As PR91790 exposed, when we have one slp node whose first_stmt_info_for_drptr
is different from first_stmt_info, it's possible that the first_stmt DR isn't
initialized yet before stmt SLP_TREE_SCALAR_STMTS[0] of slp node. So we
shouldn't use first_stmt_info for vect_setup_realignment, instead we can use
the one based on first_stmt_info_for_drptr DR with additional adjustment by
bumping the distance from first_stmt DR.
gcc/ChangeLog
2019-11-27 Kewen Lin <linkw@gcc.gnu.org>
PR tree-optimization/91790
* gcc/tree-vect-stmts.c (vectorizable_load): Use the adjusted DR for
vect_setup_realignment when first_stmt_info is different from
first_stmt_info_for_drptr.
From-SVN: r278760
2019-11-27 Richard Biener <rguenther@suse.de>
PR tree-optimization/92645
* tree-ssa-forwprop.c (simplify_vector_constructor): Handle
CTORs with just a subset of the original vectors.
* gcc.target/i386/pr92645-2.c: New testcase.
* gcc.target/i386/pr92645-3.c: Likewise.
From-SVN: r278758
2019-11-27 Richard Biener <rguenther@suse.de>
PR middle-end/92674
* tree-inline.c (expand_call_inline): Delay purging EH/abnormal
edges and instead record blocks in bitmap.
(gimple_expand_calls_inline): Adjust.
(fold_marked_statements): Delay EH cleanup until all folding is
done.
(optimize_inline_calls): Do EH/abnormal cleanup for calls after
inlining finished.
From-SVN: r278757
I noticed that tsubst of a TEMPLATE_ID_EXPR was losing the location
information from its EXPR_LOCATION. Then I noticed that
cxx_eval_constant_expression was also throwing away location information for
variable references.
* pt.c (tsubst_copy_and_build) [TEMPLATE_ID_EXPR]: Remember the
location of a variable template-id.
* constexpr.c (cxx_eval_constant_expression): Get expr location
before stripping location wrappers.
(non_const_var_error): Take location argument.
From-SVN: r278755
gcc/testsuite/ChangeLog:
PR tree-optimization/92683
* gcc.dg/strcmpopt_8.c: New test.
* gcc.dg/strcmpopt_9.c: New test.
gcc/ChangeLog:
PR tree-optimization/92683
* gimple-fold.c (gimple_fold_builtin_string_compare): Restore a test
inadvertently removed in a previous change. Rename local variable
for clarity.
From-SVN: r278742
This patch updates the vect_char_add target selector to use its own
cache instead of the vect_int cache.
This was causing a situation where bb-slp-40.c would fail on sparc when
run after other tests that use the vect_int target selector, but pass
when run on its own.
2019-11-26 Joel Hutton <Joel.Hutton@arm.com>
gcc/testsuite/
PR testsuite/92391
* lib/target-supports.exp (check_effective_target_vect_char_add):
Use a separate cache entry from vect_int.
From-SVN: r278738
This fixes the s390 bootstrap by undefining existing defines before
redefining them.
gcc/ChangeLog:
2019-11-26 Robin Dapp <rdapp@linux.ibm.com>
* config/s390/linux.h: Add undef for MUSL_DYNAMIC_LINKERxx.
From-SVN: r278721
2019-11-26 Richard Biener <rguenther@suse.de>
PR tree-optimization/92645
* tree-vect-slp.c (vect_build_slp_tree_2): For unary ops
do not build the operation from scalars if the operand is.
* gcc.target/i386/pr92645.c: New testcase.
From-SVN: r278719
2019-11-25 Harald Anlauf <anlauf@gmx.de>
PR fortran/92629
* simplify.c (convert_mpz_to_unsigned): Skip assert for argument
range when -fno-range-check is specified.
PR fortran/92629
* gfortran.dg/pr92629.f90: New testcase.
From-SVN: r278699
This implements most of the remaining C++20 additions to the <iterator>
header.
* include/bits/iterator_concepts.h (ranges::iter_swap): Fix parameter
types of poison pill overload. Use remove_reference_t when checking
constraints.
* include/bits/stl_iterator.h (move_sentinel): Define for C++20.
(move_iterator): Adjust definitions of nested types for C++20. Add
hidden friends for move_sentinel operations, iter_move and iter_swap.
(common_iterator, counted_iterator): Define for C++20.
* testsuite/24_iterators/move_iterator/cust.cc: New test.
* testsuite/24_iterators/move_iterator/sentinel.cc: New test.
* testsuite/24_iterators/common_iterator/1.cc: New test.
* testsuite/24_iterators/counted_iterator/1.cc: New test.
From-SVN: r278698
Code that directly uses _Decimal* types on architectures not
supporting DFP is properly diagnosed ("error: decimal floating-point
not supported for this target"), via a call to
targetm.decimal_float_supported_p, if the _Decimal32, _Decimal64 or
_Decimal128 keywords are used to access it. Use via mode attributes
is also diagnosed ("unable to emulate 'SD'"); so is use of the
FLOAT_CONST_DECIMAL64 pragma. However, it is possible to access those
types via typeof applied to constants or built-in functions without
such an error. I expect that there are ways to get an ICE from this;
certainly it uses a completely undefined ABI.
This patch arranges for the types not to exist in the compiler at all
when DFP is not supported. As is done with unsupported _FloatN /
_FloatNx types, the global tree nodes are left as NULL_TREE, and the
built-in function machinery is made to use error_mark_node for them in
that case in builtin-types.def, so that the built-in functions are
unavailable. Code handling constants is adjusted to give an error,
and other code that might not work with the global tree nodes being
NULL_TREE is also updated.
Bootstrapped with no regressions for x86_64-pc-linux-gnu. Also tested
with no regressions for cross to aarch64-linux-gnu, as a configuration
without DFP support.
PR c/91985
gcc:
* builtin-types.def (BT_DFLOAT32, BT_DFLOAT64, BT_DFLOAT128)
(BT_DFLOAT32_PTR, BT_DFLOAT64_PTR, BT_DFLOAT128_PTR): Define to
error_mark_node if corresponding global tree node is NULL.
* tree.c (build_common_tree_nodes): Do not initialize
dfloat32_type_node, dfloat64_type_node or dfloat128_type_node if
decimal floating-point not supported.
gcc/c:
* c-decl.c (finish_declspecs): Use int instead of decimal
floating-point types if decimal floating-point not supported.
gcc/c-family:
* c-common.c (c_common_type_for_mode): Handle decimal
floating-point types being NULL_TREE.
* c-format.c (get_format_for_type_1): Handle specified types being
NULL_TREE.
* c-lex.c (interpret_float): Give an error for decimal
floating-point constants when decimal floating-point not
supported.
gcc/lto:
* lto-lang.c (lto_type_for_mode): Handle decimal floating-point
types being NULL_TREE.
gcc/testsuite:
* gcc.dg/c2x-no-dfp-1.c, gcc.dg/gnu2x-builtins-no-dfp-1.c: New
tests.
* gcc.dg/fltconst-pedantic-dfp.c: Expect errors when decimal
floating-point not supported.
From-SVN: r278684
attribs.c has code to ignore all scoped attributes appertaining to
types except when they are part of the definition of that type.
I think the premise of that code is incorrect, and its presence is a
bug; such attributes are clearly valid in both C and C++, which
explicitly specify that attributes in certain syntactic positions
appertain to a particular type, only for that use of that type and not
for other uses of the same type specifiers without that attribute
specified, and while the standard attributes in C2x aren't relevant in
such contexts, some gnu:: attributes certainly are. Where some
attributes are invalid on some types in such contexts, that's a matter
for the individual attribute handlers to diagnose (or the front end if
the requirements on a standard attribute in the standard are more
strict than those of a handler shared with a GNU attribute).
Thus, this patch removes the bogus code to allow such attributes to be
used. Doing so (and adding tests for attributes in such positions)
shows up that the logic in the C front end for creating the
c_declarator structures for such attributes put them in the wrong
place relative to the structures for function and array types, and the
logic for postfix attributes on a list of declaration specifiers
failed to handle some cases, so those bugs are also fixed in this
patch.
Bootstrapped with no regressions for x86_64-pc-linux-gnu.
gcc:
* attribs.c (decl_attributes): Do not ignore C++11 attributes on
types.
gcc/c:
* c-tree.h (struct c_declarator): Use a structure for id member.
* c-decl.c (grokdeclarator): Extract attributes from cdk_id
declarators at the start, not when handling individual declarators
later. Use u.id.id instead of u.id.
(grokfield): Use u.id.id instead of u.id.
(build_id_declarator): Set u.id.id and u.id.attrs.
(finish_declspecs): Handle postfix attributes in case of typedef
name or typeof used.
* c-parser.c (c_parser_direct_declarator)
(c_parser_direct_declarator_inner): Place declarator for
attributes inside that for function or array, not outside. Set
u.id.attrs for identifiers.
(c_parser_parameter_declaration): Use u.id.id instead of u.id.
* gimple-parser.c (c_parser_gimple_declaration): Use u.id.id
instead of u.id.
gcc/testsuite:
* gcc.dg/gnu2x-attrs-1.c: Do not expect message about attributes
appertaining to types.
* gcc.dg/gnu2x-attrs-2.c: New test.
* g++.dg/cpp0x/gen-attrs-1.C, g++.dg/cpp0x/gen-attrs-22.C,
g++.dg/cpp0x/gen-attrs-4.C, g++.dg/cpp0x/lambda/lambda-attr1.C:
Update expected diagnostics.
From-SVN: r278683
This removes the call to fesetround as FE_TONEAREST
is the default and so the guard can be removed as well.
The test will then run as long as there's int128 support.
gcc/testsuite/ChangeLog:
* gcc.dg/torture/fp-int-convert-timode-1.c: Always run if int128.
From-SVN: r278680
2019-11-25 Richard Biener <rguenther@suse.de>
* tree-vect-slp.c (vect_detect_hybrid_slp_stmts): Add assertion.
(vect_detect_hybrid_slp): Swap lane and instance iteration,
properly re-building the visited hash-map for each lane.
From-SVN: r278679
PR ada/92362
* gcc-interface/trans.c (gnat_to_gnu) <N_Attribute_Definition_Clause>:
Use a temporary instead of clobbering the result with a freeze node.
From-SVN: r278675
The -fdec option enables the use of Hollerith comparisons in comparisons
with INTEGER, COMPLEX, REAL and CHARACTER expressions.
Mark Eggleston <mark.eggleston@codethink.com>
Jim MacArthur <jim.macarthur@codethink.co.uk>
* gfortran.texi: Update Hollerith constants support for character types
and use in comparisons.
* invoke.texi: Tidy up list of options. Update description of
-fdec-char-conversions.
* resolve.c (is_character_based): New.
(Convert_hollerith_to_character): New. (convert_to_numeric): New.
(resolve_operator): If both sides are character based and -fdec is
enabled convert Hollerith to character. If an operand is Hollerith, the
other is numeric and -fdec is enabled convert to numeric.
(resolve_ordinary_assign): Add check for -fdec-char-conversions for
assignment of character literals.
Mark Eggleston <mark.eggleston@codethink.com>
Jim MacArthur <jim.macarthur@codethink.co.uk>
* gfortran.dg/dec-comparison-character_1.f90: New test.
* gfortran.dg/dec-comparison-character_2.f90: New test.
* gfortran.dg/dec-comparison-character_3.f90: New test.
* gfortran.dg/dec-comparison-complex_1.f90: New test.
* gfortran.dg/dec-comparison-complex_2.f90: New test.
* gfortran.dg/dec-comparison-complex_3.f90: New test.
* gfortran.dg/dec-comparison-int_1.f90: New test.
* gfortran.dg/dec-comparison-int_2.f90: New test.
* gfortran.dg/dec-comparison-int_3.f90: New test.
* gfortran.dg/dec-comparison-real_1.f90: New test.
* gfortran.dg/dec-comparison-real_2.f90: New test.
* gfortran.dg/dec-comparison-real_3.f90: New test.
* gfortran.dg/dec-comparison.f90: New test.
Co-Authored-By: Jim MacArthur <jim.macarthur@codethink.co.uk>
From-SVN: r278674
2019-11-25 Martin Jambor <mjambor@suse.cz>
PR ipa/92109
* cgraph.h (cgraph_node::remove_from_clone_tree): Declare.
* cgraphclones.c (cgraph_node::remove_from_clone_tree): New method.
(cgraph_materialize_clone): Move removel from clone tree to the
the new method and use it instead.
* ipa.c (symbol_table::remove_unreachable_nodes): When removing
bodies of clones, also remove it from the clone tree.
From-SVN: r278670
2019-11-25 Martin Jambor <mjambor@suse.cz>
PR ipa/91956
* ipa-sra.c (process_isra_node_results): Put the new node to the
same comdat group as the original node.
testsuite/
* g++.dg/ipa/pr91956.C: New test.
From-SVN: r278669
gcc/
Build double32 / long-double32 multilibs if needed.
PR target/92055
* config/avr/t-avr:
(HAVE_DOUBLE_MULTILIB, HAVE_LONG_DOUBLE_MULTILIB): Remove vars.
(HAVE_DOUBLE32, HAVE_LONG_DOUBLE32, WITH_LONG_DOUBLE)
(HAVE_DOUBLE64, HAVE_LONG_DOUBLE64, WITH_DOUBLE): Set from
tm_defines and pass to genmultilib.awk.
* config/avr/genmultilib.awk: Use these variables to add double32
and / or long-double32 multilib(s) as needed.
* config/avr/driver-avr.c (avr_double_lib): Adjust comment.
From-SVN: r278668
2019-11-25 Richard Biener <rguenther@suse.de>
* cfgloop.h (get_loop_exit_edges): Add extra parameter denoting
loop body, defaulted to NULL.
(single_likely_exit): Add exit vector argument
* tree-ssa-loop-niter.h (loop_only_exit_p): Add loop body argument.
(number_of_iterations_exit): Likewise.
(number_of_iterations_exit_assumptions): Likewise.
* cfgloop.c (get_loop_exit_edges): Use passed in loop body
if not NULL.
* cfgloopanal.c (single_likely_exit): Use passed in exit vector.
* tree-ssa-loop-ivcanon.c (canonicalize_loop_induction_variables):
Compute exit vector around call to single_likely_exit.
* tree-ssa-loop-ivopts.c (tree_ssa_iv_optimize_loop): Pass down
loop body to loop_only_exit_p.
* tree-ssa-loop-niter.c (loop_only_exit_p): Get loop body from
caller.
(number_of_iterations_exit_assumptions): Get loop body from caller
if not NULL.
(number_of_iterations_exit): Pass through new loop body arg.
(infer_loop_bounds_from_undefined): Get loop body from caller.
(estimate_numbers_of_iterations): Compute loop body once.
From-SVN: r278667
2019-11-25 Richard Biener <rguenther@suse.de>
* cfganal.c (pre_and_rev_post_order_compute_fn): Use an
auto_bb_flag instead of an sbitmap for visited handling.
From-SVN: r278666
This is a subsequent patch to refactor the existing float point
vector comparison operator supports. The patch to fix PR92132
supplemented vector float point comparison by exposing the names
for unordered/ordered/uneq/ltgt and adding ungt/unge/unlt/unle/
ne. As Segher pointed out, some patterns can be refactored
together. The main link on this is:
https://gcc.gnu.org/ml/gcc-patches/2019-11/msg00452.html
gcc/ChangeLog
2019-11-25 Kewen Lin <linkw@gcc.gnu.org>
* config/rs6000/vector.md (vector_fp_comparison_simple): New code iterator.
(vector_fp_comparison_complex): Likewise.
(vector_<code><mode> for VEC_F and vector_fp_comparison_simple): New
define_and_split.
(vector_<code><mode> for VEC_F and vector_fp_comparison_complex): Likewise.
(vector_lt<mode> for VEC_F): Refactor with vector_fp_comparison_simple.
(vector_le<mode> for VEC_F): Likewise.
(vector_unge<mode> for VEC_F): Likewise.
(vector_unle<mode> for VEC_F): Likewise.
(vector_ne<mode> for VEC_F): Likewise.
(vector_ungt<mode> for VEC_F): Likewise.
(vector_unlt<mode> for VEC_F): Likewise.
(vector_ltgt<mode> for VEC_F): Refactor with vector_fp_comparison_complex.
(vector_ordered<mode> for VEC_F): Likewise.
(vector_uneq<mode> for VEC_F): Likewise.
(vector_unordered<mode> for VEC_F): Likewise.
From-SVN: r278665
