Jan Hubicka [Sat, 9 Nov 2019 17:45:29 +0000 (18:45 +0100)]
ipa-fnsummary.c (ipa_call_context::estimate_size_and_time): Fix calculation of min_size.
* ipa-fnsummary.c (ipa_call_context::estimate_size_and_time): Fix
calculation of min_size.
(ipa_update_overall_fn_summary): Likewise.
From-SVN: r278006
Jan Hubicka [Sat, 9 Nov 2019 17:37:38 +0000 (18:37 +0100)]
ipa-fnsummary.c (estimate_edge_size_and_time): Do not call estimate_edge_devirt_benefit when not computing hints...
* ipa-fnsummary.c (estimate_edge_size_and_time): Do not call
estimate_edge_devirt_benefit when not computing hints;
do not compute time when not asked for.
(estimate_calls_size_and_time): Pass NULL hints and time when
these are not computed; do not evaluate hint predicates when these are
not computed.
(ipa_merge_fn_summary_after_inlining): Do not re-evaluate edge
frequency.
From-SVN: r278005
Jakub Jelinek [Sat, 9 Nov 2019 17:09:44 +0000 (18:09 +0100)]
re PR tree-optimization/92401 (ICE in fold_ternary_loc, at fold-const.c:11698)
PR tree-optimization/92401
* gimple-match-head.c (gimple_resimplify1): Call const_unop only
if res_op->code is an expression with code length 1.
* gimple-match-head.c (gimple_resimplify2): Call const_binop only
if res_op->code is an expression with code length 2.
* gimple-match-head.c (gimple_resimplify3): Call fold_ternary only
if res_op->code is an expression with code length 3.
* g++.dg/opt/pr92401.C: New test.
From-SVN: r278004
Thomas Koenig [Sat, 9 Nov 2019 14:54:19 +0000 (14:54 +0000)]
Commit symbol for external BLAS routine when translating MATMUL to *GEMM.
2019-11-09 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/92321
* frontend-passes.c (call_external_blas): Commit symbol for
external BLAS routine.
2019-11-09 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/92321
* gfortran.dg/matmul_blas_2.f90: New test.
From-SVN: r278003
Iain Sandoe [Sat, 9 Nov 2019 13:43:04 +0000 (13:43 +0000)]
[Darwin, machopic 10/n] Rework X86 mcount stub code.
When a stub is used to call the mcount function, the code is already
marking it as used unconditionally; This is the only use of the so-
called validation outside darwin.{h,c}. This moves the 'validation'
into darwin.c which is a step towards making validation routine local.
gcc/
2019-11-09 Iain Sandoe <iain@sandoe.co.uk>
* config/darwin.c (machopic_mcount_stub_name): Validate the
symbol stub name when it is created.
* config/i386/darwin.h (FUNCTION_PROFILER): Remove the symbol
stub validation.
From-SVN: r278000
GCC Administrator [Sat, 9 Nov 2019 00:16:19 +0000 (00:16 +0000)]
Daily bump.
From-SVN: r277999
Jakub Jelinek [Fri, 8 Nov 2019 23:56:37 +0000 (00:56 +0100)]
symtab.c: Fix comment typos.
* symtab.c: Fix comment typos.
* cgraphunit.c: Likewise.
* cgraph.h: Likewise.
* cgraphclones.c: Likewise.
* cgraph.c: Likewise.
* varpool.c: Likewise.
* tree-ssa-strlen.c: Likewise.
* ipa-sra.c: Likewise.
(scan_expr_access, check_all_callers_for_issues): Fix typo
in a dump message.
From-SVN: r277995
Jakub Jelinek [Fri, 8 Nov 2019 23:55:45 +0000 (00:55 +0100)]
dec_char_conversion_in_assignment_4.f90: Use dg-do compile instead of dg-do run.
* gfortran.dg/dec_char_conversion_in_assignment_4.f90: Use
dg-do compile instead of dg-do run.
* gfortran.dg/dec_char_conversion_in_data_3.f90: Likewise.
From-SVN: r277994
Iain Sandoe [Fri, 8 Nov 2019 22:09:30 +0000 (22:09 +0000)]
[Darwin] Add include guard to darwin-protos.h
The Darwin protos header is missing an include guard, this adds one.
gcc/ChangeLog:
2019-11-08 Iain Sandoe <iain@sandoe.co.uk>
* config/darwin-protos.h: Add include quard.
From-SVN: r277993
Marek Polacek [Fri, 8 Nov 2019 21:48:47 +0000 (21:48 +0000)]
PR c++/92215 - flawed diagnostic for bit-field with non-integral type.
I noticed that for code like
struct S {
int *foo : 3;
};
we generate nonsensical
r.C:2:8: error: function definition does not declare parameters
2 | int *foo : 3;
It talks about a function because after parsing the declspecs of 'foo' we don't
see either ':' or "name :", so we think it's not a bit-field decl. So we parse
the declarator and since a ctor-initializer begins with a ':', we try to parse
it as a function body, generating the awful diagnostic. With this patch, we
issue:
r.C:2:8: error: bit-field ‘foo’ has non-integral type ‘int*’
2 | int *foo : 3;
* parser.c (cp_parser_member_declaration): Add a diagnostic for
bit-fields with non-integral types.
* g++.dg/diagnostic/bitfld4.C: New test.
From-SVN: r277991
Marek Polacek [Fri, 8 Nov 2019 21:40:45 +0000 (21:40 +0000)]
PR c++/92058 - constinit malfunction in static data member.
* g++.dg/cpp2a/constinit15.C: New test.
From-SVN: r277990
Andrew MacLeod [Fri, 8 Nov 2019 17:51:40 +0000 (17:51 +0000)]
Modify range_operator::fold_range() and wi_fold () to return via reference.
2019-11-08 Andrew MacLeod <amacleod@redhat.com>
* range-op.h (range_operator::fold_range): Return result in a
reference parameter instead of by value.
(range_operator::wi_fold): Same.
* range-op.cc (range_operator::wi_fold): Return result in a reference
parameter instead of by value.
(range_operator::fold_range): Same.
(value_range_from_overflowed_bounds): Same.
(value_range_with_overflow): Same
(create_possibly_reversed_range): Same.
(operator_equal::fold_range): Same.
(operator_not_equal::fold_range): Same.
(operator_lt::fold_range): Same.
(operator_le::fold_range): Same.
(operator_gt::fold_range): Same.
(operator_ge::fold_range): Same.
(operator_plus::wi_fold): Same.
(operator_plus::op1_range): Change call to fold_range.
(operator_plus::op2_range): Change call to fold_range.
(operator_minus::wi_fold): Return result via reference parameter.
(operator_minus::op1_range): Change call to fold_range.
(operator_minus::op2_range): Change call to fold_range.
(operator_min::wi_fold): Return result via reference parameter.
(operator_max::wi_fold): Same.
(cross_product_operator::wi_cross_product): Same.
(operator_mult::wi_fold): Same.
(operator_div::wi_fold): Same.
(operator_div op_floor_div): Fix whitespace.
(operator_exact_divide::op1_range): Change call to fold_range.
(operator_lshift::fold_range): Return result via reference parameter.
(operator_lshift::wi_fold): Same.
(operator_rshift::fold_range): Same.
(operator_rshift::wi_fold): Same.
(operator_cast::fold_range): Same.
(operator_cast::op1_range): Change calls to fold_range.
(operator_logical_and::fold_range): Return result via reference.
(wi_optimize_and_or): Adjust call to value_range_with_overflow.
(operator_bitwise_and::wi_fold): Return result via reference.
(operator_logical_or::fold_range): Same.
(operator_bitwise_or::wi_fold): Same.
(operator_bitwise_xor::wi_fold): Same.
(operator_trunc_mod::wi_fold): Same.
(operator_logical_not::fold_range): Same.
(operator_bitwise_not::fold_range): Same.
(operator_bitwise_not::op1_range): Change call to fold_range.
(operator_cst::fold_range): Return result via reference.
(operator_identity::fold_range): Same.
(operator_abs::wi_fold): Same.
(operator_absu::wi_fold): Same.
(operator_negate::fold_range): Same.
(operator_negate::op1_range): Change call to fold_range.
(operator_addr_expr::fold_range): Return result via reference.
(operator_addr_expr::op1_range): Change call to fold_range.
(operator_pointer_plus::wi_fold): Return result via reference.
(operator_pointer_min_max::wi_fold): Same.
(operator_pointer_and::wi_fold): Same.
(operator_pointer_or::wi_fold): Same.
(range_op_handler): Change call to fold_range.
(range_cast): Same.
* tree-vrp.c (range_fold_binary_symbolics_p): Change call to
fold_range.
(range_fold_unary_symbolics_p): Same.
(range_fold_binary_expr): Same.
(range_fold_unary_expr): Same.
From-SVN: r277979
Jakub Jelinek [Fri, 8 Nov 2019 16:48:41 +0000 (17:48 +0100)]
* arith.c (character2representation): Change i type to size_t.
From-SVN: r277978
Richard Sandiford [Fri, 8 Nov 2019 16:08:03 +0000 (16:08 +0000)]
Use correct vector type in neutral_op_for_slp_reduction
With the new reduction vectype handling, neutral_op_for_slp_reduction
needs to know whether the caller is using STMT_VINFO_REDUC_VECTYPE
(for an epilogue value) or STMT_VINFO_VECTYPE (for a PHI argument).
This fixes various gcc.target/aarch64/sve/slp_* tests.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-loop.c (neutral_op_for_slp_reduction): Take the
vector type as an argument rather than reading it from the
stmt_vec_info.
(vect_create_epilog_for_reduction): Update accordingly.
(vectorizable_reduction): Likewise.
(vect_transform_cycle_phi): Likewise.
From-SVN: r277977
Segher Boessenkool [Fri, 8 Nov 2019 14:53:42 +0000 (15:53 +0100)]
rs6000: Fix branch_comparison_operator
* config/rs6000/predicates.md (branch_comparison_operator): Allow only
the comparison codes that make sense for the mode used, and only the
codes that can be done with a single branch instruction.
From-SVN: r277976
Mark Eggleston [Fri, 8 Nov 2019 14:28:57 +0000 (14:28 +0000)]
Allow CHARACTER literals in assignments and data statements.
Allows character literals to used to assign values to non-character variables
in the same way that Hollerith constants are used. In addition character
literals can be used in data statements just like Hollerith constants.
Warnings of such use are output to discourage this usage as it is a non-standard
legacy feature and must be explicitly enabled.
Enabled by -fdec and -fdec-char-conversions.
Co-Authored-By: Jim MacArthur <jim.macarthur@codethink.co.uk>
From-SVN: r277975
Andre Vieira [Fri, 8 Nov 2019 13:52:56 +0000 (13:52 +0000)]
[vect] PR 92351: When peeling for alignment make alignment of epilogues unknown
gcc/ChangeLog:
2019-11-08 Andre Vieira <andre.simoesdiasvieira@arm.com>
PR tree-optimization/92351
* tree-vect-data-refs.c (vect_compute_data_ref_alignment): When we are
peeling the main loop for alignment, make sure to set the misalignment
of the epilogue's data references to DR_MISALIGNMENT_UNKNOWN.
gcc/testsuite/ChangeLog:
2019-11-08 Andre Vieira <andre.simoesdiasvieira@arm.com>
PR tree-optimization/92351
* gcc.dg/vect/vect-peel-2.c: Disable epilogue vectorization and
split the source of this test to...
* gcc.dg/vect/vect-peel-2-src.c: ... This.
* gcc.dg/vect/vect-peel-2-epilogues.c: New test.
From-SVN: r277974
Richard Biener [Fri, 8 Nov 2019 13:16:28 +0000 (13:16 +0000)]
dbgcnt.def (ivopts_loop): Add.
2019-11-08 Richard Biener <rguenther@suse.de>
* dbgcnt.def (ivopts_loop): Add.
* tree-ssa-loop-ivopts.c (tree_ssa_iv_optimize): Check
ivopts_loop before optimizing a loop.
From-SVN: r277973
Richard Biener [Fri, 8 Nov 2019 13:15:40 +0000 (13:15 +0000)]
re PR ipa/92409 (r277920 causes ICE in gcc.dg/cast-function-1.c)
2019-11-08 Richard Biener <rguenther@suse.de>
PR ipa/92409
* tree-inline.c (declare_return_variable): Properly handle
type mismatches for the return slot.
From-SVN: r277972
Eric Botcazou [Fri, 8 Nov 2019 12:45:24 +0000 (12:45 +0000)]
overflow-1.c: Add -fno-pie to the options.
* gcc.target/sparc/overflow-1.c: Add -fno-pie to the options.
* gcc.target/sparc/overflow-2.c: Likewise.
From-SVN: r277969
Eric Botcazou [Fri, 8 Nov 2019 12:30:47 +0000 (12:30 +0000)]
re PR target/92095 (internal error with -O1 -mcpu=niagara2 -fPIE)
PR target/92095
* config/sparc/sparc-protos.h (output_load_pcrel_sym): Declare.
* config/sparc/sparc.c (sparc_cannot_force_const_mem): Revert latest
change.
(got_helper_needed): New static variable.
(output_load_pcrel_sym): New function.
(get_pc_thunk_name): Remove after inlining...
(load_got_register): ...here. Rework the initialization of the GOT
register and of the GOT helper.
(save_local_or_in_reg_p): Test the REGNO of the GOT register.
(sparc_file_end): Test got_helper_needed to decide whether the GOT
helper must be emitted. Use output_asm_insn instead of fprintf.
(sparc_init_pic_reg): In PIC mode, always initialize the PIC register
if optimization is enabled.
* config/sparc/sparc.md (load_pcrel_sym<P:mode>): Emit the assembly
by calling output_load_pcrel_sym.
From-SVN: r277966
Richard Sandiford [Fri, 8 Nov 2019 11:58:45 +0000 (11:58 +0000)]
Fix code order in tree-sra.c:create_access
If get_ref_base_and_extent returns poly_int offsets or sizes,
tree-sra.c:create_access prevents SRA from being applied to the base.
However, we haven't verified by that point that we have a valid base
to disqualify.
This originally led to an ICE on the attached testcase, but it
no longer triggers there after the introduction of IPA SRA.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-sra.c (create_access): Delay disqualifying the base
for poly_int values until we know we have a base.
gcc/testsuite/
* gcc.target/aarch64/sve/acle/general/inline_2.c: New test.
From-SVN: r277965
Andre Vieira [Fri, 8 Nov 2019 11:30:50 +0000 (11:30 +0000)]
[vect] Disable vectorization of epilogues for loops with SIMDUID set
gcc/ChangeLog:
2019-11-08 Andre Vieira <andre.simoesdiasvieira@arm.com>
* tree-vect-loop.c (vect_analyze_loop): Disable epilogue vectorization
for loops with SIMDUID set. Enable epilogue vectorization for loops
with SIMDLEN set after finding a main loop with a VF that matches it.
From-SVN: r277964
Jakub Jelinek [Fri, 8 Nov 2019 10:53:50 +0000 (11:53 +0100)]
re PR target/92038 (Extremely inefficient x86_64 code for trivally copyable types passed in registers.)
PR target/92038
* gimple-ssa-store-merging.c (find_constituent_stores): For return
value only, return non-NULL if there is a single non-clobber
constituent store even if there are constituent clobbers and return
one of clobber constituent stores if all constituent stores are
clobbers.
(split_group): Handle clobbers.
(imm_store_chain_info::output_merged_store): When computing
bzero_first, look after all clobbers at the start. Don't count
clobber stmts in orig_num_stmts, except if the first orig store is
a clobber covering the whole area and split_stores cover the whole
area, consider equal number of stmts ok. Punt if split_stores
contains only ->orig stores and their number plus number of original
clobbers is equal to original number of stmts. For ->orig, look past
clobbers in the constituent stores.
(imm_store_chain_info::output_merged_stores): Don't remove clobber
stmts.
(rhs_valid_for_store_merging_p): Don't return false for clobber stmt
rhs.
(store_valid_for_store_merging_p): Allow clobber stmts.
(verify_clear_bit_region_be): Fix up a thinko in function comment.
* g++.dg/opt/store-merging-1.C: New test.
* g++.dg/opt/store-merging-2.C: New test.
* g++.dg/opt/store-merging-3.C: New test.
From-SVN: r277963
Jakub Jelinek [Fri, 8 Nov 2019 10:52:50 +0000 (11:52 +0100)]
re PR middle-end/92384 (Empty class instances have different equal testing result among GCC versions)
PR c++/92384
* function.c (assign_parm_setup_block, assign_parm_setup_stack): Don't
copy TYPE_EMPTY_P arguments from data->entry_parm to data->stack_parm
slot.
(assign_parms): For TREE_ADDRESSABLE parms with TYPE_EMPTY_P type
force creation of a unique data.stack_parm slot.
* g++.dg/torture/pr92384.C: New test.
From-SVN: r277962
Richard Biener [Fri, 8 Nov 2019 10:21:43 +0000 (10:21 +0000)]
genmatch.c (expr::gen_transform): Use the resimplify member function instead of hard-coding the...
2019-11-08 Richard Biener <rguenther@suse.de>
* genmatch.c (expr::gen_transform): Use the resimplify
member function instead of hard-coding the gimple_resimplifyN variant.
(dt_simplify::gen_1): Likewise.
From-SVN: r277961
Martin Liska [Fri, 8 Nov 2019 09:59:09 +0000 (10:59 +0100)]
Rename identifiers in a test-case.
2019-11-08 Martin Liska <mliska@suse.cz>
* g++.dg/pr92339.C: Rename identifiers to something
more readable.
From-SVN: r277960
Richard Sandiford [Fri, 8 Nov 2019 09:43:44 +0000 (09:43 +0000)]
Handle POLY_INT_CST in copy_reference_ops_from_ref
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-ssa-sccvn.c (copy_reference_ops_from_ref): Handle
POLY_INT_CST.
gcc/testsuite/
* gcc.target/aarch64/sve/acle/general/deref_2.c: New test.
* gcc.target/aarch64/sve/acle/general/whilele_8.c: Likewise.
* gcc.target/aarch64/sve/acle/general/whilelt_4.c: Likewise.
From-SVN: r277959
Richard Biener [Fri, 8 Nov 2019 09:30:52 +0000 (09:30 +0000)]
re PR tree-optimization/92324 (ICE in expand_direct_optab_fn, at internal-fn.c:2890)
2019-11-08 Richard Biener <rguenther@suse.de>
PR tree-optimization/92324
* tree-vect-loop.c (vect_create_epilog_for_reduction): Use
STMT_VINFO_REDUC_VECTYPE for all computations, inserting
sign-conversions as necessary.
(vectorizable_reduction): Reject conversions in the chain
that are not sign-conversions, base analysis on a non-converting
stmt and its operation sign. Set STMT_VINFO_REDUC_VECTYPE.
* tree-vect-stmts.c (vect_stmt_relevant_p): Don't dump anything
for debug stmts.
* tree-vectorizer.h (_stmt_vec_info::reduc_vectype): New.
(STMT_VINFO_REDUC_VECTYPE): Likewise.
* gcc.dg/vect/pr92205.c: XFAIL.
* gcc.dg/vect/pr92324-1.c: New testcase.
* gcc.dg/vect/pr92324-2.c: Likewise.
From-SVN: r277958
Richard Sandiford [Fri, 8 Nov 2019 09:06:50 +0000 (09:06 +0000)]
Handle POLY_INT_CSTs in declare_return_value
SVE allows variable-length vectors to be returned by value,
which tripped the assert in declare_return_variable.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-inline.c (declare_return_variable): Check for poly_int_tree_p
instead of INTEGER_CST.
gcc/testsuite/
* gcc.target/aarch64/sve/acle/general/inline_1.c: New test.
From-SVN: r277956
Richard Biener [Fri, 8 Nov 2019 09:01:41 +0000 (09:01 +0000)]
re PR tree-optimization/92324 (ICE in expand_direct_optab_fn, at internal-fn.c:2890)
2019-11-08 Richard Biener <rguenther@suse.de>
PR tree-optimization/92324
* tree-vect-loop.c (vect_create_epilog_for_reduction): Use
STMT_VINFO_REDUC_VECTYPE for all computations, inserting
sign-conversions as necessary.
(vectorizable_reduction): Reject conversions in the chain
that are not sign-conversions, base analysis on a non-converting
stmt and its operation sign. Set STMT_VINFO_REDUC_VECTYPE.
* tree-vect-stmts.c (vect_stmt_relevant_p): Don't dump anything
for debug stmts.
* tree-vectorizer.h (_stmt_vec_info::reduc_vectype): New.
(STMT_VINFO_REDUC_VECTYPE): Likewise.
* gcc.dg/vect/pr92205.c: XFAIL.
* gcc.dg/vect/pr92324-1.c: New testcase.
* gcc.dg/vect/pr92324-2.c: Likewise.
From-SVN: r277955
Georg-Johann Lay [Fri, 8 Nov 2019 08:49:07 +0000 (08:49 +0000)]
re PR target/92055 ([avr] Support 64-bit double)
PR target/92055
* config/avr/avr.opt (-mdouble=, -mlong-double=):
Fix a missing '-' when displaying these options in the
help screen.
From-SVN: r277954
Richard Sandiford [Fri, 8 Nov 2019 08:40:29 +0000 (08:40 +0000)]
[AArch64] Remove unused mode iterators
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* config/aarch64/iterators.md (SVE_BH, SVE_BHS): Delete.
From-SVN: r277953
Martin Liska [Fri, 8 Nov 2019 08:39:17 +0000 (09:39 +0100)]
Make mklog more robust.
2019-11-08 Martin Liska <mliska@suse.cz>
* mklog: The script fails for patches that contain:
'---param=foo=bar xyz'.
From-SVN: r277952
Richard Sandiford [Fri, 8 Nov 2019 08:37:26 +0000 (08:37 +0000)]
[AArch64] Don't handle bswap in aarch64_builtin_vectorized_function
aarch64_builtin_vectorized_function no longer needs to handle bswap*
since we have internal functions and optabs for all supported cases.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* config/aarch64/aarch64-builtins.c
(aarch64_builtin_vectorized_function): Remove bswap handling.
From-SVN: r277951
Richard Sandiford [Fri, 8 Nov 2019 08:35:55 +0000 (08:35 +0000)]
[C] Opt out of GNU vector extensions for built-in SVE types
The AArch64 port defines built-in SVE types at start-up under names
like __SVInt8_t. These types are represented in the front end and
gimple as normal VECTOR_TYPEs and are code-generated as normal vectors.
However, we'd like to stop the frontends from treating them in the
same way as GNU-style ("vector_size") vectors, for several reasons:
(1) We allowed the GNU vector extensions to be mixed with Advanced SIMD
vector types and it ended up causing a lot of confusion on big-endian
targets. Although SVE handles big-endian vectors differently from
Advanced SIMD, there are still potential surprises; see the block
comment near the head of aarch64-sve.md for details.
(2) One of the SVE vectors is a packed one-bit-per-element boolean vector.
That isn't a combination the GNU vector extensions have supported
before. E.g. it means that vectors can no longer decompose to
arrays for indexing, and that not all elements are individually
addressable. It also makes it less clear which order the initialiser
should be in (lsb first, or bitfield ordering?). We could define
all that of course, but it seems a bit weird to go to the effort
for this case when, given all the other reasons, we don't want the
extensions anyway.
(3) The GNU vector extensions only provide full-vector operations,
which is a very artifical limitation on a predicated architecture
like SVE.
(4) The set of operations provided by the GNU vector extensions is
relatively small, whereas the SVE intrinsics provide many more.
(5) It makes it easier to ensure that (with default options) code is
portable between compilers without the GNU vector extensions having
to become an official part of the SVE intrinsics spec.
(6) The length of the SVE types is usually not fixed at compile time,
whereas the GNU vector extension is geared around fixed-length
vectors.
It's possible to specify the length of an SVE vector using the
command-line option -msve-vector-bits=N, but in principle it should
be possible to have functions compiled for different N in the same
translation unit. This isn't supported yet but would be very useful
for implementing ifuncs. Once mixing lengths in a translation unit
is supported, the SVE types should represent the same type throughout
the translation unit, just as GNU vector types do.
However, when -msve-vector-bits=N is in effect, we do allow conversions
between explicit GNU vector types of N bits and the corresponding SVE
types. This doesn't undermine the intent of (5) because in this case
the use of GNU vector types is explicit and intentional. It also doesn't
undermine the intent of (6) because converting between the types is just
a conditionally-supported operation. In other words, the types still
represent the same types throughout the translation unit, it's just that
conversions between them are valid in cases where a certain precondition
is known to hold. It's similar to the way that the SVE vector types are
defined throughout the translation unit but can only be used in functions
for which SVE is enabled.
The patch adds a new flag to tree_type_common to select this behaviour.
(We currently have 17 bits free.) To avoid making the flag too specific
to vectors, I called it TYPE_INDIVISIBLE_P, to mean that the frontend
should not allow the components of the type to be accessed directly.
This could perhaps be useful in future for hiding the fact that a
type is an array, or for hiding the fields of a record or union.
The actual frontend changes are very simple, mostly just replacing
VECTOR_TYPE_P with gnu_vector_type_p in selected places.
One interesting case is:
/* Need to convert condition operand into a vector mask. */
if (VECTOR_TYPE_P (TREE_TYPE (ifexp)))
{
tree vectype = TREE_TYPE (ifexp);
tree elem_type = TREE_TYPE (vectype);
tree zero = build_int_cst (elem_type, 0);
tree zero_vec = build_vector_from_val (vectype, zero);
tree cmp_type = build_same_sized_truth_vector_type (vectype);
ifexp = build2 (NE_EXPR, cmp_type, ifexp, zero_vec);
}
in build_conditional_expr. This appears to be trying to support
elementwise conditions like "vec1 ? vec2 : vec3", which is something
the C++ frontend supports. However, this code can never trigger AFAICT,
because "vec1" does not survive c_objc_common_truthvalue_conversion:
case VECTOR_TYPE:
error_at (location, "used vector type where scalar is required");
return error_mark_node;
Even if it did, the operation should be a VEC_COND_EXPR rather
than a COND_EXPR.
I've therefore left that condition as-is, but added tests for the
"vec1 ? vec2 : vec3" case to make sure that we don't accidentally
allow it for SVE vectors in future.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-core.h (tree_type_common::indivisible_p): New member variable.
* tree.h (TYPE_INDIVISIBLE_P): New macro.
* config/aarch64/aarch64-sve-builtins.cc (register_builtin_types):
Treat the vector types as indivisible.
gcc/c-family/
* c-common.h (gnu_vector_type_p): New function.
* c-common.c (c_build_vec_perm_expr): Require __builtin_shuffle
vectors to satisfy gnu_vector_type_p.
(c_build_vec_convert): Likewise __builtin_convertvector.
(convert_vector_to_array_for_subscript): Likewise when applying
implicit vector to array conversion.
(scalar_to_vector): Likewise when converting vector-scalar
operations to vector-vector operations.
gcc/c/
* c-convert.c (convert): Only handle vector conversions if one of
the types satisfies gnu_vector_type_p or if -flax-vector-conversions
allows it.
* c-typeck.c (build_array_ref): Only allow vector indexing if the
vectors satisfy gnu_vector_type_p.
(build_unary_op): Only allow unary operators to be applied to
vectors if they satisfy gnu_vector_type_p.
(digest_init): Only allow by-element initialization of vectors
if they satisfy gnu_vector_type_p.
(really_start_incremental_init): Likewise.
(push_init_level): Likewise.
(pop_init_level): Likewise.
(process_init_element): Likewise.
(build_binary_op): Only allow binary operators to be applied to
vectors if they satisfy gnu_vector_type_p.
gcc/testsuite/
* gcc.target/aarch64/sve/acle/general-c/gnu_vectors_1.c: New test.
* gcc.target/aarch64/sve/acle/general-c/gnu_vectors_2.c: Likewise.
From-SVN: r277950
Richard Sandiford [Fri, 8 Nov 2019 08:32:19 +0000 (08:32 +0000)]
Generalise gather and scatter optabs
The gather and scatter optabs required the vector offset to be
the integer equivalent of the vector mode being loaded or stored.
This patch generalises them so that the two vectors can have different
element sizes, although they still need to have the same number of
elements.
One consequence of this is that it's possible (if unlikely)
for two IFN_GATHER_LOADs to have the same arguments but different
return types. E.g. the same scalar base and vector of 32-bit offsets
could be used to load 8-bit elements and to load 16-bit elements.
From just looking at the arguments, we could wrongly deduce that
they're equivalent.
I know we saw this happen at one point with IFN_WHILE_ULT,
and we dealt with it there by passing a zero of the return type
as an extra argument. Doing the same here also makes the load
and store functions have the same argument assignment.
For now this patch should be a no-op, but later SVE patches take
advantage of the new flexibility.
2019-11-08 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* optabs.def (gather_load_optab, mask_gather_load_optab)
(scatter_store_optab, mask_scatter_store_optab): Turn into
conversion optabs, with the offset mode given explicitly.
* doc/md.texi: Update accordingly.
* config/aarch64/aarch64-sve-builtins-base.cc
(svld1_gather_impl::expand): Likewise.
(svst1_scatter_impl::expand): Likewise.
* internal-fn.c (gather_load_direct, scatter_store_direct): Likewise.
(expand_scatter_store_optab_fn): Likewise.
(direct_gather_load_optab_supported_p): Likewise.
(direct_scatter_store_optab_supported_p): Likewise.
(expand_gather_load_optab_fn): Likewise. Expect the mask argument
to be argument 4.
(internal_fn_mask_index): Return 4 for IFN_MASK_GATHER_LOAD.
(internal_gather_scatter_fn_supported_p): Replace the offset sign
argument with the offset vector type. Require the two vector
types to have the same number of elements but allow their element
sizes to be different. Treat the optabs as conversion optabs.
* internal-fn.h (internal_gather_scatter_fn_supported_p): Update
prototype accordingly.
* optabs-query.c (supports_at_least_one_mode_p): Replace with...
(supports_vec_convert_optab_p): ...this new function.
(supports_vec_gather_load_p): Update accordingly.
(supports_vec_scatter_store_p): Likewise.
* tree-vectorizer.h (vect_gather_scatter_fn_p): Take a vec_info.
Replace the offset sign and bits parameters with a scalar type tree.
* tree-vect-data-refs.c (vect_gather_scatter_fn_p): Likewise.
Pass back the offset vector type instead of the scalar element type.
Allow the offset to be wider than the memory elements. Search for
an offset type that the target supports, stopping once we've
reached the maximum of the element size and pointer size.
Update call to internal_gather_scatter_fn_supported_p.
(vect_check_gather_scatter): Update calls accordingly.
When testing a new scale before knowing the final offset type,
check whether the scale is supported for any signed or unsigned
offset type. Check whether the target supports the source and
target types of a conversion before deciding whether to look
through the conversion. Record the chosen offset_vectype.
* tree-vect-patterns.c (vect_get_gather_scatter_offset_type): Delete.
(vect_recog_gather_scatter_pattern): Get the scalar offset type
directly from the gs_info's offset_vectype instead. Pass a zero
of the result type to IFN_GATHER_LOAD and IFN_MASK_GATHER_LOAD.
* tree-vect-stmts.c (check_load_store_masking): Update call to
internal_gather_scatter_fn_supported_p, passing the offset vector
type recorded in the gs_info.
(vect_truncate_gather_scatter_offset): Update call to
vect_check_gather_scatter, leaving it to search for a valid
offset vector type.
(vect_use_strided_gather_scatters_p): Convert the offset to the
element type of the gs_info's offset_vectype.
(vect_get_gather_scatter_ops): Get the offset vector type directly
from the gs_info.
(vect_get_strided_load_store_ops): Likewise.
(vectorizable_load): Pass a zero of the result type to IFN_GATHER_LOAD
and IFN_MASK_GATHER_LOAD.
* config/aarch64/aarch64-sve.md (gather_load<mode>): Rename to...
(gather_load<mode><v_int_equiv>): ...this.
(mask_gather_load<mode>): Rename to...
(mask_gather_load<mode><v_int_equiv>): ...this.
(scatter_store<mode>): Rename to...
(scatter_store<mode><v_int_equiv>): ...this.
(mask_scatter_store<mode>): Rename to...
(mask_scatter_store<mode><v_int_equiv>): ...this.
From-SVN: r277949
Tobias Burnus [Fri, 8 Nov 2019 08:14:40 +0000 (09:14 +0100)]
Fortran] PR91253 fix continuation-line handling with -pre_include
PR fortran/91253
* scanner.c (skip_fixed_comments): Move comment
lines to next if block.
(gfc_next_char_literal): Fix continue_line setting.
(get_file): Remove bogus ATTRIBUTE_UNUSED.
From-SVN: r277948
Kewen Lin [Fri, 8 Nov 2019 07:37:07 +0000 (07:37 +0000)]
[rs6000]Fix PR92132 by adding vec_cmp and vcond_mask supports
To support full condition reduction vectorization, we have to define
vec_cmp* and vcond_mask_*. This patch is to add related expands.
Also add the missing vector fp comparison RTL pattern supports
like: ungt, unge, unlt, unle, ne, lt and le.
gcc/ChangeLog
2019-11-08 Kewen Lin <linkw@gcc.gnu.org>
PR target/92132
* config/rs6000/predicates.md
(signed_or_equality_comparison_operator): New predicate.
(unsigned_or_equality_comparison_operator): Likewise.
* config/rs6000/rs6000.md (one_cmpl<mode>2): Remove expand.
(one_cmpl<mode>3_internal): Rename to one_cmpl<mode>2.
* config/rs6000/vector.md
(vcond_mask_<mode><mode> for VEC_I and VEC_I): New expand.
(vec_cmp<mode><mode> for VEC_I and VEC_I): Likewise.
(vec_cmpu<mode><mode> for VEC_I and VEC_I): Likewise.
(vcond_mask_<mode><VEC_int> for VEC_F): New expand for float
vector modes and same-size integer vector modes.
(vec_cmp<mode><VEC_int> for VEC_F): Likewise.
(vector_lt<mode> for VEC_F): New expand.
(vector_le<mode> for VEC_F): Likewise.
(vector_ne<mode> for VEC_F): Likewise.
(vector_unge<mode> for VEC_F): Likewise.
(vector_ungt<mode> for VEC_F): Likewise.
(vector_unle<mode> for VEC_F): Likewise.
(vector_unlt<mode> for VEC_F): Likewise.
(vector_uneq<mode>): Expose name.
(vector_ltgt<mode>): Likewise.
(vector_unordered<mode>): Likewise.
(vector_ordered<mode>): Likewise.
gcc/testsuite/ChangeLog
2019-11-08 Kewen Lin <linkw@gcc.gnu.org>
PR target/92132
* gcc.target/powerpc/pr92132-fp-1.c: New test.
* gcc.target/powerpc/pr92132-fp-2.c: New test.
* gcc.target/powerpc/pr92132-int-1.c: New test.
* gcc.target/powerpc/pr92132-int-2.c: New test.
From-SVN: r277947
Hongtao Liu [Fri, 8 Nov 2019 05:34:25 +0000 (05:34 +0000)]
Fix inefficient vector constructor.
Changelog
gcc/
PR target/92295
* config/i386/i386-expand.c (ix86_expand_vector_init_concat)
Enhance ix86_expand_vector_init_concat.
gcc/testsuite
* gcc.target/i386/pr92295.c: New test.
From-SVN: r277946
Joseph Myers [Fri, 8 Nov 2019 01:21:40 +0000 (01:21 +0000)]
Handle removal of old-style function definitions in C2x.
C2x removes support for old-style function definitions with identifier
lists, changing () in function definitions to be equivalent to (void)
(while () in declarations that are not definitions still gives an
unprototyped type).
This patch updates GCC accordingly. The new semantics for () are
implemented for C2x mode (meaning () in function definitions isn't
diagnosed by -Wold-style-definition in that mode).
-Wold-style-definition is enabled by default, and turned into a
pedwarn, for C2x.
Bootstrapped with no regressions on x86_64-pc-linux-gnu.
gcc:
* doc/invoke.texi (-Wold-style-definition): Document () not being
considered an old-style definition for C2x.
gcc/c:
* c-decl.c (grokparms): Convert () in a function definition to
(void) for C2x.
(store_parm_decls_oldstyle): Pedwarn for C2x.
(store_parm_decls): Update comment about () not generating a
prototype.
gcc/c-family:
* c.opt (Wold-style-definition): Initialize to -1.
* c-opts.c (c_common_post_options): Set warn_old_style_definition
to flag_isoc2x if not set explicitly.
gcc/testsuite:
* gcc.dg/c11-old-style-definition-1.c,
gcc.dg/c11-old-style-definition-2.c,
gcc.dg/c2x-old-style-definition-1.c,
gcc.dg/c2x-old-style-definition-2.c,
gcc.dg/c2x-old-style-definition-3.c,
gcc.dg/c2x-old-style-definition-4.c,
gcc.dg/c2x-old-style-definition-5.c,
gcc.dg/c2x-old-style-definition-6.c: New tests.
From-SVN: r277945
Jonathan Wakely [Fri, 8 Nov 2019 00:37:17 +0000 (00:37 +0000)]
libsupc++: add <compare> to precompiled header
Also process it with Doxygen.
* doc/doxygen/user.cfg.in (INPUT): Add <compare> header.
* include/precompiled/stdc++.h: Include <compare> header.
From-SVN: r277944
Jonathan Wakely [Fri, 8 Nov 2019 00:37:08 +0000 (00:37 +0000)]
libstdc++: define std::common_comparison_category for C++20
* libsupc++/compare (common_comparison_category)
(common_comparison_category_t): Define for C++20.
* testsuite/18_support/comparisons/common/1.cc: New test.
From-SVN: r277943
Peter Bergner [Fri, 8 Nov 2019 00:34:09 +0000 (00:34 +0000)]
Add another test case to exercise the previous MODE_PARTIAL_INT change.
gcc/testsuite/
PR other/92090
* gcc.target/powerpc/pr92090-2.c: New test.
From-SVN: r277942
John David Anglin [Fri, 8 Nov 2019 00:27:06 +0000 (00:27 +0000)]
pa.md (memory_barrier): Revise to use ldcw barriers.
* config/pa/pa.md (memory_barrier): Revise to use ldcw barriers.
Enhance comment.
(memory_barrier_coherent, memory_barrier_64, memory_barrier_32): New
insn patterns using ldcw instruction.
(memory_barrier): Remove insn pattern using sync instruction.
* config/pa/pa.opt (coherent-ldcw): New option.
(ordered): New option.
From-SVN: r277941
GCC Administrator [Fri, 8 Nov 2019 00:16:18 +0000 (00:16 +0000)]
Daily bump.
From-SVN: r277940
Segher Boessenkool [Thu, 7 Nov 2019 23:58:11 +0000 (00:58 +0100)]
rs6000: Remove no longer correct assert
After the simplify-rtx patch, we can now be asked about conditions we
wouldn't be asked about before. This is perfectly fine, except we
have a little over-eager assert. Remove that one.
* config/rs6000/rs6000.c (validate_condition_mode): Don't assert for
valid conditions.
From-SVN: r277936
Joseph Myers [Thu, 7 Nov 2019 23:54:49 +0000 (23:54 +0000)]
Expand C2x attribute parsing support and factor out from TM attributes.
There is one place in the C parser that already handles a subset of
the C2x [[]] attribute syntax: c_parser_transaction_attributes.
This patch factors C2x attribute parsing out of there, extending it to
cover the full C2x attribute syntax (although currently only called
from that one place in the parser - so this is another piece of
preparation for supporting C2x attributes in the places where C2x says
they are valid, not the patch that actually enables such support).
The new C2X attribute parsing code uses the same representation for
scoped attributes as C++ does, so requiring parse_tm_stmt_attr to
handle the scoped attributes representation (C++ currently
special-cases TM attributes "to avoid the pedwarn in C++98 mode"; in C
I'm using an argument to c_parser_std_attribute_specifier to disable
the pedwarn_c11 call in the TM case).
Parsing of arguments to known attributes is shared by GNU and C2x
attributes. C2x specifies that unknown attributes are ignored (GCC
practice in such a case is to warn along with ignoring the attribute)
and gives a very general balanced-token-sequence syntax for arguments
to unknown attributes (known ones each have their own syntax which is
a subset of balanced-token-sequence), so support is added for parsing
and ignoring such balanced-token-sequences as arguments of unknown
attributes.
Some limited tests are added of different attribute usages in the TM
attribute case. The cases that become valid in the TM case include
extra commas inside [[]], and an explicit "gnu" namespace, as the
extra commas have no semantic effect for C2x attributes, while
accepting the "gnu" namespace seems appropriate because the attribute
in question is accepted inside __attribute__ (()), which is considered
equivalent to the "gnu" namespace inside [[]].
Bootstrapped with no regressions on x86_64-pc-linux-gnu.
gcc/c:
* c-parser.c (c_parser_attribute_arguments): New function.
Factored out of c_parser_gnu_attribute.
(c_parser_gnu_attribute): Use c_parser_attribute_arguments.
(c_parser_balanced_token_sequence, c_parser_std_attribute)
(c_parser_std_attribute_specifier): New functions.
(c_parser_transaction_attributes): Use
c_parser_std_attribute_specifier.
gcc/c-family:
* c-attribs.c (parse_tm_stmt_attr): Handle scoped attributes.
gcc/testsuite:
* gcc.dg/tm/attrs-1.c: New test.
* gcc.dg/tm/props-5.c: New test. Based on props-4.c.
From-SVN: r277935
Jakub Jelinek [Thu, 7 Nov 2019 23:28:08 +0000 (00:28 +0100)]
spaceship-scalar1-neg.C: Change dg-do from run to compile.
* g++.dg/cpp2a/spaceship-scalar1-neg.C: Change dg-do from run to
compile.
From-SVN: r277934
Jakub Jelinek [Thu, 7 Nov 2019 23:22:41 +0000 (00:22 +0100)]
ipa-utils.c (ipa_merge_profiles): Fix fprintf format string typo - mistmatch -> mismatch.
* ipa-utils.c (ipa_merge_profiles): Fix fprintf format string
typo - mistmatch -> mismatch.
* ipa-profile.c (ipa_profile): Likewise.
* ipa-devirt.c (compare_virtual_tables): Fix a comment typo
- mistmatch -> mismatch.
cp/
* init.c (build_vec_delete_1): Fix a comment typo - mist -> must.
From-SVN: r277933
Jonathan Wakely [Thu, 7 Nov 2019 23:10:45 +0000 (23:10 +0000)]
libstdc++: make negative count safe with std::for_each_n
The Library Working Group have approved a change to std::for_each_n that
requires it to handle negative N gracefully, which we were not doing for
random access iterators.
* include/bits/stl_algo.h (for_each_n): Handle negative count.
* testsuite/25_algorithms/for_each/for_each_n_debug.cc: New test.
From-SVN: r277932
Segher Boessenkool [Thu, 7 Nov 2019 22:08:49 +0000 (23:08 +0100)]
simplify-rtx: simplify_logical_relational_operation
This introduces simplify_logical_relational_operation. Currently the
only thing implemented it can simplify is the IOR of two CONDs of the
same arguments.
* simplify-rtx.c (comparison_to_mask): New function.
(mask_to_comparison): New function.
(simplify_logical_relational_operation): New function.
(simplify_binary_operation_1): Call
simplify_logical_relational_operation.
From-SVN: r277931
Iain Sandoe [Thu, 7 Nov 2019 21:17:31 +0000 (21:17 +0000)]
[Darwin, X86, testsuite] Fix pr92258.c.
This test uses -masm=intel, which isn't supported by Darwin. Add the
necessary dg-require-effective-target.
gcc/testsuite/ChangeLog:
2019-11-07 Iain Sandoe <iain@sandoe.co.uk>
* gcc.target/i386/pr92258.c: Add dg-requires for masm_intel.
From-SVN: r277930
Jakub Jelinek [Thu, 7 Nov 2019 20:24:38 +0000 (21:24 +0100)]
PR c++/91370 - Implement P1041R4 and P1139R2 - Stronger Unicode reqs
PR c++/91370 - Implement P1041R4 and P1139R2 - Stronger Unicode reqs
* charset.c (narrow_str_to_charconst): Add TYPE argument. For
CPP_UTF8CHAR diagnose whenever number of chars is > 1, using
CPP_DL_ERROR instead of CPP_DL_WARNING.
(wide_str_to_charconst): For CPP_CHAR16 or CPP_CHAR32, use
CPP_DL_ERROR instead of CPP_DL_WARNING when multiple char16_t
or char32_t chars are needed.
(cpp_interpret_charconst): Adjust narrow_str_to_charconst caller.
* g++.dg/cpp1z/utf8-neg.C: Expect errors rather than -Wmultichar
warnings.
* g++.dg/ext/utf16-4.C: Expect errors rather than warnings.
* g++.dg/ext/utf32-4.C: Likewise.
* g++.dg/cpp2a/ucn2.C: New test.
From-SVN: r277929
Peter Bergner [Thu, 7 Nov 2019 18:48:45 +0000 (18:48 +0000)]
Allow MODE_PARTIAL_INT modes for integer constant input operands.
gcc/
PR other/92090
* config/rs6000/predicates.md (input_operand): Allow MODE_PARTIAL_INT
modes for integer constants.
gcc/testsuite/
PR other/92090
* gcc.target/powerpc/pr92090.c: New test.
From-SVN: r277928
Jan Hubicka [Thu, 7 Nov 2019 17:08:11 +0000 (18:08 +0100)]
re PR lto/92406 (ICE in ipa_call_summary at ipa-fnsummary.h:253 with lto and pgo)
PR ipa/92406
* ipa-fnsummary.c (analyze_function_body): Use get_create to copy
summary.
From-SVN: r277927
Jan Hubicka [Thu, 7 Nov 2019 17:06:43 +0000 (18:06 +0100)]
optc-save-gen.awk: Generate cl_target_option_free and cl_optimization_option_free.
* optc-save-gen.awk: Generate cl_target_option_free
and cl_optimization_option_free.
* opth-en.awk: Declare cl_target_option_free
and cl_optimization_option_free.
* tree.c (free_node): Use it.
From-SVN: r277926
Jason Merrill [Thu, 7 Nov 2019 17:06:09 +0000 (12:06 -0500)]
Implement D1959R0, remove weak_equality and strong_equality.
Shortly after I finished implementing the previous semantics, the
committee decided to remove the *_equality comparison categories, because
they were largely obsoleted by the earlier change that separated operator==
from its original dependency on operator<=>.
gcc/cp/
* method.c (enum comp_cat_tag, comp_cat_info): Remove *_equality.
(genericize_spaceship, common_comparison_type): Likewise.
* typeck.c (cp_build_binary_op): Move SPACESHIP_EXPR to be with the
relational operators, exclude other types no longer supported.
libstdc++-v3/
* libsupc++/compare: Remove strong_equality and weak_equality.
From-SVN: r277925
Jan Hubicka [Thu, 7 Nov 2019 17:06:04 +0000 (17:06 +0000)]
lto-streamer-in.c: Include alloc-pool.h.
* lto-streamer-in.c: Include alloc-pool.h.
(freeing_string_slot_hasher): Remove.
(string_slot_allocator): New object allocator.
(file_name_hash_table): Turn to hash_table<string_slot_hasher>.
(file_name_obstack): New obstack.
(canon_file_name): Allocate in obstack and allocator.
(lto_reader_init): Initialize obstack and allocator.
(lto_free_file_name_hash): New function.
* lto-streamer.h (lto_free_file_name_hash): New.
* lto.c (do_whole_program_analysis): Call lto_free_file_name_hash.
From-SVN: r277924
Feng Xue [Thu, 7 Nov 2019 15:43:01 +0000 (15:43 +0000)]
Loop split on semi-invariant conditional statement
2019-11-07 Feng Xue <fxue@os.amperecomputing.com>
PR tree-optimization/89134
* doc/invoke.texi (min-loop-cond-split-prob): Document new --params.
* params.def: Add min-loop-cond-split-prob.
* tree-ssa-loop-split.c (split_loop): Remove niter parameter, move some
outside checks on loop into the function.
(split_info): New class.
(find_vdef_in_loop, get_control_equiv_head_block): New functions.
(find_control_dep_blocks, vuse_semi_invariant_p): Likewise.
(ssa_semi_invariant_p, loop_iter_phi_semi_invariant_p): Likewise.
(control_dep_semi_invariant_p, stmt_semi_invariant_p_1): Likewise.
(stmt_semi_invariant_p, branch_removable_p): Likewise.
(get_cond_invariant_branch, compute_added_num_insns): Likewise.
(get_cond_branch_to_split_loop, do_split_loop_on_cond): Likewise.
(split_loop_on_cond): Likewise.
(tree_ssa_split_loops): Add loop split on conditional statement.
2019-11-07 Feng Xue <fxue@os.amperecomputing.com>
PR tree-optimization/89134
* gcc.dg/tree-ssa/loop-cond-split-1.c: New test.
* g++.dg/tree-ssa/loop-cond-split-1.C: New test.
* gcc.dg/torture/pr55107.c: Add -fno-split-loops.
From-SVN: r277923
Andreas Krebbel [Thu, 7 Nov 2019 11:52:05 +0000 (11:52 +0000)]
IBM Z: Add pattern for load truth value of comparison into reg
The RTXs used to express an overflow condition check in add/sub/mul are
too complex for if conversion. However, there is code in
noce_emit_store_flag which generates a simple CC compare as the base
for using a conditional load. All we have to do is to provide a
pattern to store the truth value of a CC compare into a GPR.
Done with the attached patch.
2019-11-07 Andreas Krebbel <krebbel@linux.ibm.com>
* config/s390/s390.md ("*cstorecc<mode>_z13"): New insn_and_split
pattern.
gcc/testsuite/ChangeLog:
2019-11-07 Andreas Krebbel <krebbel@linux.ibm.com>
* gcc.target/s390/addsub-signed-overflow-1.c: Expect lochi
instructions to be used.
* gcc.target/s390/addsub-signed-overflow-2.c: Likewise.
* gcc.target/s390/mul-signed-overflow-1.c: Likewise.
* gcc.target/s390/mul-signed-overflow-2.c: Likewise.
* gcc.target/s390/vector/vec-scalar-cmp-1.c: Check for 32 and 64
bit variant of lochi. Swap the values for the lochi's.
* gcc.target/s390/zvector/vec-cmp-1.c: Likewise.
From-SVN: r277922
Richard Biener [Thu, 7 Nov 2019 11:49:09 +0000 (11:49 +0000)]
re PR tree-optimization/92405 (ICE in vect_get_vec_def_for_stmt_copy, at tree-vect-stmts.c:1683)
2019-11-07 Richard Biener <rguenther@suse.de>
PR tree-optimization/92405
* tree-vect-loop.c (vectorizable_reduction): Appropriately
restrict lane-reducing ops to single stmt chains.
From-SVN: r277921
Martin Jambor [Thu, 7 Nov 2019 10:55:43 +0000 (11:55 +0100)]
Remove gimple_call_types_likely_match_p (PR 70929)
2019-11-07 Martin Jambor <mjambor@suse.cz>
PR lto/70929
* cif-code.def (MISMATCHED_ARGUMENTS): Removed.
* cgraph.h (gimple_check_call_matching_types): Remove
* cgraph.c (gimple_check_call_args): Likewise.
(gimple_check_call_matching_types): Likewise.
(symbol_table::create_edge): Do not call
gimple_check_call_matching_types.
(cgraph_edge::make_direct): Likewise.
(cgraph_edge::redirect_call_stmt_to_callee): Likewise.
* value-prof.h (check_ic_target): Remove.
* value-prof.c (check_ic_target): Remove.
(gimple_ic_transform): Do nat call check_ic_target.
* auto-profile.c (function_instance::find_icall_target_map): Likewise.
(afdo_indirect_call): Likewise.
* ipa-prop.c (update_indirect_edges_after_inlining): Do not call
gimple_check_call_matching_types.
* ipa-inline.c (early_inliner): Likewise.
testsuite/
* g++.dg/lto/pr70929_[01].C: New test.
* gcc.dg/winline-10.c: Adjust for the fact that inlining happens.
From-SVN: r277920
Kyrylo Tkachov [Thu, 7 Nov 2019 10:50:23 +0000 (10:50 +0000)]
[arm][6/X] Add support for __[us]sat16 intrinsics
This last patch adds the the __ssat16 and __usat16 intrinsics that perform
"clipping" to a particular bitwidth on packed SIMD values, setting the Q bit
as appropriate.
* config/arm/arm.md (arm_<simd32_op>): New define_expand.
(arm_<simd32_op><add_clobber_q_name>_insn): New define_insn.
* config/arm/arm_acle.h (__ssat16, __usat16): Define.
* config/arm/arm_acle_builtins.def: Define builtins for the above.
* config/arm/iterators.md (USSAT16): New int_iterator.
(simd32_op): Handle UNSPEC_SSAT16, UNSPEC_USAT16.
(sup): Likewise.
* config/arm/predicates.md (ssat16_imm): New predicate.
(usat16_imm): Likewise.
* config/arm/unspecs.md (UNSPEC_SSAT16, UNSPEC_USAT16): Define.
* gcc.target/arm/acle/simd32.c: Update test.
From-SVN: r277919
Kyrylo Tkachov [Thu, 7 Nov 2019 10:49:06 +0000 (10:49 +0000)]
[arm][5/X] Implement Q-bit-setting SIMD32 intrinsics
This patch implements some more Q-setting intrinsics of the
multiply-accumulate
variety, but these are in the SIMD32 family in that they treat their
operands as packed SIMD values, but that's not important at the RTL level.
* config/arm/arm.md (arm_<simd32_op><add_clobber_q_name>_insn):
New define_insns.
(arm_<simd32_op>): New define_expands.
* config/arm/arm_acle.h (__smlad, __smladx, __smlsd, __smlsdx,
__smuad, __smuadx): Define.
* config/arm/arm_acle_builtins.def: Define builtins for the above.
* config/arm/iterators.md (SIMD32_TERNOP_Q): New int_iterator.
(SIMD32_BINOP_Q): Likewise.
(simd32_op): Handle the above.
* config/arm/unspecs.md: Define unspecs for the above.
* gcc.target/arm/acle/simd32.c: Update test.
From-SVN: r277918
Kyrylo Tkachov [Thu, 7 Nov 2019 10:46:05 +0000 (10:46 +0000)]
[arm][4/X] Add initial support for GE-setting SIMD32 intrinsics
This patch adds in plumbing for the ACLE intrinsics that set the GE bits in
APSR. These are special SIMD instructions in Armv6 that pack bytes or
halfwords into the 32-bit general-purpose registers and set the GE bits in
APSR to indicate if some of the "lanes" of the result have overflowed or
have some other instruction-specific property.
These bits can then be used by the SEL instruction (accessed through the
__sel intrinsic) to select lanes for further processing.
This situation is similar to the Q-setting intrinsics: we have to track
the GE fake register, detect when a function reads it through __sel and restrict
existing patterns that may generate GE-clobbering instruction from
straight-line C code when reading the GE bits matters.
* config/arm/aout.h (REGISTER_NAMES): Add apsrge.
* config/arm/arm.md (APSRGE_REGNUM): Define.
(arm_<simd32_op>): New define_insn.
(arm_sel): Likewise.
* config/arm/arm.h (FIXED_REGISTERS): Add entry for apsrge.
(CALL_USED_REGISTERS): Likewise.
(REG_ALLOC_ORDER): Likewise.
(FIRST_PSEUDO_REGISTER): Update value.
(ARM_GE_BITS_READ): Define.
* config/arm/arm.c (arm_conditional_register_usage): Clear
APSRGE_REGNUM from operand_reg_set.
(arm_ge_bits_access): Define.
* config/arm/arm-builtins.c (arm_check_builtin_call): Handle
ARM_BUIILTIN_sel.
* config/arm/arm-protos.h (arm_ge_bits_access): Declare prototype.
* config/arm/arm-fixed.md (add<mode>3): Convert to define_expand.
FAIL if ARM_GE_BITS_READ.
(*arm_add<mode>3): New define_insn.
(sub<mode>3): Convert to define_expand. FAIL if ARM_GE_BITS_READ.
(*arm_sub<mode>3): New define_insn.
* config/arm/arm_acle.h (__sel, __sadd8, __ssub8, __uadd8, __usub8,
__sadd16, __sasx, __ssax, __ssub16, __uadd16, __uasx, __usax,
__usub16): Define.
* config/arm/arm_acle_builtins.def: Define builtins for the above.
* config/arm/iterators.md (SIMD32_GE): New int_iterator.
(simd32_op): Handle the above.
* config/arm/unspecs.md (UNSPEC_GE_SET): Define.
(UNSPEC_SEL, UNSPEC_SADD8, UNSPEC_SSUB8, UNSPEC_UADD8, UNSPEC_USUB8,
UNSPEC_SADD16, UNSPEC_SASX, UNSPEC_SSAX, UNSPEC_SSUB16, UNSPEC_UADD16,
UNSPEC_UASX, UNSPEC_USAX, UNSPEC_USUB16): Define.
* gcc.target/arm/acle/simd32.c: Update test.
* gcc.target/arm/acle/simd32_sel.c: New test.
From-SVN: r277917
Kyrylo Tkachov [Thu, 7 Nov 2019 10:43:19 +0000 (10:43 +0000)]
[arm][3/X] Implement __smla* intrinsics (Q-setting)
This patch implements some more Q-setting intrinsics form the SMLA* group.
These can set the saturation bit on overflow in the accumulation step.
Like earlier, these have non-Q-setting RTL forms as well for when the
Q-bit read
is not needed.
* config/arm/arm.md (arm_smlabb_setq): New define_insn.
(arm_smlabb): New define_expand.
(*maddhisi4tb): Rename to...
(maddhisi4tb): ... This.
(*maddhisi4tt): Rename to...
(maddhisi4tt): ... This.
(arm_smlatb_setq): New define_insn.
(arm_smlatb): New define_expand.
(arm_smlatt_setq): New define_insn.
(arm_smlatt): New define_expand.
(arm_<smlaw_op><add_clobber_name>_insn): New define_insn.
(arm_<smlaw_op>): New define_expand.
* config/arm/arm_acle.h (__smlabb, __smlatb, __smlabt, __smlatt,
__smlawb, __smlawt): Define.
* config/arm_acle_builtins.def: Define builtins for the above.
* config/arm/iterators.md (SMLAWBT): New int_iterator.
(slaw_op): New int_attribute.
* config/arm/unspecs.md (UNSPEC_SMLAWB, UNSPEC_SMLAWT): Define.
* gcc.target/arm/acle/dsp_arith.c: Update test.
From-SVN: r277916
Kyrylo Tkachov [Thu, 7 Nov 2019 10:41:21 +0000 (10:41 +0000)]
[arm][2/X] Implement __qadd, __qsub, __qdbl intrinsics
This patch implements some more Q-bit-setting intrinsics from ACLE.
With the plumbing from patch 1 in place they are a simple builtin->RTL
affair.
* config/arm/arm.md (arm_<ss_op>): New define_expand.
(arm_<ss_op><add_clobber_q_name>_insn): New define_insn.
* config/arm/arm_acle.h (__qadd, __qsub, __qdbl): Define.
* config/arm/arm_acle_builtins.def: Add builtins for qadd, qsub.
* config/arm/iterators.md (SSPLUSMINUS): New code iterator.
(ss_op): New code_attr.
* gcc.target/arm/acle/dsp_arith.c: New test.
From-SVN: r277915
Kyrylo Tkachov [Thu, 7 Nov 2019 10:39:39 +0000 (10:39 +0000)]
[arm][1/X] Add initial support for saturation intrinsics
This patch adds the plumbing for and an implementation of the saturation
intrinsics from ACLE, in particular the __ssat, __usat intrinsics.
These intrinsics set the Q sticky bit in APSR if an overflow occurred.
ACLE allows the user to read that bit (within the same function, it's not
defined across function boundaries) using the __saturation_occurred
intrinsic
and reset it using __set_saturation_occurred.
Thus, if the user cares about the Q bit they would be using a flow such as:
__set_saturation_occurred (0); // reset the Q bit
...
__ssat (...) // Do some calculations involving __ssat
...
if (__saturation_occurred ()) // if Q bit set handle overflow
...
For the implementation this has a few implications:
* We must track the Q-setting side-effects of these instructions to make
sure
saturation reading/writing intrinsics are ordered properly.
This is done by introducing a new "apsrq" register (and associated
APSRQ_REGNUM) in a similar way to the "fake"" cc register.
* The RTL patterns coming out of these intrinsics can have two forms:
one where they set the APSRQ_REGNUM and one where they don't.
Which one is used depends on whether the function cares about reading the Q
flag. This is detected using the TARGET_CHECK_BUILTIN_CALL hook on the
__saturation_occurred, __set_saturation_occurred occurrences.
If no Q-flag read is present in the function we'll use the simpler
non-Q-setting form to allow for more aggressive scheduling and such.
If a Q-bit read is present then the Q-setting form is emitted.
To avoid adding two patterns for each intrinsic to the MD file we make
use of define_subst to auto-generate the Q-setting forms
* Some existing patterns already produce instructions that may clobber the
Q bit, but they don't model it (as we didn't care about that bit up till
now).
Since these patterns can be generated from straight-line C code they can
affect
the Q-bit reads from intrinsics. Therefore they have to be disabled when
a Q-bit read is present. These are mostly patterns in arm-fixed.md that are
not very common anyway, but there are also a couple of widening
multiply-accumulate patterns in arm.md that can set the Q-bit during
accumulation.
There are more Q-setting intrinsics in ACLE, but these will be
implemented in
a more mechanical fashion once the infrastructure in this patch goes in.
* config/arm/aout.h (REGISTER_NAMES): Add apsrq.
* config/arm/arm.md (APSRQ_REGNUM): Define.
(add_setq): New define_subst.
(add_clobber_q_name): New define_subst_attr.
(add_clobber_q_pred): Likewise.
(maddhisi4): Change to define_expand. Split into mult and add if
ARM_Q_BIT_READ.
(arm_maddhisi4): New define_insn.
(*maddhisi4tb): Disable for ARM_Q_BIT_READ.
(*maddhisi4tt): Likewise.
(arm_ssat): New define_expand.
(arm_usat): Likewise.
(arm_get_apsr): New define_insn.
(arm_set_apsr): Likewise.
(arm_saturation_occurred): New define_expand.
(arm_set_saturation): Likewise.
(*satsi_<SAT:code>): Rename to...
(satsi_<SAT:code><add_clobber_q_name>): ... This.
(*satsi_<SAT:code>_shift): Disable for ARM_Q_BIT_READ.
* config/arm/arm.h (FIXED_REGISTERS): Mark apsrq as fixed.
(CALL_USED_REGISTERS): Mark apsrq.
(FIRST_PSEUDO_REGISTER): Update value.
(REG_ALLOC_ORDER): Add APSRQ_REGNUM.
(machine_function): Add q_bit_access.
(ARM_Q_BIT_READ): Define.
* config/arm/arm.c (TARGET_CHECK_BUILTIN_CALL): Define.
(arm_conditional_register_usage): Clear APSRQ_REGNUM from
operand_reg_set.
(arm_q_bit_access): Define.
* config/arm/arm-builtins.c: Include stringpool.h.
(arm_sat_binop_imm_qualifiers,
arm_unsigned_sat_binop_unsigned_imm_qualifiers,
arm_sat_occurred_qualifiers, arm_set_sat_qualifiers): Define.
(SAT_BINOP_UNSIGNED_IMM_QUALIFIERS,
UNSIGNED_SAT_BINOP_UNSIGNED_IMM_QUALIFIERS, SAT_OCCURRED_QUALIFIERS,
SET_SAT_QUALIFIERS): Likewise.
(arm_builtins): Define ARM_BUILTIN_SAT_IMM_CHECK.
(arm_init_acle_builtins): Initialize __builtin_sat_imm_check.
Handle 0 argument expander.
(arm_expand_acle_builtin): Handle ARM_BUILTIN_SAT_IMM_CHECK.
(arm_check_builtin_call): Define.
* config/arm/arm.md (ssmulsa3, usmulusa3, usmuluha3,
arm_ssatsihi_shift, arm_usatsihi): Disable when ARM_Q_BIT_READ.
* config/arm/arm-protos.h (arm_check_builtin_call): Declare prototype.
(arm_q_bit_access): Likewise.
* config/arm/arm_acle.h (__ssat, __usat, __ignore_saturation,
__saturation_occurred, __set_saturation_occurred): Define.
* config/arm/arm_acle_builtins.def: Define builtins for ssat, usat,
saturation_occurred, set_saturation_occurred.
* config/arm/unspecs.md (UNSPEC_Q_SET): Define.
(UNSPEC_APSR_READ): Likewise.
(VUNSPEC_APSR_WRITE): Likewise.
* config/arm/arm-fixed.md (ssadd<mode>3): Convert to define_expand.
(*arm_ssadd<mode>3): New define_insn.
(sssub<mode>3): Convert to define_expand.
(*arm_sssub<mode>3): New define_insn.
(ssmulsa3): Convert to define_expand.
(*arm_ssmulsa3): New define_insn.
(usmulusa3): Convert to define_expand.
(*arm_usmulusa3): New define_insn.
(ssmulha3): FAIL if ARM_Q_BIT_READ.
(arm_ssatsihi_shift, arm_usatsihi): Disable for ARM_Q_BIT_READ.
* config/arm/iterators.md (qaddsub_clob_q): New mode attribute.
* gcc.target/arm/acle/saturation.c: New test.
* gcc.target/arm/acle/sat_no_smlatb.c: Likewise.
* lib/target-supports.exp (check_effective_target_arm_qbit_ok_nocache):
Define..
(check_effective_target_arm_qbit_ok): Likewise.
(add_options_for_arm_qbit): Likewise.
From-SVN: r277914
Martin Liska [Thu, 7 Nov 2019 09:44:21 +0000 (10:44 +0100)]
Clear version_info_node in delete_function_version.
2019-11-07 Martin Liska <mliska@suse.cz>
PR c++/92354
* cgraph.c (delete_function_version): Clear global
variable version_info_node if equal to deleted
function.
2019-11-07 Martin Liska <mliska@suse.cz>
PR c++/92354
* g++.target/i386/pr92354.C: New test.
From-SVN: r277913
Martin Liska [Thu, 7 Nov 2019 09:44:02 +0000 (10:44 +0100)]
Add CONSTRUCTOR_NO_CLEARING to operand_equal_p.
2019-11-07 Martin Liska <mliska@suse.cz>
* fold-const.c (operand_compare::operand_equal_p): Add comparison
of CONSTRUCTOR_NO_CLEARING.
(operand_compare::hash_operand): Likewise.
From-SVN: r277912
Martin Liska [Thu, 7 Nov 2019 09:34:42 +0000 (09:34 +0000)]
Update LOCAL_PATCHES.
From-SVN: r277911
Martin Liska [Thu, 7 Nov 2019 09:34:14 +0000 (10:34 +0100)]
Reapply all revisions mentioned in LOCAL_PATCHES.
2019-11-07 Martin Liska <mliska@suse.cz>
* all source files: Reapply all revisions mentioned in LOCAL_PATCHES.
From-SVN: r277910
Martin Liska [Thu, 7 Nov 2019 09:33:54 +0000 (10:33 +0100)]
Libsanitizer: merge from trunk
2019-11-07 Martin Liska <mliska@suse.cz>
* merge.sh: Update to use llvm-project git repository.
* all source files: Merge from upstream
82588e05cc32bb30807e480abd4e689b0dee132a.
From-SVN: r277909
Georg-Johann Lay [Thu, 7 Nov 2019 09:19:31 +0000 (09:19 +0000)]
Support 64-bit double and 64-bit long double configurations.
gcc/
Support 64-bit double and 64-bit long double configurations.
PR target/92055
* config.gcc (tm_defines) [avr]: Set from --with-double=,
--with-long-double=.
* config/avr/t-multilib: Remove.
* config/avr/t-avr: Output of genmultilib.awk is now fully
dynamically generated and no more part of the repo.
(HAVE_DOUBLE_MULTILIB, HAVE_LONG_DOUBLE_MULTILIB): New variables.
Pass them down to...
* config/avr/genmultilib.awk: ...here and handle them.
* gcc/config/avr/avr.opt (-mdouble=, avr_double). New option and var.
(-mlong-double=, avr_long_double). New option and var.
* common/config/avr/avr-common.c (opts.h, diagnostic.h): Include.
(TARGET_OPTION_OPTIMIZATION_TABLE) <-mdouble=, -mlong-double=>:
Set default as requested by --with-double=
(TARGET_HANDLE_OPTION): Define to this...
(avr_handle_option): ...new hook worker.
* config/avr/avr.h (DOUBLE_TYPE_SIZE): Define to avr_double.
(LONG_DOUBLE_TYPE_SIZE): Define to avr_long_double.
(avr_double_lib): New proto for spec function.
(EXTRA_SPEC_FUNCTIONS) <double-lib>: Add.
(DRIVER_SELF_SPECS): Call %:double-lib.
* config/avr/avr.c (avr_option_override): Assert
sizeof(long double) >= sizeof(double) for the target.
* config/avr/avr-c.c (avr_cpu_cpp_builtins)
[__HAVE_DOUBLE_MULTILIB__, __HAVE_LONG_DOUBLE_MULTILIB__]
[__HAVE_DOUBLE64__, __HAVE_DOUBLE32__, __DEFAULT_DOUBLE__=]
[__HAVE_LONG_DOUBLE64__, __HAVE_LONG_DOUBLE32__]
[__HAVE_LONG_DOUBLE_IS_DOUBLE__, __DEFAULT_LONG_DOUBLE__=]:
New built-in define depending on --with-double=, --with-long-double=.
* config/avr/driver-avr.c (avr_double_lib): New spec function.
* doc/invoke.tex (AVR Options) <-mdouble=,-mlong-double=>: Doc.
* doc/install.texi (Cross-Compiler-Specific Options)
<--with-double=, --with-long-double=>: Doc.
libgcc/
Support 64-bit double and 64-bit long double configurations.
PR target/92055
* config/avr/t-avr (HOST_LIBGCC2_CFLAGS): Only add -DF=SF if
long double is a 32-bit type.
* config/avr/t-avrlibc: Copy double64 and long-double64
multilib(s) from the vanilla one.
* config/avr/t-copy-libgcc: New Makefile snip.
From-SVN: r277908
Richard Biener [Thu, 7 Nov 2019 07:36:39 +0000 (07:36 +0000)]
dbgcnt.def (gimple_unroll): New.
2019-11-07 Richard Biener <rguenther@suse.de>
* dbgcnt.def (gimple_unroll): New.
* tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Check
gimple_unroll debug counter before applying transform.
(try_peel_loop): Likewise.
From-SVN: r277907
Prathamesh Kulkarni [Thu, 7 Nov 2019 06:27:39 +0000 (06:27 +0000)]
Adjust pr92163.c test to require effective target fopenacc.
2019-11-07 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
testsuite/
* gcc.dg/tree-ssa/pr92163.c: Add dg-require-effective-target fopenacc.
From-SVN: r277906
Jerry DeLisle [Thu, 7 Nov 2019 03:06:20 +0000 (03:06 +0000)]
re PR libfortran/90374 (Fortran 2018: Support d0.d, e0.d, es0.d, en0.d, g0.d and ew.d e0 edit descriptors for output)
2019-11-06 Jerry DeLisle <jvdelisle@gcc.ngu.org>
PR fortran/90374
* io.c (check_format): Allow zero width for D, E, EN, and ES
specifiers as default and when -std=F2018 is given. Retain
existing errors when using the -fdec family of flags.
* libgfortran/io/format.c (parse_format_list): Relax format checking for
zero width as default and when -std=f2018.
io/format.h (format_token): Move definition to io.h.
io/io.h (format_token): Add definition here to allow access to
this definition at higher levels. Rename the declaration of
write_real_g0 to write_real_w0 and add a new format_token
argument, allowing higher level functions to pass in the
token for handling of g0 vs the other zero width specifiers.
io/transfer.c (formatted_transfer_scalar_write): Add checks for
zero width and call write_real_w0 to handle it.
io/write.c (write_real_g0): Remove.
(write_real_w0): Add new, same as previous write_real_g0 except
check format token to handle the g0 case.
* gfortran.dg/fmt_error_10.f: Modify for new constraints.
* gfortran.dg/fmt_error_7.f: Add dg-options "-std=f95".
* gfortran.dg/fmt_error_9.f: Modify for new constraints.
* gfortran.dg/fmt_zero_width.f90: New test.
From-SVN: r277905
Xiong Hu Luo [Thu, 7 Nov 2019 01:24:03 +0000 (01:24 +0000)]
Fix copy-paste typo syntax error by r277872
gcc/testsuite/ChangeLog:
2019-11-07 Xiong Hu Luo <luoxhu@linux.ibm.com>
* gcc.target/powerpc/pr72804.c: Move inline options from
dg-require-effective-target to dg-options.
From-SVN: r277904
Joseph Myers [Thu, 7 Nov 2019 01:01:07 +0000 (01:01 +0000)]
Move string concatenation for C into the parser.
This patch is another piece of preparation for C2x attributes support.
C2x attributes require unbounded lookahead in the parser, because the
token sequence '[[' that starts a C2x attribute is also valid in
Objective-C in some of the same contexts, so it is necessary to see
whether the matching ']]' are consecutive tokens or not to determine
whether those tokens start an attribute.
Unbounded lookahead means lexing an unbounded number of tokens before
they are parsed. c_lex_one_token does various context-sensitive
processing of tokens that cannot be done at that lookahead time,
because it depends on information (such as whether particular
identifiers are typedefs) that may be different at the time it is
relevant than at the time the lookahead is needed (recall that more or
less arbitrary C code, including declarations and statements, can
appear inside expressions in GNU C).
Most of that context-sensitive processing is not a problem, simply
because it is not needed for lookahead purposes so can be deferred
until the tokens lexed during lookahead are parsed. However, the
earliest piece of context-sensitive processing is the handling of
string literals based on flags passed to c_lex_with_flags, which
determine whether adjacent literals are concatenated and whether
translation to the execution character set occurs.
Because the choice of whether to translate to the execution character
set is context-sensitive, this means that unbounded lookahead requires
the C parser to move to the approach used by the C++ parser, where
string literals are generally not translated or concatenated from
within c_lex_with_flags, but only later in the parser once it knows
whether translation is needed. (Translation requires the tokens in
their form before concatenation.)
Thus, this patch makes that change to the C parser. Flags in the
parser are still used for two special cases similar to C++: the
handling of an initial #pragma pch_preprocess, and arranging for
strings inside attributes not to be translated (the latter is made
more logically correct by saving and restoring the flags, as in the
C++ parser, rather than assuming that the state outside the attribute
was always to translate string literals, which might not be the case
in corner cases involving declarations and attributes inside
attributes).
The consequent change to pragma_lex to use c_parser_string_literal
makes it disallow wide strings and disable translation in that
context, which also follows C++ and is more logically correct than the
previous state without special handling in that regard. Translation
to the execution character set is always disabled when string
constants are handled in the GIMPLE parser.
Although the handling of strings is now a lot closer to that in C++,
there are still some differences, in particular regarding the handling
of locations. See c-c++-common/Wformat-pr88257.c, which has different
expected multiline diagnostic output for C and C++, for example; I'm
not sure whether the C or C++ output is better there (C++ has a more
complete range than C, C mentions a macro definition location that C++
doesn't), but I tried to keep the locations the same as those
previously used by the C front end, as far as possible, to minimize
the testsuite changes needed, rather than possibly making them closer
to those used with C++.
The only changes needed for tests of user-visible diagnostics were for
the wording of one diagnostic changing to match C++ (as a consequence
of having a check for wide strings based on a flag in a general
string-handling function rather than in a function specific to asm).
However, although locations are extremely similar to what they were
before, I couldn't make them completely identical in all cases. (My
understanding of the implementation reason for the differences is as
follows: lex_string uses src_loc from each cpp_token; the C parser is
using the virtual location from cpp_get_token_with_location as called
by c_lex_with_flags, and while passing that through
linemap_resolve_location with LRK_MACRO_DEFINITION_LOCATION, as this
patch does, produces something very close to what lex_string uses,
it's not completely identical in some cases.)
This results in changes being needed to two of the gcc.dg/plugin tests
that use a plugin to test details of how string locations are handled.
Because the tests being changed are for ICEs and the only change is to
the details of the particular non-user-visible error that code gives
in cases it can't handle (one involving __FILE__, one involving a
string literal from stringizing), I think it's OK to change that
non-user-visible error and that the new errors are no worse than the
old ones. So these particular errors are now different for C and C++
(some other messages in those tests already had differences between C
and C++).
Bootstrapped with no regressions on x86_64-pc-linux-gnu.
gcc/c:
* c-parser.c (c_parser): Remove lex_untranslated_string. Add
lex_joined_string and translate_strings_p.
(c_lex_one_token): Pass 0 or C_LEX_STRING_NO_JOIN to
c_lex_with_flags.
(c_parser_string_literal): New function.
(c_parser_static_assert_declaration_no_semi): Use
c_parser_string_literal. Do not set lex_untranslated_string.
(c_parser_asm_string_literal): Use c_parser_string_literal.
(c_parser_simple_asm_expr): Do not set lex_untranslated_string.
(c_parser_gnu_attributes): Set and restore translate_strings_p
instead of lex_untranslated_string.
(c_parser_asm_statement): Do not set lex_untranslated_string.
(c_parser_asm_operands): Likewise.
(c_parser_has_attribute_expression): Set and restore
translate_strings_p instead of lex_untranslated_string.
(c_parser_postfix_expression): Use c_parser_string_literal.
(pragma_lex): Likewise.
(c_parser_pragma_pch_preprocess): Set lex_joined_string.
(c_parse_file): Set translate_strings_p.
* gimple-parser.c (c_parser_gimple_postfix_expression)
(c_parser_gimple_or_rtl_pass_list): Use c_parser_string_literal.
* c-parser.c (c_parser_string_literal): Declare function.
gcc/testsuite:
* gcc.dg/asm-wide-1.c, gcc.dg/diagnostic-token-ranges.c,
gcc.dg/plugin/diagnostic-test-string-literals-1.c,
gcc.dg/plugin/diagnostic-test-string-literals-2.c: Update expected
diagnostics.
From-SVN: r277903
Jason Merrill [Thu, 7 Nov 2019 00:50:19 +0000 (19:50 -0500)]
Implement D1907R1 "structural type".
ISO C++ paper D1907R1 proposes "structural type" as an alternative to the
current notion of "strong structural equality", which has various problems.
I'm implementing it to give people a chance to try it.
The build_base_field changes are to make it easier for structural_type_p to
see whether a base is private or protected.
* tree.c (structural_type_p): New.
* pt.c (invalid_nontype_parm_type_p): Use it.
* class.c (build_base_field_1): Take binfo. Copy TREE_PRIVATE.
(build_base_field): Pass binfo.
From-SVN: r277902
Jason Merrill [Thu, 7 Nov 2019 00:31:52 +0000 (19:31 -0500)]
PR c++/92150 - partial specialization with class NTTP.
Here unify was getting confused by the VIEW_CONVERT_EXPR we add in
finish_id_expression_1 to make class NTTP const when they're used in an
expression.
Tested x86_64-pc-linux-gnu, applying to trunk.
* pt.c (unify): Handle VIEW_CONVERT_EXPR.
From-SVN: r277901
Jason Merrill [Thu, 7 Nov 2019 00:21:44 +0000 (19:21 -0500)]
Use satisfaction with nested requirements.
gcc/cp/
2019-11-06 Andrew Sutton <asutton@lock3software.com>
* constraint.cc (build_parameter_mapping): Use
current_template_parms when the declaration is not available.
(norm_info::norm_info) Make explicit.
(normalize_constraint_expression): Factor into a separate overload
that takes arguments, and use that in the original function.
(tsubst_nested_requirement): Use satisfy_constraint instead of
trying to evaluate this as a constant expression.
(finish_nested_requirement): Keep the normalized constraint and the
original normalization arguments with the requirement.
(diagnose_nested_requirement): Use satisfy_constraint. Tentatively
implement more comprehensive diagnostics, but do not enable.
* parser.c (cp_parser_requires_expression): Relax requirement that
requires-expressions can live only inside templates.
* pt.c (any_template_parm_r): Look into type of PARM_DECL.
2019-11-06 Jason Merrill <jason@redhat.com>
* pt.c (use_pack_expansion_extra_args_p): Still do substitution if
all packs are simple pack expansions.
(add_extra_args): Check that the extra args aren't dependent.
gcc/testsuite/
* lib/prune.exp: Ignore "in requirements" in diagnostics.
* g++.dg/cpp2a/requires-18.C: New test.
* g++.dg/cpp2a/requires-19.C: New test.
From-SVN: r277900
GCC Administrator [Thu, 7 Nov 2019 00:16:37 +0000 (00:16 +0000)]
Daily bump.
From-SVN: r277899
Kwok Cheung Yeung [Thu, 7 Nov 2019 00:07:04 +0000 (00:07 +0000)]
Support using multiple registers to hold the frame pointer
When multiple hard registers are required to hold the frame pointer,
ensure that the registers after the first are marked as non-allocatable,
live and eliminable as well.
2019-11-07 Kwok Cheung Yeung <kcy@codesourcery.com>
gcc/
* ira.c (setup_alloc_regs): Setup no_unit_alloc_regs for
frame pointer in multiple registers.
(ira_setup_eliminable_regset): Setup eliminable_regset,
ira_no_alloc_regs and regs_ever_live for frame pointer in
multiple registers.
From-SVN: r277895
Kelvin Nilsen [Wed, 6 Nov 2019 23:10:51 +0000 (23:10 +0000)]
vsx.md (xxswapd_<mode>): Add support for V2DF and V2DI modes.
gcc/ChangeLog:
2019-11-06 Kelvin Nilsen <kelvin@gcc.gnu.org>
* config/rs6000/vsx.md (xxswapd_<mode>): Add support for V2DF and
V2DI modes.
From-SVN: r277893
Iain Sandoe [Wed, 6 Nov 2019 21:46:00 +0000 (21:46 +0000)]
[Darwin, testsuite] Fix framework-1.c on later Darwin.
The test works by checking that a known framework path is accessible
when the '-F' option is given. We need to find a framework path that
exists across a range of Darwin versions, and parseable by GCC. This
adjusts the test to use a header path that exists and is parsable from
Darwin9 through Darwin19.
gcc/testsuite/ChangeLog:
2019-11-06 Iain Sandoe <iain@sandoe.co.uk>
* gcc.dg/framework-1.c: Adjust test header path.
From-SVN: r277892
Jason Merrill [Wed, 6 Nov 2019 20:20:00 +0000 (15:20 -0500)]
C++20 NB CA378 - Remove constrained non-template functions.
No real use cases have ever arisen for constraints on non-templated
functions, and handling of them has never been entirely clear, so the
committee agreed to accept this national body comment proposing that we
remove them.
* decl.c (grokfndecl): Reject constraints on non-templated function.
From-SVN: r277891
Jan Hubicka [Wed, 6 Nov 2019 19:36:22 +0000 (20:36 +0100)]
ggc-common.c (ggc_prune_overhead_list): Do not delete surviving allocations.
* ggc-common.c (ggc_prune_overhead_list): Do not delete surviving
allocations.
* mem-stats.h (mem_alloc_description<T>::release_object_overhead):
Do not silently ignore summary corruptions.
From-SVN: r277890
Jonathan Wakely [Wed, 6 Nov 2019 17:53:38 +0000 (17:53 +0000)]
libstdc++: Add compare_three_way and install <compare> header
* include/Makefile.in: Regenerate.
* libsupc++/Makefile.in: Regenerate.
* libsupc++/compare (__3way_builtin_ptr_cmp): Define helper.
(compare_three_way): Add missing implementation.
From-SVN: r277889
Jonathan Wakely [Wed, 6 Nov 2019 17:53:12 +0000 (17:53 +0000)]
libstdc++: remove redundant equality operators
Now that operator<=> is supported, these operators can be generated by
the compiler.
* include/bits/iterator_concepts.h (unreachable_sentinel_t): Remove
redundant equality operators.
* testsuite/util/testsuite_iterators.h (test_range::sentinel):
Likewise.
From-SVN: r277888
Matthias Kretz [Wed, 6 Nov 2019 16:06:08 +0000 (16:06 +0000)]
Fix parser to recognize operator?:
This change lets grok_op_properties print its useful "ISO C++ prohibits
overloading operator ?:" message instead of the cryptic error message about
a missing type-specifier before '?' token.
2019-11-06 Matthias Kretz <m.kretz@gsi.de>
* parser.c (cp_parser_operator): Parse operator?: as an
attempt to overload the conditional operator.
From-SVN: r277887
Richard Sandiford [Wed, 6 Nov 2019 14:03:08 +0000 (14:03 +0000)]
Don't vectorise single-iteration epilogues
With a later patch I saw a case in which we peeled a single iteration
for gaps but didn't need to peel further iterations to make up a full
vector. We then tried to vectorise the single-iteration epilogue.
2019-11-06 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-loop.c (vect_analyze_loop): Only try to vectorize
the epilogue if there are peeled iterations for it to handle.
From-SVN: r277886
Claudiu Zissulescu [Wed, 6 Nov 2019 13:31:43 +0000 (14:31 +0100)]
[ARC] Don't split ior/mov predicated insns.
Do not split long immediate constants for predicated instructions.
gcc/
xxxx-xx-xx Claudiu Zissulescu <claziss@synopsys.com>
* config/arc/arc.c (arc_split_ior): Add asserts.
(arc_split_mov_const): Likewise.
(arc_check_ior_const): Do not match known short immediate values.
* config/arc/arc.md (movsi): Don't split predicated instructions.
(iorsi): Likewise.
testsuite/
xxxx-xx-xx Claudiu Zissulescu <claziss@synopsys.com>
Sahahb Vahedi <shahab@synopsys.com>
Cupertino Miranda <cmiranda@synopsys.com>
* gcc.target/arc/or-cnst-size2.c: New test.
Co-Authored-By: Sahahb Vahedi <shahab@synopsys.com>
From-SVN: r277885
Claudiu Zissulescu [Wed, 6 Nov 2019 13:31:32 +0000 (14:31 +0100)]
[ARC] Update mea option documentation
Update -mea option documentation.
gcc/
xxxx-xx-xx Claudiu Zissulescu <claziss@synopsys.com>
* config/arc/arc.opt (mea): Update help string.
* doc/invoke.texi(ARC): Update mea option info.
From-SVN: r277884
Claudiu Zissulescu [Wed, 6 Nov 2019 13:31:07 +0000 (14:31 +0100)]
[ARC] Cleanup sign/zero extend patterns
Clean up sign/zero extend patterns.
gcc/
xxxx-xx-xx Claudiu Zissulescu <claziss@synopsys.com>
* config/arc/arc.md (zero_extendqihi2_i): Cleanup pattern.
(zero_extendqisi2_ac): Likewise.
(zero_extendhisi2_i): Likewise.
(extendqihi2_i): Likewise.
(extendqisi2_ac): Likewise.
(extendhisi2_i): Likewise.
From-SVN: r277883
Richard Biener [Wed, 6 Nov 2019 12:56:42 +0000 (12:56 +0000)]
tree-vect-loop.c (vectorizable_reduction): Remember reduction PHI.
2019-11-06 Richard Biener <rguenther@suse.de>
* tree-vect-loop.c (vectorizable_reduction): Remember reduction
PHI. Use STMT_VINFO_REDUC_IDX to skip the reduction operand.
Simplify single_defuse_cycle condition.
From-SVN: r277882
Richard Sandiford [Wed, 6 Nov 2019 12:31:19 +0000 (12:31 +0000)]
Use scan-tree-dump instead of scan-tree-dump-times for some vect tests
With later patches, we're able to vectorise the epilogues of these tests
on AArch64 and so get two instances of "vectorizing stmts using SLP".
Although it would be possible with a bit of effort to predict when
this happens, it doesn't seem important whether we get 1 vs. 2
occurrences. All that matters is zero vs. nonzero.
2019-11-06 Richard Sandiford <richard.sandiford@arm.com>
gcc/testsuite/
* gcc.dg/vect/slp-9.c: Use scan-tree-dump rather than
scan-tree-dump-times.
* gcc.dg/vect/slp-widen-mult-s16.c: Likewise.
* gcc.dg/vect/slp-widen-mult-u8.c: Likewise.
From-SVN: r277881
Richard Sandiford [Wed, 6 Nov 2019 12:29:47 +0000 (12:29 +0000)]
Check the VF is small enough for an epilogue loop
The number of iterations of an epilogue loop is always smaller than the
VF of the main loop. vect_analyze_loop_costing was taking this into
account when deciding whether the loop is cheap enough to vectorise,
but that has no effect with the unlimited cost model. We need to use
a separate check for correctness as well.
This can happen if the sizes returned by autovectorize_vector_sizes
happen to be out of order, e.g. because the target prefers smaller
vectors. It can also happen with later patches if two vectorisation
attempts happen to end up with the same VF.
2019-11-06 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-loop.c (vect_analyze_loop_2): When vectorizing an
epilogue loop, make sure that the VF is small enough or that
the epilogue loop can be fully-masked.
From-SVN: r277880
Richard Sandiford [Wed, 6 Nov 2019 12:29:27 +0000 (12:29 +0000)]
Restructure vect_analyze_loop
Once vect_analyze_loop has found a valid loop_vec_info X, we carry
on searching for alternatives if (1) X doesn't satisfy simdlen or
(2) we want to vectorize the epilogue of X. I have a patch that
optionally adds a third reason: we want to see if there are cheaper
alternatives to X.
This patch restructures vect_analyze_loop so that it's easier
to add more reasons for continuing. There's supposed to be no
behavioural change.
If we wanted to, we could allow vectorisation of epilogues once
loop->simdlen has been reached by changing "loop->simdlen" to
"simdlen" in the new vect_epilogues condition. That should be
a separate change though.
2019-11-06 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-loop.c (vect_analyze_loop): Break out of the main
loop when we've finished, rather than returning directly from
the loop. Use a local variable to track whether we're still
searching for the preferred simdlen. Make vect_epilogues
record whether the next iteration should try to treat the
loop as an epilogue.
From-SVN: r277879