Gcc uses the same -march= for both -m32 and -m64 on x86-64 unless
--with-arch-32= is used. There is no need for -march=i486 to compile
32-bit libatomic on x86-64.
PR target/70454
* configure.tgt (XCFLAGS): Don't add -march=i486 to compile
32-bit x86 target library on x86-64.
From-SVN: r235287
* config/i386/i386.md (*lea<mode>_general_1): Rename from
*lea_general_1. Use explicit SWI12 mode interator.
(*lea<mode>_general_2): Rename from *lea_general_2.
Use explicit SWI12 mode interator.
(*lea<mode>_general_3): Rename from *lea_general_3.
Use explicit SWI12 mode interator.
(*lea<SWI12:mode>_general_4): Split from *lea_general_4.
Use explicit SWI12 mode interator.
(*lea<SWI48:mode>_general_4): Split from *lea_general_4.
Use explicit SWI48 mode interator.
From-SVN: r235286
Since mov<mode>_internal patterns handle both aligned/unaligned load
and store, we can simplify ix86_avx256_split_vector_move_misalign and
ix86_expand_vector_move_misalign.
* config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
Short-cut unaligned load and store cases. Handle all integer
vector modes.
(ix86_expand_vector_move_misalign): Short-cut unaligned load
and store cases. Call ix86_avx256_split_vector_move_misalign
directly without checking mode class.
From-SVN: r235283
2016-04-20 Bob Duff <duff@adacore.com>
* s-os_lib.ads: Minor comment fix.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* exp_ch5.adb (Expand_N_Assignment_Statement): Do no generate
a discriminant check for a type whose partial view has unknown
discriminants when the full view has discriminants with defaults.
2016-04-20 Javier Miranda <miranda@adacore.com>
* exp_util.adb (Remove_Side_Effects): When generating C code
remove side effect of type conversion of access to unconstrained
array type.
(Side_Effect_Free): Return false for the type
conversion of access to unconstrained array type when generating
C code.
* sem_res.adb (Resolved_Type_Conversion): Remove side effects
of access to unconstrained array type conversion when generating
C code.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Build_Predicate_Function_Declaration): New
function, to construct the declaration of a predicate function
at the end of the current declarative part rather than at the
(possibly later) freeze point of the type. This also allows uses
of a type with predicates in instantiations elsewhere.
(Resolve_Aspect_Expression): New procedure to detect visiblity
errors in aspect expressions, at the end of the declarative part
that includes the type declaration.
* sem_ch3.adb (Complete_Private_Subtype): Propagate properly the
predicate function from private to full view.
* einfo.adb (Predicate_Function): Refine search for predicate
function when type has a full view and predicate function may
be defined on either view.
2016-04-20 Javier Miranda <miranda@adacore.com>
* frontend.adb: Passing the root of the tree to
Unnest_Subprograms().
* exp_ch6.adb (Expand_N_Subprogram_Body): Remove code that
took care of adding subprograms to the Unest_Bodies table since
performing such action too early disables the ability to process
generic instantiations.
(Unnest_Subprograms): Adding parameter.
(Search_Unnesting_Subprograms): New subprogram.
* exp_ch6.ads (Unnest_Subrograms): Update documentation.
From-SVN: r235268
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* exp_util.adb, freeze.adb, sem_util.adb: Minor reformatting.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* exp_unst.adb (Check_Static_Type): For a private type, check
full view.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_attr.adb (Check_Type): Reject an attribute reference in
an aspect expression, when the prefix of the reference is the
current instance of the type to which the aspect applies.
From-SVN: r235267
2016-04-20 Bob Duff <duff@adacore.com>
* sem_ch6.adb (Enter_Overloaded_Entity): Do not warn about
hiding unless we're actually hiding something. The previous
code would (for example) warn about a "<" on a record type
because it incorrectly thought it was hiding the "<" on Boolean
in Standard. We need to check that the homonym S is in fact a
homograph of a predefined operator.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* exp_util.ads, exp_util.adb (Build_Procedure_Form): Moved here
from exp_ch6.adb, for use in SPARK_To_C mode when creating the
procedure equivalent to a function returning an array, when this
construction is deferred to the freeze point of the function.
* sem_util.adb (Is_Unchecked_Conversion_Instance): Include a
function that renames an instance of Unchecked_Conversion.
* freeze.adb (Freeze_Subprogram): Generate the proper procedure
declaration for a function returning an array.
* exp_ch6.adb (Build_Procedure_Form): Moved to exp_util.
From-SVN: r235266
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_util.ads, sem_util.adb (Is_Expanded_Priority_Attribute):
New predicate to determine that in a context with full run-time,
a function call is an expansion of a reference to attribute
Priority.
* sem_ch5.adb (Analyze_Function_Call): use it.
* exp_ch5.adb (Expand_N_Subprogram_Call): use it.
From-SVN: r235265
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* einfo.adb Flag286 is now used as Is_Exception_Handler.
(Is_Exception_Handler): New routine.
(Set_Is_Exception_Handler): New routine.
(Write_Entity_Flags): Output the status of Is_Exception_Handler.
* einfo.ads New attribute Is_Exception_Handler along with
occurrences in entities.
(Is_Exception_Handler): New routine along with pragma Inline.
(Set_Is_Exception_Handler): New routine along with pragma Inline.
* exp_ch7.adb (Make_Transient_Block): Ignore blocks generated
for exception handlers with a choice parameter.
* sem_ch11.adb (Analyze_Exception_Handlers): Mark the scope
generated for a choice parameter as an exception handler.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_ch3.adb (Build_Derived_Access_Type): Remove dead code.
(Constrain_Discriminated_Type): In an instance, if the type has
unknown discriminants, use its full view.
(Process_Subtype): Check that the base type is private before
adding subtype to Private_Dependents list.
2016-04-20 Bob Duff <duff@adacore.com>
* sem_ch13.adb: Minor comment fix.
From-SVN: r235264
2016-04-20 Yannick Moy <moy@adacore.com>
* sem_ch4.adb: Fix typos in comments.
* sem_res.adb (Resolve_Case_Expression): Fix type of case alternatives.
From-SVN: r235263
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Analyze_Selected_Component): A reference to the
current instance of a task type is legal if the prefix is an
expression of that task type and the selector is an entry or
entry family.
2016-04-20 Arnaud Charlet <charlet@adacore.com>
* a-cfdlli.ads (List): Type is no longer tagged, not needed. Move
varsize field at the end for efficiency.
From-SVN: r235261
2016-04-20 Vincent Celier <celier@adacore.com>
* gnatcmd.adb: Do not invoke gprls when the invocation of "gnat
ls" includes the switch -V.
* clean.adb: "<target>-gnatclean -P" now calls "gprclean
--target=<target>"
* make.adb: "<target>-gnatmake -P" now calls "gprbuild
--target=<target>"
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_ch12.adb (Qualify_Type): Do not perform
partial qualification when the immediate scope is a generic unit.
From-SVN: r235260
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch4.adb (Expand_Allocator_Expression): Ensure that the
tag assignment and adjustment preceed the accessibility check.
* exp_ch7.adb (Is_Subprogram_Call): Reimplemented.
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* exp_prag.adb (Expand_Attributes): Ensure that
the temporary used to capture the value of attribute 'Old's
prefix is properly initialized.
2016-04-20 Javier Miranda <miranda@adacore.com>
* exp_unst.ads, exp_unst.adb (Get_Level, Subp_Index): Moved to library
level.
From-SVN: r235258
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_ch12.adb (Copy_Generic_Node): Handle the special
qualification installed for universal literals that act as
operands in binary or unary operators. (Qualify_Operand): Mark
the qualification to signal the instantiation mechanism how to
handle global reference propagation.
* sinfo.adb (Is_Qualified_Universal_Literal): New routine.
(Set_Is_Qualified_Universal_Literal): New routine.
* sinfo.ads New attribute Is_Qualified_Universal_Literal along
with occurrences in nodes.
(Is_Qualified_Universal_Literal):
New routine along with pragma Inline.
(Set_Is_Qualified_Universal_Literal): New routine along with
pragma Inline.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem.adb (Do_Analyze): Save and restore Style_Max_Line_Length
so that the corresponding checks are preserved across compilations
that include System.Constants in their context.
2016-04-20 Gary Dismukes <dismukes@adacore.com>
* sem_type.adb: Minor typo fix and reformatting.
* a-conhel.ads: Update comment.
2016-04-20 Bob Duff <duff@adacore.com>
* a-cihama.adb, a-cihase.adb, a-coinve.adb (Copy): Rewrite the
code so it doesn't trigger an "uninit var" warning.
From-SVN: r235256
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_attr.ads Add new table Universal_Type_Attribute.
* sem_util.adb (Yields_Universal_Type): Use a table lookup when
checking attributes.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* exp_aggr.adb (Init_Stored_Discriminants,
Init_Visible_Discriminants): New procedures, subsidiary of
Build_Record_Aggr_Code, to handle properly the construction
of aggregates for a derived type that constrains some parent
discriminants and renames others.
From-SVN: r235255
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_ch12.adb (Qualify_Universal_Operands): New routine.
(Save_References_In_Operator): Add explicit qualifications in
the generic template for all operands of universal type.
* sem_type.adb (Disambiguate): Update the call to Matches.
(Matches): Reimplemented.
* sem_util.ads, sem_util.adb (Yields_Universal_Type): New routine.
From-SVN: r235254
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Rep_Item_Too_Late): Better error message for
an illegal aspect that freezes the entity to which it applies.
2016-04-20 Bob Duff <duff@adacore.com>
* a-stwibo.ads, a-stzbou.ads
("="): Add overriding keyword before function to avoid crash when
compiler is called with -gnatyO (check overriding indicators).
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Analyze_Pragma, case Check_Policy): If this
is a configuration pragma and it uses the ARG syntax, insert
the rewritten pragma after the current one rather than using
Insert_Actions.
2016-04-20 Arnaud Charlet <charlet@adacore.com>
* exp_aggr.adb (Backend_Processing_Possible): Add handling of
C back-end.
From-SVN: r235253
PR tree-optimization/70725
* tree-if-conv.c (is_false_predicate): New function.
(predicate_mem_writes): Use it.
* gcc.dg/pr70725.c: New test.
From-SVN: r235250
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_res.adb (Rewrite_Renamed_Operator): Do not rewrite the
renamed operator when the associated node appears within a
pre/postcondition.
* sem_util.ads, sem_util.adb (In_Pre_Post_Condition): New routine.
From-SVN: r235249
2016-04-20 Yannick Moy <moy@adacore.com>
* osint.adb (Relocate_Path): Fix test when Path is shorter than Prefix.
* einfo.adb (Set_Overridden_Operation): Add assertion.
* sem_util.adb (Unique_Entity): for renaming-as-body return the spec
entity.
2016-04-20 Javier Miranda <miranda@adacore.com>
* exp_unst.adb (Append_Unique_Call): New subprogram.
(Unnest_Subprogram): Replace the unique occurrence
of Call.Append() by Append_Unique_Call() which protects us from
adding to the Calls table duplicated entries.
2016-04-20 Arnaud Charlet <charlet@adacore.com>
* exp_attr.adb (Is_GCC_Target): Fix for C backend.
* xref_lib.ads (Dependencies_Tables): instantiate
Table package with types that guarantee its safe use.
* s-imgllu.adb, s-imgint.adb, s-imguns.adb, s-imglli.adb: Avoid nested
procedures.
From-SVN: r235248
2016-04-20 Javier Miranda <miranda@adacore.com>
* contracts.adb (Build_Postconditions_Procedure): Code cleanup.
* ghost.adb (Os_OK_Ghost_Context.Is_OK_Declaration): Handle the
declaration of the internally built _postcondition procedure.
From-SVN: r235245
2016-04-20 Arnaud Charlet <charlet@adacore.com>
* snames.ads-tmpl (Internal_Attribute_Id, Attribute_Class_Array): Fix
indentation.
* sem_util.adb (Is_Unchecked_Conversion_Instance):
defense against library-level renamings of other functions,
which are never instances of Unchecked_Conversion.
* einfo.ads: minor fix of casing in comment
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* exp_ch6.adb (Expand_N_Subprogram_Stub): Do not expand a body
that has been analyzed and expanded already. Qualify the names
in the proper body for use in the generation of C code.
From-SVN: r235244
2016-04-20 Javier Miranda <miranda@adacore.com>
* contracts.adb (Build_Postconditions_Procedure): Force its
inlining when generating C code.
* sem_attr.adb (Analyze_Attribute_Old_Result): Handle inlined
_postconditions when generating C code.
* exp_ch6.adb (Inlined_Subprogram): Inline calls to
_postconditions when generating C code.
* sinfo.ads, sinfo.adb (Corresponding_Spec, Set_Corresponding_Spec):
types of return value and argument changed from Node_Id to
Entity_Id.
2016-04-20 Vincent Celier <celier@adacore.com>
* make.adb, clean.adb, gnatname.adb: Revert previous change for now.
2016-04-20 Eric Botcazou <ebotcazou@adacore.com>
* sem_ch12.adb (Analyze_Instance_And_Renamings): Do not reset
the Has_Delayed_Freeze flag on the anonymous instance node.
From-SVN: r235243
2016-04-20 Javier Miranda <miranda@adacore.com>
* sem_ch5.adb (Analyze_Iterator_Specification): Remove transient
scope associated with the renaming object declaration.
* exp_util.adb (Insert_Actions): Remove handling of iterator
loop marked as requiring the secondary stack.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_attr.adb (Analyze_Attribute, case 'Image): Implement
AI12-0124, which extends the functionality of the attribute so it
reflects the semantics of GNAT 'Img when applied to scalar types.
* lib-xref.adb: minor whitespace layout fix.
From-SVN: r235242
2016-04-20 Vincent Celier <celier@adacore.com>
* clean.adb (Gnatclean): Fail if project file specified and
gprclean is not available.
* gnatname.adb: Fail is -P is used and gprname is not available.
* make.adb (Initialize): Fail if project file specified and
gprbuild is not available.
2016-04-20 Bob Duff <duff@adacore.com>
* sem_ch5.adb (Analyze_Iterator_Specification): Do not use secondary
stack when possible.
From-SVN: r235241
2016-04-20 Gary Dismukes <dismukes@adacore.com>
* par_sco.adb, sem_util.adb, sem_ch13.adb: Minor typo corrections and
reformatting.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Analyze_Pragma, case Default_Storage_Pool):
If the pragma comes from an aspect specification, verify that
the aspect applies to an entity with a declarative part.
* exp_ch5.adb: Code cleanup.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_res.adb (Resolve_If_Expression): If first expression is
universal, resolve subsequent ones with the corresponding class
type (Any_Integer or Any_Real).
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_ch5.adb (Analyze_Iterator_Specification): If expansion is
disabled, complete the analysis of the iterator name to ensure
that reference for entities within are properly generated.
2016-04-20 Arnaud Charlet <charlet@adacore.com>
* a-dispat.ads (Yield): add Global contract.
* a-calend.ads, a-reatim.ads: Added Initializes => Clock_Time.
* a-taside.adb: Added Initializes => Tasking_State.
From-SVN: r235240
2016-04-20 Hristian Kirtchev <kirtchev@adacore.com>
* sem_ch13.adb (Add_Invariant): Do not replace
the saved expression of an invariatn aspect when inheriting
a class-wide type invariant as this clobbers the existing
expression. Do not use New_Copy_List as it is unnecessary
and leaves the parent pointers referencing the wrong part of
the tree. Do not replace the type references for ASIS when
inheriting a class-wide type invariant as this clobbers the
existing replacement.
2016-04-20 Ed Schonberg <schonberg@adacore.com>
* sem_util.adb (Build_Explicit_Dereference): If the designated
expression is an entity name, generate reference to the entity
because it will not be resolved again.
From-SVN: r235238
2016-04-20 Richard Biener <rguenther@suse.de>
PR tree-optimization/70726
* tree-vect-stmts.c (vectorizable_shift): Do not use scalar
shift amounts from a pattern stmt operand.
* g++.dg/vect/pr70726.cc: New testcase.
From-SVN: r235236
This patches fixes a problem with stack variable accesses being
scheduled after the stack pointer restore instructions. In the
testcase this happened with the stack variable 'a' accessed through the
frame pointer.
The existing stack_tie we have in the backend is basically useless
when trying to block stack variable accesses from being scheduled
across an insn. The alias set of stack variables and the frame alias
set usually differ and hence aren't in conflict with each other. The
solution appears to be a magic MEM term with a scratch register which
is handled as a full memory barrier when analyzing scheduling
dependencies.
With the patch a (clobber (mem:BLK (scratch))) is being added to the
restore instruction in order to prevent any memory operations to be
scheduled across the insn. The patch does that only for the one case
where the stack pointer is restored from an FPR. Theoretically this
might happen also in the case where the stack pointer gets restored
using a load multiple. However, triggering that problem with
load-multiple appears to be much harder since the load-multiple will
restore the frame pointer as well. So in order to see the problem a
different call-clobbered register would need to be used as temporary
stack pointer.
Another case which needs to be handled some day is the stack pointer
allocation part. It needs to be a memory barrier as well.
Bootstrapped and regression tested with --with-arch z196 and z13 on
s390 and s390x.
-Andreas-
gcc/ChangeLog:
2016-04-20 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
PR target/70674
* config/s390/s390.c (s390_restore_gprs_from_fprs): Pick the new
stack_restore_from_fpr pattern when restoring r15.
(s390_optimize_prologue): Strip away the memory barrier in the
parallel when trying to get rid of restore insns.
* config/s390/s390.md ("stack_restore_from_fpr"): New insn
definition for loading the stack pointer from an FPR. Compared to
the normal move insn this pattern includes a full memory barrier.
gcc/testsuite/ChangeLog:
2016-04-20 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
PR target/70674
* gcc.target/s390/pr70674.c: New test.
From-SVN: r235234
PR middle-end/70680
* gimplify.c (gimplify_omp_for): Call omp_notice_variable for
implicitly linear or lastprivate iterator on the outer context.
* testsuite/libgomp.c/pr70680-1.c: New test.
* testsuite/libgomp.c/pr70680-2.c: New test.
From-SVN: r235232
From INSTRUCTION EXCEPTION SPECIFICATION section in Intel software
developer manual volume 2, only legacy SSE instructions with memory
operand not 16-byte aligned get General Protection fault. There is
no need to check 1, 2, 4, 8 byte alignments. Since x86 backend has
accurate constraints and predicates for 16-byte alignment, we can
remove alignment check in ix86_legitimate_combined_insn.
* config/i386/i386.c (ix86_legitimate_combined_insn): Remove
alignment check.
* config/i386/i386.md (ssememalign): Removed.
* config/i386/sse.md: Remove ssememalign attribute from patterns.
From-SVN: r235224
2016-04-19 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libgfortran/70684
* io/list_read (check_buffers): Add '\r' to check for end of line.
factor.
* gfortran.dg/list_read_14.f90: New test.
From-SVN: r235220
* call.c (implicit_conversion): If we choose a copy constructor
for list-initialization from the same type, the conversion is an
exact match.
From-SVN: r235219
