2017-09-19 Richard Biener <rguenther@suse.de>
* graphite-scop-detection.c (scop_detection::can_represent_loop):
Do not iterate to sibling loops but only to siblings of inner
loops.
From-SVN: r252962
* config/i386/i386.c (fold_builtin_cpu): Add M_AMDFAM17H
to processor_model and "amdfam17h" to arch_names_table.
* doc/extend.texi (__builtin_cpu_is): Document amdfam17h CPU name.
From-SVN: r252948
Previously SLP_TREE_NUMBER_OF_VEC_STMTS was calculated while scheduling
an SLP tree after analysis, but sometimes it can be useful to know the
value during analysis too. This patch moves the calculation to
vect_slp_analyze_node_operations instead.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vectorizer.h (vect_slp_analyze_operations): Replace parameters
with a vec_info *.
* tree-vect-loop.c (vect_analyze_loop_operations): Update call
accordingly.
* tree-vect-slp.c (vect_slp_analyze_node_operations): Add vec_info *
parameter. Set SLP_TREE_NUMBER_OF_VEC_STMTS here rather than in
vect_schedule_slp_instance.
(vect_slp_analyze_operations): Replace parameters with a vec_info *.
Update call to vect_slp_analyze_node_operations. Simplify return
value.
(vect_slp_analyze_bb_1): Update call accordingly.
(vect_schedule_slp_instance): Remove vectorization_factor parameter.
Don't calculate SLP_TREE_NUMBER_OF_VEC_STMTS here.
(vect_schedule_slp): Update call accordingly.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252935
The vectoriser was calling vect_get_smallest_scalar_type without
having proven that the type actually is a scalar. This seems to
be the intended behaviour: the ultimate test of whether the type
is interesting (and hence scalar) is whether an associated vector
type exists, but this is only tested later.
The patch simply makes the function cope gracefully with non-scalar
inputs.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vect-data-refs.c (vect_get_smallest_scalar_type): Cope
with types that aren't in fact scalar.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252934
Without this we'd pick an unrolling factor based purely on longs,
ignoring the ints. It's posssible that vect_get_smallest_scalar_type
should also handle shifts, but I think we'd still want this as a
belt-and-braces fix.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
* tree-vect-slp.c (vect_record_max_nunits): New function,
split out from...
(vect_build_slp_tree_1): ...here.
(vect_build_slp_tree_2): Call it for phis too.
gcc/testsuite/
* gcc.dg/vect/slp-multitypes-13.c: New test.
From-SVN: r252933
vectorizable_mask_load_store was not passing the required mask type to
vect_get_vec_def_for_operand. This doesn't matter for masks that are
defined in the loop, since their STMT_VINFO_VECTYPE will be what we need
anyway. But it's not possible to tell which mask type the caller needs
when looking at an invariant scalar boolean. As the comment above the
function says:
In case OP is an invariant or constant, a new stmt that creates a vector def
needs to be introduced. VECTYPE may be used to specify a required type for
vector invariant.
This fixes the attached testcase for SVE.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
* tree-vect-stmts.c (vectorizable_mask_load_store): Pass mask_vectype
to vect_get_vec_def_for_operand when getting the mask operand.
gcc/testsuite/
* gfortran.dg/vect/mask-store-1.f90: New test.
From-SVN: r252932
This patch changes the type of the multiplier applied by
vectorizable_live_operation from unsigned_type_node to bitsizetype,
which matches the type of TYPE_SIZE and is the type expected of a
BIT_FIELD_REF bit position. This is shown by existing tests when
SVE is added.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vect-loop.c (vectorizable_live_operation): Fix type of
bitstart.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252931
vectorizable_live_operation needs to use BIT_FIELD_REF to extract one
element of a vector. For a packed vector boolean type, the number of
bits to extract should be taken from TYPE_PRECISION rather than TYPE_SIZE.
This is shown by existing tests once SVE is added.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vect-loop.c (vectorizable_live_operation): Fix element size
calculation for vector booleans.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252930
vect_transform_stmt calls vectorizable_live_operation for
each live statement in an SLP node, but vect_analyze_stmt
only called it the once. This patch makes vect_analyze_stmt
consistent with vect_transform_stmt, which should be a bit
more robust, and also means that a later patch can use
slp_index when deciding validity.
2017-09-18 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vect-stmts.c (can_vectorize_live_stmts): New function,
split out from...
(vect_transform_stmt): ...here.
(vect_analyze_stmt): Use it instead of calling
vectorizable_live_operation directly.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252929
PR libstdc++/71187
* include/std/type_traits (__declval): New function to deduce return
type of declval.
(__declval_protector::_delegate): Remove.
(declval): Use __declval instead of add_rvalue_reference and
__declval_protector::__delegate.
* testsuite/20_util/declval/requirements/1_neg.cc: Adjust dg-error
lineno.
* testsuite/20_util/make_signed/requirements/typedefs_neg.cc:
Likewise.
* testsuite/20_util/make_unsigned/requirements/typedefs_neg.cc:
Likewise.
From-SVN: r252922
gcc/ada/
2017-09-18 Bob Duff <duff@adacore.com>
* sem_ch4.adb (Complete_Object_Operation): Do not insert 'Access for
reference types in the access-to-access case.
2017-09-18 Eric Botcazou <ebotcazou@adacore.com>
* sem_attr.adb (Analyze_Access_Attribute): Move check for the presence
of the "aliased" keyword on the prefix from here to...
(Resolve_Attribute) <Attribute_Access>: ...here. Remove useless call
to Check_No_Implicit_Aliasing.
* sinfo.ads (Non_Aliased_Prefix): Delete.
(Set_Non_Aliased_Prefix): Likewise.
* sinfo.adb (Non_Aliased_Prefix): Delete.
(Set_Non_Aliased_Prefix): Likewise.
2017-09-18 Bob Duff <duff@adacore.com>
* exp_ch5.adb (Build_Formal_Container_Iteration,
Expand_Formal_Container_Element_Loop): Convert the container to the
root type before passing it to the iteration operations, so it will be
of the right type.
2017-09-18 Bob Duff <duff@adacore.com>
* einfo.ads, validsw.ads, treepr.ads, sem_util.ads: Comment fixes.
2017-09-18 Bob Duff <duff@adacore.com>
* exp_ch3.adb (Build_Array_Init_Proc): If validity checking is enabled,
and it's a bit-packed array, pass False to the Consider_IS parameter of
Needs_Simple_Initialization.
2017-09-18 Hristian Kirtchev <kirtchev@adacore.com>
* sem_ch6.adb (Check_Inline_Pragma): Link the newly generated spec to
the preexisting body.
* sem_prag.adb (Check_Inline_Always_Placement): New routine.
(Process_Inline): Verify the placement of pragma Inline_Always. The
pragma must now appear on the initial declaration of the related
subprogram.
2017-09-18 Ed Schonberg <schonberg@adacore.com>
* sem_ch3.adb (Analyze_Declarations): In ASIS mode, At the end of the
declarative list in a subprogram body, analyze aspext specifications to
provide basic semantic information, because otherwise the aspect
specifications might only be snalyzed during expansion, when related
subprograms are generated.
2017-09-18 Bob Duff <duff@adacore.com>
* exp_ch9.adb (Is_Simple_Barrier_Name): Follow Original_Node, in case
validity checks have rewritten the tree.
2017-09-18 Bob Duff <duff@adacore.com>
* sem_util.adb: Comment fixes, and remove redundant Is_Itype check.
2017-09-18 Ed Schonberg <schonberg@adacore.com>
* sem_ch12.adb (Save_References_In_Aggregate): When constructing a
qualified exxpression for an aggregate in a generic unit, verify that
the scope of the type is itself visible and not hidden, so that the
qualified expression is correctly resolved in any instance.
gcc/testsuite/
2017-09-18 Bob Duff <duff@adacore.com>
* gnat.dg/validity_check.adb: New testcase.
2017-09-18 Eric Botcazou <ebotcazou@adacore.com>
* gnat.dg/overload.ads, gnat.dg/overload.adb: New testcase.
2017-09-18 Bob Duff <duff@adacore.com>
* gnat.dg/tagged_prefix_call.adb: New testcase.
From-SVN: r252916
2017-09-18 Bob Duff <duff@adacore.com>
* sem_ch4.adb (Analyze_Qualified_Expression): Give an error if the type
mark refers to the current instance. Set the type to Any_Type in that
case, to avoid later crashes.
2017-09-18 Ed Schonberg <schonberg@adacore.com>
* exp_ch3.adb (Replace_Discriminant_References): New procedure,
subsidiary of Build_Assignment, used to handle the initialization code
for a mutable record component whose default value is an aggregate that
sets the values of the discriminants of the components.
2017-09-18 Ed Schonberg <schonberg@adacore.com>
* gnat.dg/default_variants.adb: New testcase.
2017-09-18 Eric Botcazou <ebotcazou@adacore.com>
* sem_ch13.adb (Analyze_Attribute_Definition_Clause) <Address>: Mark
the entity as being volatile for an overlay that toggles the scalar
storage order.
2017-09-18 Fedor Rybin <frybin@adacore.com>
* doc/gnat_ugn/gnat_utility_programs.rst: Document that gnattest
options -U main and --harness-only are not compatible.
From-SVN: r252913
2017-09-18 Bob Duff <duff@adacore.com>
Alternate fix for PR ada/71358
* libgnat/g-comlin.adb (Getopt): Remove manual null access checks.
Instead, make a local copy of Config, and if it's null, allocate an
empty Command_Line_Configuration_Record, so we won't crash on null
pointer dereference.
From-SVN: r252909
rs6000_return_addr assumes that the stack link is at frame+0, which is
true for count>0. For count==0, rs6000_return_addr is called with
frame==frame_pointer_rtx and the stack link is *not* at frame+0 if
-fstack-protector-all or -fsanitize=address because rs6000.h sets
FRAME_GROWS_DOWNWARD for those options.
PR target/81996
* gcc/config/rs6000/rs6000.c (rs6000_return_addr): Use
stack_pointer_rtx for count 0. Update comments. Break up
large rtl expression.
From-SVN: r252901
gcc:
config/i386/i386.c: (xlogue_layout::STUB_NAME_MAX_LEN): Increase to 20
bytes.
(xlogue_layout::s_stub_names): Add an additional size-2 diminsion.
(xlogue_layout::get_stub_name): Modify to select the appropairate sse
or avx version of the stub.
gcc/testsuite:
gcc.target/i386/pr82196-1.c: New test.
gcc.target/i386/pr82196-2.c: Likewise.
libgcc:
config/i386/i386-asm.h (PASTE2): New macro.
(ASMNAME): Modify to use PASTE2.
(MS2SYSV_STUB_PREFIX): New macro for isa prefix.
(MS2SYSV_STUB_BEGIN, MS2SYSV_STUB_END): New macros for stub headers.
config/i386/resms64.S: Rename to a header file, use MS2SYSV_STUB_BEGIN
instead of HIDDEN_FUNC and MS2SYSV_STUB_END instead of FUNC_END.
config/i386/resms64f.S: Likewise.
config/i386/resms64fx.S: Likewise.
config/i386/resms64x.S: Likewise.
config/i386/savms64.S: Likewise.
config/i386/savms64f.S: Likewise.
config/i386/avx_resms64.S: New file that only defines a macro and
includes it's corresponding header file.
config/i386/avx_resms64f.S: Likewise.
config/i386/avx_resms64fx.S: Likewise.
config/i386/avx_resms64x.S: Likewise.
config/i386/avx_savms64.S: Likewise.
config/i386/avx_savms64f.S: Likewise.
config/i386/sse_resms64.S: Likewise.
config/i386/sse_resms64f.S: Likewise.
config/i386/sse_resms64fx.S: Likewise.
config/i386/sse_resms64x.S: Likewise.
config/i386/sse_savms64.S: Likewise.
config/i386/sse_savms64f.S: Likewise.
config/i386/t-msabi: Modified to add avx and sse versions of stubs.
From-SVN: r252896
preferred_stack_boundary may not be the minimum stack alignment. For
leaf functions without TLS access, max_used_stack_slot_alignment may be
smaller. We should update preferred_stack_boundary for leaf functions.
gcc/
PR target/82166
* config/i386/i386.c (ix86_finalize_stack_frame_flags): Properly
compute the minimum stack alignment. Also update preferred stack
boundary for leaf functions.
gcc/testsuite/
PR target/82166
* gcc.target/i386/pr82166.c: New test.
From-SVN: r252895
2017-09-17 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82173
* decl.c (gfc_get_pdt_instance): Use the component initializer
expression for the default, rather than the parameter value.
* resolve.c (resolve_pdt): New function.
(resolve_symbol): Call it. Remove false error, prohibiting
deferred type parameters for dummy arguments.
PR fortran/60483
* primary.c (gfc_match_varspec): If the type of an associate
name is unknown and yet there is a match, try resolving the
target expression and using its type.
2017-09-17 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82173
* gfortran.dg/pdt_1.f03 : Eliminate spurious error checks.
* gfortran.dg/pdt_2.f03 : The same.
* gfortran.dg/pdt_3.f03 : The same.
* gfortran.dg/pdt_4.f03 : Add 'modtype' and two new errors in
module 'bad_vars'. Add error concerning assumed parameters and
save attribute.
* gfortran.dg/pdt_11.f03 : New test.
PR fortran/60483
* gfortran.dg/associate_9.f90 : Remove XFAIL and change to run.
* gfortran.dg/associate_25.f90 : New test.
* gfortran.dg/pdt_12.f03 : New test.
From-SVN: r252894
This should have been after the early exit for non-vectorised statements.
2017-09-16 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
PR tree-optimization/82228
* tree-vect-loop.c (vectorizable_live_operation): Move initialization
of ncopies.
From-SVN: r252888
in nds32_option_optimization_table.
gcc/
* common/config/nds32/nds32-common.c
(nds32_option_optimization_table): Refine formatting.
(nds32_option_optimization_table): Use -fsched-pressure and
-fomit-frame-pointer for specific optimization level.
From-SVN: r252876
2017-09-16 Tom de Vries <tom@codesourcery.com>
PR c/81875
* c-parser.c (c_parser_omp_for_loop): Fold only operands of cond, not
cond itself.
* testsuite/libgomp.c-c++-common/pr81875.c: New test.
From-SVN: r252873