Jeff Law [Fri, 23 Oct 2015 22:34:04 +0000 (16:34 -0600)]
[PATCH] Call flush_ssaname_freelist more often
[PATCH] Call flush_ssaname_freelist more often
* passes.c (execute_function_todo): Do not call flush_ssaname_freelist
here. Instead...
(execute_todo): Call it here.
* tree-ssanames.c (make_ssa_name_fn): Unconditionally gather reuse
statistics
(pass_release_ssa_names::execute): Do not call flusH_ssaname_freelist.
Jeff Law [Fri, 23 Oct 2015 19:16:53 +0000 (13:16 -0600)]
[RFA] Fix pr67830, another type narrowing problem
PR tree-optimization/67830
* match.pd ((bit_and (plus/minus (convert @0) (convert @1)) mask)):
Explicitly verify the mask has no bits outside the type of
the innermost operands.
PR tree-optimization/67830
* gcc.dg/pr67830.c: New test.
* contracts.ads, contracts.adb: New unit.
* exp_ch6.adb Add with and use clauses for Contracts.
(Expand_Subprogram_Contract): Moved to Contracts.
* exp_ch6.ads (Expand_Subprogram_Contract): Moved to Contracts.
* sem_ch3.adb Add with and use clauses for Contracts.
(Analyze_Object_Contract): Moved to Contracts.
(Analyze_Declarations): Remove local variable Pack_Decl. Do not
capture global references in contracts. Check the hidden states
of a package body.
* sem_ch6.adb Add with and use clauses in Contracts.
(Analyze_Generic_Subprogram_Body): Do not capture global
references in contracts.
(Analyze_Subprogram_Body_Contract):
Moved to Contracts.
(Analyze_Subprogram_Body_Helper): Freeze the
contract of the nearest enclosing package body. Always analyze
the subprogram body contract. Do not expand the subprogram
body contract.
(Analyze_Subprogram_Contract): Moved to Contracts.
* sem_ch6.ads (Analyze_Subprogram_Body_Contract): Moved to Contracts.
(Analyze_Subprogram_Contract): Moved to Contracts.
* sem_ch7.adb Add with and use clauses for Contracts.
(Analyze_Package_Body_Contract): Moved to Contracts.
(Analyze_Package_Body_Helper): Freeze the contract of the
nearest enclosing package body.
(Analyze_Package_Contract): Moved to Contracts.
* sem_ch7.ads (Analyze_Package_Body_Contract): Moved to Contracts.
(Analyze_Package_Contract): Moved to Contracts.
* sem_ch10.adb Add with and use clauses for Contracts.
(Analyze_Compilation_Unit): Do not capture global references
in contracts.
(Analyze_Subprogram_Body_Stub_Contract): Moved to Contracts.
* sem_ch10.ads (Analyze_Subprogram_Body_Stub_Contract): Moved
to Contracts.
* sem_ch12.adb Add with and use clauses for Contracts.
(Analyze_Subprogram_Instantiation): Update the call to
Instantiate_Subprogram_Contract.
(Instantiate_Package_Body):
Do not copy the entity of the spec when creating an entity
for the body. Construct a brand new defining identifier for
the body and inherit the Comes_From_Source flag from the spec.
(Instantiate_Subprogram_Body): Remove Anon_Id to Act_Decl_Id
and update all occurrences. Construct a brand new defining
identifier for the body and inherit the Comes_From_Source
flag from the spec.
(Instantiate_Subprogram_Contract): Moved
to Contracts.
(Save_Global_References_In_Aspects): Moved to
the spec of Sem_Ch12.
(Save_Global_References_In_Contract):
Moved to Contracts.
* sem_ch12.ads (Save_Global_References_In_Aspects): Moved from
the body of Sem_Ch12.
(Save_Global_References_In_Contract):
Moved to Contracts.
* sem_prag.adb Add with and use clauses for Contracts.
(Add_Item): Removed. All references to this routine have been
replaced with calls to Append_New_Elmt.
(Analyze_Constituent):
Add special diagnostics for errors caused by freezing of
contracts.
(Analyze_Refined_State_In_Decl_Part): Add formal
parameter Freeze_Id. Add new global variable Freeze_Posted.
(Collect_Body_States): Removed.
(Report_Unused_States): Removed.
* sem_prag.ads (Analyze_Defined_State_In_Decl_Part): Add formal
parameter Freeze_Id and update comment on usage.
* sem_util.adb Remove with and use clauses for
Sem_Ch12.
(Add_Contract_Item): Moved to Contracts.
(Check_Unused_Body_States): New routine.
(Collect_Body_States):
New routine.
(Create_Generic_Contract): Moved to Contracts.
(Inherit_Subprogram_Contract): Moved to Contracts.
(Report_Unused_Body_States): New routine.
* sem_util.ads (Add_Contract_Item): Moved to Contracts.
(Check_Unused_Body_States): New routine.
(Collect_Body_States): New routine.
(Create_Generic_Contract): Moved to Contracts.
(Inherit_Subprogram_Contract): Moved to Contracts.
(Report_Unused_Body_States): New routine.
* sinfo.adb (Is_Expanded_Contract): New routine.
(Set_Is_Expanded_Contract): New routine.
* sinfo.ads New attribute Is_Expanded_Contract along with
placement in nodes.
(Is_Expanded_Contract): New routine along
with pragma Inline.
(Set_Is_Expanded_Contract): New routine
along with pragma Inline.
* gcc-interface/Make-lang.in: Add entry for contracts.o
2015-10-23 Bob Duff <duff@adacore.com>
* bindgen.adb, init.c, opt.ads, switch-b.adb: Implement new -Ea and
-Es switches.
* switch-b.ads: Minor comment fix.
* bindusg.adb: Document new -Ea and -Es switches.
* s-exctra.ads: Use -Es instead of -E.
Arnaud Charlet [Fri, 23 Oct 2015 12:52:53 +0000 (14:52 +0200)]
[multiple changes]
2015-10-23 Gary Dismukes <dismukes@adacore.com>
* exp_ch6.adb: Minor reformatting.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_ch12.adb (Check_Formal_Packages): A formal package whose
actual part is (others => <>) os identical to a formal package
with an actual part written as (<>).
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* a-reatim.adb ("/"): For Time_Span division convert the operands
to integers and then use integer division, which conforms to
the rounding required by Ada RM.
Ed Schonberg [Fri, 23 Oct 2015 12:51:30 +0000 (12:51 +0000)]
sem_ch6.adb (Check_Missing_Return): Do not report a missing return statement on a function body constructed to...
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Check_Missing_Return): Do not report a missing
return statement on a function body constructed to complete a
package body for a premature instantiation.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* exp_ch6.adb (Build_Procedure_Body_Form): Replace body of
original function with that of generated procedure, to simplify
processing and avoid scoping problems with local declarations.
(Rewrite_Function_Call_For_C): Handle properly the case of a
parameterless function.
Arnaud Charlet [Fri, 23 Oct 2015 12:48:46 +0000 (14:48 +0200)]
[multiple changes]
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* s-taskin.ads: Minor code clean up.
(Ada_Task_Control_Block): Move fixed size field before variable sized
ones.
* einfo.ads: Minor editing.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Check_Aggregate_Accessibility): Apply rule in RM
6.5 (8.3) to verify that access discriminants in an aggregate
in a return statement have the proper accessibility, i.e. do
not lead to dangling references.
2015-10-23 Eric Botcazou <ebotcazou@adacore.com>
* sem_ch13.adb (Analyze_Attribute_Definition_Clause): Add missing
test on Address_Clause_Overlay_Warnings to the "constant overlays
variable" warning. For the reverse case, also issue a warning if
the modification is potentially made through the initialization
of the variable.
2015-10-23 Jose Ruiz <ruiz@adacore.com>
* a-exetim-posix.adb (Clock): Use the pthread_getcpuclockid
function to have access to CPU clocks for tasks other than the
calling task.
* debug.adb: Switch -gnatd.5 is no longer in use, remove the
associated documentation.
* exp_dbug.adb (Get_External_Name): Do not add a special prefix
for ignored Ghost entities or when switch -gnatd.5 is enabled.
* exp_dbug.ads Remove the documentation concerning the encoding
of ignored Ghost entities.
2015-10-23 Bob Duff <duff@adacore.com>
* a-exextr.adb (Notify_Exception): For Unhandled_Raise_In_Main,
mimic the output from Ada.Exceptions.Last_Chance_Handler; don't
print "Exception raised".
* s-stalib.ads, s-exctra.ads, s-exctra.adb: Add
Unhandled_Raise_In_Main to types Exception_Trace_Kind/Trace_Kind.
gcc/testsuite
PR tree-optimization/65947
* lib/target-supports.exp
(check_effective_target_vect_max_reduc): Add.
* gcc.dg/vect/pr65947-1.c: New test.
* gcc.dg/vect/pr65947-2.c: New test.
* gcc.dg/vect/pr65947-3.c: New test.
* gcc.dg/vect/pr65947-4.c: New test.
* gcc.dg/vect/pr65947-5.c: New test.
* gcc.dg/vect/pr65947-6.c: New test.
* gcc.dg/vect/pr65947-7.c: New test.
* gcc.dg/vect/pr65947-8.c: New test.
* gcc.dg/vect/pr65947-9.c: New test.
* gcc.dg/vect/pr65947-10.c: New test.
* gcc.dg/vect/pr65947-11.c: New test.
* gnatname.adb: When gnatname is invoked with a project file,
specified with switch -P, if gprname is available, gnatname will
invoke gprname, with the target if it is a cross gnatname.
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* exp_ch4.adb: Fix typo.
* exp_ch6.adb: Update comment.
* exp_attr.adb (Expand_Min_Max_Attribute): Simplify expansion.
* exp_aggr.adb (Convert_To_Positional): Only convert to
positional when generating C in case of an object declaration.
(In_Object_Declaration): New.
(Late_Expansion): Adapt to trees generated by Modify_Tree_For_C.
* sinfo.ads: Update documentation.
2015-10-23 Joel Brobecker <brobecker@adacore.com brobecker>
* sigtramp.h (struct sigcontext, struct ucontext): Remove declarations,
and replace them by include of corresponding header file.
* ghost.adb (Is_Subject_To_Ghost): Check the
original node when searching for pragma Ghost to catch cases
where a source construct has been rewritten into something else.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* einfo.ads, einfo.adb (Rewritten_For_C): New flag on functions
that return a constrained array type. When generating C these
functions are rewritten as procedures with an out parameter,
and calls to such functions are rewritten accordingly.
* exp_ch6.adb (Expand_N_Subprogram_Declaration): When
Modify_Tree_For_C is set and the function returns a constrained
array type, generate a procedure declaration with an additional
out parameter. Mark original function as Rewritten_For_C.
The new declaration is inserted in tree immediately after
current declaration.
(Expand_Subprogram_Body): If entity is marked Rewritten_For_C,
generate body of corresponding procedure using declarations
and statements for function body. Replace return statements
with assignments to the out parameter, followed by a simple
return statement.
(Rewrite_Function_Call_For_C): New procedure to replace a function
call that returns an array by a procedure call.
* sem_util.adb (Denotes_Iterator): New routine.
(Is_Iterator): Code cleanup. Factor out the detection of a
predefined iterator. As a result this fixes a missing case
where a tagged type implements interface Reversible_Iterator.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_attr.adb (Eval_Attribute): Constant-fold 'Enabled if
not within a generic unit, even if expander is not active, so
that instances of container packages remain preelaborable in
-gnatc mode.
2015-10-23 Tristan Gingold <gingold@adacore.com>
* init.c (__gnat_sigtramp): New assembly function for arm64-darwin.
(__gnat_error_handler): Use trampoline for arm64.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* exp_ch3.adb (Expand_N_Object_Declaration): if the type of the
object is a class-wide limited interface type, the expression
is not restricted to the forms specified for limited types.
Arnaud Charlet [Fri, 23 Oct 2015 12:21:21 +0000 (14:21 +0200)]
[multiple changes]
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_util.adb (Is_Iterator, Is_Reversible_iterator): Use
root type to determine whether the type is a descendant of the
corresponding interface type, so take into account multiple
levels of subtypes and derivations.
2015-10-23 Olivier Hainque <hainque@adacore.com>
* tracebak.c: Refine selection of GCC/GENERIC_UNWINDER for
tracebacks on x86 & x86_64.
(x86 & x86_64): If !SJLJ, always pick the
GCC_UNWINDER for x86_64 (not only on linux).
* sem_util.ads: Minor fix in comment.
* sem_prag.adb (Analyze_Pragma): Pragma Volatile_Function should
not apply to a function instantiation.
* sem_util.adb (Has_Effectively_Volatile_Profile): New routine.
(Is_Volatile_Function): An instance of Ada.Unchecked_Conversion
is a volatile function when its profile contains an effectively
volatile type.
* sem_util.ads (Has_Effectively_Volatile_Profile): New routine.
Arnaud Charlet [Fri, 23 Oct 2015 10:55:10 +0000 (12:55 +0200)]
[multiple changes]
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* exp_unst.adb (Unnest_Subprogram): Complete previous
change and update comments.
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_util.ads, sem_util.adb (Check_Function_With_Address_Parameter):
A subprogram that has an Address parameter and is declared in a Pure
package is not considered Pure, because the parameter may be used as a
pointer and the referenced data may change even if the address value
itself does not.
* freeze.adb (Freeze_Subprogram): use it.
* exp_ch6.adb (Expand_N_Subprogram_Body): Use it.
Arnaud Charlet [Fri, 23 Oct 2015 10:52:34 +0000 (12:52 +0200)]
[multiple changes]
2015-10-23 Olivier Hainque <hainque@adacore.com>
* tracebak.c: Fallback to generic unwinder for gcc-sjlj on x86 &
x86_64 linux
* tracebak.c: Rework x86 & x86_64 sections to resort to the
generic unwinder if __USING_SJLJ_EXCEPTIONS__.
2015-10-23 Javier Miranda <miranda@adacore.com>
* sem_util.adb (Check_Function_Writable_Actuals): For function
calls restrict the check to elementary types, as requested by
RM 6.4.1(6.15/3)
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* exp_unst.adb (Unnest_Subprogram): Suppress initialization on
Decl_ARECnT since we are taking care of all initializations in
the generated code.
Arnaud Charlet [Fri, 23 Oct 2015 10:49:44 +0000 (12:49 +0200)]
[multiple changes]
2015-10-23 Ed Schonberg <schonberg@adacore.com>
* sem_dim.adb (Analyze_Dimension_Extension_Or_Record_Aggregate):
Handle properly a box-initialized aggregate component.
2015-10-23 Yannick Moy <moy@adacore.com>
* sem_prag.adb (Analyze_Pragma): Reject Volatile_Function not placed
on a function.
2015-10-23 Yannick Moy <moy@adacore.com>
* a-extiin.ads, a-reatim.ads, a-interr.ads, a-exetim-mingw.ads,
a-exetim-default.ads, a-exetim.ads, a-taside.ads: Add "Global => null"
contract on subprograms.
* lib-xref-spark_specific.adb: collect scopes for stubs of
protected objects
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* gnat1drv.adb (Adjust_Global_Switches): Enable
Back_Annotate_Rep_Info to get size information from gigi.
(Gnat1drv): Code clean ups.
* frontend.adb (Frontend): Ditto.
Bob Duff [Fri, 23 Oct 2015 10:41:13 +0000 (10:41 +0000)]
exp_ch6.adb (Expand_N_Extended_Return_Statement): Do not call SS_Release for a block statement enclosing the return statement in...
2015-10-23 Bob Duff <duff@adacore.com>
* exp_ch6.adb (Expand_N_Extended_Return_Statement): Do not call
SS_Release for a block statement enclosing the return statement in the
case where a build-in-place function return is returning
the result on the secondary stack. This is accomplished by
setting the Sec_Stack_Needed_For_Return flag on such blocks.
It was already being set for the function itself, and it was
already set correctly for blocks in the non-build-in-place case
(in Expand_Simple_Function_Return).
(Set_Enclosing_Sec_Stack_Return): New procedure to perform
the Set_Sec_Stack_Needed_For_Return calls. Called in the
build-in-place and non-build-in-place cases.
(Expand_Simple_Function_Return): Call
Set_Enclosing_Sec_Stack_Return instead of performing the loop
in line.
2015-10-23 Bob Duff <duff@adacore.com>
* scng.adb (Char_Literal_Case): If an apostrophe
follows a reserved word, treat it as a lone apostrophe, rather
than the start of a character literal. This was already done for
"all", but it needs to be done also for (e.g.) "Delta".
Arnaud Charlet [Fri, 23 Oct 2015 10:29:50 +0000 (12:29 +0200)]
[multiple changes]
2015-10-23 Bob Duff <duff@adacore.com>
* exp_strm.adb (Build_Record_Or_Elementary_Input_Function): Use
Underlying_Type for B_Typ, in case the Typ is a subtype of a type with
unknown discriminants.
* g-awk.ads: Minor style fix in comment
* debug.adb: Document the use of debug switch -gnatd.5.
* einfo.adb: Code reformatting. (Is_Ghost_Entity): Moved from ghost.adb.
* einfo.ads New synthesized attribute Is_Ghost_Enity along
with usage in nodes and pragma Inline.
(Is_Ghost_Entity: Moved from ghost.ads.
* exp_ch3.adb Code reformatting.
(Expand_Freeze_Array_Type): Capture, set and restore the Ghost mode.
(Expand_Freeze_Class_Wide_Type): Capture, set and restore the
Ghost mode.
(Expand_Freeze_Enumeration_Type): Capture, set and
restore the Ghost mode.
(Expand_Freeze_Record_Type): Capture, set and restore the Ghost mode.
* exp_ch6.adb (Expand_Subprogram_Contract): Do not expand the
contract of an ignored Ghost subprogram.
* exp_ch13.adb Add with and use clauses for Ghost.
(Expand_N_Freeze_Entity): Capture, set and restore the Ghost mode.
* exp_dbug.adb (Get_External_Name): Code reformatting. Add a
special prefix for ignored Ghost entities or when requested by
-gnatd.5 for any Ghost entity.
* exp_dbug.ads Document the use of prefix "_ghost_" for ignored
Ghost entities.
* exp_prag.adb (Expand_Pragma_Check): Capture, set and restore the
Ghost mode.
(Expand_Pragma_Loop_Variant): Use In_Assertion_Expr
to signal the original context.
* ghost.adb (Check_Ghost_Overriding): Code cleanup.
(Is_Ghost_Entity): Moved to einfo.adb. (Is_OK_Declaration):
Move the assertion expression check to the outer level.
(Is_OK_Ghost_Context): An assertion expression is a valid Ghost
context.
* ghost.ads (Is_Ghost_Entity): Moved to einfo.ads.
* sem_ch3.adb (Analyze_Object_Contract): A source Ghost object
cannot be imported or exported. Mark internally generated objects
as Ghost when applicable.
(Make_Class_Wide_Type): Inherit the ghostness of the root tagged type.
* sem_ch6.adb (Analyze_Subprogram_Body_Helper): Mark
a stand alone subprogram body as Ghost when applicable.
(Analyze_Subprogram_Declaration): Mark internally generated
subprograms as Ghost when applicable.
* sem_ch7.adb: Code cleanup.
* sem_ch13.adb (Add_Invariants): Add various formal
parameters to break dependency on global variables.
(Build_Invariant_Procedure): Code cleanup. Capture, set and
restore the Ghost mode.
* sem_res.adb (Resolve_Actuals): The actual parameter of a source
Ghost subprogram whose formal is of mode IN OUT or OUT must be
a Ghost variable.
* sem_ch8.adb Code cleanup.
(Find_Expanded_Name): Replace
the call to In_Pragmas_Depends_Or_Global with a call to
In_Abstract_View_Pragma.
(In_Abstract_View_Pragma): New routine.
(In_Pragmas_Depends_Or_Global): Removed.
* sem_prag.adb (Analyze_Part_Of): Catch a case where indicator
Part_Of denotes the abstract view of a variable.
Arnaud Charlet [Fri, 23 Oct 2015 10:25:57 +0000 (10:25 +0000)]
sem_util.ads (Unique_Defining_Entity): Document the result for tasks and entries.
2015-10-23 Arnaud Charlet <charlet@adacore.com>
* sem_util.ads (Unique_Defining_Entity): Document the result
for tasks and entries.
* sem_util.adb (Unique_Entity): Return declaration entity for task and
entry bodies
This moves the fold rules for trunc, floor, ceil, round, nearbyint and
rint in one go, since they're tested as a group. Most of the code is
supporting the f(x)->x fold when x is known to be integer-valued.
Like with the non-negative test, this is probably more elegantly handled
by tracking range information for reals, but until that happens, I think
we should handle it analogously to tree_expr_nonnegative_p.
I've incorporated the fix for PR68031 in the new version of
integer_valued_real_p. However, it seemed confusing to test for an
SSA name at the head of the function rather than the case statement,
and not fall through to tree_simple_nonnegative_warnv_p (which
conceptually shouldn't care whether an update is in progress).
But tree_simple_nonnegative_warnv_p is a no-op for SSA names,
so I simply changed it to:
Doing these folds later meant that IPA would start to use information
about the aborting sinf and floor in 20030125-1.c before the folds
kicked in. I changed them from noinline to weak to stop that.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi.
gcc/
* builtins.c (integer_valued_real_p): Move to fold-const.c.
(fold_trunc_transparent_mathfn, fold_builtin_trunc, fold_builtin_floor)
(fold_builtin_ceil, fold_builtin_round): Delete.
(fold_builtin_1): Handle constant trunc, floor, ceil and round
arguments here.
* convert.c (convert_to_real): Remove narrowing of rounding
functions.
* fold-const.h (integer_valued_real_unary_p)
(integer_valued_real_binary_p, integer_valued_real_call_p)
(integer_valued_real_single_p, integer_valued_real_p): Declare.
* fold-const.c (tree_single_nonnegative_warnv_p): Move
name_registered_for_update_p check to SSA_NAME case statement.
Don't call tree_simple_nonnegative_warnv_p for SSA names.
(integer_valued_real_unary_p, integer_valued_real_binary_p)
(integer_valued_real_call_p, integer_valued_real_single_p)
(integer_valued_real_invalid_p): New functions.
(integer_valued_real_p): Move from fold-const.c and rework
to call the functions above. Handle SSA names.
* gimple-fold.h (gimple_stmt_integer_valued_real_p): Declare.
* gimple-fold.c (gimple_assign_integer_valued_real_p)
(gimple_call_integer_valued_real_p, gimple_phi_integer_valued_real_p)
(gimple_stmt_integer_valued_real_p): New functions.
* match.pd: Fold f(f(x))->f(x) for fp->fp rounding functions f.
Fold f(x)->x for the same f if x is known to be integer-valued.
Fold f(extend(x))->extend(f'(x)) if doing so doesn't affect
the result. Canonicalize floor(x) as trunc(x) if x is
nonnegative.
gcc/testsuite/
* gcc.c-torture/execute/20030125-1.c (floor, floorf, sin, sinf):
Make weak rather than noinline.
* gcc.dg/builtins-57.c: Compile with -O.
* gcc.dg/torture/builtin-integral-1.c: Skip for -O0.
Martin Sebor [Thu, 22 Oct 2015 23:48:17 +0000 (23:48 +0000)]
re PR driver/68043 (many undocumented options, missing punctuation)
gcc/ChangeLog:
2015-10-22 Martin Sebor <msebor@redhat.com>
PR driver/68043
* config/i386/i386.opt: Add missing periods to the ends of sentences.
* config/msp430/msp430.opt: Same.
gcc/testsuite/ChangeLog:
2015-10-22 Martin Sebor <msebor@redhat.com>
PR driver/68043
* gcc.misc-tests/help.exp: Verify that option descriptions
end in periods.
* lib/options.exp (check_for_options): Use the regexp --line option.
Print unexpected match on failure.
Jeff Law [Thu, 22 Oct 2015 19:18:05 +0000 (13:18 -0600)]
[PATCH] Fix abort in write_eligible_delay
[PATCH] Fix abort in write_eligible_delay
* genattrtab.c (main): If we do not have any annul-true or annul-false
slots, then write out a dummy eligible_for_annul_true or
eligible_for_annul_false as needed.
Nick Clifton [Thu, 22 Oct 2015 16:00:38 +0000 (16:00 +0000)]
msp430.opt: Add -msilicon-errata and -msilicon-errata-warn.
* config/msp430/msp430.opt: Add -msilicon-errata and
-msilicon-errata-warn.
* config/msp430/msp430.h (ASM_SPEC): Pass new options on to
assembler.
* doc/invoke.texi: Document new options.
Richard Biener [Thu, 22 Oct 2015 13:36:46 +0000 (13:36 +0000)]
re PR tree-optimization/58497 (SLP vectorizes identical operations)
2015-10-22 Richard Biener <rguenther@suse.de>
PR tree-optimization/58497
* tree-vect-generic.c (ssa_uniform_vector_p): New helper.
(expand_vector_operations_1): Use it. Lower operations on
all uniform vectors to scalar operations if the HW supports it.
Richard Biener [Thu, 22 Oct 2015 13:33:17 +0000 (13:33 +0000)]
re PR tree-optimization/19049 (not vectorizing a fortran loop)
2015-10-22 Richard Biener <rguenther@suse.de>
PR tree-optimization/19049
PR tree-optimization/65962
* tree-vect-data-refs.c (vect_analyze_group_access_1): Fall back
to strided accesses if single-element interleaving doesn't work.
Richard Biener [Thu, 22 Oct 2015 11:44:11 +0000 (11:44 +0000)]
re PR middle-end/68046 (-ftrapv doesn't catch leaq-based overflows on x86-64)
2015-10-22 Richard Biener <rguenther@suse.de>
PR middle-end/68046
PR middle-end/61893
* optabs.c (emit_libcall_block_1): Allow a NULL_RTX equiv.
(expand_binop): For -ftrapv optabs do not record an REG_EQUAL note.
(expand_unop): Likewise.
Max Ostapenko [Thu, 22 Oct 2015 09:47:17 +0000 (12:47 +0300)]
Fix bootstrap on x86_64-apple-darwin14 after r229119.
libsanitizer/
PR bootstrap/68041
* configure.ac (link_sanitizer_common): Link against librt only if it
contains shm_open, required by sanitizers.
(CXX_ABI_NEEDED): Remove variable.
* configure: Regenerate.
* ubsan/Makefile.am (libubsan_la_LIBADD): Do not add -lc++abi anymore.
* ubsan/Makefile.in: Regenerate.
Andre Vieira [Thu, 22 Oct 2015 05:12:32 +0000 (05:12 +0000)]
[PATCH][ARM] Fix for testcase after r228661
This patch addresses PR-67948 by changing the xor-and.c test,
initially written for a simplify-rtx pattern, to make it pass post
r228661 (see
https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00676.html). This test
no longer triggered the simplify-rtx pattern it was written for prior
to r228661, though other optimizations did lead to the same assembly
the test checked for. The optimization added with r228661 matches the
pattern used in the test and optimizes it to a better and still valid
sequence. Being unable to easily change the test to trigger the
original simplify-rtx pattern, I chose to change it to pass with the
new produced assembly sequence.
This is correct because the transformation is valid and it yields a
more efficient pattern. However, as I pointed out before this test
doesn't test the optimization it originally was intended for.
Tested by running regression tests for armv6.
Is this OK to commit?
Thanks, Andre
2015-10-22 Andre Vieira <andre.simoesdiasvieira@arm.com>
PR testsuite/67948
* gcc.target/arm/xor-and.c: Adjust test to check for eor
instead of orr.
[Patch AArch64 63304] Fix issue with global state.
Jiong pointed out privately that there was a thinko
in the way in which the global state was being
set and reset. I don't like adding such
global state but ....
Martin Sebor [Wed, 21 Oct 2015 22:24:41 +0000 (22:24 +0000)]
Improve --help output to generate references to option aliases.
gcc/
PR driver/68043
* opts.c (undocumented_msg, use_diagnosed_msg): New globals.
(print_filtered_help): Reference aliased option's name and encourage
readers to use it in preference to the alias if the former is not
documented. Mention when using an option is diagnosed.
* gcc.c (display_help): End each sentence with a period.
Aditya Kumar [Wed, 21 Oct 2015 21:18:27 +0000 (21:18 +0000)]
a scalar depending on vdefs in the current region is not invariant
When a scalar is defined in function of an array reference in the current scop,
it does variate. Graphite cannot represent the condition in scop-11.c, as a[*]
variates in the current region, and it is not an affine condition:
for (j = 0; j <= 20; j++)
a[j] = b + i;
if (a[12] == 23)
b = 3;
else
b = 1;
for (j = 0; j <= 20; j++)
a[j] = b + i;
* graphite-scop-detection.c (parameter_index_in_region): Update call to
invariant_in_sese_p_rec.
* graphite-sese-to-poly.c (extract_affine): Same.
* sese.c (invariant_in_sese_p_rec): Pass in an extra parameter has_vdefs.
(scalar_evolution_in_region): Return chrec_dont_know when the scalar variable
depends on virtual definitions in the current region.
* sese.h (invariant_in_sese_p_rec): Update declaration.
Aditya Kumar [Wed, 21 Oct 2015 21:18:17 +0000 (21:18 +0000)]
add an upper limit on the number of array references
* graphite-scop-detection.c (build_scops): Do not handle scops with more
than PARAM_GRAPHITE_MAX_ARRAYS_PER_SCOP arrays.
* params.def (PARAM_GRAPHITE_MAX_ARRAYS_PER_SCOP): New.
Co-Authored-By: Sebastian Pop <s.pop@samsung.com>
From-SVN: r229150
Jan Hubicka [Wed, 21 Oct 2015 21:14:06 +0000 (23:14 +0200)]
re PR ipa/67056 (Wrong code generated)
PR ipa/67056
* ipa-polymorphic-call.c (possible_placement_new): If cur_offset
is negative we don't know the type.
(check_stmt_for_type_change): Skip constructors of non-polymorphic
types as those won't help devirutalization.
* g++.dg/ipa/pr67056.C: New testcase.
Relax ADDR_SPACE_GENERIC_P checks for default address space hooks
If all address spaces use the same modes and forms, we would
be forced to replicate these hooks in the backend. Which would
then require the creation of a new hook to replace
target_default_pointer_address_modes_p.
Change default of non-overlapping address space conversion
The current default of making all undefined coversions being
set to null is not useful. It has caused all users to lie
and say that spaces are subsets when they are not, just so
that they can override the conversion.
* expr.c (expand_expr_real_2): Use convert_modes on disjoint
address spaces.
The fold code also expanded cabs(x+yi) to fsqrt(x*x+y*y) when optimising
for speed. tree-ssa-math-opts.c has this transformation too, but unlike
the fold code, it first checks whether the target implements the sqrt
optab. The patch simply removes the fold code and keeps the
tree-ssa-math-opts.c logic the same.
gcc.dg/lto/20110201-1_0.c was relying on us replacing cabs
with fsqrt even on targets where fsqrt is itself a library call.
The discussion leading up to that patch suggested that we only
want to test the fold on targets with a square root instruction,
so it would be OK to skip the test on other targets:
The patch does that using the sqrt_insn effective target.
It's possible that removing the tree folds renders the LTO trick
unnecessary, but since the test was originally for an ICE, it seems
better to leave it as-is.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi. 20110201-1_0.c passes on all three.
projects / thirdparty / gcc.git / log
