2019-07-09  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

	* libgnat/s-memory.adb: Disable calls to Abort defer/undefer
	when ZCX_By_Default.

From-SVN: r273284

The compiler does not report an error on a generic type that has a
representation clause when its ultimate parent is not a generic formal.

2019-07-09  Javier Miranda  <miranda@adacore.com>

gcc/ada/

	* sem_ch13.adb (Rep_Item_Too_Early): Representation clauses are
	not allowed for a derivation of a generic type. Extend the
	current test to check that none of the parents is a generic
	type.

gcc/testsuite/

	* gnat.dg/rep_clause8.adb: New testcase.

From-SVN: r273283

2019-07-09  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

	* exp_ch9.adb, exp_util.adb, repinfo.adb, sem_ch12.adb,
	sem_prag.adb, sem_res.adb, sem_spark.adb, sem_util.adb: Minor
	reformatting.

From-SVN: r273282

This patch fixes a code generation error on an equality operation one of
whose operands is an overloaded call, and several equality operators are
visible. The resolution would succes but in some cases the wrong entity
was lwfton the equality node, leading to expansion with the wrong
interpretation.  If the equality operation is the operand of a negation,
the resolution of the negation must make direct use of the equality
resolution,

2019-07-09  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

	* sem_res.adb (Resolve_Equality_Op): If the node was overloaded,
	set properly the entity to which the node has been resolved. The
	original entity is the first one found during analysis, and is
	not necessarily the resolved one.
	(Resolve_Op_Not): If the argument of negation is an overloaded
	equality operation, call its resolution directly given that the
	context type does not participate in overload resolution.

gcc/testsuite/

	* gnat.dg/equal7.adb, gnat.dg/equal7_pkg.adb,
	gnat.dg/equal7_pkg.ads: New testcase.

From-SVN: r273281

This patch enables the elaboration order v4.0 as the default elaboration
order in GNATbind. The previous v3.0 elaboration order is now referred
to as the "legacy elaboration order mechanism" and is available using
binder switch -H.

2019-07-09  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

	* bindo.adb: Remove with and use clauses for Debug.  Add with
	and use clauses for Opt.
	(Find_Elaboration_Order): Enable the v4.0 elaboration order. The
	v3.0 mechanism is now available under binder switch -H.
	* bindusg.adb (Display): Enable switch -H.
	* debug.adb: Free compiler switch -gnatd_G.  Free binder switch
	-d_N.
	* sem_elab.adb: Update the section on switches to remove
	-gnatd_G.
	(Invocation_Graph_Recording_OK): The invocation graph is now
	unconditionally recorded in ALI files.
	* switch-b.adb (Scan_Binder_Switches): Scan switch -H.
	* doc/gnat_ugn/building_executable_programs_with_gnat.rst:
	Update the documentation on compiler switches related to
	elaboration.  Update the documentation on binder switches to
	include switch -H.
	* doc/gnat_ugn/elaboration_order_handling_in_gnat.rst: Update
	the documentation on elaboration order handling in GNAT.
	* gnat_ugn.texi: Regenerate.

From-SVN: r273280

In Ada, the name of operators contains a pair of double quotes, which
need to be properly escaped when the name appears in the JSON output of
-gnatR.

The change also ensures that formal parameters are not listed in the
layout information, since this information is not back-annotated for
them.

2019-07-09  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

	* repinfo.adb (List_Entities): Disregard formals altogether.
	(List_Name): Properly escape the double quote in the JSON
	output.

From-SVN: r273279

Under validity checking mode the compiler may silently skip generating
code to perform runtime range checks.

2019-07-09  Javier Miranda  <miranda@adacore.com>

gcc/ada/

	* exp_util.adb (Remove_Side_Effects): Preserve the
	Do_Range_Check flag.

gcc/testsuite/

	* gnat.dg/range_check3.adb, gnat.dg/range_check3_pkg.adb,
	gnat.dg/range_check3_pkg.ads: New testcase.

From-SVN: r273278

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* sinfo.ads: Refine comment for Do_Range_Check.

From-SVN: r273277

In the special GNATprove mode for proof of programs, expand the Enum_Rep
attribute reference so that a suitable static integer is in the AST
where required by the rest of analysis.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* exp_spark.adb (Expand_SPARK_N_Attribute_Reference): Expand
	attribute reference on Enum_Rep.

From-SVN: r273276

This patch removes a spurious bug on the use of the current instance of
a generic package G as the actual in a nested instantiation of a generic
unit GU that has a formal package whose generic_package name is G. This
is only legal if G has no generic formal part, and the formal package
declaration is declared with a box or without a formal_paxkage_actual
part.

2019-07-09  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

	* sem_ch12.adb (Instantiate_Formal_Package): Handle properly the
	case where the actual for a formal package in an instance is the
	current instance of an enclosing generic package.
	(Check_Formal_Packages): If the formal package declaration is
	box-initialized or lacks associations altogether, no internal
	instance was created to verify conformance, and there is no
	validating package to remove from tree.

gcc/testsuite/

	* gnat.dg/generic_inst5.adb, gnat.dg/generic_inst6.adb,
	gnat.dg/generic_inst6_g1-c.adb, gnat.dg/generic_inst6_g1-c.ads,
	gnat.dg/generic_inst6_g1.ads, gnat.dg/generic_inst6_i1.ads,
	gnat.dg/generic_inst6_i2.ads, gnat.dg/generic_inst6_x.ads: New
	testcases.

From-SVN: r273275

In GNATprove mode, subprograms with a body to inline should have been
filtered in Analyze_Subprogram_Body_Helper to match the conditions for
inlining subprograms in GNATprove. Prevent a call to Set_Body_To_Inline
in GNATprove mode that did not go through this filtering.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* freeze.adb (Build_Renamed_Body): Do not set body to inline in
	GNATprove mode.

From-SVN: r273274

In the special mode for GNATprove, expand the type of static expressions
like done during compilation, to both get suitable legality checks and
increase the precision of the formal analysis.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* exp_util.adb (Expand_Subtype_From_Expr): Still expand the type
	of static expressions in GNATprove_Mode.
	* sem_ch3.adb (Analyze_Object_Declaration): Remove obsolete
	special case for GNATprove_Mode.

From-SVN: r273273

2019-07-09  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

	* doc/gnat_rm/the_gnat_library.rst,
	doc/gnat_ugn/building_executable_programs_with_gnat.rst,
	erroutc.adb, libgnat/g-comlin.adb, libgnat/g-comlin.ads,
	libgnat/g-regexp.ads, libgnat/g-regpat.ads,
	libgnat/g-spipat.ads, libgnat/s-os_lib.ads,
	libgnat/s-regexp.ads: Reword "wild card" to "wildcard".
	* gnat_rm.texi, gnat_ugn.texi: Regenerate.

From-SVN: r273272

Allocator expressions and sub-expressions of (extension) aggregates are
implicitly the source of assignments in Ada. Thus, they should be moved
when of a deep type when checking ownership rules in SPARK.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* sem_spark.adb (Check_Expression): Handle correctly implicit
	assignments as part of allocators and (extension) aggregates.
	(Get_Root_Object): Adapt for new path expressions.
	(Is_Path_Expression): Return True for (extension) aggregate.

From-SVN: r273271

2019-07-09  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

	* einfo.ads: Fix a typo.

From-SVN: r273270

No impact on compilation with GCC.

2019-07-09  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

	* sem_util.adb (Scope_Within_Or_Same): Handle properly task
	bodies and protected bodies, so that local variables within have
	their proper scopes after these constructs have been rewritten
	during expansion. This patch resembles but is not identical to
	the code in Scope_Within.

From-SVN: r273269

2019-07-09  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

	* gnat1drv.adb (Adjust_Global_Switches): Set
	Dynamic_Elaboration_Checks to True in CodePeer mode.

From-SVN: r273268

Check for declaration of global variables prior to use in the ownership
checking for SPARK.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* sem_spark.adb (Get_Perm_Or_Tree): Issue an error when
	encountering unknown global variable.

From-SVN: r273267

Checking of the readable status of sub-expressions occurring in the
target path of an assignment should occur before the right-hand-side is
moved or borrowed or observed.

There is no impact on compilation.

2019-07-09  Yannick Moy  <moy@adacore.com>

gcc/ada/

	* sem_spark.adb (Check_Expression): Change signature to take an
	Extended_Checking_Mode, for handling read permission checking of
	sub-expressions in an assignment.
	(Check_Parameter_Or_Global): Adapt to new behavior of
	Check_Expression for mode Assign.
	(Check_Safe_Pointers): Do not analyze generic bodies.
	(Check_Assignment): Separate checking of the target of an
	assignment.

From-SVN: r273266

This changes the -gnatRj output from a concatenation of entities to an
array of entities, thus making it strictly conforming JSON and easier to
be parsed by means of GNATColl or Python.

2019-07-09  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

	* repinfo.ads (JSON format): Adjust.
	* repinfo.adb (Need_Blank_Line): Rename to...
	(Need_Separator): ...this.
	(Blank_Line): Rename to...
	(Write_Separator): ...this and add JSON specific handling.
	(List_Array_Info): Adjust to above renaming.
	(List_Object_Info): Likewise.
	(List_Record_Info): Likewise.
	(List_Subprogram_Info): Likewise.
	(List_Type_Info): Likewise.
	(List_Entities): Do not set Need_Blank_Line.
	(List_Rep_Info): Set Need_Separator and add JSON specific
	handling. Output a single JSON stream in the  normal case.

From-SVN: r273265

2019-07-09  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

	* doc/gnat_ugn/the_gnat_compilation_model.rst: Update doc on
	-fdump-ada-spec now that we generate Ada 2012.
	* gnat_ugn.texi: Regenerate.

From-SVN: r273264

2019-07-09  Martin Liska  <mliska@suse.cz>

	* lto-compress.c (lto_normalized_zstd_level): Do not use
	ZSTD_CLEVEL_DEFAULT as it is not default in old releases
	of libzstd.  One can use 0 as a default compression level.

From-SVN: r273263

2019-07-09  Martin Liska  <mliska@suse.cz>

	* doc/invoke.texi: Add link from -fprofile-dir option.
	Use better wording for 'gcno filename'.

From-SVN: r273262

PR middle-end/71924 - missing -Wreturn-local-addr returning alloca result
PR middle-end/90549 - missing -Wreturn-local-addr maybe returning an address of a local array plus offset

gcc/ChangeLog:

	PR middle-end/71924
	PR middle-end/90549
	* gimple-ssa-isolate-paths.c (isolate_path): Add attribute.  Update
	comment.
	(args_loc_t): New type.
	(args_loc_t, locmap_t): same.
	(diag_returned_locals): New function.
	(is_addr_local): Same.
	(handle_return_addr_local_phi_arg, warn_return_addr_local): Same.
	(find_implicit_erroneous_behavior): Call warn_return_addr_local_phi_arg.
	(find_explicit_erroneous_behavior): Call warn_return_addr_local.

gcc/testsuite/ChangeLog:

	PR middle-end/71924
	PR middle-end/90549
	* gcc.c-torture/execute/return-addr.c: New test.
	* gcc.dg/Wreturn-local-addr-2.c: New test.
	* gcc.dg/Wreturn-local-addr-4.c: New test.
	* gcc.dg/Wreturn-local-addr-5.c: New test.
	* gcc.dg/Wreturn-local-addr-6.c: New test.
	* gcc.dg/Wreturn-local-addr-7.c: New test.
	* gcc.dg/Wreturn-local-addr-8.c: New test.
	* gcc.dg/Wreturn-local-addr-9.c: New test.
	* gcc.dg/Wreturn-local-addr-10.c: New test.
	* gcc.dg/Walloca-4.c: Handle expected warnings.
	* gcc.dg/pr41551.c: Same.
	* gcc.dg/pr59523.c: Same.
	* gcc.dg/tree-ssa/pr88775-2.c: Same.
	* gcc.dg/tree-ssa/alias-37.c: Same.
	* gcc.dg/winline-7.c: Same.

From-SVN: r273261

From-SVN: r273260

tree-vect-stmts.c (scan_operand_equal_p): Look through MEM_REF with SSA_NAME address of POINTER_PLUS_EXPR.

	* tree-vect-stmts.c (scan_operand_equal_p): Look through MEM_REF
	with SSA_NAME address of POINTER_PLUS_EXPR.  Handle MULT_EXPR
	and casts in offset when different, both through gimple stmts
	and through trees.  Rewritten using loops to minimize code duplication
	for each operand.

	* g++.dg/vect/simd-6.cc: Replace xfail with target x86.
	* g++.dg/vect/simd-9.cc: Likewise.

	* testsuite/libgomp.c++/scan-13.C: Replace xfail with target x86.
	* testsuite/libgomp.c++/scan-16.C: Likewise.

From-SVN: r273249

re PR c++/91110 (ICE: tree check: expected class 'type', have 'exceptional' (error_mark) in cp_omp_mappable_type_1, at cp/decl2.c:1421)

	PR c++/91110
	* decl2.c (cp_omp_mappable_type_1): Don't emit any note for
	error_mark_node type.

	* g++.dg/gomp/pr91110.C: New test.

From-SVN: r273248

	* emit-rtl.c (set_insn_locations): New function moved from...
	* function.c (set_insn_locations): ...here.
	* ira-emit.c (emit_moves): Propagate location of the first instruction
	to the inserted move instructions.
	* reg-stack.c (compensate_edge): Set the location if the sequence is
	inserted on the edge.
	* rtl.h (set_insn_locations): Declare.

From-SVN: r273247

The function rs6000_machine_from_flags chooses what .machine string to
used based on the rs6000_isa_flags flags.  For that it checks for each
ISA level if something for its ISA_*_MASKS is selected.

This does not work for GFXOPT and GPOPT: these are set as flags in
ISA_2_5_MASKS_SERVER, but they aren't actually new there, they just
are not selected by default for older ISAs (they were optional).

This patch makes OPTION_MASK_PPC_GFXOPT and OPTION_MASK_PPC_GPOPT not
influence the .machine selection.


	* config/rs6000/rs6000.c (rs6000_machine_from_flags): Ignore
	OPTION_MASK_PPC_GFXOPT and OPTION_MASK_PPC_GPOPT for selecting the
	.machine string.

From-SVN: r273246

This testcase tests that with -mcpu=power8 we do not generate any
mtvsr* instructions, and we do the copy with {l,st}xvd2x.


gcc/testsuite/
	PR rtl-optimization/88233
	* gcc.target/powerpc/pr88233.c: New testcase.

From-SVN: r273245

Currently the second lower-subreg pass is run right before RA.  This
is much too late to be very useful.  At least for targets that do not
have RTL patterns for operations on multi-register modes it is a lot
better to split patterns earlier, before combine and all related
passes.

This adds an option -fsplit-wide-types-early that does that, and
enables it by default for rs6000.


	PR rtl-optimization/88233
	* common.opt (fsplit-wide-types-early): New option.
	* common/config/rs6000/rs6000-common.c
	(rs6000_option_optimization_table): Add OPT_fsplit_wide_types_early for
	OPT_LEVELS_ALL.
	* doc/invoke.texi (Optimization Options): Add -fsplit-wide-types-early.
	* lower-subreg.c (pass_lower_subreg2::gate): Add test for
	flag_split_wide_types_early.
	(pass_data_lower_subreg3): New.
	(pass_lower_subreg3): New.
	(make_pass_lower_subreg3): New.
	* passes.def (pass_lower_subreg2): Move after the loop passes.
	(pass_lower_subreg3): New, inserted where pass_lower_subreg2 was.
	* tree-pass.h (make_pass_lower_subreg2): Move up, to its new place in
	the pass pipeline; its previous place is taken by ...
	(make_pass_lower_subreg3): ... this.

From-SVN: r273240

The gcc.c-torture/execute/builtins/lib directory contains a reimplementation
of many C library string functions, which causes non-trivial register allocation
bugs with LTO and static linked libraries.  To fix this long-standing test
issue, turn off ipa-ra which avoids the register corruption across calls.  All
builtin torture tests now pass on aarch64-none-elf.  Committed as obvious.

    testsuite/
	PR testsuite/91059
	PR testsuite/78529
	* gcc.c-torture/execute/builtins/builtins.exp: Add -fno-ipa-ra.

From-SVN: r273238

Define s390_shift_truncation_mask to allow the optabs optimization

    sh = (64 - sh)
 -> sh = -sh

for a rotation operation.

gcc/ChangeLog:

2019-07-08  Robin Dapp  <rdapp@linux.ibm.com>

	* config/s390/s390.c (s390_shift_truncation_mask): Define.
	(TARGET_SHIFT_TRUNCATION_MASK): Define.

gcc/testsuite/ChangeLog:

2019-07-08  Robin Dapp  <rdapp@linux.ibm.com>

	* gcc.target/s390/rotate-truncation-mask.c: New test.

From-SVN: r273237

Add s390_valid_shift_count to determine the validity of a
shift-count operand.  This is used to replace increasingly
complex substitutions that should have allowed address-style
shift-count handling, an and mask as well as no-op subregs
on the operand.

gcc/ChangeLog:

2019-07-08  Robin Dapp  <rdapp@linux.ibm.com>

        * config/s390/constraints.md: Add new jsc constraint.
        * config/s390/predicates.md: New predicates.
        * config/s390/s390-protos.h (s390_valid_shift_count): New function.
        * config/s390/s390.c (s390_valid_shift_count): New function.
        (print_shift_count_operand): Use s390_valid_shift_count.
        (print_operand): Likewise.
        * config/s390/s390.md: Use new predicate.
        * config/s390/subst.md: Remove addr_style_op and masked_op substs.
        * config/s390/vector.md: Use new predicate.

2019-07-08  Robin Dapp  <rdapp@linux.ibm.com>

        * gcc.target/s390/combine-rotate-modulo.c: New test.
        * gcc.target/s390/combine-shift-rotate-add-mod.c: New test.
        * gcc.target/s390/vector/combine-shift-vec.c: New test.

From-SVN: r273236

2019-07-08  Kito Cheng  <kito.cheng@sifive.com>

	* MAINTAINERS (Write After Approval): Remove myself, already listed in
	RISC-V port maitainer.

From-SVN: r273235

2019-07-08  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/91108
	* tree-ssa-sccvn.c: Include builtins.h.
	(vn_reference_lookup_3): Use only alignment constraints to
	verify same-valued store disambiguation.

	* gcc.dg/tree-ssa/ssa-fre-61.c: Adjust back.
	* gcc.dg/tree-ssa/ssa-fre-78.c: New testcase.

From-SVN: r273232

	* testsuite/gcc.dg/vect/slp-reduc-sad.c (uint32_t):
	Remove unused declaration.

From-SVN: r273231

Fixes github.com/riscv/riscv-gcc issue #161.  We were accidentally using
BITS_PER_WORD to compute shift counts when we should have been using the
bitsize of the operand modes.  This was wrong when we had an SImode shift
and a 64-bit target.

	Andrew Waterman  <andrew@sifive.com>
	gcc/
	* config/riscv/riscv.md (lshrsi3_zero_extend_3+1): Use operands[1]
	bitsize	instead of BITS_PER_WORD.
	gcc/testsuite/
	* gcc.target/riscv/shift-shift-2.c: Add one more test.

From-SVN: r273230

2019-07-08  Martin Liska  <mliska@suse.cz>

	* collect2.c (defined): Revert to before r254460.
	(scan_prog_file): Revert to before r254460.

From-SVN: r273229

2019-07-08  Paolo Carlini  <paolo.carlini@oracle.com>

	PR c++/65143
	* g++.dg/tree-ssa/final2.C: New.
	* g++.dg/tree-ssa/final3.C: Likewise.

From-SVN: r273228