, merged from ...
(do_class_using_decl): ... here.  Call it.  And ...
(finish_nonmember_using_decl): ... here.  Call it.

gcc/testsuite/
* g++.dg/cpp0x/using-enum-2.C: Adjust expected error text.
* g++.dg/cpp0x/using-enum-3.C: Likewise.
* g++.dg/lookup/using4.C: Likewise.
* g++.dg/lookup/using7.C: Likewise.
* g++.dg/template/using12.C: Likewise.
* g++.dg/template/using18.C: Likewise.
* g++.dg/template/using22.C: Likewise.

From-SVN: r278106

re PR fortran/81651 (Enhancement request: have f951 print out fully qualified module file name)

2019-11-12  Harald Anlauf  <anlauf@gmx.de>

PR fortran/81651
* module.c (gzopen_included_file, gzopen_included_file_1)
(gzopen_intrinsic_module, bad_module, gfc_use_module): Use fully
qualified module path for error reporting.

From-SVN: r278105

testsuite: Add testcases for PR92449

PR target/92449
* gcc.c-torture/compile/pr92449.c: New test.
* gcc.target/powerpc/pr92449-1.c: New test.

From-SVN: r278104

rs6000: Handle unordered for xscmpexp[dq]p without NaNs (PR92449)

* config/rs6000/vsx.md (xscmpexpdp_<code> for CMP_TEST): Handle
UNORDERED if !HONOR_NANS (DFmode).
(xscmpexpqp_<code>_<mode> for CMP_TEST and IEEE128): Handle UNORDERED
if !HONOR_NANS (<MODE>mode).

From-SVN: r278103

ipa-cp.c (ignore_edge_p): Also look for optimize flag.

* ipa-cp.c (ignore_edge_p): Also look for optimize flag.
(ipcp_verify_propagated_values): Likewise.
(propagate_constants_across_call): Likewise.
(propagate_constants_topo): Likewise.
(ipcp_propagate_stage): Likewise.

From-SVN: r278102

PR fortran/92470 Fixes for CFI_address

        libgfortran/
        PR fortran/92470
        * runtime/ISO_Fortran_binding.c (CFI_address): Handle non-zero
        lower_bound; update error message.
        (CFI_allocate): Fix comment typo.
        (CFI_establish): Fix identation, fix typos, don't check values of 'dv'
        argument.

        gcc/testsuite/
        PR fortran/92470
        * gfortran.dg/ISO_Fortran_binding_17.c: New.
        * gfortran.dg/ISO_Fortran_binding_17.f90: New.
        * gfortran.dg/ISO_Fortran_binding_1.c (elemental_mult_c, allocate_c,
        section_c, select_part_c): Update for CFI_{address} changes;
        add asserts.

From-SVN: r278101

re PR ipa/92471 ([ICE] lto1 segmentation fault: ipa-profile.c ipa_get_cs_argument_count (args=0x0))

PR ipa/92471
* ipa-profile.c (check_argument_count): Break out from ...;
watch for missing summaries.
(ipa_profile): Here.

From-SVN: r278100

PR tree-optimization/92412 - excessive errno aliasing assumption defeats optimization

gcc/ChangeLog:

PR tree-optimization/92412
* targhooks.c (default_ref_may_alias_errno): Errono can only alias
extern variables.

gcc/testsuite/ChangeLog:

PR tree-optimization/92412
* gcc.dg/strlenopt-91.c: New test.

From-SVN: r278099

PR middle-end/83688 - check if buffers may overlap when copying strings using sprintf

gcc/ChangeLog:

PR middle-end/83688
* gimple-ssa-sprintf.c (format_result::alias_info): New struct.
(directive::argno): New member.
(format_result::aliases, format_result::alias_count): New data members.
(format_result::append_alias): New member function.
(fmtresult::dst_offset): New data member.
(pass_sprintf_length::call_info::dst_origin): New data member.
(pass_sprintf_length::call_info::dst_field, dst_offset): Same.
(char_type_p, array_elt_at_offset, field_at_offset): New functions.
(get_origin_and_offset): Same.
(format_string): Call it.
(format_directive): Call append_alias and set directive argument
number.
(maybe_warn_overlap): New function.
(pass_sprintf_length::compute_format_length): Call it.
(pass_sprintf_length::handle_gimple_call): Initialize new members.
* gcc/tree-ssa-strlen.c (): Also enable when -Wrestrict is on.

gcc/testsuite/ChangeLog:

PR tree-optimization/35503
* gcc.dg/tree-ssa/builtin-sprintf-warn-23.c: New test.

From-SVN: r278098

[PR c++/6936] Delete duplicate test

https://gcc.gnu.org/ml/gcc-patches/2019-11/msg00926.html
* g++.dg/lookup/pr6936.C: Delete, identical to using38.C

From-SVN: r278096

Free dominance info at the beginning of pass_jump_after_combine

try_forward_edges does not update dominance info, and merge_blocks
relies on it being up-to-date.  In PR92430 stale dominance info makes
merge_blocks produce a loop in the dominator tree, which in turn makes
delete_basic_block loop forever.

Fix by freeing dominance info at the beginning of cleanup_cfg.

gcc/ChangeLog:

2019-11-12  Ilya Leoshkevich  <iii@linux.ibm.com>

PR rtl-optimization/92430
* cfgcleanup.c (pass_jump_after_combine::execute): Free
dominance info at the beginning.

gcc/testsuite/ChangeLog:

2019-11-12  Ilya Leoshkevich  <iii@linux.ibm.com>

PR rtl-optimization/92430
* gcc.dg/pr92430.c: New test (from Arseny Solokha).

From-SVN: r278095

re PR tree-optimization/92460 (ICE: verify_ssa failed (error: definition in block 13 does not dominate use in block 22))

2019-11-12  Richard Biener  <rguenther@suse.de>

PR tree-optimization/92460
* tree-vect-stmts.c (vectorizable_simd_clone_call): Unshare
expression before gimplifying.

From-SVN: r278094

re PR tree-optimization/92461 (ICE: verify_ssa failed (error: excess use operand for statement))

2019-11-12  Richard Biener  <rguenther@suse.de>

PR tree-optimization/92461
* tree-vect-loop.c (vect_create_epilog_for_reduction): Update
stmt after propagation.

* gcc.dg/torture/pr92461.c: New testcase.

From-SVN: r278093

Use more SET_OPTION_IF_UNSET.

2019-11-12  Martin Liska  <mliska@suse.cz>

* config/i386/i386-options.c (ix86_recompute_optlev_based_flags):
Use SET_OPTION_IF_UNSET.
(ix86_option_override_internal): Likewise.
* opts.c (default_options_optimization): Likewise.
(finish_options): Likewise.
(enable_fdo_optimizations): Likewise.
(common_handle_option): Likewise.
2019-11-12  Martin Liska  <mliska@suse.cz>

* c-opts.c (c_common_post_options):
Use SET_OPTION_IF_UNSET.
2019-11-12  Martin Liska  <mliska@suse.cz>

* options.c (gfc_init_options):
Use SET_OPTION_IF_UNSET.
2019-11-12  Martin Liska  <mliska@suse.cz>

* go-lang.c (go_langhook_post_options):
Use SET_OPTION_IF_UNSET.

From-SVN: r278091

Remove option_default_params and option_validate_param hooks.

2019-11-12  Martin Liska  <mliska@suse.cz>

* common/common-target.def: Remove option_validate_param and
option_default_params.
* common/common-targhooks.c (default_option_validate_param):
Remove.
* common/common-targhooks.h (default_option_validate_param):
Remove.
* common/config/aarch64/aarch64-common.c (TARGET_OPTION_DEFAULT_PARAMS):
Remove usage of this.
(TARGET_OPTION_VALIDATE_PARAM): Likewise.
(aarch64_option_validate_param): Likewise.
(aarch64_option_default_params): Likewise
* common/config/bpf/bpf-common.c (bpf_option_default_params): Likewise.
(TARGET_OPTION_DEFAULT_PARAMS): Likewise.
* common/config/ia64/ia64-common.c (ia64_option_default_params): Likewise.
(TARGET_OPTION_DEFAULT_PARAMS): Likewise.
* common/config/powerpcspe/powerpcspe-common.c (rs6000_option_default_params): Likewise.
(TARGET_OPTION_DEFAULT_PARAMS): Likewise.
* common/config/rs6000/rs6000-common.c (rs6000_option_default_params): Likewise.
(TARGET_OPTION_DEFAULT_PARAMS): Likewise.
* common/config/sh/sh-common.c (sh_option_default_params): Likewise.
(TARGET_OPTION_DEFAULT_PARAMS): Likewise.
* config/aarch64/aarch64.c (aarch64_override_options_internal): Validate
guard_size here.
* doc/tm.texi: Remove option_default_params and option_validate_param.
* doc/tm.texi.in: Likewise.

From-SVN: r278090

Fix test-suite fallout.

2019-11-12  Martin Liska  <mliska@suse.cz>

* gcc.dg/completion-3.c: Append = to all expected
results and sort expected output.
* gcc.dg/pr83620.c: Update error message.
* gcc.dg/spellcheck-params-2.c: Likewise.
* gcc.dg/spellcheck-params.c: Likewise.
* gcc.misc-tests/help.exp: Update expected output.

From-SVN: r278089

Remove set_default_param_value from documentation.

2019-11-12  Martin Liska  <mliska@suse.cz>

* common/common-target.def:
Do not mention set_default_param_value
and set_param_value.
* doc/tm.texi: Likewise.

From-SVN: r278088

Remove last leftover usage of params* files.

2019-11-12  Martin Liska  <mliska@suse.cz>

* common.opt: Remove param_values.
* config/i386/i386-options.c (ix86_valid_target_attribute_p):
Remove finalize_options_struct.
* gcc.c (driver::decode_argv): Do not call global_init_params
and finish_params.
(driver::finalize): Do not call params_c_finalize
and finalize_options_struct.
* opt-suggestions.c (option_proposer::get_completions): Remove
special casing of params.
(option_proposer::find_param_completions): Remove.
(test_completion_partial_match): Update expected output.
* opt-suggestions.h: Remove find_param_completions.
* opts-common.c (add_misspelling_candidates): Add
--param with a space.
* opts.c (handle_param): Remove.
(init_options_struct):. Remove init_options_struct and
similar calls.
(finalize_options_struct): Remove.
(common_handle_option): Use SET_OPTION_IF_UNSET.
* opts.h (finalize_options_struct): Remove.
* toplev.c (general_init): Do not call global_init_params.
(toplev::finalize): Do not call params_c_finalize and
finalize_options_struct.

From-SVN: r278087

Remove gcc/params.* files.

2019-11-12  Martin Liska  <mliska@suse.cz>

* Makefile.in: Remove PARAMS_H and params.list
and params.options.
* params-enum.h: Remove.
* params-list.h: Remove.
* params-options.h: Remove.
* params.c: Remove.
* params.def: Remove.
* params.h: Remove.
* asan.c: Do not include params.h.
* auto-profile.c: Likewise.
* bb-reorder.c: Likewise.
* builtins.c: Likewise.
* cfgcleanup.c: Likewise.
* cfgexpand.c: Likewise.
* cfgloopanal.c: Likewise.
* cgraph.c: Likewise.
* combine.c: Likewise.
* common/config/aarch64/aarch64-common.c: Likewise.
* common/config/gcn/gcn-common.c: Likewise.
* common/config/ia64/ia64-common.c: Likewise.
* common/config/powerpcspe/powerpcspe-common.c: Likewise.
* common/config/rs6000/rs6000-common.c: Likewise.
* common/config/sh/sh-common.c: Likewise.
* config/aarch64/aarch64.c: Likewise.
* config/alpha/alpha.c: Likewise.
* config/arm/arm.c: Likewise.
* config/avr/avr.c: Likewise.
* config/csky/csky.c: Likewise.
* config/i386/i386-builtins.c: Likewise.
* config/i386/i386-expand.c: Likewise.
* config/i386/i386-features.c: Likewise.
* config/i386/i386-options.c: Likewise.
* config/i386/i386.c: Likewise.
* config/ia64/ia64.c: Likewise.
* config/rs6000/rs6000-logue.c: Likewise.
* config/rs6000/rs6000.c: Likewise.
* config/s390/s390.c: Likewise.
* config/sparc/sparc.c: Likewise.
* config/visium/visium.c: Likewise.
* coverage.c: Likewise.
* cprop.c: Likewise.
* cse.c: Likewise.
* cselib.c: Likewise.
* dse.c: Likewise.
* emit-rtl.c: Likewise.
* explow.c: Likewise.
* final.c: Likewise.
* fold-const.c: Likewise.
* gcc.c: Likewise.
* gcse.c: Likewise.
* ggc-common.c: Likewise.
* ggc-page.c: Likewise.
* gimple-loop-interchange.cc: Likewise.
* gimple-loop-jam.c: Likewise.
* gimple-loop-versioning.cc: Likewise.
* gimple-ssa-split-paths.c: Likewise.
* gimple-ssa-sprintf.c: Likewise.
* gimple-ssa-store-merging.c: Likewise.
* gimple-ssa-strength-reduction.c: Likewise.
* gimple-ssa-warn-alloca.c: Likewise.
* gimple-ssa-warn-restrict.c: Likewise.
* graphite-isl-ast-to-gimple.c: Likewise.
* graphite-optimize-isl.c: Likewise.
* graphite-scop-detection.c: Likewise.
* graphite-sese-to-poly.c: Likewise.
* graphite.c: Likewise.
* haifa-sched.c: Likewise.
* hsa-gen.c: Likewise.
* ifcvt.c: Likewise.
* ipa-cp.c: Likewise.
* ipa-fnsummary.c: Likewise.
* ipa-inline-analysis.c: Likewise.
* ipa-inline.c: Likewise.
* ipa-polymorphic-call.c: Likewise.
* ipa-profile.c: Likewise.
* ipa-prop.c: Likewise.
* ipa-split.c: Likewise.
* ipa-sra.c: Likewise.
* ira-build.c: Likewise.
* ira-conflicts.c: Likewise.
* loop-doloop.c: Likewise.
* loop-invariant.c: Likewise.
* loop-unroll.c: Likewise.
* lra-assigns.c: Likewise.
* lra-constraints.c: Likewise.
* modulo-sched.c: Likewise.
* opt-suggestions.c: Likewise.
* opts.c: Likewise.
* postreload-gcse.c: Likewise.
* predict.c: Likewise.
* reload.c: Likewise.
* reorg.c: Likewise.
* resource.c: Likewise.
* sanopt.c: Likewise.
* sched-deps.c: Likewise.
* sched-ebb.c: Likewise.
* sched-rgn.c: Likewise.
* sel-sched-ir.c: Likewise.
* sel-sched.c: Likewise.
* shrink-wrap.c: Likewise.
* stmt.c: Likewise.
* targhooks.c: Likewise.
* toplev.c: Likewise.
* tracer.c: Likewise.
* trans-mem.c: Likewise.
* tree-chrec.c: Likewise.
* tree-data-ref.c: Likewise.
* tree-if-conv.c: Likewise.
* tree-inline.c: Likewise.
* tree-loop-distribution.c: Likewise.
* tree-parloops.c: Likewise.
* tree-predcom.c: Likewise.
* tree-profile.c: Likewise.
* tree-scalar-evolution.c: Likewise.
* tree-sra.c: Likewise.
* tree-ssa-ccp.c: Likewise.
* tree-ssa-dom.c: Likewise.
* tree-ssa-dse.c: Likewise.
* tree-ssa-ifcombine.c: Likewise.
* tree-ssa-loop-ch.c: Likewise.
* tree-ssa-loop-im.c: Likewise.
* tree-ssa-loop-ivcanon.c: Likewise.
* tree-ssa-loop-ivopts.c: Likewise.
* tree-ssa-loop-manip.c: Likewise.
* tree-ssa-loop-niter.c: Likewise.
* tree-ssa-loop-prefetch.c: Likewise.
* tree-ssa-loop-unswitch.c: Likewise.
* tree-ssa-math-opts.c: Likewise.
* tree-ssa-phiopt.c: Likewise.
* tree-ssa-pre.c: Likewise.
* tree-ssa-reassoc.c: Likewise.
* tree-ssa-sccvn.c: Likewise.
* tree-ssa-scopedtables.c: Likewise.
* tree-ssa-sink.c: Likewise.
* tree-ssa-strlen.c: Likewise.
* tree-ssa-structalias.c: Likewise.
* tree-ssa-tail-merge.c: Likewise.
* tree-ssa-threadbackward.c: Likewise.
* tree-ssa-threadedge.c: Likewise.
* tree-ssa-uninit.c: Likewise.
* tree-switch-conversion.c: Likewise.
* tree-vect-data-refs.c: Likewise.
* tree-vect-loop.c: Likewise.
* tree-vect-slp.c: Likewise.
* tree-vrp.c: Likewise.
* tree.c: Likewise.
* value-prof.c: Likewise.
* var-tracking.c: Likewise.
2019-11-12  Martin Liska  <mliska@suse.cz>

* gimple-parser.c: Do not include params.h.
2019-11-12  Martin Liska  <mliska@suse.cz>

* name-lookup.c: Do not include params.h.
* typeck.c: Likewise.
2019-11-12  Martin Liska  <mliska@suse.cz>

* lto-common.c: Do not include params.h.
* lto-partition.c: Likewise.
* lto.c: Likewise.

From-SVN: r278086

Apply mechanical replacement (generated patch).

2019-11-12  Martin Liska  <mliska@suse.cz>

* asan.c (asan_sanitize_stack_p): Replace old parameter syntax
with the new one, include opts.h if needed.  Use SET_OPTION_IF_UNSET
macro.
(asan_sanitize_allocas_p): Likewise.
(asan_emit_stack_protection): Likewise.
(asan_protect_global): Likewise.
(instrument_derefs): Likewise.
(instrument_builtin_call): Likewise.
(asan_expand_mark_ifn): Likewise.
* auto-profile.c (auto_profile): Likewise.
* bb-reorder.c (copy_bb_p): Likewise.
(duplicate_computed_gotos): Likewise.
* builtins.c (inline_expand_builtin_string_cmp): Likewise.
* cfgcleanup.c (try_crossjump_to_edge): Likewise.
(try_crossjump_bb): Likewise.
* cfgexpand.c (defer_stack_allocation): Likewise.
(stack_protect_classify_type): Likewise.
(pass_expand::execute): Likewise.
* cfgloopanal.c (expected_loop_iterations_unbounded): Likewise.
(estimate_reg_pressure_cost): Likewise.
* cgraph.c (cgraph_edge::maybe_hot_p): Likewise.
* combine.c (combine_instructions): Likewise.
(record_value_for_reg): Likewise.
* common/config/aarch64/aarch64-common.c (aarch64_option_validate_param): Likewise.
(aarch64_option_default_params): Likewise.
* common/config/ia64/ia64-common.c (ia64_option_default_params): Likewise.
* common/config/powerpcspe/powerpcspe-common.c (rs6000_option_default_params): Likewise.
* common/config/rs6000/rs6000-common.c (rs6000_option_default_params): Likewise.
* common/config/sh/sh-common.c (sh_option_default_params): Likewise.
* config/aarch64/aarch64.c (aarch64_output_probe_stack_range): Likewise.
(aarch64_allocate_and_probe_stack_space): Likewise.
(aarch64_expand_epilogue): Likewise.
(aarch64_override_options_internal): Likewise.
* config/alpha/alpha.c (alpha_option_override): Likewise.
* config/arm/arm.c (arm_option_override): Likewise.
(arm_valid_target_attribute_p): Likewise.
* config/i386/i386-options.c (ix86_option_override_internal): Likewise.
* config/i386/i386.c (get_probe_interval): Likewise.
(ix86_adjust_stack_and_probe_stack_clash): Likewise.
(ix86_max_noce_ifcvt_seq_cost): Likewise.
* config/ia64/ia64.c (ia64_adjust_cost): Likewise.
* config/rs6000/rs6000-logue.c (get_stack_clash_protection_probe_interval): Likewise.
(get_stack_clash_protection_guard_size): Likewise.
* config/rs6000/rs6000.c (rs6000_option_override_internal): Likewise.
* config/s390/s390.c (allocate_stack_space): Likewise.
(s390_emit_prologue): Likewise.
(s390_option_override_internal): Likewise.
* config/sparc/sparc.c (sparc_option_override): Likewise.
* config/visium/visium.c (visium_option_override): Likewise.
* coverage.c (get_coverage_counts): Likewise.
(coverage_compute_profile_id): Likewise.
(coverage_begin_function): Likewise.
(coverage_end_function): Likewise.
* cse.c (cse_find_path): Likewise.
(cse_extended_basic_block): Likewise.
(cse_main): Likewise.
* cselib.c (cselib_invalidate_mem): Likewise.
* dse.c (dse_step1): Likewise.
* emit-rtl.c (set_new_first_and_last_insn): Likewise.
(get_max_insn_count): Likewise.
(make_debug_insn_raw): Likewise.
(init_emit): Likewise.
* explow.c (compute_stack_clash_protection_loop_data): Likewise.
* final.c (compute_alignments): Likewise.
* fold-const.c (fold_range_test): Likewise.
(fold_truth_andor): Likewise.
(tree_single_nonnegative_warnv_p): Likewise.
(integer_valued_real_single_p): Likewise.
* gcse.c (want_to_gcse_p): Likewise.
(prune_insertions_deletions): Likewise.
(hoist_code): Likewise.
(gcse_or_cprop_is_too_expensive): Likewise.
* ggc-common.c: Likewise.
* ggc-page.c (ggc_collect): Likewise.
* gimple-loop-interchange.cc (MAX_NUM_STMT): Likewise.
(MAX_DATAREFS): Likewise.
(OUTER_STRIDE_RATIO): Likewise.
* gimple-loop-jam.c (tree_loop_unroll_and_jam): Likewise.
* gimple-loop-versioning.cc (loop_versioning::max_insns_for_loop): Likewise.
* gimple-ssa-split-paths.c (is_feasible_trace): Likewise.
* gimple-ssa-store-merging.c (imm_store_chain_info::try_coalesce_bswap): Likewise.
(imm_store_chain_info::coalesce_immediate_stores): Likewise.
(imm_store_chain_info::output_merged_store): Likewise.
(pass_store_merging::process_store): Likewise.
* gimple-ssa-strength-reduction.c (find_basis_for_base_expr): Likewise.
* graphite-isl-ast-to-gimple.c (class translate_isl_ast_to_gimple): Likewise.
(scop_to_isl_ast): Likewise.
* graphite-optimize-isl.c (get_schedule_for_node_st): Likewise.
(optimize_isl): Likewise.
* graphite-scop-detection.c (build_scops): Likewise.
* haifa-sched.c (set_modulo_params): Likewise.
(rank_for_schedule): Likewise.
(model_add_to_worklist): Likewise.
(model_promote_insn): Likewise.
(model_choose_insn): Likewise.
(queue_to_ready): Likewise.
(autopref_multipass_dfa_lookahead_guard): Likewise.
(schedule_block): Likewise.
(sched_init): Likewise.
* hsa-gen.c (init_prologue): Likewise.
* ifcvt.c (bb_ok_for_noce_convert_multiple_sets): Likewise.
(cond_move_process_if_block): Likewise.
* ipa-cp.c (ipcp_lattice::add_value): Likewise.
(merge_agg_lats_step): Likewise.
(devirtualization_time_bonus): Likewise.
(hint_time_bonus): Likewise.
(incorporate_penalties): Likewise.
(good_cloning_opportunity_p): Likewise.
(ipcp_propagate_stage): Likewise.
* ipa-fnsummary.c (decompose_param_expr): Likewise.
(set_switch_stmt_execution_predicate): Likewise.
(analyze_function_body): Likewise.
(compute_fn_summary): Likewise.
* ipa-inline-analysis.c (estimate_growth): Likewise.
* ipa-inline.c (caller_growth_limits): Likewise.
(inline_insns_single): Likewise.
(inline_insns_auto): Likewise.
(can_inline_edge_by_limits_p): Likewise.
(want_early_inline_function_p): Likewise.
(big_speedup_p): Likewise.
(want_inline_small_function_p): Likewise.
(want_inline_self_recursive_call_p): Likewise.
(edge_badness): Likewise.
(recursive_inlining): Likewise.
(compute_max_insns): Likewise.
(early_inliner): Likewise.
* ipa-polymorphic-call.c (csftc_abort_walking_p): Likewise.
* ipa-profile.c (ipa_profile): Likewise.
* ipa-prop.c (determine_known_aggregate_parts): Likewise.
(ipa_analyze_node): Likewise.
(ipcp_transform_function): Likewise.
* ipa-split.c (consider_split): Likewise.
* ipa-sra.c (allocate_access): Likewise.
(process_scan_results): Likewise.
(ipa_sra_summarize_function): Likewise.
(pull_accesses_from_callee): Likewise.
* ira-build.c (loop_compare_func): Likewise.
(mark_loops_for_removal): Likewise.
* ira-conflicts.c (build_conflict_bit_table): Likewise.
* loop-doloop.c (doloop_optimize): Likewise.
* loop-invariant.c (gain_for_invariant): Likewise.
(move_loop_invariants): Likewise.
* loop-unroll.c (decide_unroll_constant_iterations): Likewise.
(decide_unroll_runtime_iterations): Likewise.
(decide_unroll_stupid): Likewise.
(expand_var_during_unrolling): Likewise.
* lra-assigns.c (spill_for): Likewise.
* lra-constraints.c (EBB_PROBABILITY_CUTOFF): Likewise.
* modulo-sched.c (sms_schedule): Likewise.
(DFA_HISTORY): Likewise.
* opts.c (default_options_optimization): Likewise.
(finish_options): Likewise.
(common_handle_option): Likewise.
* postreload-gcse.c (eliminate_partially_redundant_load): Likewise.
(if): Likewise.
* predict.c (get_hot_bb_threshold): Likewise.
(maybe_hot_count_p): Likewise.
(probably_never_executed): Likewise.
(predictable_edge_p): Likewise.
(predict_loops): Likewise.
(expr_expected_value_1): Likewise.
(tree_predict_by_opcode): Likewise.
(handle_missing_profiles): Likewise.
* reload.c (find_equiv_reg): Likewise.
* reorg.c (redundant_insn): Likewise.
* resource.c (mark_target_live_regs): Likewise.
(incr_ticks_for_insn): Likewise.
* sanopt.c (pass_sanopt::execute): Likewise.
* sched-deps.c (sched_analyze_1): Likewise.
(sched_analyze_2): Likewise.
(sched_analyze_insn): Likewise.
(deps_analyze_insn): Likewise.
* sched-ebb.c (schedule_ebbs): Likewise.
* sched-rgn.c (find_single_block_region): Likewise.
(too_large): Likewise.
(haifa_find_rgns): Likewise.
(extend_rgns): Likewise.
(new_ready): Likewise.
(schedule_region): Likewise.
(sched_rgn_init): Likewise.
* sel-sched-ir.c (make_region_from_loop): Likewise.
* sel-sched-ir.h (MAX_WS): Likewise.
* sel-sched.c (process_pipelined_exprs): Likewise.
(sel_setup_region_sched_flags): Likewise.
* shrink-wrap.c (try_shrink_wrapping): Likewise.
* targhooks.c (default_max_noce_ifcvt_seq_cost): Likewise.
* toplev.c (print_version): Likewise.
(process_options): Likewise.
* tracer.c (tail_duplicate): Likewise.
* trans-mem.c (tm_log_add): Likewise.
* tree-chrec.c (chrec_fold_plus_1): Likewise.
* tree-data-ref.c (split_constant_offset): Likewise.
(compute_all_dependences): Likewise.
* tree-if-conv.c (MAX_PHI_ARG_NUM): Likewise.
* tree-inline.c (remap_gimple_stmt): Likewise.
* tree-loop-distribution.c (MAX_DATAREFS_NUM): Likewise.
* tree-parloops.c (MIN_PER_THREAD): Likewise.
(create_parallel_loop): Likewise.
* tree-predcom.c (determine_unroll_factor): Likewise.
* tree-scalar-evolution.c (instantiate_scev_r): Likewise.
* tree-sra.c (analyze_all_variable_accesses): Likewise.
* tree-ssa-ccp.c (fold_builtin_alloca_with_align): Likewise.
* tree-ssa-dse.c (setup_live_bytes_from_ref): Likewise.
(dse_optimize_redundant_stores): Likewise.
(dse_classify_store): Likewise.
* tree-ssa-ifcombine.c (ifcombine_ifandif): Likewise.
* tree-ssa-loop-ch.c (ch_base::copy_headers): Likewise.
* tree-ssa-loop-im.c (LIM_EXPENSIVE): Likewise.
* tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Likewise.
(try_peel_loop): Likewise.
(tree_unroll_loops_completely): Likewise.
* tree-ssa-loop-ivopts.c (avg_loop_niter): Likewise.
(CONSIDER_ALL_CANDIDATES_BOUND): Likewise.
(MAX_CONSIDERED_GROUPS): Likewise.
(ALWAYS_PRUNE_CAND_SET_BOUND): Likewise.
* tree-ssa-loop-manip.c (can_unroll_loop_p): Likewise.
* tree-ssa-loop-niter.c (MAX_ITERATIONS_TO_TRACK): Likewise.
* tree-ssa-loop-prefetch.c (PREFETCH_BLOCK): Likewise.
(L1_CACHE_SIZE_BYTES): Likewise.
(L2_CACHE_SIZE_BYTES): Likewise.
(should_issue_prefetch_p): Likewise.
(schedule_prefetches): Likewise.
(determine_unroll_factor): Likewise.
(volume_of_references): Likewise.
(add_subscript_strides): Likewise.
(self_reuse_distance): Likewise.
(mem_ref_count_reasonable_p): Likewise.
(insn_to_prefetch_ratio_too_small_p): Likewise.
(loop_prefetch_arrays): Likewise.
(tree_ssa_prefetch_arrays): Likewise.
* tree-ssa-loop-unswitch.c (tree_unswitch_single_loop): Likewise.
* tree-ssa-math-opts.c (gimple_expand_builtin_pow): Likewise.
(convert_mult_to_fma): Likewise.
(math_opts_dom_walker::after_dom_children): Likewise.
* tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
(hoist_adjacent_loads): Likewise.
(gate_hoist_loads): Likewise.
* tree-ssa-pre.c (translate_vuse_through_block): Likewise.
(compute_partial_antic_aux): Likewise.
* tree-ssa-reassoc.c (get_reassociation_width): Likewise.
* tree-ssa-sccvn.c (vn_reference_lookup_pieces): Likewise.
(vn_reference_lookup): Likewise.
(do_rpo_vn): Likewise.
* tree-ssa-scopedtables.c (avail_exprs_stack::lookup_avail_expr): Likewise.
* tree-ssa-sink.c (select_best_block): Likewise.
* tree-ssa-strlen.c (new_stridx): Likewise.
(new_addr_stridx): Likewise.
(get_range_strlen_dynamic): Likewise.
(class ssa_name_limit_t): Likewise.
* tree-ssa-structalias.c (push_fields_onto_fieldstack): Likewise.
(create_variable_info_for_1): Likewise.
(init_alias_vars): Likewise.
* tree-ssa-tail-merge.c (find_clusters_1): Likewise.
(tail_merge_optimize): Likewise.
* tree-ssa-threadbackward.c (thread_jumps::profitable_jump_thread_path): Likewise.
(thread_jumps::fsm_find_control_statement_thread_paths): Likewise.
(thread_jumps::find_jump_threads_backwards): Likewise.
* tree-ssa-threadedge.c (record_temporary_equivalences_from_stmts_at_dest): Likewise.
* tree-ssa-uninit.c (compute_control_dep_chain): Likewise.
* tree-switch-conversion.c (switch_conversion::check_range): Likewise.
(jump_table_cluster::can_be_handled): Likewise.
* tree-switch-conversion.h (jump_table_cluster::case_values_threshold): Likewise.
(SWITCH_CONVERSION_BRANCH_RATIO): Likewise.
(param_switch_conversion_branch_ratio): Likewise.
* tree-vect-data-refs.c (vect_mark_for_runtime_alias_test): Likewise.
(vect_enhance_data_refs_alignment): Likewise.
(vect_prune_runtime_alias_test_list): Likewise.
* tree-vect-loop.c (vect_analyze_loop_costing): Likewise.
(vect_get_datarefs_in_loop): Likewise.
(vect_analyze_loop): Likewise.
* tree-vect-slp.c (vect_slp_bb): Likewise.
* tree-vectorizer.h: Likewise.
* tree-vrp.c (find_switch_asserts): Likewise.
(vrp_prop::check_mem_ref): Likewise.
* tree.c (wide_int_to_tree_1): Likewise.
(cache_integer_cst): Likewise.
* var-tracking.c (EXPR_USE_DEPTH): Likewise.
(reverse_op): Likewise.
(vt_find_locations): Likewise.
2019-11-12  Martin Liska  <mliska@suse.cz>

* gimple-parser.c (c_parser_parse_gimple_body): Replace old parameter syntax
with the new one, include opts.h if needed.  Use SET_OPTION_IF_UNSET
macro.
2019-11-12  Martin Liska  <mliska@suse.cz>

* name-lookup.c (namespace_hints::namespace_hints): Replace old parameter syntax
with the new one, include opts.h if needed.  Use SET_OPTION_IF_UNSET
macro.
* typeck.c (comptypes): Likewise.
2019-11-12  Martin Liska  <mliska@suse.cz>

* lto-partition.c (lto_balanced_map): Replace old parameter syntax
with the new one, include opts.h if needed.  Use SET_OPTION_IF_UNSET
macro.
* lto.c (do_whole_program_analysis): Likewise.

From-SVN: r278085

Include new generated gcc/params.opt file.

2019-11-12  Martin Liska  <mliska@suse.cz>

* Makefile.in: Include params.opt.
* flag-types.h (enum parloops_schedule_type): Add
parloops_schedule_type used in params.opt.
* params.opt: New file.

From-SVN: r278084

Param to options conversion.

2019-11-12  Martin Liska  <mliska@suse.cz>

* common.opt: Remove --param and --param= options.
* opt-functions.awk: Mark CL_PARAMS for options
that have Param keyword.
* opts-common.c (decode_cmdline_options_to_array):
Replace --param key=value with --param=key=value.
* opts.c (print_filtered_help): Remove special
printing of params.
(print_specific_help): Update title for params.
(common_handle_option): Do not handle OPT__param.
opts.h (SET_OPTION_IF_UNSET): New macro.
* doc/options.texi: Document Param keyword.

From-SVN: r278083

Add OpenACC 2.6 `serial' construct support

    The `serial' construct (cf. section 2.5.3 of the OpenACC 2.6 standard)
    is equivalent to a `parallel' construct with clauses `num_gangs(1)
    num_workers(1) vector_length(1)' implied.
    These clauses are therefore not supported with the `serial'
    construct. All the remaining clauses accepted with `parallel' are also
    accepted with `serial'.

    The `serial' construct is implemented like `parallel', except for
    hardcoding dimensions rather than taking them from the relevant
    clauses, in `expand_omp_target'.

    Separate codes are used to denote the `serial' construct throughout the
    middle end, even though the mapping of `serial' to an equivalent
    `parallel' construct could have been done in the individual language
    frontends. In particular, this allows to distinguish between compute
    constructs in warnings, error messages, dumps etc.

    2019-11-12  Maciej W. Rozycki  <macro@codesourcery.com>
                Tobias Burnus  <tobias@codesourcery.com>
                Frederik Harwath  <frederik@codesourcery.com>
                Thomas Schwinge  <thomas@codesourcery.com>

    gcc/
    * gimple.h (gf_mask): Add GF_OMP_TARGET_KIND_OACC_SERIAL
    enumeration constant.
    (is_gimple_omp_oacc): Handle GF_OMP_TARGET_KIND_OACC_SERIAL.
    (is_gimple_omp_offloaded): Likewise.
    * gimplify.c (omp_region_type): Add ORT_ACC_SERIAL enumeration
    constant.  Adjust the value of ORT_NONE accordingly.
    (is_gimple_stmt): Handle OACC_SERIAL.
    (oacc_default_clause): Handle ORT_ACC_SERIAL.
    (gomp_needs_data_present): Likewise.
    (gimplify_adjust_omp_clauses): Likewise.
    (gimplify_omp_workshare): Handle OACC_SERIAL.
    (gimplify_expr): Likewise.
    * omp-expand.c (expand_omp_target):
    Handle GF_OMP_TARGET_KIND_OACC_SERIAL.
    (build_omp_regions_1, omp_make_gimple_edges): Likewise.
    * omp-low.c (is_oacc_parallel): Rename function to...
    (is_oacc_parallel_or_serial): ... this.
    Handle GF_OMP_TARGET_KIND_OACC_SERIAL.
    (scan_sharing_clauses): Adjust accordingly.
    (scan_omp_for): Likewise.
    (lower_oacc_head_mark): Likewise.
    (convert_from_firstprivate_int): Likewise.
    (lower_omp_target): Likewise.
    (check_omp_nesting_restrictions): Handle
    GF_OMP_TARGET_KIND_OACC_SERIAL.
    (lower_oacc_reductions): Likewise.
    (lower_omp_target): Likewise.
    * tree.def (OACC_SERIAL): New tree code.
    * tree-pretty-print.c (dump_generic_node): Handle OACC_SERIAL.

    * doc/generic.texi (OpenACC): Document OACC_SERIAL.

    gcc/c-family/
    * c-pragma.h (pragma_kind): Add PRAGMA_OACC_SERIAL enumeration
    constant.
    * c-pragma.c (oacc_pragmas): Add "serial" entry.

    gcc/c/
    * c-parser.c (OACC_SERIAL_CLAUSE_MASK): New macro.
    (c_parser_oacc_kernels_parallel): Rename function to...
    (c_parser_oacc_compute): ... this.  Handle PRAGMA_OACC_SERIAL.
    (c_parser_omp_construct): Update accordingly.

    gcc/cp/
    * constexpr.c (potential_constant_expression_1): Handle
    OACC_SERIAL.
    * parser.c (OACC_SERIAL_CLAUSE_MASK): New macro.
    (cp_parser_oacc_kernels_parallel): Rename function to...
    (cp_parser_oacc_compute): ... this.  Handle PRAGMA_OACC_SERIAL.
    (cp_parser_omp_construct): Update accordingly.
    (cp_parser_pragma): Handle PRAGMA_OACC_SERIAL.  Fix alphabetic
    order.
    * pt.c (tsubst_expr): Handle OACC_SERIAL.

    gcc/fortran/
    * gfortran.h (gfc_statement): Add ST_OACC_SERIAL_LOOP,
    ST_OACC_END_SERIAL_LOOP, ST_OACC_SERIAL and ST_OACC_END_SERIAL
    enumeration constants.
    (gfc_exec_op): Add EXEC_OACC_SERIAL_LOOP and EXEC_OACC_SERIAL
    enumeration constants.
    * match.h (gfc_match_oacc_serial): New prototype.
    (gfc_match_oacc_serial_loop): Likewise.
    * dump-parse-tree.c (show_omp_node, show_code_node): Handle
    EXEC_OACC_SERIAL_LOOP and EXEC_OACC_SERIAL.
    * match.c (match_exit_cycle): Handle EXEC_OACC_SERIAL_LOOP.
    * openmp.c (OACC_SERIAL_CLAUSES): New macro.
    (gfc_match_oacc_serial_loop): New function.
    (gfc_match_oacc_serial): Likewise.
    (oacc_is_loop): Handle EXEC_OACC_SERIAL_LOOP.
    (resolve_omp_clauses): Handle EXEC_OACC_SERIAL.
    (oacc_code_to_statement): Handle EXEC_OACC_SERIAL and
    EXEC_OACC_SERIAL_LOOP.
    (gfc_resolve_oacc_directive): Likewise.
    * parse.c (decode_oacc_directive) <'s'>: Add case for "serial"
    and "serial loop".
    (next_statement): Handle ST_OACC_SERIAL_LOOP and ST_OACC_SERIAL.
    (gfc_ascii_statement): Likewise.  Handle ST_OACC_END_SERIAL_LOOP
    and ST_OACC_END_SERIAL.
    (parse_oacc_structured_block): Handle ST_OACC_SERIAL.
    (parse_oacc_loop): Handle ST_OACC_SERIAL_LOOP and
    ST_OACC_END_SERIAL_LOOP.
    (parse_executable): Handle ST_OACC_SERIAL_LOOP and
    ST_OACC_SERIAL.
    (is_oacc): Handle EXEC_OACC_SERIAL_LOOP and EXEC_OACC_SERIAL.
    * resolve.c (gfc_resolve_blocks, gfc_resolve_code): Likewise.
    * st.c (gfc_free_statement): Likewise.
    * trans-openmp.c (gfc_trans_oacc_construct): Handle
    EXEC_OACC_SERIAL.
    (gfc_trans_oacc_combined_directive): Handle
    EXEC_OACC_SERIAL_LOOP.
    (gfc_trans_oacc_directive): Handle EXEC_OACC_SERIAL_LOOP and
    EXEC_OACC_SERIAL.
    * trans.c (trans_code): Likewise.

    gcc/testsuite/
    * c-c++-common/goacc/parallel-dims.c: New test.
    * gfortran.dg/goacc/parallel-dims.f90: New test.

    libgomp/
    * testsuite/libgomp.oacc-c-c++-common/parallel-dims.c: New test.
    * testsuite/libgomp.oacc-fortran/parallel-dims-aux.c: New test.
    * testsuite/libgomp.oacc-fortran/parallel-dims.f89: New test.
    * testsuite/libgomp.oacc-fortran/parallel-dims-2.f90: New test.

Reviewed-by: Thomas Schwinge <thomas@codesourcery.com>
Co-Authored-By: Frederik Harwath <frederik@codesourcery.com>
Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
Co-Authored-By: Tobias Burnus <tobias@codesourcery.com>
From-SVN: r278082

re PR target/92449 (ICE in extract_insn, at recog.c:2311)

PR target/92449
* tree-complex.c (expand_complex_multiplication): If !HONOR_NANS,
don't emit UNORDERED_EXPR guarded libcall.  Formatting fixes.

From-SVN: r278081

re PR tree-optimization/92452 (ICE in vrp_prop::check_array_ref at tree-vrp.c:4153)

PR tree-optimization/92452
* tree-vrp.c (vrp_prop::check_array_ref): If TRUNC_DIV_EXPR folds
into NULL_TREE, set up_bound to NULL_TREE instead of computing
MINUS_EXPR on it.

* c-c++-common/pr92452.c: New test.

From-SVN: r278080

re PR tree-optimization/92347 (ICE in vect_get_vec_def_for_operand_1, at tree-vect-stmts.c:1537)

2019-11-11  Andre Vieira  <andre.simoesdiasvieira@arm.com>

        PR tree-optimization/92347
        * tree-vect-loop.c (vect_transform_loop): Don't overwrite epilogues
        safelen with 0.

        * gcc.dg/vect/pr92347.c: New test.

From-SVN: r278079

Implement P1946R0, Allow defaulting comparisons by value.

* method.c (early_check_defaulted_comparison): Accept by-value,
reject mixed by-value and by-reference parms.
* decl.c (grokdeclarator): Set funcdef_flag for defaulted friend.
* decl2.c (grokfield): Don't SET_DECL_FRIEND_CONTEXT.

From-SVN: r278078

* typeck.c (cp_build_binary_op): Sorry about <=> on VECTOR_TYPE.

From-SVN: r278077

[RS6000] PC-relative TLS support

Supporting TLS for -mpcrel turns out to be relatively simple.  The
existing TLSGD and TLSLD unspecs happily can have their GOT pointer
reg element replaced with zero, refelecting the fact that optimisation
of calls to __tls_get_addr when pc-rel won't use the GOT pointer.
Some other insns also can be reused, and just a few added.

* config/rs6000/predicates.md (unspec_tls): Allow const0_rtx for got
element of unspec vec.
* config/rs6000/rs6000.c (rs6000_legitimize_tls_address): Support
PC-relative TLS.
* config/rs6000/rs6000.md (UNSPEC_TLSTLS_PCREL): New unspec.
(tls_gd_pcrel, tls_ld_pcrel): New insns.
(tls_dtprel, tls_tprel): Set attr prefixed when tls_size is not 16.
(tls_got_tprel_pcrel, tls_tls_pcrel): New insns.

From-SVN: r278076

[RS6000] Remove TARGET_TLS_MARKERS and require binutils 2.20

This patch removes !TARGET_TLS_MARKERS support.  -mtls-markers (and
-mno-tls-markers) disappear as valid options too, because I figure
they haven't been used too much except by people testing the
compiler.

* config/rs6000/rs6000.opt (mtls-markers): Delete.
* config/rs6000/rs6000.h (TARGET_TLS_MARKERS): Don't define.
(IS_NOMARK_TLSGETADDR): Likewise.
* config/rs6000/rs6000-protos.h (rs6000_output_tlsargs): Delete.
* config/rs6000/rs6000.c (rs6000_output_tlsargs): Delete.
(rs6000_legitimize_tls_address): Remove !TARGET_TLS_MARKERS code.
(rs6000_call_template_1): Delete TARGET_TLS_MARKERS test and
allow other UNSPECs besides UNSPEC_TLSGD and UNSPEC_TLSLD.
(rs6000_indirect_call_template_1): Likewise.
(rs6000_pltseq_template): Likewise.
(rs6000_opt_vars): Remove "tls-markers" entry.
* config/rs6000/rs6000.md (tls_gd<bits>): Replace TARGET_TLS_MARKERS
with TARGET_ELF.
(tls_gd_high<bits>, tls_gd_low<bits>): Likewise.
(tls_ld<bits>, tls_ld_high<bits>, tls_ld_low<bits>): Likewise.
(pltseq_plt_pcrel<mode>): Likewise.
(call_value_local32): Remove IS_NOMARK_TLSGETADDR predicate test.
(call_value_local64): Likewise.
(call_value_indirect_nonlocal_sysv<mode>): Remove IS_NOMARK_TLSGETADDR
output and length attribute sub-expression.
(call_value_nonlocal_sysv<mode>),
(call_value_nonlocal_sysv_secure<mode>),
(call_value_local_aix<mode>, call_value_nonlocal_aix<mode>),
(call_value_indirect_aix<mode>, call_value_indirect_elfv2<mode>),
(call_value_indirect_pcrel<mode>): Likewise.
* doc/install.texi (powerpc-*-*): Require binutils-2.20.
* configure.ac (HAVE_AS_TLS_MARKERS): Delete test.
* configure: Regenerate.
* config.in: Regenerate.

From-SVN: r278075

Daily bump.

From-SVN: r278074

libgo: fix DejaGNU testsuite compiler when using build sysroot

    Patch by Maciej W. Rozycki.

    Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/206577

From-SVN: r278070

Add prefixed insn support for stack_protect_setdi & stack_protect_testdi

2019-11-11  Michael Meissner  <meissner@linux.ibm.com>

* config/rs6000/predicates.md (prefixed_memory): New predicate.
* config/rs6000/rs6000.md (stack_protect_setdi): Deal with either
address being a prefixed load/store.
(stack_protect_testdi): Deal with either address being a prefixed
load.

From-SVN: r278069

re PR c++/92447 (ICE in poplevel, at cp/decl.c:585)

PR c++/92447
* decl.c (finish_function): Move ctype initialization before
DECL_DELETED_FN handling.

* g++.dg/cpp0x/pr92447.C: New test.

From-SVN: r278068

re PR bootstrap/92433 (r276645 breaks bootstrap on powerpc)

PR bootstrap/92433
* config/rs6000/rs6000-c.c (altivec_build_resolved_builtin): Guard
ALTIVEC_BUILTIN_VEC_VCMPGE_P argument swapping with n == 3 check.  Use
std::swap.

From-SVN: r278066

c-parser.c (c_parser_translation_unit): Diagnose declare target without corresponding end declare target.

* c-parser.c (c_parser_translation_unit): Diagnose declare target
without corresponding end declare target.

* semantics.c (finish_translation_unit): Diagnose declare target
without corresponding end declare target.

* c-c++-common/gomp/declare-target-5.c: New test.

From-SVN: r278065

Fix SLP downward group access classification (PR92420)

This PR was caused by the SLP handling in get_group_load_store_type
returning VMAT_CONTIGUOUS rather than VMAT_CONTIGUOUS_REVERSE for
downward groups.

A more elaborate fix would be to try to combine the reverse permutation
into SLP_TREE_LOAD_PERMUTATION for loads, but that's really a follow-on
optimisation and not backport material.  It might also not necessarily
be a win, if the target supports (say) reversing and odd/even swaps
as independent permutes but doesn't recognise the combined form.

2019-11-11  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
PR tree-optimization/92420
* tree-vect-stmts.c (get_negative_load_store_type): Move further
up file.
(get_group_load_store_type): Use it for reversed SLP accesses.

gcc/testsuite/
* gcc.dg/vect/pr92420.c: New test.

From-SVN: r278064

ipa-prop.c (ipa_propagate_indirect_call_infos): Remove ipcp summary.

* ipa-prop.c (ipa_propagate_indirect_call_infos): Remove ipcp
summary.
(ipcp_transformation_t::duplicate): Break out from ...
(ipa_node_params_t::duplicate): ... here; add copying of agg
replacements.
* ipa-prop.h (ipcp_transformation): Add constructor and destructor.
(ipcp_transformation_t): Add duplicate.

From-SVN: r278062

Bump minimum MPFR version to 3.1.0

Bump the minimum MPFR version to 3.1.0, released 2011-10-03. With this
requirement one can still build GCC with the operating system provided
MPFR on old but still supported operating systems like SLES 12 (MPFR
3.1.2) or RHEL/CentOS 7.x (MPFR 3.1.1).

This allows removing some code in the Fortran frontend, as well as
fixing PR 91828.

ChangeLog:

2019-11-11  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/91828
* configure.ac: Bump minimum MPFR to 3.1.0, recommended to 3.1.6+.
* configure: Regenerated.

gcc/ChangeLog:

2019-11-11  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/91828
* doc/install.texi: Document that the minimum MPFR version is
3.1.0.

gcc/fortran/ChangeLog:

2019-11-11  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/91828
* simplify.c (gfc_simplify_fraction): Remove fallback path for
MPFR < 3.1.0.

From-SVN: r278058

[ARC] Fix movsi_ne pattern.

The movsi_ne variants are in a wrong order, leading to wrong
computation of the internal attribute "cond". Hence, to errors when
outputting annul-true or annul-false instructions.

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

* config/arc/arc.md (movsi_ne): Reorder instruction variants.

testsuite/
xxxx-xx-xx  Shahab Vahedi  <shahab@synopsys.com>

* gcc.target/arc/delay-slot-limm.c: New test.

From-SVN: r278057

[ARC] Fix legitimize pic address.

There are cases when an pic address gets complicated, and it needs to
be resolved via force_reg function found in
prepare_move_operands. When this happens, we need to disambiguate the
pic address and re-legitimize it.

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

* config/arc/arc.c (arc_legitimize_pic_address): Consider UNSPECs
as well, if interesting recover the symbol and re-legitimize the
pic address.

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

* gcc.target/arc/pic-2.c: New file.

From-SVN: r278056

Fix commit for PR fortran/92142 - CFI_setpointer corrupts descriptor

2019-11-11  Tobias Burnus  <tobias@codesourcery.com>
            Mark Eggleston  <mark.eggleston@codethink.com>

        PR fortran/92142
        * gcc/testsuite/gfortran.dg/ISO_Fortran_binding_16.f90:
        Correct dg-additional-sources.

Co-Authored-By: Mark Eggleston <mark.eggleston@codethink.com>
From-SVN: r278055

Relax lto-dump.o dependency.

2019-11-11  Martin Liska  <mliska@suse.cz>

* Make-lang.in: Relax dependency of lto-dump.o to
LTO_OBJS which will allow faster linking (mainly with LTO).

From-SVN: r278054

Sort debug counter names.

2019-11-11  Martin Liska  <mliska@suse.cz>

* dbgcnt.def (DEBUG_COUNTER): Sort counters
alphabetically.

From-SVN: r278053

Update link to "Why not LGPL".

* doc/xml/gnu/gpl-3.0.xml: Adjust link to "Why not LGPL".

From-SVN: r278051

[vect]Account for epilogue's peeling for gaps when checking if we have enough
niters for epilogue

gcc/ChangeLog:
2019-11-11  Andre Vieira  <andre.simoesdiasvieira@arm.com>

* tree-vect-loop-manip.c (vect_do_peeling): Take epilogue gaps into
account when checking if there are enough iterations to vectorize
epilogue.

gcc/testsuite/ChangeLog:
2019-11-11  Andre Vieira  <andre.simoesdiasvieira@arm.com>

* gcc.dg/vect/vect-reduc-epilogue-gaps.c: New test.

From-SVN: r278049

PR fortran/92142 - CFI_setpointer corrupts descriptor

2019-11-11  José Rui Faustino de Sousa  <jrfsousa@gmail.com>

        libgfortran/
        PR fortran/92142
        * runtime/ISO_Fortran_binding.c (CFI_setpointer): Don't
        override descriptor attribute; with -fcheck, check that
        it is a pointer.

        gcc/testsuite/
        PR fortran/92142
        * gcc/testsuite/gfortran.dg/ISO_Fortran_binding_16.c: New.
        * gcc/testsuite/gfortran.dg/ISO_Fortran_binding_16.f90: New.
        * gcc/testsuite/gfortran.dg/ISO_Fortran_binding_10.c: Correct
        upper bounds for case 0.

From-SVN: r278048