pmderodat [Tue, 11 Dec 2018 11:11:06 +0000 (11:11 +0000)]
[Ada] Uintp: add a new UI_From_Integral generic constructor
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* uintp.ads, uintp.adb (UI_From_Integral): New generic function,
to simplify construction of Universal_Integer representations
from any Integer type. If type is small enough the code is
equivalent to a call to UI_To_Int with appropriate conversions;
otherwise the routine uses the existing mechanism of building a
vector of suitable integer values and calling Vector_To_Uint.
The routine must not be applied to a biased type.
pmderodat [Tue, 11 Dec 2018 11:11:00 +0000 (11:11 +0000)]
[Ada] Crash on compilation unit function that builds in place
This patch fixes a crash on a function that builds its limited result in
place. Previously this was handled properly only if the function was a
child unit.
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* sem_ch3.adb (Build_Itype_Reference): Handle properly an itype
reference created for a function that is a compilation unit, for
example if the function builds in place an object of a limited
type.
gcc/testsuite/
* gnat.dg/bip_cu.adb, gnat.dg/bip_cu_constructor.adb,
gnat.dg/bip_cu_constructor.ads, gnat.dg/bip_cu_t.adb,
gnat.dg/bip_cu_t.ads: New testcase.
pmderodat [Tue, 11 Dec 2018 11:10:42 +0000 (11:10 +0000)]
[Ada] Crash on ignored Ghost expression function
This patch updates freezing to ensure that freeze nodes are inserted
into the tree when the entity being frozen is non-Ghost, and the context
is an ignored Ghost spec expression.
* exp_util.adb (Insert_Action): Add new formal parameter
Spec_Expr_OK.
(Insert_Actions): Add new formal parameter Spec_Expr_OK. Update
all calls to Insert_Actions where relevant. Honour an insertion
from a spec expression context when requested by the caller.
* exp_util.ads (Insert_Action): Add new formal parameter
Spec_Expr_OK.
(Insert_Actions): Add new formal parameter Spec_Expr_OK.
* freeze.adb (Add_To_Result): Force the insertion of the freeze
node even when the context is a spec expression.
gcc/testsuite/
* gnat.dg/ghost2.adb, gnat.dg/ghost2.ads: New testcase.
pmderodat [Tue, 11 Dec 2018 11:10:27 +0000 (11:10 +0000)]
[Ada] Complete implementation of RM C.6(19) clause
This ensures that the compiler fully implements the C.6(19) clause of
the Ada Reference Manual and gives a warning when the clause does change
the passing mechanism of the affected parameter.
2018-12-11 Eric Botcazou <ebotcazou@adacore.com>
gcc/ada/
* fe.h (Is_Atomic_Object): Declare.
(Is_Volatile_Object): Likewise.
* gcc-interface/trans.c (atomic_or_volatile_copy_required_p):
New.
(Call_to_gnu): Generate a copy for an actual parameter passed by
reference if the conditions set forth by RM C.6(19) are met and
specificially deal with an atomic actual parameter.
gcc/testsuite/
* gnat.dg/atomic11.adb, gnat.dg/atomic11_pkg1.ads,
gnat.dg/atomic11_pkg2.ads: New testcase.
pmderodat [Tue, 11 Dec 2018 11:10:22 +0000 (11:10 +0000)]
[Ada] Stubs that complete generic subprogram do have a "prior declaration"
The intuition behind the Is_Subprogram_Stub_Without_Prior_Declaration
utility routine is to detect stubs that act as subprogram declarations
and False on stubs that act as completions. This behaviour is now fixed
for stubs that correspond to generic subprogram declarations.
This patch affects a routine that is only used in GNATprove, so no
frontend test provided. An example where the result changed from True to
False is:
-----------
-- p.ads --
-----------
package P is
generic
procedure Proc;
end P;
-----------
-- p.adb --
-----------
package body P is
procedure Proc is separate; -- now we return False for this stub
end P;
pmderodat [Tue, 11 Dec 2018 11:10:17 +0000 (11:10 +0000)]
[Ada] GNATprove: improve proofs for uninitialized constrained objects
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* sem_ch4.adb (Analyze_Allocator): In GNATprove mode build a
subtype declaration if the allocator has a subtype indication
with a constraint. This allows additional proofs to be applied
to allocators that designate uninitialized constrained objects.
pmderodat [Tue, 11 Dec 2018 11:10:12 +0000 (11:10 +0000)]
[Ada] Support access types in GNATprove
SPARK RM has been updated to support access types in SPARK. Part of this
support is that now SPARK RM 3.1 lists access types as having full
default initialization. Now updated.
There is no impact on compilation.
2018-12-11 Yannick Moy <moy@adacore.com>
gcc/ada/
* sem_util.adb (Has_Full_Default_Initialization): Consider
access types as having full default initialization.
pmderodat [Tue, 11 Dec 2018 11:10:07 +0000 (11:10 +0000)]
[Ada] Better error message from GNATprove on illegal switch
When a compilation switch is wrongly passed to GNATprove without the
leading hyphen, this patch issues a clear error message instead of the
obscure 'usage' message previously displayed.
There is no impact on compilation.
2018-12-11 Yannick Moy <moy@adacore.com>
gcc/ada/
* gnat1drv.adb (Gnat1drv): Issue specific error message in
GNATprove mode when multiple file names on the command line.
* osint.adb, osint.ads (Dump_Command_Line_Source_File_Names):
New procedure to print file names on the command line.
pmderodat [Tue, 11 Dec 2018 11:10:02 +0000 (11:10 +0000)]
[Ada] Fix runtime stack overflow for out/in-out actuals without discr.
2018-12-11 Javier Miranda <miranda@adacore.com>
gcc/ada/
* exp_aggr.adb (Convert_To_Assignments): When gnerating C, do
not declare a temporary to initialize an aggregate assigned to
Out or In_Out parameters whose type has no discriminants. This
avoids stack overflow errors at runtime.
pmderodat [Tue, 11 Dec 2018 11:09:57 +0000 (11:09 +0000)]
[Ada] Additionsal trasformations for unnesting in package bodies
This patch extends the previous algorithm for creating an explicit
elaboration procedure for a package body when expansion generates
subprograms in the statement part of the body. For unnesting to work
properly, these subprograms must appear within an explicit subprogram
body so that uplevel references can be placed in the proper activation
record.
Ongoing work for LLVM generation.
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* exp_ch7.adb (Check_Unnesting_Elaboration_Code): Extend
algorithm to cover subprograms generated in nested loops and in
exception handlers, in order to build an explicit elaboration
procedure in more complex cases.
pmderodat [Tue, 11 Dec 2018 11:09:46 +0000 (11:09 +0000)]
[Ada] Spurious visibility error on aspect Predicate
The GNAT-defined aspect Predicate has the same semantics as the Ada
aspect Dynamic_Predicate, including direct visibility to the components
of a record type to which the aspect applies.
The following must compile quietly:
gcc -c integer_stacks.ads
----
pragma SPARK_Mode (On);
with Bounded_Stacks;
package Integer_Stacks is
new Bounded_Stacks (Element => Integer, Default_Element => 0);
----
generic
type Element is private;
Default_Element : Element;
package Bounded_Stacks is
type Stack (Capacity : Positive) is tagged private
with Default_Initial_Condition => Empty (Stack);
function "=" (Left, Right : Stack) return Boolean;
function Extent (This : Stack) return Natural;
function Empty (This : Stack) return Boolean;
function Full (This : Stack) return Boolean;
procedure Reset (This : out Stack) with
Post'Class => Empty (This) and
not Full (This),
Global => null,
Depends => (This =>+ null);
procedure Push (This : in out Stack; Item : Element) with
Pre'Class => not Full (This) and
Extent (This) < This.Capacity,
Post'Class => not Empty (This) and
Extent (This) = Extent (This'Old) + 1,
Global => null,
Depends => (This =>+ Item);
procedure Pop (This : in out Stack; Item : out Element) with
Pre'Class => not Empty (This),
Post'Class => not Full (This) and
Extent (This) = Extent (This'Old) - 1,
Global => null,
Depends => (This =>+ null, Item => This);
function Peek (This : Stack) return Element with
Pre'Class => not Empty (This),
Global => null,
Depends => (Peek'Result => This);
private
type Contents is array (Positive range <>) of Element;
type Stack (Capacity : Positive) is tagged record
Values : Contents (1 .. Capacity); -- := (others => Default_Element);
-- Top : Natural;
Top : Natural := 0;
end record with Predicate => Top <= Capacity,
Annotate => (GNATprove,
Intentional,
"type ""Stack"" is not fully initialized",
"Because zeroing Top is sufficient");
end Bounded_Stacks;
----
package body Bounded_Stacks is
------------
-- Extent --
------------
function Extent (This : Stack) return Natural is
(This.Top);
-----------
-- Empty --
-----------
function Empty (This : Stack) return Boolean is
(This.Top = 0);
----------
-- Full --
----------
function Full (This : Stack) return Boolean is
(This.Top = This.Capacity);
-----------
-- Reset --
-----------
procedure Reset (This : out Stack) is
begin
This := (This.Capacity, Top => 0, others => <>);
-- This.Top := 0;
end Reset;
----------
-- Push --
----------
procedure Push (This : in out Stack; Item : Element) is
begin
This.Top := This.Top + 1;
This.Values (This.Top) := Item;
end Push;
---------
-- Pop --
---------
procedure Pop (This : in out Stack; Item : out Element) is
begin
Item := This.Values (This.Top);
This.Top := This.Top - 1;
end Pop;
----------
-- Peek --
----------
function Peek (This : Stack) return Element is
(This.Values (This.Top));
---------
-- "=" --
---------
function "=" (Left, Right : Stack) return Boolean is
begin
if Left.Top /= Right.Top then
return False;
else
for K in 1 .. Left.Top loop
if Left.Values (K) /= Right.Values (K) then
return False;
end if;
end loop;
return True;
end if;
end "=";
end Bounded_Stacks;
----
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* sem_ch13.adb (Check_Aspect_At_End_Of_Declarations,
Freeze_Entity_Checks): Process aspect Predicate in the same
fashion as aspect Dynamic_Predicate.
pmderodat [Tue, 11 Dec 2018 11:09:29 +0000 (11:09 +0000)]
[Ada] Fix an ACATS regression with CCG
2018-12-11 Javier Miranda <miranda@adacore.com>
gcc/ada/
* exp_aggr.adb (In_Object_Declaration): Removed.
(Is_CCG_Supported_Aggregate): New subprogram that replaces
In_Object_Declaration extending its functionality to indicate if
an aggregate is in a context supported by the CCG backend.
pmderodat [Tue, 11 Dec 2018 11:09:19 +0000 (11:09 +0000)]
[Ada] Spurious errors on aspect specifications in generic units
This patch fixes spurious errors on aspect specifications on record
types when the aspect expression references a component of the type that
is not a discriminant. The patch also cleans up the legality checks on
aspect specifications, and improves error message on illegal aspect
specifications whose expressions are not conformant between
specification and freeze point, because of changes in visibility.
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* sem_ch13.adb (Push_Type, Pop_Type): New procedures, used for
analysis of aspect expressions for record types, whose
components (not only discriminants) may be referenced in aspect
expressions.
(Analyze_Aspect_Specifications, Analyze_Aspects_At_Freeze_Point,
Analyze_Aspect_At_End-Of_Declarations,
Resolve_Aspect_Expressions): Use the new subprograms.
(Check_Aspect_At_End_Of_Declarations): Improve error message.
(Build_Predicate_Functions): Do not build their bodies in a
generic unit.
(Is_Derived_Type_With_Constraint): New subprogram to uncover and
reject aspect specificationss on types that appear after the
type is frozen.
* sem_ch13.ads (Push_Scope_And_Install_Discriminants,
Uninstall_Discriminants_And_Pop_Scope): Remove.
* sem_ch6.adb, sem_ch6.ads (Fully_Conformant_Expressions):
Additional parameter to improve error message on illegal aspect
specifications whose resolution differ between aspect
specification and freeze point.
* freeze.adb: Remove references to
Install/Uninstall_Discriminants.
gcc/testsuite/
* gnat.dg/aspect1.adb, gnat.dg/aspect1_horizontal.adb,
gnat.dg/aspect1_horizontal.ads, gnat.dg/aspect1_vectors_2d.ads:
New testcase.
* gnat.dg/static_pred1.adb: Expect an error message.
pmderodat [Tue, 11 Dec 2018 11:09:08 +0000 (11:09 +0000)]
[Ada] Crash on nesting of subunits with bodies acting as specs
This patch corrects an issue whereby a set of nested subunits including
subprogram subunits acting as bodies would cause a crash when a child
subunit "withs" an ansestor in certain instances due to a mismanagment
of the scope stack.
------------
-- Source --
------------
-- w.ads
package W is
end;
-- w-b.ads
package W.B is
pragma Elaborate_Body;
end;
-- w-b.adb
with W.B.C;
package body w.B is end;
-- w-b-c.adb
with W;
procedure W.B.C is
package D is
procedure E;
end;
package body D is separate;
begin
null;
end;
-- w-b-c-d.adb
separate (W.B.C)
package body D is
procedure E is separate;
end;
-- w-b-c-d-e.adb
with W;
separate (W.B.C.D)
procedure E is
begin
null;
end;
* sem_ch10.adb (Analyze_Subunit): Modify conditional to fully
remove parent contexts from library-level subprogram bodies in
addition to package bodies.
pmderodat [Tue, 11 Dec 2018 11:09:04 +0000 (11:09 +0000)]
[Ada] Suppress call to Initial_Condition when the annotation is ignored
This patch suppresses the generation of the Initial_Condition procedure
tasked with verifying the run-time semantics of the pragma when the
pragma is ignored by means of -gnata, pragma Assertion_Policy, etc.
* exp_prag.adb (Expand_Pragma_Initial_Condition): Do not
generate an Initial_Condition procedure and a call to it when
the associated pragma is ignored.
* sem_ch10.adb (Analyze_Compilation_Unit): Minor cleanup.
pmderodat [Tue, 11 Dec 2018 11:08:45 +0000 (11:08 +0000)]
[Ada] Almost always inline init. procedure of small and simple records
2018-12-11 Eric Botcazou <ebotcazou@adacore.com>
gcc/ada/
* fe.h (Debug_Generated_Code): Declare.
* gcc-interface/gigi.h (enum inline_status_t): Rename
is_disabled to is_default, is_enabled to is_requested and add
is_prescribed.
* gcc-interface/decl.c (inline_status_for_subprog): New function.
(gnat_to_gnu_entity) <E_Subprogram_Type>: Use it to get the
inlining status of the subprogram.
* gcc-interface/trans.c (gigi): Adjust to above renaming.
(build_raise_check): Likewise.
(Compilation_Unit_to_gnu): Likewise.
(gnat_to_gnu): Likewise.
* gcc-interface/utils.c (create_subprog_decl): Likewise. Deal
with is_prescribed status by setting
DECL_DISREGARD_INLINE_LIMITS. Do not set the
DECL_NO_INLINE_WARNING_P flag if Debug_Generated_Code is true.
segher [Tue, 11 Dec 2018 08:30:36 +0000 (08:30 +0000)]
rs6000: Don't use rs6000_isa_flags_explicit for soft float tests (PR88145)
...specifically, those for builtins. Soft float can be enabled
implicitly, too (for certain CPUs for example). We should use
rs6000_isa_flags instead, to decide whether to expand a builtin or
to bail out with an error instead.
PR target/88145
* config/rs6000/rs6000.c (rs6000_expand_zeroop_builtin): Use
rs6000_isa_flags instead of rs6000_isa_flags_explicit to decide
whether soft float is enabled.
(rs6000_expand_mtfsb_builtin): Ditto.
(rs6000_expand_set_fpscr_rn_builtin): Ditto.
(rs6000_expand_set_fpscr_drn_builtin): Ditto.
kargl [Mon, 10 Dec 2018 18:05:37 +0000 (18:05 +0000)]
2018-12-10 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/88269
* io.c (io_constraint): Update macro. If locus line buffer is NULL,
use gfc_current_locus in error messages.
(check_io_constraints): Catch missing IO UNIT in write and read
statements. io_constraint macro is incompatible here.
2018-12-10 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/88269
* gfortran.dg/pr88269.f90: New test.
uros [Mon, 10 Dec 2018 15:47:16 +0000 (15:47 +0000)]
PR target/88418
* config/i386/i386.c (ix86_expand_sse_cmp): For vector modes,
check operand 1 with vector_operand predicate.
(ix86_expand_sse_movcc): For vector modes, check op_true with
vector_operand, not nonimmediate_operand.
testsuite/ChangeLog:
PR target/88418
* gcc.target/i386/pr88418.c: New test.
redi [Mon, 10 Dec 2018 15:25:02 +0000 (15:25 +0000)]
Make test for Filesystem TS actually use the Filesystem TS
This test was copied from 27_io/filesystem/path/query/is_absolute.cc but
should have been modified to test the path type from the TS instead of
std::filesystem::path.
* testsuite/experimental/filesystem/path/query/is_absolute.cc: Fix
test to use TS, not C++17.
rguenth [Mon, 10 Dec 2018 14:20:18 +0000 (14:20 +0000)]
2018-12-10 Richard Biener <rguenther@suse.de>
* tree-affine.c (tree_to_aff_combination): Remove unreachable
MEM_REF case.
(aff_combination_expand): Cache on SSA names, not possibly
on conversion trees. Avoid expanding cases we do not handle.
rguenth [Mon, 10 Dec 2018 13:56:51 +0000 (13:56 +0000)]
2018-12-10 Richard Biener <rguenther@suse.de>
PR tree-optimization/88427
* vr-values.c (vr_values::extract_range_from_phi_node):
Handle symbolic ranges conservatively when trying to drop
to Inf +- 1.
ebotcazou [Mon, 10 Dec 2018 11:16:47 +0000 (11:16 +0000)]
* c-c++-common/patchable_function_entry-decl.c: Pass -mcpu=gr6 for
Visium and remove other specific handling.
* c-c++-common/patchable_function_entry-default.c: Likewise.
* c-c++-common/patchable_function_entry-definition.c: Likewise.
amker [Mon, 10 Dec 2018 06:54:02 +0000 (06:54 +0000)]
* auto-profile.c (string_table::get_index_by_decl)
(function_instance::get_function_instance_by_decl): Avoid infinite
recursion by using DECL_FROM_INLINE.
ibuclaw [Sun, 9 Dec 2018 19:55:54 +0000 (19:55 +0000)]
Generate and scan documentation output in Ddoc tests.
The tests in gdc.test/compilable/ddoc*.d don't require the module to be
compiled all the way down to object code. Instead, only compile the
test sources with -fdoc, and scan the generated html content.
gcc/testsuite/ChangeLog:
PR d/88039
* gdc.test/gdc-test.exp (gdc-convert-args): Handle -D.
(dmd2dg): Check generated html in ddoc tests.
(gdc-do-test): Set dg-do-what-default to compile for ddoc tests.
tkoenig [Sun, 9 Dec 2018 18:54:47 +0000 (18:54 +0000)]
2018-12-09 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/88411
* io/transfer.c (dta_transfer_init): Do not treat as an
asynchronous statement unless the statement has
ASYNCHRONOUS="YES".
(st_write_done): Likewise.
(st_read_done): Do not perform async_wait for synchronous I/O
on an async unit.
(st_read_done): Likewise.
2018-12-09 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/88411
* testsuite/libgomp.fortran/async_io_8.f90: New test.
Backports VRP fixes from the D front-end implementation to the C++ port,
and fixes errors reported by ubsan build where the conversion from D
didn't include adjusting integer suffixes from 'UL' to 'ULL'.
tschwinge [Sun, 9 Dec 2018 12:49:20 +0000 (12:49 +0000)]
Correct the reported line number in Fortran combined OpenACC directives
gcc/fortran/
* trans-openmp.c (gfc_trans_oacc_combined_directive): Set the
location of combined acc loops.
gcc/testsuite/
* gfortran.dg/goacc/combined-directives-3.f90: New file.
Reviewed-by: Thomas Schwinge <thomas@codesourcery.com>
git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@266924 138bc75d-0d04-0410-961f-82ee72b054a4
tschwinge [Sun, 9 Dec 2018 12:48:26 +0000 (12:48 +0000)]
Correct the reported line number in C++ combined OpenACC directives
The C++ FE doesn't set the expr_location of the split acc loop in combined acc
parallel/kernels loop directives. This only happens for with combined
directives, otherwise cp_parser_omp_construct would be responsible for setting
the location. After fixing this bug, I was able to resolve a couple of long
standing diagnostics discrepancies between the C/C++ FEs in the test suite.
tschwinge [Sun, 9 Dec 2018 12:47:47 +0000 (12:47 +0000)]
Split up "gfortran.dg/goacc/loop-2.f95"
gcc/testsuite/
* gfortran.dg/goacc/loop-2.f95: Split into...
* gfortran.dg/goacc/loop-2-kernels-nested.f95: ... this new
file...
* gfortran.dg/goacc/loop-2-kernels-tile.f95: ..., and this new
file...
* gfortran.dg/goacc/loop-2-kernels.f95: ..., and this new file...
* gfortran.dg/goacc/loop-2-parallel-3.f95: ..., and this new
file...
* gfortran.dg/goacc/loop-2-parallel-nested.f95: ..., and this new
file...
* gfortran.dg/goacc/loop-2-parallel-tile.f95: ..., and this new
file...
* gfortran.dg/goacc/loop-2-parallel.f95: ..., and this new file.
tschwinge [Sun, 9 Dec 2018 12:47:35 +0000 (12:47 +0000)]
[PR88420] Fortran OpenACC "Clause SEQ conflicts with INDEPENDENT"
The Fortran front end declares that the OpenACC "Clause SEQ conflicts with
INDEPENDENT". While that combination doesn't make too much sense indeed, it's
still valid; these are orthogonal concepts.
jakub [Sat, 8 Dec 2018 13:37:49 +0000 (13:37 +0000)]
PR fortran/88304
* tree-nested.c (convert_local_reference_stmt): Handle clobbers where
lhs is not a decl normally, don't call use_pointer_in_frame on that
lhs.
* gfortran.fortran-torture/compile/pr88304-2.f90: New test.
ebotcazou [Sat, 8 Dec 2018 11:07:53 +0000 (11:07 +0000)]
PR rtl-optimization/88390
* dwarf2cfi.c (struct dw_cfi_row): Add window_save field.
(cfi_row_equal_p): Compare it.
(dwarf2out_frame_debug_cfa_window_save): Add FAKE parameter.
If FAKE is false, set window_save of the current row.
(dwarf2out_frame_debug) <REG_CFA_TOGGLE_RA_MANGLE>: Call above
function with FAKE parameter set to true.
<REG_CFA_WINDOW_SAVE>: Likewise but with FAKE parameter set to false.
(change_cfi_row): Emit a DW_CFA_GNU_window_save if necessary.
jakub [Sat, 8 Dec 2018 08:58:24 +0000 (08:58 +0000)]
PR libgomp/87995
* testsuite/libgomp.c-c++-common/cancel-taskgroup-3.c: Require
tls_runtime effective target.
(t): New threadprivate variable.
(main): Set t in threads which execute iterations of the worksharing
loop. Propagate that to the task after the loop and don't abort
if the current taskgroup hasn't been cancelled.
paolo [Fri, 7 Dec 2018 17:48:39 +0000 (17:48 +0000)]
/cp
2018-12-07 Paolo Carlini <paolo.carlini@oracle.com>
* decl2.c (grokbitfield): Use DECL_SOURCE_LOCATION in error messages
about bit-fields with function type, warn_if_not_aligned type, and
static bit-fields; avoid DECL_NAME for unnamed declarations.
/testsuite
2018-12-07 Paolo Carlini <paolo.carlini@oracle.com>
* g++.dg/other/bitfield7.C: New.
* g++.dg/parse/bitfield8.C: Likewise.
* g++.dg/parse/bitfield9.C: Likewise.
* g++.dg/pr53037-4.C: Test the locations too.
bergner [Fri, 7 Dec 2018 17:33:55 +0000 (17:33 +0000)]
gcc/
PR target/87496
* config/rs6000/rs6000.c (rs6000_option_override_internal): Disallow
-mabi=ieeelongdouble and -mabi=ibmlongdouble without -mlong-double-128.
Do not error for -mabi=ibmlongdouble and no ISA 2.06 support.
* doc/invoke.texi: Document -mabi=ibmlongdouble and -mabi=ieeelongdouble
require -mlong-double-128.
gcc/testsuite/
PR target/87496
* gcc.target/powerpc/pr87496.c: Rename from this...
* gcc.target/powerpc/pr87496-1.c: ...to this. Update comment.
* gcc.target/powerpc/pr87496-2.c: New test.
* gcc.target/powerpc/pr87496-3.c: New test.
ktkachov [Fri, 7 Dec 2018 17:05:21 +0000 (17:05 +0000)]
[AArch64][2/2] Add sve_width -moverride tunable
On top of the previous patch that implements TARGET_ESTIMATED_POLY_VALUE
and adds an sve_width tuning field to the CPU structs, this patch implements
an -moverride knob to adjust this sve_width field to allow for experimentation.
Again, reminder that this only has an effect when compiling for VLA-SVE that is,
without msve-vector-bits=<foo>. This just adjusts tuning heuristics in the compiler,,
like profitability thresholds for vectorised versioned loops, and others.
It can be used, for example like -moverride=sve_width=256 to set the sve_width
tuning field to 256. Widths outside of the accepted SVE widths [128 - 2048] are rejected
as you'd expect.
law [Fri, 7 Dec 2018 17:02:11 +0000 (17:02 +0000)]
* gimple-ssa-evrp-analyze.h (class evrp_range_analyzer): Add
m_update_global_ranges member. Add corresponding argument to ctor.
* gimple-ssa-evrp-analyze.c
(evrp_range_analyzer::evrp_range_analyzer): Add new argument and
initialize m_update_global_ranges.
(evrp_range_analyzer::set_ssa_range_info): Assert that we are
updating global ranges.
(evrp_range_analyzer::record_ranges_from_incoming_edge): Only
update global ranges if explicitly requested.
(evrp_range_analyzer::record_ranges_from_phis): Similarly.
(evrp_range_analyzer::record_ranges_from_stmt): Similarly.
* gimple-ssa-evrp.c (evrp_dom_walker): Pass new argument to
evrp_range_analyzer ctor.
* gimple-ssa-sprintf.c (sprintf_dom_walker): Similarly.
* tree-ssa-dom.c (dom_opt_dom_walker): Similarly.
* gcc.c-torture/builtins/strnlen.x: New file to filter -Og from
options to test.
The hook TARGET_ESTIMATED_POLY_VALUE allows a target to give an estimate for a poly_int run-time value.
It is used exclusively in tuning decisions, things like estimated loop iterations, probabilities etc.
It is not relied on for correctness.
If we know the SVE width implemented in hardware we can make more more
informed decisions in the implementation of TARGET_ESTIMATED_POLY_VALUE,
even when compiling for VLA vectorisation.
This patch adds an sve_width field to our tuning structs and sets it for
the current CPU tunings.
A new value is introduced to the aarch64_sve_vector_bits_enum enum that indicates
that SVE is not available: SVE_NOT_IMPLEMENTED. I set it to the same value as SVE_SCALABLE
so that parts of the aarch64 backend that follow the pattern:
if (vector_width == SVE_SCALABLE)
do_vla_friendly_action ()
else
assume_specific_width_for_correctness ()
continue to work without change, but the CPU tuning structs can use a more
appropriate moniker for indicating the absence of SVE.
This sets sve_width to SVE_NOT_IMPLEMENTED for all cores.
I aim to add an -moverride switch in the next patch that allows a power user to experiment
with different values of it for investigations.
rsandifo [Fri, 7 Dec 2018 15:03:15 +0000 (15:03 +0000)]
[AArch64][SVE] Remove unnecessary PTRUEs from integer arithmetic
When using the unpredicated immediate forms of MUL, LSL, LSR and ASR,
the rtl patterns would still have the predicate operand we created for
the other forms. This patch splits the patterns after reload in order
to get rid of the predicate, like we already do for WHILE.
2018-12-07 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* config/aarch64/aarch64-sve.md (*mul<mode>3, *v<optab><mode>3):
Split the patterns after reload if we don't need the predicate
operand.
(*post_ra_mul<mode>3, *post_ra_v<optab><mode>3): New patterns.
gcc/testsuite/
* gcc.target/aarch64/sve/pred_elim_2.c: New test.
rsandifo [Fri, 7 Dec 2018 15:01:47 +0000 (15:01 +0000)]
[AArch64][SVE] Remove unnecessary PTRUEs from FP arithmetic
When using the unpredicated all-register forms of FADD, FSUB and FMUL,
the rtl patterns would still have the predicate operand we created for
the other forms. This patch splits the patterns after reload in order
to get rid of the predicate, like we already do for WHILE.
2018-12-07 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* config/aarch64/iterators.md (SVE_UNPRED_FP_BINARY): New code
iterator.
(sve_fp_op): Handle minus and mult.
* config/aarch64/aarch64-sve.md (*add<mode>3, *sub<mode>3)
(*mul<mode>3): Split the patterns after reload if we don't
need the predicate operand.
(*post_ra_<sve_fp_op><mode>3): New pattern.
gcc/testsuite/
* gcc.target/aarch64/sve/pred_elim_1.c: New test.
nickc [Fri, 7 Dec 2018 10:33:30 +0000 (10:33 +0000)]
Add a recursion limit to libiberty's demangling code. The limit is enabled by default, but can be disabled via a new demangling option.
include * demangle.h (DMGL_NO_RECURSE_LIMIT): Define.
(DEMANGLE_RECURSION_LIMIT): Define
PR 87681
PR 87675
PR 87636
PR 87350
PR 87335
libiberty * cp-demangle.h (struct d_info): Add recursion_level field.
* cp-demangle.c (d_function_type): Add recursion counter.
If the recursion limit is reached and the check is not disabled,
then return with a failure result.
(cplus_demangle_init_info): Initialise the recursion_level field.
(d_demangle_callback): If the recursion limit is enabled, check
for a mangled string that is so long that there is not enough
stack space for the local arrays.
* cplus-dem.c (struct work): Add recursion_level field.
(squangle_mop_up): Set the numb and numk fields to zero.
(work_stuff_copy_to_from): Handle the case where a btypevec or
ktypevec field is NULL.
(demangle_nested_args): Add recursion counter. If
the recursion limit is not disabled and reached, return with a
failure result.
jakub [Thu, 6 Dec 2018 23:29:04 +0000 (23:29 +0000)]
PR fortran/88377
* trans-openmp.c (gfc_omp_clause_default_ctor,
gfc_omp_clause_copy_ctor, gfc_omp_clause_assign_op,
gfc_omp_clause_linear_ctor, gfc_omp_clause_dtor): Only consider
GFC_DECL_GET_SCALAR_ALLOCATABLE vars as scalar allocatables if they
have pointer type.
jakub [Thu, 6 Dec 2018 23:28:04 +0000 (23:28 +0000)]
PR c/88367
* tree-vrp.c (extract_range_from_binary_expr): For POINTER_PLUS_EXPR
with -fno-delete-null-pointer-checks, set_nonnull only if the pointer
is non-NULL and offset is known to have most significant bit clear.
* vr-values.c (vr_values::vrp_stmt_computes_nonzero): For ADDR_EXPR
of MEM_EXPR, return true if the MEM_EXPR has non-zero offset with
most significant bit clear. If offset does have most significant bit
set and -fno-delete-null-pointer-checks, don't return true even if
the base pointer is non-NULL.
paolo [Thu, 6 Dec 2018 23:20:16 +0000 (23:20 +0000)]
/cp
2018-12-06 Paolo Carlini <paolo.carlini@oracle.com>
* class.c (check_bitfield_decl): In error message about non-integral
type print the type itself too.
* decl.c (grokdeclarator): Do not ICE on unnamed bit-fields declared
friends; when calling build_decl for a FIELD_DECL possibly pass the
declarator->id_loc.
/testsuite
2018-12-06 Paolo Carlini <paolo.carlini@oracle.com>
projects / thirdparty / gcc.git / log
