platform/upstream/gcc.git
6 years ago[Ada] Minor comment improvement
Ed Schonberg [Thu, 31 May 2018 10:46:39 +0000 (10:46 +0000)]
[Ada] Minor comment improvement

2018-05-31  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* einfo.ads: Improve documentation for the Is_Unsigned_Type entity
flag.

From-SVN: r261004

6 years ago[Ada] Simplify call to Unique_Defining_Entity on protected entry declarations
Piotr Trojanek [Thu, 31 May 2018 10:46:34 +0000 (10:46 +0000)]
[Ada] Simplify call to Unique_Defining_Entity on protected entry declarations

Calling Unique_Defining_Entity on protectected entry declarations is
equivalent to calling a simpler Defining_Entity; use the simpler routine.

Simplification only; semantics unaffected, so no test provided.

2018-05-31  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

* sem_prag.adb (Analyze_Pragma): Replace call to Unique_Defining_Entity
with a semantically equivalent call to Defining_Entity.

From-SVN: r261003

6 years ago[Ada] Set Etype on rewriteen Max_Queue_Length expressions
Piotr Trojanek [Thu, 31 May 2018 10:46:17 +0000 (10:46 +0000)]
[Ada] Set Etype on rewriteen Max_Queue_Length expressions

Rewriting of Max_Queue_Length expression into N_Integer_Literal should probably
be done in expansion and not in analysis, but anyway it should not strip the
expression from its Etype because backends (e.g. GNATprove) expect that Etype
to be present.

No frontend test is provided, because GNAT doesn't care about the missing
Etype decoration. This patch allows to simplify AST processing in GNATprove.

2018-05-31  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

* sem_prag.adb (Analyze_Pragma): Set Etype on the rewritten
Max_Queue_Length expression.

From-SVN: r261002

6 years ago[Ada] Update gnatelim documentation wrt. compilation of main unit
Sergey Rybin [Thu, 31 May 2018 10:46:12 +0000 (10:46 +0000)]
[Ada] Update gnatelim documentation wrt. compilation of main unit

2018-05-31  Sergey Rybin  <rybin@adacore.com>

gcc/ada/

* doc/gnat_ugn/gnat_and_program_execution.rst: gnatelim does not need
that the main unit to be built before the call, now it computes the
closure of the main unit itself.
* gnat_ugn.texi: Regenerate.

From-SVN: r261001

6 years ago[Ada] Fix tracking of the first item in Repinfo
Eric Botcazou [Thu, 31 May 2018 10:46:07 +0000 (10:46 +0000)]
[Ada] Fix tracking of the first item in Repinfo

2018-05-31  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* repinfo.adb (List_Structural_Record_Layout): Set First to false
after having listed the fields of the parent type, if any.

From-SVN: r261000

6 years ago[Ada] Fix strange behavior of Object_Size for packed array subtype
Eric Botcazou [Thu, 31 May 2018 10:46:02 +0000 (10:46 +0000)]
[Ada] Fix strange behavior of Object_Size for packed array subtype

2018-05-31  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* gcc-interface/decl.c (gnat_to_gnu_entity) <discrete_type>: Do not
deal with the ___XP suffix for packed array types here...
<E_Array_Subtype>: ...or here when processing the implementation type
but when processing the original type instead.  Do not reuse the DECL
of the implementation type for the original type.  Tidy up.

From-SVN: r260999

6 years ago[Ada] Post warning on object size clause for subtype
Eric Botcazou [Thu, 31 May 2018 10:45:57 +0000 (10:45 +0000)]
[Ada] Post warning on object size clause for subtype

This ensures that a warning for an object size clause present on a subtype
is posted on the clause and not on a size clause present on the type.

2018-05-31  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* einfo.ads (Object_Size_Clause): Declare.
* einfo.adb (Object_Size_Clause): New function.
* gcc-interface/utils.c (maybe_pad_type): Test Has_Size_Clause before
retrieving Size_Clause and post the warning on the object size clause
if Has_Object_Size_Clause is true.

gcc/testsuite/

* gnat.dg/size_clause1.adb: New testcase.

From-SVN: r260998

6 years ago[Ada] Fix compiler crash for tagged private types
Javier Miranda [Thu, 31 May 2018 10:45:51 +0000 (10:45 +0000)]
[Ada] Fix compiler crash for tagged private types

2018-05-31  Javier Miranda  <miranda@adacore.com>

gcc/ada/

* sem_util.ads, sem_util.adb (Find_Primitive_Eq): New subprogram.
* exp_ch4.adb (Expand_Composite_Equality): Use the new subprogram
Find_Primitive_Eq to search for the primitive of types whose underlying
type is a tagged type.

gcc/testsuite/

* gnat.dg/tagged1.adb, gnat.dg/tagged1.ads: New testcase.

From-SVN: r260997

6 years ago[Ada] Fix check on placement of multiple loop (in)variant pragmas
Yannick Moy [Thu, 31 May 2018 10:45:05 +0000 (10:45 +0000)]
[Ada] Fix check on placement of multiple loop (in)variant pragmas

Loop (in)variants should appear next to each other, which is checked by GNAT
frontend. As statements inserted during expansion may break this contiguity,
GNAT recognizes specially such statements which originate in loop pragmas. In
some cases, this special treatment was not properly put in place, which lead to
spurious errors being issued.

2018-05-31  Yannick Moy  <moy@adacore.com>

gcc/ada/

* sem_prag.adb (Analyze_Pragma.Check_Loop_Pragma_Placement): Inverse
order of treatment between nodes recognized as loop pragmas (or
generated from one) and block statements.

From-SVN: r260996

6 years ago[Ada] Posix 2008: reimplement System.OS_Primitives.Clock using clock_gettime
Doug Rupp [Thu, 31 May 2018 10:44:56 +0000 (10:44 +0000)]
[Ada] Posix 2008: reimplement System.OS_Primitives.Clock using clock_gettime

gettimeofday is deprecated in Posix 2008, clock_gettime is the recommended
replacement.

2018-05-31  Doug Rupp  <rupp@adacore.com>

gcc/ada/

* libgnat/s-osprim__posix2008.adb (Clock): Implement using
clock_gettime.

From-SVN: r260995

6 years ago[Ada] Unnesting: add a predicate to help expansion of task and protected types
Ed Schonberg [Thu, 31 May 2018 10:44:51 +0000 (10:44 +0000)]
[Ada] Unnesting: add a predicate to help expansion of task and protected types

2018-05-31  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.ads, exp_unst.adb (In_Synchronized_Unit): New predicate to
mark subprograms that cannot be eliminated because they must be treated
as reachable from outside the current unit. This includes entry bodies
and protected operations.

From-SVN: r260994

6 years ago[Ada] Convert operands of generated code to Standard.Unsigned for modular ops.
Ed Schonberg [Thu, 31 May 2018 10:44:45 +0000 (10:44 +0000)]
[Ada] Convert operands of generated code to Standard.Unsigned for modular ops.

2018-05-31  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_ch4.adb (Expand_Modular_Addition, Expand_Modular_Subtraction):
Convert all operands of the generated code to Standard.Unsigned, to
prevent spurious visibility errors. This is needed when the generated
expansion involves a modular index type of an array type, and the index
type is not immediately visible.

From-SVN: r260993

6 years agore PR sanitizer/86012 (libsanitizer build failure on sparc64-linux-gnu)
Matthias Klose [Thu, 31 May 2018 09:57:33 +0000 (09:57 +0000)]
re PR sanitizer/86012 (libsanitizer build failure on sparc64-linux-gnu)

2018-05-31  Matthias Klose  <doko@ubuntu.com>

        PR sanitizer/86012
        * sanitizer_common/sanitizer_platform_limits_posix.cc: Define
        SIZEOF_STRUCT_USTAT for 32bit sparc.

From-SVN: r260990

6 years agoPatch implementing vld1_*_x3, vst1_*_x2 and vst1_*_x3 intrinsics for AARCH64 for...
Sameera Deshpande [Thu, 31 May 2018 08:46:20 +0000 (08:46 +0000)]
Patch implementing vld1_*_x3, vst1_*_x2 and vst1_*_x3 intrinsics for AARCH64 for all types.

From-SVN: r260989

6 years agoDaily bump.
GCC Administrator [Thu, 31 May 2018 00:16:46 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r260988

6 years agomsp430.c (msp430_output_labelref): Prepend user_label_prefix to name.
Jozef Lawrynowicz [Wed, 30 May 2018 23:33:49 +0000 (23:33 +0000)]
msp430.c (msp430_output_labelref): Prepend user_label_prefix to name.

* config/msp430/msp430.c (msp430_output_labelref): Prepend
user_label_prefix to name.

From-SVN: r260983

6 years agotree-core.h: Update comment about the format of NAME string passed to handler in...
Jozef Lawrynowicz [Wed, 30 May 2018 23:32:31 +0000 (23:32 +0000)]
tree-core.h: Update comment about the format of NAME string passed to handler in attribute_spec.

* tree-core.h: Update comment about the format of NAME string
passed to handler in attribute_spec.

From-SVN: r260982

6 years agomsp430.md: Remove erroneous subreg expression from zero_extendqisi2 insn pattern.
Jozef Lawrynowicz [Wed, 30 May 2018 23:26:03 +0000 (23:26 +0000)]
msp430.md: Remove erroneous subreg expression from zero_extendqisi2 insn pattern.

* gcc/config/msp430/msp430.md: Remove erroneous subreg expression from
zero_extendqisi2 insn pattern. Remove msp430x ISA restriction on
zero_extend{q,h}isi2.

From-SVN: r260981

6 years ago* crtstuff.c: Remove declaration of _Jv_RegisterClasses.
Rasmus Villemoes [Wed, 30 May 2018 23:21:48 +0000 (01:21 +0200)]
* crtstuff.c: Remove declaration of _Jv_RegisterClasses.

From-SVN: r260980

6 years agoPR c++/77777 improve location for diagnostic
Jonathan Wakely [Wed, 30 May 2018 23:13:48 +0000 (00:13 +0100)]
PR c++/77777 improve location for diagnostic

Pass in the location of the invalid expression, not the next input
location (which might be a comma or closing parenthesis on a different
line).

gcc/cp:

PR c++/77777
* call.c (resolve_args): Use location of expression, not current input
location.

gcc/testsuite:

PR c++/77777
* g++.dg/diagnostic/pr77777.C: New test.

From-SVN: r260979

6 years agore PR c/55976 (-Werror=return-type should error on returning a value from a void...
David Pagan [Wed, 30 May 2018 22:55:38 +0000 (22:55 +0000)]
re PR c/55976 (-Werror=return-type should error on returning a value from a void function)

2018-05-30  David Pagan  <dave.pagan@oracle.com>

PR c/55976
* c-decl.c (grokdeclarator): Update check for return type warnings.
(start_function): Likewise.
(finish_function): Likewise.
* c-typeck.c (c_finish_return): Update check for return type warnings.
Pass OPT_Wreturn_type to pedwarn when appropriate.

PR c/55976
* c-opts.c (c_common_post_options): Set default for warn_return_type
for C++/C++ with ObjC extensions only. For C, makes it possible to
differentiate between default (no option), -Wreturn-type, and
-Wno-return-type.

PR c/55976
* gcc.dg/noncompile/pr55976-1.c: New test.
* gcc.dg/noncompile/pr55976-2.c: New test.

From-SVN: r260978

6 years agoextend.texi: Document some architecture specific constraints and sort entries.
Borislav Petkov [Wed, 30 May 2018 22:49:25 +0000 (22:49 +0000)]
extend.texi: Document some architecture specific constraints and sort entries.

* doc/extend.texi: Document some architecture specific
constraints and sort entries.

From-SVN: r260977

6 years agoPR middle-end/85369 - no -Wstringop-overflow for a strcpy / stpcpy call with a nonstr...
Martin Sebor [Wed, 30 May 2018 22:24:43 +0000 (22:24 +0000)]
PR middle-end/85369 - no -Wstringop-overflow for a strcpy / stpcpy call with a nonstring pointer when providing movstr pattern

gcc/ChangeLog:

PR middle-end/85369
* builtins.c (expand_builtin_stpcpy_1): New function.
(expand_builtin_stpcpy): Call it, and call maybe_warn_nonstring_arg
only if the former succeeds.

From-SVN: r260976

6 years agoChangeLog fix
Jason Merrill [Wed, 30 May 2018 19:42:52 +0000 (15:42 -0400)]
ChangeLog fix

From-SVN: r260974

6 years agoDo not warn about zero-as-null when NULL is used.
Ville Voutilainen [Wed, 30 May 2018 19:33:38 +0000 (22:33 +0300)]
Do not warn about zero-as-null when NULL is used.

gcc/cp/

Do not warn about zero-as-null when NULL is used.
* call.c (conversion_null_warnings): Check for pointer
types converted from zero constants.
(convert_like_real): Add a warning sentinel at the end.
* tree.c (maybe_warn_zero_as_null_pointer_constant): Also
check null_node_p.

testsuite/

Do not warn about zero-as-null when NULL is used.
* g++.dg/warn/Wzero-as-null-pointer-constant-7.C: New.

From-SVN: r260973

6 years agoPR c++/85807 - ICE with call in template NSDMI.
Jason Merrill [Wed, 30 May 2018 19:31:11 +0000 (15:31 -0400)]
PR c++/85807 - ICE with call in template NSDMI.

* init.c (get_nsdmi): Use push_to/pop_from_top_level.
* tree.c (bot_manip): Don't set_flags_from_callee in a template.

From-SVN: r260972

6 years agopatch to add support of ARMv8.4 in saphira
Sameera Deshpande [Wed, 30 May 2018 18:40:03 +0000 (18:40 +0000)]
patch to add support of ARMv8.4 in saphira

From-SVN: r260971

6 years agosimple-object.h (simple_object_copy_lto_debug_sections): Add rename parameter.
Jan Hubicka [Wed, 30 May 2018 17:57:50 +0000 (19:57 +0200)]
simple-object.h (simple_object_copy_lto_debug_sections): Add rename parameter.

* simple-object.h (simple_object_copy_lto_debug_sections): Add rename
parameter.

From-SVN: r260970

6 years agoPR c++/85873 - constant initializer_list array not in .rodata.
Jason Merrill [Wed, 30 May 2018 17:33:06 +0000 (13:33 -0400)]
PR c++/85873 - constant initializer_list array not in .rodata.

* tree.c (build_target_expr): Set TREE_READONLY.
* call.c (set_up_extended_ref_temp): Set TREE_READONLY.

From-SVN: r260969

6 years agoImprove error recovery for structured binding in condition.
Jason Merrill [Wed, 30 May 2018 17:32:54 +0000 (13:32 -0400)]
Improve error recovery for structured binding in condition.

* parser.c (cp_parser_check_condition_declarator): Handle
cp_error_declarator.

From-SVN: r260968

6 years agoRemove accidental commit.
Jan Hubicka [Wed, 30 May 2018 17:10:03 +0000 (17:10 +0000)]
Remove accidental commit.

From-SVN: r260967

6 years ago20111213-1_0.c: Fix template.
Jan Hubicka [Wed, 30 May 2018 17:07:36 +0000 (19:07 +0200)]
20111213-1_0.c: Fix template.

* gcc.dg/lto/20111213-1_0.c: Fix template.
* gcc.dg/lto/20100603-1_0.c: Fix template.
* gcc.dg/lto/20100603-2_0.c: Fix template.
* gcc.dg/lto/20100603-3_0.c: Fix template.

From-SVN: r260966

6 years ago* doc/invoke.texi (-flinker-output): Document
Jan Hubicka [Wed, 30 May 2018 16:43:59 +0000 (18:43 +0200)]
* doc/invoke.texi (-flinker-output): Document

From-SVN: r260965

6 years agopasses.c (ipa_write_summaries): Only modify statements if body is in memory.
Jan Hubicka [Wed, 30 May 2018 16:43:11 +0000 (16:43 +0000)]
passes.c (ipa_write_summaries): Only modify statements if body is in memory.

* passes.c (ipa_write_summaries): Only modify statements if body
is in memory.
* cgraphunit.c (ipa_passes): Also produce intermeidate code when
incrementally linking.
(ipa_passes): Likewise.
* lto-cgraph.c (lto_output_node): When incrementally linking do not
pass down resolution info.
* common.opt (flag_incremental_link): Update info.
* gcc.c (plugin specs): Turn flinker-output=* to
-plugin-opt=-linker-output-known
* toplev.c (compile_file): Also cut compilation when doing incremental
link.
* flag-types. (enum lto_partition_model): Add
LTO_LINKER_OUTPUT_NOLTOREL.
(invoke.texi): Add -flinker-output docs.
* ipa.c (symbol_table::remove_unreachable_nodes): Handle LTO incremental
link same way as WPA; do not stream in dead initializers.

From-SVN: r260964

6 years agopasses.c (ipa_write_summaries): Only modify statements if body is in memory.
Jan Hubicka [Wed, 30 May 2018 16:42:41 +0000 (18:42 +0200)]
passes.c (ipa_write_summaries): Only modify statements if body is in memory.

* passes.c (ipa_write_summaries): Only modify statements if body
is in memory.
* cgraphunit.c (ipa_passes): Also produce intermeidate code when
incrementally linking.
(ipa_passes): Likewise.
* lto-cgraph.c (lto_output_node): When incrementally linking do not
pass down resolution info.
* common.opt (flag_incremental_link): Update info.
* gcc.c (plugin specs): Turn flinker-output=* to
-plugin-opt=-linker-output-known
* toplev.c (compile_file): Also cut compilation when doing incremental
link.
* flag-types. (enum lto_partition_model): Add
LTO_LINKER_OUTPUT_NOLTOREL.
(invoke.texi): Add -flinker-output docs.
* ipa.c (symbol_table::remove_unreachable_nodes): Handle LTO incremental
link same way as WPA; do not stream in dead initializers.

* dwarf2out.c (dwarf2out_die_ref_for_decl,
darf2out_register_external_decl): Support incremental link.

* lang.opt (lto_linker_output): Add nolto-rel.
* lto-lang.c (lto_post_options): Handle LTO_LINKER_OUTPUT_REL
and LTO_LINKER_OUTPUT_NOLTOREL.
(lto_init): Generate lto when doing incremental link.
* lto.c (lto_precess_name): Add lto1-inclink.

* testsuite/g++.dg/lto/20081109-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081118_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081119-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081120-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081120-2_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081123_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081204-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20081219_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20090302_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20090313_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20091002-2_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20091002-3_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20091026-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20100724-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20101010-4_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20101015-2_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/20110311-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr45621_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr48042_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr48354-1_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr54625-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr54625-2_0.c: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/lto/pr68811_0.C: Add -flinker-output=nolto-rel.
* testsuite/g++.dg/torture/pr43760.C: New test. Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081120-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081120-2_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081126_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081204-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081204-2_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081212-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20081224_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20090116_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20090126-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20090126-2_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20090206-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20090219_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20091013-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20091014-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20091015-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20091016-1_0.c: Add -flinker-output=nolto-rel.
* testsuite/gcc.dg/lto/20091020-1_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20091020-2_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20091027-1_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20100426_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20100430-1_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20100603-1_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20100603-2_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20100603-3_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/20111213-1_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr45736_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr52634_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr54702_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr59323-2_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr59323_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr60820_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr81406_0.c: Add -flinker-output-nolto-rel.
* testsuite/gcc.dg/lto/pr83388_0.c: Add -flinker-output-nolto-rel.
* testsuite/gfortran.dg/lto/20091016-1_0.f90: Add -flinker-output-nolto-rel.
* testsuite/gfortran.dg/lto/20091028-1_0.f90: Add -flinker-output-nolto-rel.
* testsuite/gfortran.dg/lto/20091028-2_0.f90: Add -flinker-output-nolto-rel.
* testsuite/gfortran.dg/lto/pr46911_0.f: Add -flinker-output-nolto-rel.
* testsuite/gfortran.dg/lto/pr47839_0.f90: Add -flinker-output-nolto-rel.

From-SVN: r260963

6 years ago* lto-opts.c (lto_write_options): Skip OPT_dumpdir, OPT_fresolution_.
Jan Hubicka [Wed, 30 May 2018 16:37:53 +0000 (18:37 +0200)]
* lto-opts.c (lto_write_options): Skip OPT_dumpdir, OPT_fresolution_.

From-SVN: r260962

6 years agolto-wrapper.c (debug_objcopy): Add rename parameter; pass it down to simple_object_co...
Jan Hubicka [Wed, 30 May 2018 16:36:55 +0000 (18:36 +0200)]
lto-wrapper.c (debug_objcopy): Add rename parameter; pass it down to simple_object_copy_lto_debug_sections.

* lto-wrapper.c (debug_objcopy): Add rename parameter; pass
it down to simple_object_copy_lto_debug_sections.
(run_gcc): Determine incremental LTO link time and configure
lto1 into non-wpa mode, disable renaming of debug sections.

From-SVN: r260961

6 years agolto-plugin.c: (non_claimed_files): New static var.
Jan Hubicka [Wed, 30 May 2018 16:34:54 +0000 (18:34 +0200)]
lto-plugin.c: (non_claimed_files): New static var.

* lto-plugin.c: (non_claimed_files): New static var.
(linker_ouput_known): New static var.
(all_symbols_read_handler): When user specifies linker output do not
imply it; output warning when nonlto-rel mode is forced.
(claim_file_header): Record number of nonclaimed files.
(process_option): Remember if linker output is known

From-SVN: r260960

6 years agoextend.texi (PowerPC AltiVec Built-in Functions): Remove descriptions of various...
Kelvin Nilsen [Wed, 30 May 2018 16:20:16 +0000 (16:20 +0000)]
extend.texi (PowerPC AltiVec Built-in Functions): Remove descriptions of various incorrectly documented functions.

gcc/ChangeLog:

2018-05-30  Kelvin Nilsen  <kelvin@gcc.gnu.org>

* doc/extend.texi (PowerPC AltiVec Built-in Functions): Remove
descriptions of various incorrectly documented functions.

From-SVN: r260959

6 years agoFixing ChangeLog entry for earlier commit.
Andre Vieira [Wed, 30 May 2018 16:04:03 +0000 (16:04 +0000)]
Fixing ChangeLog entry for earlier commit.

From-SVN: r260958

6 years agoReverting r260635
Andre Vieira [Wed, 30 May 2018 15:59:14 +0000 (15:59 +0000)]
Reverting r260635

    gcc
    2018-05-30  Andre Vieira  <andre.simoesdiasvieira@arm.com>

2018-05-24  Andre Vieira  <andre.simoesdiasvieira@arm.com>

        PR target/83009
Revert:
        * config/aarch64/predicates.md (aarch64_mem_pair_lanes_operand): Make
        address check not strict.

    gcc/testsuite
    2018-05-30  Andre Vieira  <andre.simoesdiasvieira@arm.com>

2018-05-24  Andre Vieira  <andre.simoesdiasvieira@arm.com>
Revert
        PR target/83009
        * gcc/target/aarch64/store_v2vec_lanes.c: Add extra tests.
    git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@260635 138bc75d-0d04-0410-961f-82ee72b054a4

From-SVN: r260957

6 years agosimple-object.h (simple_object_copy_lto_debug_sections): Add rename parameter.
Jan Hubicka [Wed, 30 May 2018 15:26:26 +0000 (17:26 +0200)]
simple-object.h (simple_object_copy_lto_debug_sections): Add rename parameter.

* simple-object.h (simple_object_copy_lto_debug_sections): Add rename
parameter.
* simple-object.c (handle_lto_debug_sections): Add rename parameter.
(handle_lto_debug_sections_rename): New function.
(handle_lto_debug_sections_norename): New function.
(simple_object_copy_lto_debug_sections): Add rename parameter.

From-SVN: r260956

6 years agore PR tree-optimization/85964 (compile time hog w/ -O3 -ftracer -fno-guess-branch...
Richard Biener [Wed, 30 May 2018 14:55:04 +0000 (14:55 +0000)]
re PR tree-optimization/85964 (compile time hog w/ -O3 -ftracer -fno-guess-branch-probability)

2018-05-30  Richard Biener  <rguenther@suse.de>

PR tree-optimization/85964
* tracer.c (better_p): Drop initialized count check, we only
call the function with initialized counts now.
(find_best_successor): Do find a best edge if one
has uninitialized count.
(find_best_predecessor): Likewise.  Do BB frequency check only
if count is initialized.

From-SVN: r260954

6 years ago[AArch64] Improve LDP/STP generation that requires a base register
Jackson Woodruff [Wed, 30 May 2018 10:59:25 +0000 (10:59 +0000)]
[AArch64] Improve LDP/STP generation that requires a base register

This patch generalizes the formation of LDP/STP that require a base register.

In AArch64, LDP/STP instructions have different sized immediate offsets than
normal LDR/STR instructions. This part of the backend attempts to spot groups
of four LDR/STR instructions that can be turned into LDP/STP instructions by
using a base register.

Previously, we would only accept address pairs that were ordered in ascending
or descending order, and only strictly sequential loads/stores. In fact, the
instructions that we generate from this should be able to consider any order
of loads or stores (provided that they can be re-ordered). They should also be
able to accept non-sequential loads and stores provided that the two pairs of
addresses are amenable to pairing. The current code is also overly restrictive
on the range of addresses that are accepted, as LDP/STP instructions may take
negative offsets as well as positive ones.

This patch improves that by allowing us to accept all orders of loads/stores
that are valid, and extending the range that the LDP/STP addresses can reach.

2017-05-30  Jackson Woodruff  <jackson.woodruff@arm.com>

* config/aarch64/aarch64.c (aarch64_host_wide_int_compare): New.
(aarch64_ldrstr_offset_compare): New.
(aarch64_operands_adjust_ok_for_ldpstp): Update to consider all
load/store orderings.
(aarch64_gen_adjusted_ldpstp): Likewise.

* gcc.target/aarch64/simd/ldp_stp_9: New.
* gcc.target/aarch64/simd/ldp_stp_10: New.
* gcc.target/aarch64/simd/ldp_stp_11: New.
* gcc.target/aarch64/simd/ldp_stp_12: New.

From-SVN: r260952

6 years ago[AArch64] Fix aarch64_ira_change_pseudo_allocno_class
Wilco Dijkstra [Wed, 30 May 2018 10:31:21 +0000 (10:31 +0000)]
[AArch64] Fix aarch64_ira_change_pseudo_allocno_class

A recent commit removing '*' from the md files caused a large regression in
h264ref.  It turns out aarch64_ira_change_pseudo_allocno_class is no longer
effective after the SVE changes, and the combination results in the regression.
This patch fixes it by explicitly checking for a subset of GENERAL_REGS and
FP_REGS.  Add a missing ? to aarch64_get_lane to fix a failure in the testsuite.

    gcc/
* config/aarch64/aarch64.c (aarch64_ira_change_pseudo_allocno_class):
Check for subset of GENERAL_REGS and FP_REGS.
* config/aarch64/aarch64-simd.md (aarch64_get_lane): Increase cost of
r=w alternative.

From-SVN: r260951

6 years ago[Ada] Update documentation of Sinfo.Package_Instantiation field
Piotr Trojanek [Wed, 30 May 2018 08:59:03 +0000 (08:59 +0000)]
[Ada] Update documentation of Sinfo.Package_Instantiation field

This was forgotten when renaming Get_Package_Instantiation_Node to
Get_Unit_Instantiation_Node.

2018-05-30  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

* einfo.ads (Package_Instantiation): Update comment after a routine
that uses this field has been renamed.

From-SVN: r260950

6 years ago[Ada] Unnesting: always transform local Raise statements
Ed Schonberg [Wed, 30 May 2018 08:58:57 +0000 (08:58 +0000)]
[Ada] Unnesting: always transform local Raise statements

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_ch11.adb (Replace_Raise_By_Goto): The transfomation is legal even
if the local raise statement includes a string expression. This
expression might be picked up by an outer handler or discarded, but
plays no role in this transformation.

From-SVN: r260949

6 years ago[Ada] Minor reformatting
Hristian Kirtchev [Wed, 30 May 2018 08:58:51 +0000 (08:58 +0000)]
[Ada] Minor reformatting

2018-05-30  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

* exp_aggr.adb, exp_unst.adb, freeze.adb, libgnat/a-direct.adb: Minor
reformatting.

From-SVN: r260948

6 years ago[Ada] Unnesting: handle the semantic of Procedure_To_Call field
Ed Schonberg [Wed, 30 May 2018 08:58:46 +0000 (08:58 +0000)]
[Ada] Unnesting: handle the semantic of Procedure_To_Call field

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.adb (Visit_Node): Handle the semantic Procedure_To_Call
field in relevant nodes: Allocate, Free, and return statements.

From-SVN: r260947

6 years ago[Ada] Unnesting: skip unanalyzed Component_Assocation
Ed Schonberg [Wed, 30 May 2018 08:58:40 +0000 (08:58 +0000)]
[Ada] Unnesting: skip unanalyzed Component_Assocation

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.adb (Visit_Node): Do not traverse a Component_Association
that has not been analyzed, as will be the case for a nested aggregate
that is expanded into individual assignments.

From-SVN: r260946

6 years ago[Ada] Implement pragma Max_Entry_Queue_Depth
Justin Squirek [Wed, 30 May 2018 08:58:33 +0000 (08:58 +0000)]
[Ada] Implement pragma Max_Entry_Queue_Depth

This patch implements AI12-0164-1 for the aspect/pragma Max_Entry_Queue_Depth.
Previously, the GNAT specific pragma Max_Queue_Length fulfilled this role, but
was not named to match the standard and thus was insufficent.

------------
-- Source --
------------

--  pass.ads

with System;
package Pass is

   SOMETHING : constant Integer := 5;
   Variable : Boolean := False;

   protected type Protected_Example is

      entry A (Item : Integer)
         with Max_Entry_Queue_Depth => 2;            --  OK

      entry B (Item : Integer);
      pragma Max_Entry_Queue_Depth (SOMETHING);      --  OK

      entry C (Item : Integer);                      --  OK

      entry D (Item : Integer)
         with Max_Entry_Queue_Depth => 4;            --  OK

      entry D (Item : Integer; Item_B : Integer)
         with Max_Entry_Queue_Depth => Float'Digits; --  OK

      entry E (Item : Integer);
      pragma Max_Entry_Queue_Depth (SOMETHING * 2);  --  OK

      entry E (Item : Integer; Item_B : Integer);
      pragma Max_Entry_Queue_Depth (11);             --  OK

      entry F (Item : Integer; Item_B : Integer);
      pragma Pre (Variable = True);
      pragma Max_Entry_Queue_Depth (11);             --  OK

      entry G (Item : Integer; Item_B : Integer)
         with Pre => (Variable = True),
              Max_Entry_Queue_Depth => 11;           --  OK

   private
      Data : Boolean := True;
   end Protected_Example;

   Prot_Ex  : Protected_Example;

end Pass;

--  fail.ads

package Fail is

   --  Not near entry

   pragma Max_Entry_Queue_Depth (40);                                --  ERROR

   --  Task type

   task type Task_Example is

      entry Insert (Item : in Integer)
         with Max_Entry_Queue_Depth => 10;                           --  ERROR

      -- Entry family in task type

      entry A (Positive) (Item : in Integer)
         with Max_Entry_Queue_Depth => 10;                           --  ERROR

   end Task_Example;

   Task_Ex : Task_Example;

   --  Aspect applied to protected type

   protected type Protected_Failure_0
      with Max_Entry_Queue_Depth => 50 is                            --  ERROR

      entry A (Item : Integer);
   private
      Data : Integer := 0;
   end Protected_Failure_0;

   Protected_Failure_0_Ex : Protected_Failure_0;

   protected type Protected_Failure is
      pragma Max_Entry_Queue_Depth (10);                             --  ERROR

      --  Duplicates

      entry A (Item : Integer)
         with Max_Entry_Queue_Depth => 10;                           --  OK
      pragma Max_Entry_Queue_Depth (4);                              --  ERROR

      entry B (Item : Integer);
      pragma Max_Entry_Queue_Depth (40);                             --  OK
      pragma Max_Entry_Queue_Depth (4);                              --  ERROR

      entry C (Item : Integer)
         with Max_Entry_Queue_Depth => 10,                           --  OK
              Max_Entry_Queue_Depth => 40;                           --  ERROR

      -- Duplicates with the same value

      entry AA (Item : Integer)
         with Max_Entry_Queue_Depth => 10;                           --  OK
      pragma Max_Entry_Queue_Depth (10);                             --  ERROR

      entry BB (Item : Integer);
      pragma Max_Entry_Queue_Depth (40);                             --  OK
      pragma Max_Entry_Queue_Depth (40);                             --  ERROR

      entry CC (Item : Integer)
         with Max_Entry_Queue_Depth => 10,                           --  OK
              Max_Entry_Queue_Depth => 10;                           --  ERROR

      --  On subprogram

      procedure D (Item : Integer)
         with Max_Entry_Queue_Depth => 10;                           --  ERROR

      procedure E (Item : Integer);
      pragma Max_Entry_Queue_Depth (4);                              --  ERROR

      function F (Item : Integer) return Integer
         with Max_Entry_Queue_Depth => 10;                           --  ERROR

      function G (Item : Integer) return Integer;
      pragma Max_Entry_Queue_Depth (4);                              --  ERROR

      --  Bad parameters

      entry H (Item : Integer)
         with Max_Entry_Queue_Depth => 0;                            --  ERROR

      entry I (Item : Integer)
         with Max_Entry_Queue_Depth => -1;                           --  ERROR

      entry J (Item : Integer)
         with Max_Entry_Queue_Depth => 16#FFFF_FFFF_FFFF_FFFF_FFFF#; --  ERROR

      entry K (Item : Integer)
         with Max_Entry_Queue_Depth => False;                        --  ERROR

      entry L (Item : Integer)
         with Max_Entry_Queue_Depth => "JUNK";                       --  ERROR

      entry M (Item : Integer)
         with Max_Entry_Queue_Depth => 1.0;                          --  ERROR

      entry N (Item : Integer)
         with Max_Entry_Queue_Depth => Long_Integer'(3);             --  ERROR

      -- Entry family

      entry O (Boolean) (Item : Integer)
         with Max_Entry_Queue_Depth => 5;                            --  ERROR

   private
      Data : Integer := 0;
   end Protected_Failure;

   I : Positive := 1;

   Protected_Failure_Ex : Protected_Failure;

end Fail;

--  dtest.adb

with Ada.Text_IO; use Ada.Text_IO;

procedure Dtest is
   protected Prot is
      entry Wait;
        pragma Max_Entry_Queue_Depth (2);
      procedure Wakeup;
   private
      Barrier : Boolean := False;
   end Prot;

   protected body Prot is
      entry Wait when Barrier is
      begin
         null;
      end Wait;

      procedure Wakeup is
      begin
         Barrier := True;
      end Wakeup;
   end Prot;

   task type T;

   task body T is
   begin
      Put_Line ("Waiting...");
      Prot.Wait;
   exception
      when others =>
         Put_Line ("Got exception");
   end T;

   T1, T2 : T;
begin
   delay 0.1;

   Prot.Wait;
   Put_Line ("Done");
exception
   when others =>
      Put_Line ("Main got exception");
      Prot.Wakeup;
end Dtest;

----------------------------
-- Compilation and output --
----------------------------

& gcc -c -g -gnatDG pass.ads
& gcc -c -g fail.ads
& grep -c "(2, 5, 0, 4, 6, 10, 11, 11, 11)" pass.ads.dg
& gnatmake -g -q dtest
fail.ads:5:04: pragma "Max_Queue_Length" must apply to a protected entry
fail.ads:12:15: aspect "Max_Queue_Length" cannot apply to task entries
fail.ads:17:15: aspect "Max_Queue_Length" cannot apply to task entries
fail.ads:26:12: aspect "Max_Queue_Length" must apply to a protected entry
fail.ads:36:07: pragma "Max_Queue_Length" must apply to a protected entry
fail.ads:42:07: pragma "Max_Queue_Length" duplicates aspect declared at line 41
fail.ads:46:07: pragma "Max_Queue_Length" duplicates pragma declared at line 45
fail.ads:50:15: aspect "Max_Queue_Length" for "C" previously given at line 49
fail.ads:56:07: pragma "Max_Queue_Length" duplicates aspect declared at line 55
fail.ads:60:07: pragma "Max_Queue_Length" duplicates pragma declared at line 59
fail.ads:64:15: aspect "Max_Queue_Length" for "CC" previously given at line 63
fail.ads:69:15: aspect "Max_Queue_Length" must apply to a protected entry
fail.ads:72:07: pragma "Max_Queue_Length" must apply to a protected entry
fail.ads:75:15: aspect "Max_Queue_Length" must apply to a protected entry
fail.ads:78:07: pragma "Max_Queue_Length" must apply to a protected entry
fail.ads:83:35: entity for aspect "Max_Queue_Length" must be positive
fail.ads:86:35: entity for aspect "Max_Queue_Length" must be positive
fail.ads:89:35: entity for aspect "Max_Queue_Length" out of range of Integer
fail.ads:92:35: expected an integer type
fail.ads:92:35: found type "Standard.Boolean"
fail.ads:95:35: expected an integer type
fail.ads:95:35: found a string type
fail.ads:98:35: expected an integer type
fail.ads:98:35: found type universal real

2018-05-30  Justin Squirek  <squirek@adacore.com>

gcc/ada/

* aspects.adb, aspects.ads: Register new aspect.
* par-prag.adb (Prag): Register new pragma.
* sem_ch13.adb (Analyze_Aspect_Specifications): Add processing for new
aspect similar to Aspect_Max_Queue_Length.
* sem_prag.adb, sem_prag.ads (Analyze_Pragma): Register new pragma and
set it to use the same processing as Pragma_Max_Queue_Length.
* snames.ads-tmpl: Move definition of Name_Max_Entry_Queue_Depth so
that it can be processed as a pragma in addition to a restriction and
add an entry for the pragma itself.

From-SVN: r260945

6 years ago[Ada] Extend the applicability of Thread_Local_Storage to composite types
Ed Schonberg [Wed, 30 May 2018 08:58:27 +0000 (08:58 +0000)]
[Ada] Extend the applicability of Thread_Local_Storage to composite types

This patch allows the GNAT-specific Thread_Local_Storage to be applied
to variables of a composite type initiallized with an aggregate with
static components that requires no elaboration code.

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* freeze.adb (Freeze_Object_Declaration): A pragma Thread_Local_Storage
is now legal on a variable of composite type initialized with an
aggregate that is fully static and requires no elaboration code.
* exp_aggr.adb (Convert_To_Positional): Recognize additional cases of
nested aggregates that are compile-time static, so they can be used to
initialize variables declared with Threqd_Local_Storage.
* doc/gnat_rm/implementation_defined_pragmas.rst: Add documentation on
Thread_Local_Storage.
* gnat_rm.texi: Regenerate.

gcc/testsuite/

* gnat.dg/tls1.adb, gnat.dg/tls1_pkg.ads: New testcase.

From-SVN: r260944

6 years ago[Ada] Correctly ignore Assertion_Policy in modes CodePeer and GNATprove
Yannick Moy [Wed, 30 May 2018 08:58:17 +0000 (08:58 +0000)]
[Ada] Correctly ignore Assertion_Policy in modes CodePeer and GNATprove

In the modes for static analysis with CodePeer or formal verification with
GNATprove, the value of Assertion_Policy for a given policy is ignored if
it's not Disable, as CodePeer/GNATprove are meant to check assertions even
when not enabled at run time. This was not done consistently, which could
lead to spurious errors on policy mismatch on ghost code inside assertions.

There is no impact on compilation.

2018-05-30  Yannick Moy  <moy@adacore.com>

gcc/ada/

* sem_util.adb (Policy_In_Effect): Take into account CodePeer and
GNATprove modes.

From-SVN: r260943

6 years ago[Ada] ACATS 4.1G - CXAG003 - Name_Case_Equivalence doesn't exist
Justin Squirek [Wed, 30 May 2018 08:58:12 +0000 (08:58 +0000)]
[Ada] ACATS 4.1G - CXAG003 - Name_Case_Equivalence doesn't exist

Implement a missing portion of Ada 2005's AI05-0049-1 for subprogram
Ada.Directories.Name_Case_Equivalence so that user programs can account for
operating system differences in case sensitivity.

------------
-- Source --
------------

--  main.adb

with Ada.Directories; use Ada.Directories;
with Ada.Text_IO;     use Ada.Text_IO;
procedure Main is
begin

  --  Directory layout:
  --     /empty +-- Nothing...
  --
  --     /mutliplefiles +-- "TEST1.TXT"
  --                    |
  --                "test1.txt"
  --
  --     /singlefile +-- "test1.txt"
  --
  --     /noncasable +-- "!"
  --

  Put_Line (Name_Case_Equivalence ("./empty")'Image);
  Put_Line (Name_Case_Equivalence ("./multiplefiles")'Image);
  Put_Line (Name_Case_Equivalence ("./singlefile")'Image);
  Put_Line (Name_Case_Equivalence ("./multiplefiles/test1.txt")'Image);
  Put_Line (Name_Case_Equivalence ("./singlefile/test1.txt")'Image);
  Put_Line (Name_Case_Equivalence ("./noncaseable/!")'Image);
end;

----------------------------
-- Compilation and Output --
----------------------------

& gnatmake -q main.adb
& main
CASE_SENSITIVE
CASE_SENSITIVE
CASE_SENSITIVE
CASE_SENSITIVE
CASE_SENSITIVE
CASE_SENSITIVE

2018-05-30  Justin Squirek  <squirek@adacore.com>

gcc/ada/

* libgnat/a-direct.adb, libgnat/a-direct.ads (Name_Case_Equivalence):
Add implementation.
(Start_Search): Modify to use Start_Search_Internal
(Start_Search_Internal): Add to break out an extra flag for searching
case insensative due to the potential for directories within the same
OS to allow different casing schemes.
* sysdep.c (__gnat_name_case_equivalence): Add as a default fallback
for when the more precise solution fails.

From-SVN: r260942

6 years ago[Ada] Minor reformatting
Hristian Kirtchev [Wed, 30 May 2018 08:58:05 +0000 (08:58 +0000)]
[Ada] Minor reformatting

2018-05-30  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

* checks.adb, exp_ch5.adb, exp_ch7.adb, exp_unst.adb, sem_eval.adb:
Minor reformatting.

From-SVN: r260941

6 years ago[Ada] Add support for Define_Switch with a callback in GNAT.Command_Line
Pascal Obry [Wed, 30 May 2018 08:58:00 +0000 (08:58 +0000)]
[Ada] Add support for Define_Switch with a callback in GNAT.Command_Line

Add support for Define_Switch with a callback in GNAT.Command_Line.

The callback is called for every instance of the switch found on the
command line. This make it possible to have full control over the
switch value and chain multiple actions if needed.

2018-05-30  Pascal Obry  <obry@adacore.com>

gcc/ada/

* libgnat/g-comlin.ads (Value_Callback, Define_Switch): New.
* libgnat/g-comlin.adb: Add corresponding implementation.

From-SVN: r260940

6 years ago[Ada] Fix several typos
Gary Dismukes [Wed, 30 May 2018 08:57:55 +0000 (08:57 +0000)]
[Ada] Fix several typos

2018-05-30  Gary Dismukes  <dismukes@adacore.com>

gcc/ada/

* sem_res.adb, sem_util.adb: Fix several typos.

From-SVN: r260939

6 years ago[Ada] Craft .ctors/.dtors sections manually for VxWorks EH registration
Olivier Hainque [Wed, 30 May 2018 08:57:50 +0000 (08:57 +0000)]
[Ada] Craft .ctors/.dtors sections manually for VxWorks EH registration

Temporary change for experimental purposes. Further cleanups will
be needed if this sheme works as we hope.

2018-05-30  Olivier Hainque  <hainque@adacore.com>

gcc/ada/

* vx_crtbegin_attr.c (CTOR_ATTRIBUTE, DTOR_ATTRIBUTE): Empty.
(eh_registration_ctors, eh_registration_tors): New static variables,
forced in a .ctors/.dtors section, respectively, with priority.

From-SVN: r260938

6 years ago[Ada] Spell preanalysis, preanalyze correctly
Bob Duff [Wed, 30 May 2018 08:57:44 +0000 (08:57 +0000)]
[Ada] Spell preanalysis, preanalyze correctly

2018-05-30  Bob Duff  <duff@adacore.com>

gcc/ada/

* aspects.ads, contracts.adb, exp_util.adb, expander.adb, expander.ads,
freeze.adb, inline.adb, lib-xref.adb, sem.ads, sem_aggr.adb,
sem_attr.adb, sem_ch13.adb, sem_ch3.adb, sem_ch5.adb, sem_ch6.adb,
sem_ch8.adb, sem_dim.adb, sem_elab.adb, sem_res.adb, sem_res.ads,
sinfo.ads: Spell preanalysis, preanalyze correctly.

From-SVN: r260937

6 years ago[Ada] Minor comment rework in GNAT.Secondary_Stack_Info
Bob Duff [Wed, 30 May 2018 08:57:38 +0000 (08:57 +0000)]
[Ada] Minor comment rework in GNAT.Secondary_Stack_Info

2018-05-30  Bob Duff  <duff@adacore.com>

gcc/ada/

* libgnat/g-sestin.ads: Rework documentation comments.

From-SVN: r260936

6 years ago[Ada] Reuse Is_Rewrite_Substitution where possible
Piotr Trojanek [Wed, 30 May 2018 08:57:33 +0000 (08:57 +0000)]
[Ada] Reuse Is_Rewrite_Substitution where possible

Use a high-level Is_Rewrite_Substitution instead of a low-level inequality,
with the intention to improve the code easier to read. Semantics unaffected,
so no test provided.

2018-05-30  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

* errout.adb, exp_aggr.adb, exp_ch7.adb, exp_util.adb, lib.adb,
sem_ch13.adb, sem_ch4.adb, sem_res.adb, sem_util.adb
(Has_Original_Node): Refactor to use Is_Rewrite_Substitution.

From-SVN: r260935

6 years ago[Ada] Simplify conditions by removing always true conjuncts
Piotr Trojanek [Wed, 30 May 2018 08:57:26 +0000 (08:57 +0000)]
[Ada] Simplify conditions by removing always true conjuncts

Checking "Original_Node (N) /= N" is equivalent to "Is_Rewrite_Substitution
(N)", which is checked just two lines before. Trivial simplification,
semantics unaffected.

2018-05-30  Piotr Trojanek  <trojanek@adacore.com>

gcc/ada/

* sem_prag.adb (Analyze_Pragma): Remove conjuncts that are always true.

From-SVN: r260934

6 years ago[Ada] Spurious error on legal synchronized constituent
Hristian Kirtchev [Wed, 30 May 2018 08:57:21 +0000 (08:57 +0000)]
[Ada] Spurious error on legal synchronized constituent

This patch corrects the predicate which determines whether an entity denotes a
synchronized object as per SPARK RM 9.1. to account for a case where the object
is not atomic, but its type is.

The patch also cleans up various atomic object-related predicates.

2018-05-30  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

* sem_util.adb (Is_Atomic_Object): Cleaned up. Split the entity logic
in a separate routine.
(Is_Atomic_Object_Entity): New routine.
(Is_Atomic_Prefix): Cleaned up.
(Is_Synchronized_Object): Check that the object is atomic, or its type
is atomic.
(Object_Has_Atomic_Components): Removed.
* sem_util.ads (Is_Atomic_Object): Reword the comment on usage.
(Is_Atomic_Object_Entity): New routine.

gcc/testsuite/

* gnat.dg/synchronized1.adb, gnat.dg/synchronized1.ads: New testcase.

From-SVN: r260933

6 years ago[Ada] Refine logic to set Needs_Activation_Record on subprogram types
Ed Schonberg [Wed, 30 May 2018 08:57:11 +0000 (08:57 +0000)]
[Ada] Refine logic to set Needs_Activation_Record on subprogram types

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* sem_ch3.adb (Access_Subprogram_Declaration): The flag
Needs_Activation_Record is only needed on a subprogram type, not on a
pointer to such.
* sem_res.adb (Resolve_Selected_Component): If the context type and the
component type are anonymous access to subprograms, use the component
type to obtain the proper value of Needs_Activation_Record flag for the
expression.

From-SVN: r260932

6 years ago[Ada] Minor comment addition
Eric Botcazou [Wed, 30 May 2018 08:57:05 +0000 (08:57 +0000)]
[Ada] Minor comment addition

2018-05-30  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* gcc-interface/trans.c (gnat_to_gnu) <N_Assignment_Statement>: Add
comment about the memset case.

From-SVN: r260931

6 years ago[Ada] Secondary stack leak in "for X of ..." loops
Bob Duff [Wed, 30 May 2018 08:56:59 +0000 (08:56 +0000)]
[Ada] Secondary stack leak in "for X of ..." loops

This patch fixes a memory leak bug. In particular, in a loop of the form "for X
of ...", with a type that has the Iterable aspect specified, if the result of
the Element function is returned on the secondary stack (e.g. the result
subtype has caller-unknown size), then memory for the secondary stack could
leak.

2018-05-30  Bob Duff  <duff@adacore.com>

gcc/ada/

* exp_ch5.adb (Expand_Formal_Container_Element_Loop): Remove the code
to analyze the Elmt_Decl, because it gets analyzed in the wrong scope.
We need to analyze it as part of analyzing the block, so that if the
call to Element that initializes Elmt_Decl returns on the secondary
stack, the block will ss_mark/ss_release. This block is inside the
loop; we don't want to leak memory until the loop exits.  The purpose
of analyzing Elmt_Decl first was to catch the error of modifying it,
which is illegal because it's a loop parameter. The above causes us to
miss that error.  Therefore, we add a flag Is_Loop_Parameter, and set
it on the Element entity, so we end up with an E_Variable node with the
flag set.
* einfo.ads, einfo.adb (Is_Loop_Parameter): New flag.
* sem_ch5.adb (Diagnose_Non_Variable_Lhs): Give the "assignment to loop
parameter not allowed" error if Is_Loop_Parameter.
* sem_util.adb (Is_Variable): Return False if Is_Loop_Parameter, to
trigger the call to Diagnose_Non_Variable_Lhs.

From-SVN: r260930

6 years ago[Ada] Ignore out of range values for System.Priority in CodePeer mode
Arnaud Charlet [Wed, 30 May 2018 08:56:54 +0000 (08:56 +0000)]
[Ada] Ignore out of range values for System.Priority in CodePeer mode

2018-05-30  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

* checks.adb (Apply_Scalar_Range_Check):
* sem_eval.adb (Check_Non_Static_Context, Out_Of_Range): Ignore out of
range values for System.Priority in CodePeer mode since the actual
target compiler may provide a wider range.

From-SVN: r260929

6 years ago[Ada] Unnesting: look for specification of main unit
Ed Schonberg [Wed, 30 May 2018 08:56:49 +0000 (08:56 +0000)]
[Ada] Unnesting: look for specification of main unit

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.adb: Search specification of main unit as well, for
unnesting.

From-SVN: r260928

6 years ago[Ada] Unnesting: properly handle local subprogram in declare blocks
Ed Schonberg [Wed, 30 May 2018 08:56:44 +0000 (08:56 +0000)]
[Ada] Unnesting: properly handle local subprogram in declare blocks

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_ch7.adb (Check_Unnesting_Elaboration_Code): The statement part of
a package body that is a compilation unit may contain blocks that
declare local subprograms.  In Subprogram_Unnesting Mode such
subprograms must be handled as nested inside the (implicit) elaboration
procedure that executes that statement part. To handle properly uplevel
references we construct that subprogram explicitly, to contain blocks
and inner subprograms, The statement part of the compilation unit
becomes a call to this subprogram. This is only done if blocks are
present in the statement list of the body.

From-SVN: r260927

6 years ago[Ada] Minor comment fix
Bob Duff [Wed, 30 May 2018 08:56:35 +0000 (08:56 +0000)]
[Ada] Minor comment fix

2018-05-30  Bob Duff  <duff@adacore.com>

gcc/ada/

* exp_ch7.adb: Minor comment fix.

From-SVN: r260926

6 years ago[Ada] Unnesting: properly handle subprogram instantiations
Ed Schonberg [Wed, 30 May 2018 08:56:29 +0000 (08:56 +0000)]
[Ada] Unnesting: properly handle subprogram instantiations

2018-05-30  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.adb (Visit_Node): Handle properly subprogram instantiations
that have no corresponding body and appear as attributes of the
corresponding wrapper package declaration.
(Register_Subprogram): New subprogram, used for subprogram bodies and
for subprogram instantiations to enter callable entity into Subp table.

From-SVN: r260925

6 years ago[Ada] Secondary stack implementation clean up
Hristian Kirtchev [Wed, 30 May 2018 08:56:23 +0000 (08:56 +0000)]
[Ada] Secondary stack implementation clean up

This patch reimplements the secondary stack runtime support as follows:

   * The compiler interface remains unchanged. This applies to both types and
     subprograms used by the compiler to create and manage secondary stacks.

   * The secondary stack is no longer a doubly linked list of chunks.

   * Various allocation scenarios are now handled by the same mechanism.

In addition, the patch introduces a lightweight private interface for testing
purposes.

------------
-- Source --
------------

--  comparator.ads

generic
   type Field_Typ is private;
   --  The type of the field being compared

   with function Image (Val : Field_Typ) return String;
   --  Field-to-String converted

procedure Comparator
  (Field_Nam    : String;
   Actual_Val   : Field_Typ;
   Expected_Val : Field_Typ);
--  Compare actual value Actual_Val against expected value Expected_Val for
--  field Field_Nam. Emit an error if this is not the case.

--  comparator.adb

with Ada.Text_IO; use Ada.Text_IO;

procedure Comparator
  (Field_Nam    : String;
   Actual_Val   : Field_Typ;
   Expected_Val : Field_Typ)
is
begin
   if Actual_Val /= Expected_Val then
      Put_Line (Field_Nam);
      Put_Line ("  Actual   :" & Image (Actual_Val));
      Put_Line ("  Expected :" & Image (Expected_Val));
   end if;
end Comparator;

--  debugger.ads

package Debugger is

   Verbouse : constant Boolean := False;
   --  Set to True in order to obtain verbouse output

   procedure Output (Msg : String);
   --  Emit Msg to standard output if Verbouse is True

end Debugger;

--  debugger.adb

with Ada.Text_IO; use Ada.Text_IO;

package body Debugger is

   ------------
   -- Output --
   ------------

   procedure Output (Msg : String) is
   begin
      if Verbouse then
         Put_Line (Msg);
      end if;
   end Output;
end Debugger;

--  s-sestte.ads

package System.Secondary_Stack.Tester is

   procedure Test_Dynamic_Stack_Dynamic_Chunks;
   --  Test various properties of a dynamic stack's dynamic chunks

   procedure Test_Dynamic_Stack_Illegal_Allocations;
   --  Test whether illegal allocations on a dynamic stack are properly
   --  detected and reported.

   procedure Test_Dynamic_Stack_Static_Chunk;
   --  Test various properties of a dynamic stack's static chunk

   procedure Test_Dynamic_Stack_Zero_Chunk_Size;
   --  Test various properties of a dynamic stack with default chunk size of
   --  zero.

   procedure Test_Static_Stack_Illegal_Allocations;
   --  Test whether illegal allocations on a static stack are properly
   --  detected and reported.

   procedure Test_Static_Stack_Overflow;
   --  Test whether overflow of a static stack's static chunk is properly
   --  detected and reported.

   procedure Test_Static_Stack_Static_Chunk;
   --  Test various properties of a static chunk's static chunk

end System.Secondary_Stack.Tester;

--  s-sestte.adb

with Ada.Assertions;          use Ada.Assertions;
with Ada.Text_IO;             use Ada.Text_IO;
with System;                  use System;
with System.Parameters;       use System.Parameters;
with System.Soft_Links;       use System.Soft_Links;
with System.Storage_Elements; use System.Storage_Elements;

with Comparator;
with Debugger;                use Debugger;

package body System.Secondary_Stack.Tester is

   Units : constant := Standard'Maximum_Alignment;
   --  Each allocation of the secondary stack is rouded up to the nearest
   --  multiple of the maximum alignment. This value is called a "unit" in
   --  order to facilitate further allocations.

   -----------------------
   -- Local subprograms --
   -----------------------

   procedure Compare_Boolean is
     new Comparator
           (Field_Typ => Boolean,
            Image     => Boolean'Image);

   procedure Compare_Chunk_Count is
     new Comparator
           (Field_Typ => Chunk_Count,
            Image     => Chunk_Count'Image);

   procedure Compare_Chunk_Id is
     new Comparator
           (Field_Typ => Chunk_Id,
            Image     => Chunk_Id'Image);

   procedure Compare_Memory_Index is
     new Comparator
           (Field_Typ => Memory_Index,
            Image     => Memory_Index'Image);

   procedure Compare_Memory_Size is
     new Comparator
           (Field_Typ => Memory_Size,
            Image     => Memory_Size'Image);

   procedure Compare_MSWI is
     new Comparator
           (Field_Typ => Memory_Size_With_Invalid,
            Image     => Memory_Size_With_Invalid'Image);

   procedure Initialize_Stack (Size : Memory_Size);
   --  Create a new secondary stack for the calling task where the default
   --  chunk size is Size.

   procedure Match_Chunk
     (Match_Nam : String;
      Actual    : Chunk_Info;
      Expected  : Chunk_Info);
   --  Check whether actual chunk info Actual matches expected chunk info
   --  Expected. Match_Nam is the name of the match.

   procedure Match_Pointer
     (Actual    : Stack_Pointer_Info;
      Expected  : Stack_Pointer_Info);
   --  Check whether actual pointer info Actual matches expected pointer info
   --  Expected.

   procedure Match_Stack
     (Match_Nam : String;
      Actual    : Stack_Info;
      Expected  : Stack_Info);
   --  Check whether actual stack info Stack matches expected stack info
   --  Expected. Match_Nam is the name of the match.

   procedure Test_Static_Chunk (Def_Chunk_Size : Memory_Size);
   --  Common testing for properties of the static chunk for both static and
   --  dynamic secondary stacks. Def_Chunk_Size denotes the default size of a
   --  secondary stack chunk. This routine assumes that the secondary stack
   --  can fit 12 * Units.

   ----------------------
   -- Initialize_Stack --
   ----------------------

   procedure Initialize_Stack (Size : Memory_Size) is
      Stack : SS_Stack_Ptr;

   begin
      --  Obtain the secondary stack of the calling task

      Stack := Get_Sec_Stack.all;

      --  If the calling task has an existing secodnary stack, destroy it
      --  because this scenario utilizes a custom secondary stack.

      if Stack /= null then

         --  Destroy the existing secondary stack because it will be replaced
         --  with a new one.

         SS_Free (Stack);
         pragma Assert (Stack = null);
      end if;

      --  Create a brand new empty secondary stack

      SS_Init (Stack, Size);
      pragma Assert (Stack /= null);

      --  Associate the secondary stack with the calling task

      Set_Sec_Stack (Stack);
   end Initialize_Stack;

   -----------------
   -- Match_Chunk --
   -----------------

   procedure Match_Chunk
     (Match_Nam : String;
      Actual    : Chunk_Info;
      Expected  : Chunk_Info)
   is
   begin
      Output (Match_Nam);

      Compare_MSWI
        ("Size",               Actual.Size,
                               Expected.Size);
      Compare_MSWI
        ("Size_Up_To_Chunk",   Actual.Size_Up_To_Chunk,
                               Expected.Size_Up_To_Chunk);
   end Match_Chunk;

   -------------------
   -- Match_Pointer --
   -------------------

   procedure Match_Pointer
     (Actual    : Stack_Pointer_Info;
      Expected  : Stack_Pointer_Info)
   is
   begin
      Compare_Memory_Index
        ("Byte",               Actual.Byte,
                               Expected.Byte);
      Compare_Chunk_Id
        ("Chunk",              Actual.Chunk,
                               Expected.Chunk);
   end Match_Pointer;

   -----------------
   -- Match_Stack --
   -----------------

   procedure Match_Stack
     (Match_Nam : String;
      Actual    : Stack_Info;
      Expected  : Stack_Info)
   is
   begin
      Output (Match_Nam);

      Compare_Memory_Size
        ("Default_Chunk_Size", Actual.Default_Chunk_Size,
                               Expected.Default_Chunk_Size);
      Compare_Boolean
        ("Freeable",           Actual.Freeable,
                               Expected.Freeable);
      Compare_Memory_Size
        ("High_Water_Mark",    Actual.High_Water_Mark,
                               Expected.High_Water_Mark);
      Compare_Chunk_Count
        ("Number_Of_Chunks",   Actual.Number_Of_Chunks,
                               Expected.Number_Of_Chunks);

      Match_Pointer (Actual.Top, Expected.Top);
   end Match_Stack;

   ---------------------------------------
   -- Test_Dynamic_Stack_Dynamic_Chunks --
   ---------------------------------------

   procedure Test_Dynamic_Stack_Dynamic_Chunks is
      Def_Chunk_Size : constant Memory_Size := 4 * Units;

      Dummy_1 : Address;
      Dummy_2 : Address;
      Dummy_3 : Address;
      Dummy_4 : Address;
      Mark    : Mark_Id;

   begin
      Output ("#### Test_DSDCs ####");

      --  Create a brand new empty secondary stack
      --
      --       1  2  3  4
      --    +------------+
      --    |            |
      --    +------------+

      Initialize_Stack (Def_Chunk_Size);

      Match_Stack
        (Match_Nam => "Empty stack",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 0,
            Number_Of_Chunks   => 1,
            Top                => (Byte => 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "Empty stack, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --       Mark
      --       |
      --       1  2  3  4
      --    +------------+
      --    |            |
      --    +------------+

      Mark := SS_Mark;

      --       Mark                           Top.Byte
      --       |                              |
      --       1  2  3  4      1  2  3  4  5  6
      --    +------------+  +---------------+
      --    |            |->|###############|
      --    +------------+  +---------------+
      --       1  2  3  4      5  6  7  8  9
      --                                   |
      --                                   HWM

      SS_Allocate (Dummy_1, 5 * Units);

      Match_Stack
        (Match_Nam => "After 5u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 9 * Units,
            Number_Of_Chunks   => 2,
            Top                => (Byte => (5 * Units) + 1, Chunk => 2)));

      Match_Chunk
        (Match_Nam => "After 5u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 5u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 5 * Units,
            Size_Up_To_Chunk => 4 * Units));

      --       Mark                                     Top.Byte
      --       |                                        |
      --       1  2  3  4      1  2  3  4  5      1  2  3  4
      --    +------------+  +---------------+  +------------+
      --    |            |->|###############|->|######      |
      --    +------------+  +---------------+  +------------+
      --       1  2  3  4      5  6  7  8  9     10 11 12 13
      --                                             |
      --                                             HWM
      --
      --  Note that the size of Chunk 3 defaults to 4 because the request is
      --  smaller than the default chunk size.

      SS_Allocate (Dummy_2, 2 * Units);

      Match_Stack
        (Match_Nam => "After 2u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 3,
            Top                => (Byte => (2 * Units) + 1, Chunk => 3)));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 5 * Units,
            Size_Up_To_Chunk => 4 * Units));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 3",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 3),
         Expected  =>
           (Size             => 4 * Units,
            Size_Up_To_Chunk => 9 * Units));

      --       Top.Byte
      --       |
      --       1  2  3  4         1  2  3  4  5         1  2  3  4
      --    +------------+     +---------------+     +------------+
      --    |            | --> |###############| --> |######      |
      --    +------------+     +---------------+     +------------+
      --       1  2  3  4         5  6  7  8  9        10 11 12 13
      --                                                   |
      --                                                   HWM

      SS_Release (Mark);

      --                Top.Byte
      --                |
      --       1  2  3  4         1  2  3  4  5         1  2  3  4
      --    +------------+     +---------------+     +------------+
      --    |#########   | --> |###############| --> |######      |
      --    +------------+     +---------------+     +------------+
      --       1  2  3  4         5  6  7  8  9         10 11 12 13
      --                                                   |
      --                                                   HWM

      SS_Allocate (Dummy_3, 3 * Units);

      Match_Stack
        (Match_Nam => "After 3u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 3,
            Top                => (Byte => (3 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After 3u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 3u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 5 * Units,
            Size_Up_To_Chunk => 4 * Units));

      Match_Chunk
        (Match_Nam => "After 3u allocation, chunk 3",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 3),
         Expected  =>
           (Size             => 4 * Units,
            Size_Up_To_Chunk => 9 * Units));

      --                                                  Top.Byte
      --                                                  |
      --       1  2  3  4         1  2  3  4  5  6  7  8  9
      --    +------------+     +------------------------+
      --    |#########   | --> |########################|
      --    +------------+     +------------------------+
      --       1  2  3  4         5  6  7  8  9 10 11 12
      --                                               |
      --                                               HWM

      SS_Allocate (Dummy_4, 8 * Units);

      Match_Stack
        (Match_Nam => "After 8u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 12 * Units,
            Number_Of_Chunks   => 2,
            Top                => (Byte => (8 * Units) + 1, Chunk => 2)));

      Match_Chunk
        (Match_Nam => "After 8u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 8u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 8 * Units,
            Size_Up_To_Chunk => 4 * Units));

   exception
      when others =>
         Put_Line ("Test_DSDCs: unexpected exception");
   end Test_Dynamic_Stack_Dynamic_Chunks;

   --------------------------------------------
   -- Test_Dynamic_Stack_Illegal_Allocations --
   --------------------------------------------

   procedure Test_Dynamic_Stack_Illegal_Allocations is
      Def_Chunk_Size : constant Memory_Size := 4 * Units;

      Dummy_1 : Address;
      Dummy_2 : Address;

   begin
      Output ("#### Test_DSIA ####");

      --  Create a brand new empty secondary stack
      --
      --       1  2  3  4
      --    +------------+
      --    |            |
      --    +------------+

      Initialize_Stack (Def_Chunk_Size);

      Match_Stack
        (Match_Nam => "Empty stack",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 0,
            Number_Of_Chunks   => 1,
            Top                => (Byte => 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "Empty stack, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --  It should not be possible to allocate an object of size zero

      Zero_Allocation : begin
         SS_Allocate (Dummy_1, 0);
         Put_Line ("Test_DSIA: ERROR: zero allocation succeeded");

      exception
         when Assertion_Error =>
            Match_Stack
              (Match_Nam => "After zero allocation",
               Actual    => Get_Stack_Info (Get_Sec_Stack.all),
               Expected  =>
                 (Default_Chunk_Size => Def_Chunk_Size,
                  Freeable           => True,
                  High_Water_Mark    => 0,
                  Number_Of_Chunks   => 1,
                  Top                => (Byte => 1, Chunk => 1)));

            Match_Chunk
              (Match_Nam => "After zero allocation",
               Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
               Expected  =>
                 (Size             => Def_Chunk_Size,
                  Size_Up_To_Chunk => 0));

         when others =>
            Put_Line ("Test_DSIA: zero allocation: unexpected exception");
      end Zero_Allocation;

      --  It should not be possible to allocate an object of negative size

      Negative_Allocation : begin
         SS_Allocate (Dummy_2, -8);
         Put_Line ("Test_DSIA: ERROR: negative allocation succeeded");

      exception
         when Assertion_Error =>
            Match_Stack
              (Match_Nam => "After negative allocation",
               Actual    => Get_Stack_Info (Get_Sec_Stack.all),
               Expected  =>
                 (Default_Chunk_Size => Def_Chunk_Size,
                  Freeable           => True,
                  High_Water_Mark    => 0,
                  Number_Of_Chunks   => 1,
                  Top                => (Byte => 1, Chunk => 1)));

            Match_Chunk
              (Match_Nam => "After negative allocation",
               Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
               Expected  =>
                 (Size             => Def_Chunk_Size,
                  Size_Up_To_Chunk => 0));

         when others =>
            Put_Line ("Test_DSIA: negative allocation: unexpected exception");
      end Negative_Allocation;

   exception
      when others =>
         Put_Line ("Test_DSIA: unexpected exception");
   end Test_Dynamic_Stack_Illegal_Allocations;

   -------------------------------------
   -- Test_Dynamic_Stack_Static_Chunk --
   -------------------------------------

   procedure Test_Dynamic_Stack_Static_Chunk is
      Def_Chunk_Size : constant Memory_Size := 12 * Units;

      Dummy_1 : Address;
      Dummy_2 : Address;
      Dummy_3 : Address;
      Dummy_4 : Address;
      Mark_1  : Mark_Id;
      Mark_2  : Mark_Id;

   begin
      Output ("#### Test_DSSC ####");

      --  Create a brand new empty secondary stack
      --
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------+
      --    |                                    |
      --    +------------------------------------+

      Initialize_Stack  (Def_Chunk_Size);
      Test_Static_Chunk (Def_Chunk_Size);

   exception
      when others =>
         Put_Line ("Test_DSSC: unexpected exception");
   end Test_Dynamic_Stack_Static_Chunk;

   ----------------------------------------
   -- Test_Dynamic_Stack_Zero_Chunk_Size --
   ----------------------------------------

   procedure Test_Dynamic_Stack_Zero_Chunk_Size is
      Def_Chunk_Size : constant Memory_Size := 0;

      Dummy_1 : Address;
      Dummy_2 : Address;
      Mark    : Mark_Id;

   begin
      Output ("#### Test_DSZCS ####");

      --  Create a brand new empty secondary stack
      --
      --    ++
      --    ||
      --    ++

      Initialize_Stack (Def_Chunk_Size);

      Match_Stack
        (Match_Nam => "Empty stack",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 0,
            Number_Of_Chunks   => 1,
            Top                => (Byte => 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "Empty stack, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --       Mark
      --       |
      --       1
      --    ++
      --    ||
      --    ++

      Mark := SS_Mark;

      --       Mark         Top.Byte
      --       |            |
      --       |   1  2  3  4
      --    ++  +---------+
      --    ||->|#########|
      --    ++  +---------+
      --           1  2  3
      --                 |
      --                 HWM

      SS_Allocate (Dummy_1, 3 * Units);

      Match_Stack
        (Match_Nam => "After 3u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 3 * Units,
            Number_Of_Chunks   => 2,
            Top                => (Byte => (3 * Units) + 1, Chunk => 2)));

      Match_Chunk
        (Match_Nam => "After 3u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 3u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 3 * Units,
            Size_Up_To_Chunk => 0));

      --       Mark                   Top.Byte
      --       |                      |
      --       |   1  2  3      1  2  3
      --    ++  +---------+  +------+
      --    ||->|#########|->|######|
      --    ++  +---------+  +------+
      --           1  2  3      4  5
      --                           |
      --                           HWM

      SS_Allocate (Dummy_2, 2 * Units);

      Match_Stack
        (Match_Nam => "After 2u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 5 * Units,
            Number_Of_Chunks   => 3,
            Top                => (Byte => (2 * Units) + 1, Chunk => 3)));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 3 * Units,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 3",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 3),
         Expected  =>
           (Size             => 2 * Units,
            Size_Up_To_Chunk => 3 * Units));

      --       Top.Byte
      --       |
      --       |   1  2  3      1  2
      --    ++  +---------+  +------+
      --    ||->|#########|->|######|
      --    ++  +---------+  +------+
      --           1  2  3      4  5
      --                           |
      --                           HWM

      SS_Release (Mark);

      --                             Top.Byte
      --                             |
      --           1  2  3  4  5  6  7
      --    ++  +------------------+
      --    ||->|##################|
      --    ++  +------------------+
      --           1  2  3  4  5  6
      --                          |
      --                          HWM

      SS_Allocate (Dummy_2, 6 * Units);

      Match_Stack
        (Match_Nam => "After 6u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 6 * Units,
            Number_Of_Chunks   => 2,
            Top                => (Byte => (6 * Units) + 1, Chunk => 2)));

      Match_Chunk
        (Match_Nam => "After 6u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      Match_Chunk
        (Match_Nam => "After 6u allocation, chunk 2",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 2),
         Expected  =>
           (Size             => 6 * Units,
            Size_Up_To_Chunk => 0));

   exception
      when others =>
         Put_Line ("Test_DSZCS: unexpected exception");
   end Test_Dynamic_Stack_Zero_Chunk_Size;

   -----------------------
   -- Test_Static_Chunk --
   -----------------------

   procedure Test_Static_Chunk (Def_Chunk_Size : Memory_Size) is
      Dummy_1 : Address;
      Dummy_2 : Address;
      Dummy_3 : Address;
      Dummy_4 : Address;
      Mark_1  : Mark_Id;
      Mark_2  : Mark_Id;

   begin
      --  This routine assumes an empty secondary stack

      Match_Stack
        (Match_Nam => "Empty stack",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 0,
            Number_Of_Chunks   => 1,
            Top                => (Byte => 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "Empty stack, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                   Top.Byte
      --                   |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |############
      --    +------------------------------------. . .
      --                |
      --                HWM

      SS_Allocate (Dummy_1, 4 * Units);

      Match_Stack
        (Match_Nam => "After 4u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 4 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (4 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After 4u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                   Mark_1
      --                   Top.Byte
      --                   |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |############
      --    +------------------------------------. . .
      --                |
      --                HWM

      Mark_1 := SS_Mark;

      --                   Mark_1
      --                   |              Top.Byte
      --                   |              |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |###########################
      --    +------------------------------------. . .
      --                               |
      --                               HWM

      SS_Allocate (Dummy_2, 5 * Units);

      Match_Stack
        (Match_Nam => "After 5u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 9 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (9 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After 5u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                   Mark_1         Mark_2
      --                   |              Top.Byte
      --                   |              |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |###########################
      --    +------------------------------------. . .
      --                               |
      --                               HWM

      Mark_2 := SS_Mark;

      --                   Mark_1         Mark_2
      --                   |              |     Top.Byte
      --                   |              |     |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |#################################
      --    +------------------------------------. . .
      --                                     |
      --                                     HWM

      SS_Allocate (Dummy_3, 2 * Units);

      Match_Stack
        (Match_Nam => "After 2u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (11 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After 2u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                   Mark_1
      --                   |              Top.Byte
      --                   |              |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |#################################
      --    +------------------------------------. . .
      --                                     |
      --                                     HWM

      SS_Release (Mark_2);

      Match_Stack
        (Match_Nam => "After Mark_2 release",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (9 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After Mark_2 release, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                   Top.Byte
      --                   |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |#################################
      --    +------------------------------------. . .
      --                                     |
      --                                     HWM

      SS_Release (Mark_1);

      Match_Stack
        (Match_Nam => "After Mark_1 release",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (4 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After Mark_1 release, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));

      --                                        Top.Byte
      --                                        |
      --       1  2  3  4  5  6  7  8  9 10 11 12
      --    +------------------------------------. . .
      --    |#################################
      --    +------------------------------------. . .
      --                                     |
      --                                     HWM

      SS_Allocate (Dummy_4, 6 * Units);

      Match_Stack
        (Match_Nam => "After 6u allocation",
         Actual    => Get_Stack_Info (Get_Sec_Stack.all),
         Expected  =>
           (Default_Chunk_Size => Def_Chunk_Size,
            Freeable           => True,
            High_Water_Mark    => 11 * Units,
            Number_Of_Chunks   => 1,
            Top                => (Byte => (10 * Units) + 1, Chunk => 1)));

      Match_Chunk
        (Match_Nam => "After 6u allocation, chunk 1",
         Actual    => Get_Chunk_Info (Get_Sec_Stack.all, 1),
         Expected  =>
           (Size             => Def_Chunk_Size,
            Size_Up_To_Chunk => 0));
   end Test_Static_Chunk;

   -------------------------------------------
   -- Test_Static_Stack_Illegal_Allocations --
   -------------------------------------------

   procedure Test_Static_Stack_Illegal_Allocations is
      Dummy_1 : Address;
      Dummy_2 : Address;

   begin
      Output ("#### Test_SSIA ####");

      --  It should not be possible to allocate an object of size zero

      Zero_Allocation : begin
         SS_Allocate (Dummy_1, 0);
         Put_Line ("Test_SSIA: ERROR: zero allocation succeeded");

      exception
         when Assertion_Error =>
            Output ("After zero allocation");

         when others =>
            Put_Line ("Test_SSIA: zero allocation: unexpected exception");
      end Zero_Allocation;

      --  It should not be possible to allocate an object of negative size

      Negative_Allocation : begin
         SS_Allocate (Dummy_2, -8);
         Put_Line ("Test_SSIA: ERROR: negative allocation succeeded");

      exception
         when Assertion_Error =>
            Output ("After negative allocation");

         when others =>
            Put_Line ("Test_SSIA: negative allocation: unexpected exception");
      end Negative_Allocation;

   exception
      when others =>
         Put_Line ("Test_SSIA: unexpected exception");
   end Test_Static_Stack_Illegal_Allocations;

   --------------------------------
   -- Test_Static_Stack_Overflow --
   --------------------------------

   procedure Test_Static_Stack_Overflow is
      Info  : constant Stack_Info := Get_Stack_Info (Get_Sec_Stack.all);
      Dummy : Address;

   begin
      Output ("#### Test_SSO ####");

      --  Try to overflow the static chunk

      Overflow : begin
         SS_Allocate (Dummy, Storage_Offset (Info.Default_Chunk_Size));
         Put_Line ("Test_SSO: ERROR: Overflow not detected");

      exception
         when Storage_Error =>
            Output ("After overflow");

         when others =>
            Put_Line ("Test_SSO: overflow: unexpected exception");
      end Overflow;

   exception
      when others =>
         Put_Line ("Test_SSO: unexpected exception");
   end Test_Static_Stack_Overflow;

   ------------------------------------
   -- Test_Static_Stack_Static_Chunk --
   ------------------------------------

   procedure Test_Static_Stack_Static_Chunk is
      Info : Stack_Info;

   begin
      Output ("#### Test_SSSC ####");

      Info := Get_Stack_Info (Get_Sec_Stack.all);
      Test_Static_Chunk (Info.Default_Chunk_Size);

   exception
      when others =>
         Put_Line ("Test_SSSC: unexpected exception");
   end Test_Static_Stack_Static_Chunk;

end System.Secondary_Stack.Tester;

--  main.adb

with Ada.Text_IO;                   use Ada.Text_IO;
with System.Parameters;             use System.Parameters;
with System.Secondary_Stack.Tester; use System.Secondary_Stack.Tester;

procedure Main is
   task Tester;

   --  The various scenarios are tested within a task because this guarantees
   --  that on a normal compilation, the task's secondary stack is created on
   --  the heap and can be safely freed and replaced with a custom one.

   task body Tester is
   begin
      if Sec_Stack_Dynamic then
         Test_Dynamic_Stack_Static_Chunk;
         Test_Dynamic_Stack_Dynamic_Chunks;
         Test_Dynamic_Stack_Zero_Chunk_Size;
         Test_Dynamic_Stack_Illegal_Allocations;
      else
         Test_Static_Stack_Static_Chunk;
         Test_Static_Stack_Overflow;
         Test_Static_Stack_Illegal_Allocations;
      end if;
   end Tester;

begin null; end Main;

-----------------
-- Compilation --
-----------------

$ gnatmake -a -f -q -gnata -gnatws main.adb
$ ./main

2018-05-30  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

* libgnat/s-secsta.adb: Reimplement the secondary stack support.
* libgnat/s-secsta.ads: Update the documentation of all routines in the
public part of the package.  Reimplement the private part of the
package to account for the new secondary stack structure.  Add types
and subprograms for testing purposes.  Add several documentation
sections.

From-SVN: r260924

6 years ago[Ada] Minor reformatting
Hristian Kirtchev [Wed, 30 May 2018 08:56:18 +0000 (08:56 +0000)]
[Ada] Minor reformatting

2018-05-30  Hristian Kirtchev  <kirtchev@adacore.com>

gcc/ada/

* exp_aggr.adb, exp_ch3.adb, exp_ch4.adb, exp_ch7.adb, exp_unst.adb,
exp_util.adb, exp_util.ads, libgnat/a-calcon.adb, libgnat/a-calcon.ads,
libgnat/s-os_lib.adb, repinfo.adb, sem_ch3.adb, sem_disp.adb,
sem_disp.ads, sem_util.adb: Minor reformatting.

From-SVN: r260923

6 years ago[Ada] Move special flags for Ada runtime files from Makefile.in to Makefile.rtl
Arnaud Charlet [Wed, 30 May 2018 08:56:12 +0000 (08:56 +0000)]
[Ada] Move special flags for Ada runtime files from Makefile.in to Makefile.rtl

2018-05-30  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

* gcc-interface/Makefile.in: Move special flags for Ada runtime files
from here...
* Makefile.rtl: ... to here.  Update comments.  Protect call to
"GCC_FOR_TARGET" in case target_os isn't defined.

From-SVN: r260922

6 years ago[Ada] Move target pair settings in Makefiles
Arnaud Charlet [Wed, 30 May 2018 08:56:05 +0000 (08:56 +0000)]
[Ada] Move target pair settings in Makefiles

2018-05-30  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

* gcc-interface/Makefile.in: Move target pair settings from here...
* Makefile.rtl: ... to here.
(setup-rts): New target.

From-SVN: r260921

6 years agoReplace dead store with early return
Jonathan Wakely [Wed, 30 May 2018 08:18:33 +0000 (09:18 +0100)]
Replace dead store with early return

* typeck.c (cxx_sizeof_or_alignof_type): Return size_one_node instead
of using it in dead store.

From-SVN: r260920

6 years agoUse poly_int tree accessors
Richard Sandiford [Wed, 30 May 2018 06:31:47 +0000 (06:31 +0000)]
Use poly_int tree accessors

This patch generalises various places that used hwi tree accessors so
that they can handle poly_ints instead.  In many cases these changes
are by inspection rather than because something had shown them to be
necessary.

I think the alias.c part is a minor bug fix: previously we used
fits_uhwi_p for a signed HOST_WIDE_INT (which the caller does
treat as signed rather than unsigned).  We also checked whether
each individual offset overflowed but didn't check whether the
sum did.

2018-05-30  Richard Sandiford  <richard.sandiford@linaro.org>

gcc/
* alias.c (adjust_offset_for_component_ref): Use poly_int_tree_p
and wi::to_poly_offset.  Add the current offset and then check
whether the sum fits, rather than using an unchecked addition of
a checked term.  Check for a shwi rather than a uhwi.
* expr.c (get_bit_range): Use tree_to_poly_uint64.
(store_constructor): Use poly_int_tree_p.
(expand_expr_real_1): Likewise.
* function.c (assign_temp): Likewise.
* fold-const.c (const_binop): Use poly_int_tree_p and
wi::to_poly_offset.
(fold_indirect_ref_1): Likewise.  Use multiple_p to attempt an exact
division.
* ipa-icf-gimple.c (func_checker::compare_operand): Use
to_poly_offset for MEM offsets.
* ipa-icf.c (sem_variable::equals): Likewise.
* stor-layout.c (compute_record_mode): Use poly_int_tree_p.
* tree-ssa-sccvn.c (ao_ref_init_from_vn_reference): Use
wi::to_poly_offset for BIT_FIELD_REF offsets.
(vn_reference_maybe_forwprop_address): Use poly_int_tree_p and
wi::to_poly_offset.
* var-tracking.c (emit_note_insn_var_location): Use
tree_to_poly_uint64.

From-SVN: r260914

6 years agocmd/go, cmd/vet: make vet work with gccgo
Ian Lance Taylor [Wed, 30 May 2018 00:16:58 +0000 (00:16 +0000)]
cmd/go, cmd/vet: make vet work with gccgo

    Backport https://golang.org/cl/113715 and https://golang.org/cl/113716:

    cmd/go: don't pass -compiler flag to vet

    Without this running go vet -compiler=gccgo causes vet to fail.
    The vet tool does need to know the compiler, but it is passed in
    vetConfig.Compiler.

    cmd/go, cmd/vet, go/internal/gccgoimport: make vet work with gccgo

    When using gccgo/GoLLVM, there is no package file for a standard
    library package. Since it is impossible for the go tool to rebuild the
    package, and since the package file exists only in the form of a .gox
    file, this seems like the best choice. Unfortunately it was confusing
    vet, which wanted to see a real file. This caused vet to report errors
    about missing package files for standard library packages. The
    gccgoimporter knows how to correctly handle this case. Fix this by

    1) telling vet which packages are standard;
    2) letting vet skip those packages;
    3) letting the gccgoimporter handle this case.

    As a separate required fix, gccgo/GoLLVM has no runtime/cgo package,
    so don't try to depend on it (as it happens, this fixes golang/go#25324).

    The result is that the cmd/go vet tests pass when using -compiler=gccgo.

    Reviewed-on: https://go-review.googlesource.com/114516

From-SVN: r260913

6 years agoDaily bump.
GCC Administrator [Wed, 30 May 2018 00:16:38 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r260912

6 years agocrypto/x509: specify path to AIX certificate file
Ian Lance Taylor [Wed, 30 May 2018 00:16:02 +0000 (00:16 +0000)]
crypto/x509: specify path to AIX certificate file

    Reviewed-on: https://go-review.googlesource.com/113179

From-SVN: r260908

6 years agoRISC-V: Fix a comment typo.
Jim Wilson [Tue, 29 May 2018 22:31:17 +0000 (22:31 +0000)]
RISC-V: Fix a comment typo.

* config/riscv/riscv.c (riscv_interrupt_type): Fix comment typo.

From-SVN: r260907

6 years agoPR c++/67445 - returning temporary initializer_list.
Jason Merrill [Tue, 29 May 2018 20:04:52 +0000 (16:04 -0400)]
PR c++/67445 - returning temporary initializer_list.

PR c++/67711 - assigning from temporary initializer_list.
PR c++/48562 - new initializer_list.
* typeck.c (maybe_warn_about_returning_address_of_local): Also warn
about returning local initializer_list.
* cp-tree.h (AUTO_TEMP_NAME, TEMP_NAME_P): Remove.
* call.c (build_over_call): Warn about assignment from temporary
init_list.
* init.c (build_new_1): Warn about 'new std::initializer_list'.
(find_list_begin, maybe_warn_list_ctor): New.
(perform_member_init): Use maybe_warn_list_ctor.

From-SVN: r260905

6 years agore PR target/85950 (Unsafe-math-optimizations regresses optimization using SSE4.1...
Uros Bizjak [Tue, 29 May 2018 18:35:34 +0000 (20:35 +0200)]
re PR target/85950 (Unsafe-math-optimizations regresses optimization using SSE4.1 roundss)

PR target/85950
* config/i386/i386.md (l<rounding_insn><MODEF:mode><SWI48:mode>2):
Enable for TARGET_SSE4_1 and generate rounds{s,d} and cvtts{s,d}2si{,q}
sequence.
(sse4_1_round<mode>2): Use nonimmediate_operand
for operand 1 predicate.

testsuite/ChangeLog:

PR target/85950
* gcc.target/i386/pr85950.c: New test.

From-SVN: r260903

6 years agoPR middle-end/85888 - New test case c-c++-common/attr-nonstring-6.c from r260541...
Martin Sebor [Tue, 29 May 2018 18:29:04 +0000 (18:29 +0000)]
PR middle-end/85888 - New test case c-c++-common/attr-nonstring-6.c from r260541 fails with excess errors

2018-05-29  Martin Sebor  <msebor@redhat.com>
    Richard Biener  <rguenther@suse.de>

PR testsuite/85888
* calls.c (get_size_range): Call determine_value_range instead
of get_value_range..
* tree-vrp.h (determine_value_range): Declared new function.
* tree-vrp.c (determine_value_range_1, determine_value_range): New.

Co-Authored-By: Richard Biener <rguenther@suse.de>
From-SVN: r260902

6 years agore PR c++/85883 (class template argument deduction fails in new-expression)
Marek Polacek [Tue, 29 May 2018 17:44:07 +0000 (17:44 +0000)]
re PR c++/85883 (class template argument deduction fails in new-expression)

PR c++/85883
* init.c (build_new): Handle deducing a class with new
with more than one argument.

* g++.dg/cpp1z/class-deduction55.C: New test.
* g++.dg/cpp1z/class-deduction56.C: New test.
* g++.dg/cpp1z/class-deduction57.C: New test.

From-SVN: r260901

6 years agoQualify another call in <variant>
Jonathan Wakely [Tue, 29 May 2018 17:21:31 +0000 (18:21 +0100)]
Qualify another call in <variant>

* include/std/variant (__erased_dtor): Qualify call to __get.

From-SVN: r260900

6 years agore PR c++/85952 (Bogus -Wunused-but-set-variable warning with array structured binding)
Jakub Jelinek [Tue, 29 May 2018 17:07:57 +0000 (19:07 +0200)]
re PR c++/85952 (Bogus -Wunused-but-set-variable warning with array structured binding)

PR c++/85952
* init.c (build_aggr_init): For structured binding initialized from
array call mark_rvalue_use on the initializer.

* g++.dg/warn/Wunused-var-33.C: New test.

From-SVN: r260899

6 years agotree-vect-data-refs.c (vect_preserves_scalar_order_p): Make sure to use non-pattern...
Richard Biener [Tue, 29 May 2018 14:35:28 +0000 (14:35 +0000)]
tree-vect-data-refs.c (vect_preserves_scalar_order_p): Make sure to use non-pattern stmts for get_earlier_stmt arguments.

2018-05-29  Richard Biener  <rguenther@suse.de>

* tree-vect-data-refs.c (vect_preserves_scalar_order_p): Make
sure to use non-pattern stmts for get_earlier_stmt arguments.
* tree-vectorizer.h (get_earlier_stmt): Assert we do not get
called on pattern stmts.
(get_later_stmt): Likewise.

From-SVN: r260896

6 years agolibgcov: report about a different timestamp (PR gcov-profile/85759).
Martin Liska [Tue, 29 May 2018 12:11:21 +0000 (14:11 +0200)]
libgcov: report about a different timestamp (PR gcov-profile/85759).

2018-05-29  Martin Liska  <mliska@suse.cz>

        PR gcov-profile/85759
* doc/gcov.texi: Document GCOV_ERROR_FILE and GCOV_EXIT_AT_ERROR
env variables.
2018-05-29  Martin Liska  <mliska@suse.cz>

        PR gcov-profile/85759
* libgcov-driver-system.c (gcov_error): Introduce usage of
        GCOV_EXIT_AT_ERROR env. variable.
* libgcov-driver.c (merge_one_data): Print error that we
        overwrite a gcov file with a different timestamp.

From-SVN: r260895

6 years agotree-cfg.c (verify_gimple_assign_unary): Add checking for VEC_UNPACK_*_EXPR.
Jakub Jelinek [Tue, 29 May 2018 12:01:26 +0000 (14:01 +0200)]
tree-cfg.c (verify_gimple_assign_unary): Add checking for VEC_UNPACK_*_EXPR.

* tree-cfg.c (verify_gimple_assign_unary): Add checking for
VEC_UNPACK_*_EXPR.
(verify_gimple_assign_binary): Check TYPE_VECTOR_SUBPARTS for
VEC_PACK_*_EXPR.

From-SVN: r260894

6 years agore PR target/85918 (Conversions to/from [unsigned] long long are not vectorized for...
Jakub Jelinek [Tue, 29 May 2018 11:58:24 +0000 (13:58 +0200)]
re PR target/85918 (Conversions to/from [unsigned] long long are not vectorized for AVX512DQ target)

PR target/85918
* tree.def (VEC_UNPACK_FIX_TRUNC_HI_EXPR, VEC_UNPACK_FIX_TRUNC_LO_EXPR,
VEC_PACK_FLOAT_EXPR): New tree codes.
* tree-pretty-print.c (op_code_prio): Handle
VEC_UNPACK_FIX_TRUNC_HI_EXPR and VEC_UNPACK_FIX_TRUNC_LO_EXPR.
(dump_generic_node): Handle VEC_UNPACK_FIX_TRUNC_HI_EXPR,
VEC_UNPACK_FIX_TRUNC_LO_EXPR and VEC_PACK_FLOAT_EXPR.
* tree-inline.c (estimate_operator_cost): Likewise.
* gimple-pretty-print.c (dump_binary_rhs): Handle VEC_PACK_FLOAT_EXPR.
* fold-const.c (const_binop): Likewise.
(const_unop): Handle VEC_UNPACK_FIX_TRUNC_HI_EXPR and
VEC_UNPACK_FIX_TRUNC_LO_EXPR.
* tree-cfg.c (verify_gimple_assign_unary): Likewise.
(verify_gimple_assign_binary): Handle VEC_PACK_FLOAT_EXPR.
* cfgexpand.c (expand_debug_expr): Handle VEC_UNPACK_FIX_TRUNC_HI_EXPR,
VEC_UNPACK_FIX_TRUNC_LO_EXPR and VEC_PACK_FLOAT_EXPR.
* expr.c (expand_expr_real_2): Likewise.
* optabs.def (vec_packs_float_optab, vec_packu_float_optab,
vec_unpack_sfix_trunc_hi_optab, vec_unpack_sfix_trunc_lo_optab,
vec_unpack_ufix_trunc_hi_optab, vec_unpack_ufix_trunc_lo_optab): New
optabs.
* optabs.c (expand_widen_pattern_expr): For
VEC_UNPACK_FIX_TRUNC_HI_EXPR and VEC_UNPACK_FIX_TRUNC_LO_EXPR use
sign from result type rather than operand's type.
(expand_binop_directly): For vec_packu_float_optab and
vec_packs_float_optab allow result type to be different from operand's
type.
* optabs-tree.c (optab_for_tree_code): Handle
VEC_UNPACK_FIX_TRUNC_HI_EXPR, VEC_UNPACK_FIX_TRUNC_LO_EXPR and
VEC_PACK_FLOAT_EXPR.  Formatting fixes.
* tree-vect-generic.c (expand_vector_operations_1):  Handle
VEC_UNPACK_FIX_TRUNC_HI_EXPR, VEC_UNPACK_FIX_TRUNC_LO_EXPR and
VEC_PACK_FLOAT_EXPR.
* tree-vect-stmts.c (supportable_widening_operation): Handle
FIX_TRUNC_EXPR.
(supportable_narrowing_operation): Handle FLOAT_EXPR.
* config/i386/i386.md (fixprefix, floatprefix): New code attributes.
* config/i386/sse.md (*float<floatunssuffix>v2div2sf2): Rename to ...
(float<floatunssuffix>v2div2sf2): ... this.  Formatting fix.
(vpckfloat_concat_mode, vpckfloat_temp_mode, vpckfloat_op_mode): New
mode attributes.
(vec_pack<floatprefix>_float_<mode>): New expander.
(vunpckfixt_mode, vunpckfixt_model, vunpckfixt_extract_mode): New mode
attributes.
(vec_unpack_<fixprefix>fix_trunc_lo_<mode>,
vec_unpack_<fixprefix>fix_trunc_hi_<mode>): New expanders.
* doc/md.texi (vec_packs_float_@var{m}, vec_packu_float_@var{m},
vec_unpack_sfix_trunc_hi_@var{m}, vec_unpack_sfix_trunc_lo_@var{m},
vec_unpack_ufix_trunc_hi_@var{m}, vec_unpack_ufix_trunc_lo_@var{m}):
Document.
* doc/generic.texi (VEC_UNPACK_FLOAT_HI_EXPR,
VEC_UNPACK_FLOAT_LO_EXPR): Fix pasto in description.
(VEC_UNPACK_FIX_TRUNC_HI_EXPR, VEC_UNPACK_FIX_TRUNC_LO_EXPR,
VEC_PACK_FLOAT_EXPR): Document.

* gcc.target/i386/avx512dq-pr85918.c: Add -mprefer-vector-width=512
and -fno-vect-cost-model options.  Add aligned(64) attribute to the
arrays.  Add suffix 1 to all functions and use 4 iterations rather
than N.  Add functions with conversions to and from float.
Add new set of functions with 8 iterations and another one
with 16 iterations, expect 24 vectorized loops instead of just 4.
* gcc.target/i386/avx512dq-pr85918-2.c: New test.

From-SVN: r260893

6 years agotree-vectorizer.h (struct vec_info): Add stmt_vec_infos member.
Richard Biener [Tue, 29 May 2018 11:49:44 +0000 (11:49 +0000)]
tree-vectorizer.h (struct vec_info): Add stmt_vec_infos member.

2018-05-29  Richard Biener  <rguenther@suse.de>

* tree-vectorizer.h (struct vec_info): Add stmt_vec_infos
member.
(stmt_vec_info_vec): Make pointer.
(init_stmt_vec_info_vec): Remove.
(free_stmt_vec_info_vec): Likewise.
(set_stmt_vec_info_vec): New function.
(free_stmt_vec_infos): Likewise.
(vinfo_for_stmt): Adjust for stmt_vec_info_vec indirection.
(set_vinfo_for_stmt): Likewise.
(get_earlier_stmt): Likewise.
(get_later_stmt): Likewise.
* tree-vectorizer.c (stmt_vec_info_vec): Make pointer.
(vec_info::vec_info): Allocate stmt_vec_infos and set the global.
(vec_info::~vec_info): Free stmt_vec_infos.
(vectorize_loops): Set the global stmt_vec_info_vec to NULL.
Remove old init_stmt_vec_info_vec/free_stmt_vec_info_vec calls.
(pass_slp_vectorize::execute): Likewise.
* tree-vect-stmts.c (init_stmt_vec_info_vec): Remove.
(free_stmt_vec_info_vec): Likewise.
(set_stmt_vec_info_vec): New function.
(free_stmt_vec_infos): Likewise.
* tree-vect-loop.c (_loop_vec_info::~_loop_vec_info): Set
the global stmt_vec_info_vec.
* tree-parloops.c (gather_scalar_reductions): Use
set_stmt_vec_info_vec/free_stmt_vec_infos and maintain a local
vector.

From-SVN: r260892

6 years agodominance.c (iterate_fix_dominators): Push/pop TV_DOMINANCE.
Richard Biener [Tue, 29 May 2018 11:44:16 +0000 (11:44 +0000)]
dominance.c (iterate_fix_dominators): Push/pop TV_DOMINANCE.

2018-05-29  Richard Biener  <rguenther@suse.de>

* dominance.c (iterate_fix_dominators): Push/pop TV_DOMINANCE.

From-SVN: r260891

6 years agoAdd vec::reverse.
Martin Liska [Tue, 29 May 2018 09:55:02 +0000 (11:55 +0200)]
Add vec::reverse.

2018-05-29  Martin Liska  <mliska@suse.cz>
    David Malcolm  <dmalcolm@redhat.com>

* vec.c (test_reverse): New.
(vec_c_tests): Add new test.
* vec.h (vl_ptr>::reverse): New function.

Co-Authored-By: David Malcolm <dmalcolm@redhat.com>
From-SVN: r260890

6 years ago[Ada] Adjust documentation of -gnatn switch
Eric Botcazou [Tue, 29 May 2018 09:42:50 +0000 (09:42 +0000)]
[Ada] Adjust documentation of -gnatn switch

This changes the wording in the documentation of the -gnatn switch to make
it use "units" rather than "modules" and also adjusts the usage message.

No functional changes.

2018-05-29  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* doc/gnat_ugn/building_executable_programs_with_gnat.rst (Alphabetical
List of All Switches): Replace "modules" with "units".
(Subprogram Inlining Control): Likewise.
* gnat_ugn.texi: Regenerate.
* usage.adb (Usage): Fix description of -gnatn switch.

From-SVN: r260889

6 years ago[Ada] Fix typos in Makefile.in
Arnaud Charlet [Tue, 29 May 2018 09:42:45 +0000 (09:42 +0000)]
[Ada] Fix typos in Makefile.in

2018-05-29  Arnaud Charlet  <charlet@adacore.com>

gcc/ada/

* gcc-interface/Makefile.in: Fix typos.

From-SVN: r260888

6 years ago[Ada] Attach reference to finalizers to tree
Ed Schonberg [Tue, 29 May 2018 09:42:39 +0000 (09:42 +0000)]
[Ada] Attach reference to finalizers to tree

2018-05-29  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_ch7.adb (Build_Finalizer_Call): Attach to tree the reference to
the finalizer procedure in the At_End handler, for use in LLVM
generation.

From-SVN: r260887

6 years ago[Ada] Wrong equality on untagged private type
Javier Miranda [Tue, 29 May 2018 09:42:34 +0000 (09:42 +0000)]
[Ada] Wrong equality on untagged private type

When a private type declaration T1 is completed with a derivation of an
untagged private type that overrides the predefined equality primitive, and the
full view of T2 is a derivation of another private type T2 whose full view is a
tagged type, the compiler may generate code that references the wrong equality
primitive when processing comparisons of objects of type T1.

2018-05-29  Javier Miranda  <miranda@adacore.com>

gcc/ada/

* exp_ch4.adb (Expand_N_Op_Eq, Expand_Composite_Equality): Use the new
subprogram Inherits_From_Tagged_Full_View to identify more reliably
untagged private types completed with a derivation of an untagged
private whose full view is a tagged type.
* sem_util.ads, sem_util.adb (Inherits_From_Tagged_Full_View): New
subprogram.
(Collect_Primitive_Operations): Handle untagged private types completed
with a derivation of an untagged private type whose full view is a
tagged type. In such case, collecting the list of primitives we may
find two equality primitives: one associated with the untagged private
and another associated with the ultimate tagged type (and we must
remove from the returned list this latter one).

gcc/testsuite/

* gnat.dg/equal2.adb: New testcase.

From-SVN: r260886

6 years ago[Ada] Unnesting: handle statement sequences that include an At_End handler
Ed Schonberg [Tue, 29 May 2018 09:42:29 +0000 (09:42 +0000)]
[Ada] Unnesting: handle statement sequences that include an At_End handler

2018-05-29  Ed Schonberg  <schonberg@adacore.com>

gcc/ada/

* exp_unst.adb (Visit_Node): Handle statement sequences that include an
At_End handler.

From-SVN: r260885

6 years ago[Ada] Plug small hole in -gnatR output
Eric Botcazou [Tue, 29 May 2018 09:42:16 +0000 (09:42 +0000)]
[Ada] Plug small hole in -gnatR output

The -gnatR switch outputs representation information for locally-defined
types but it was skipping those defined in blocks without label, unlike
those defined in named blocks.  This change plugs this small hole.

The following procedure:

procedure P is
begin
  declare
    type R is record
      I : Integer;
    end record;
  begin
    null;
  end;
end;

must now generate the following output with -gnatR:

Representation information for unit P (body)
--------------------------------------------

for B_1.R'Size use 32;
for B_1.R'Alignment use 4;
for B_1.R use record
   I at 0 range  0 .. 31;
end record;

2018-05-29  Eric Botcazou  <ebotcazou@adacore.com>

gcc/ada/

* repinfo.adb (List_Entities): Also recurse into blocks without label.

From-SVN: r260884